Scientific basis for erosion and sediment control ... · Scientific basis for erosion and sediment...

Scientific basis for erosion and sediment control practices in New Zealand


Les Basher

Landcare Research

Jonathan Moores

NIWA

Gregor McLean

Southern Skies Environmental Ltd

Prepared for:

Tasman District Council

189 Queen St Private Bag 4 Richmond, Nelson 7050 New Zealand

July 2016

Landcare Research, 1st Floor, 24 Nile Street, Private Bag 6, Nelson 7042, New Zealand, Ph +64 3 545 7700, www.landcareresearch.co.nz

NIWA, 41 Market Place, Viaduct Harbour, Private Bag 99940, Auckland 1149

Southern Skies Environmental Ltd, Unit 107b, 19 Surrey Crescent, P O Box 46-188, Auckland 1147

http://www.landcareresearch.co.nz/

Reviewed by: Approved for release by:

Ian Lynn Capability Leader Landcare Research

Chris Phillips Portfolio Leader – Managing Land and Water Landcare Research

Landcare Research Contract Report: LC2562

Disclaimer

This report has been prepared by Landcare Research for Tasman District Council. If used by other parties, no warranty or representation is given as to its accuracy and no liability is accepted for loss or damage arising directly or indirectly from reliance on the information in it.

© Landcare Research New Zealand Ltd and Tasman District Council 2016

No part of this work covered by copyright may be reproduced or copied in any form or by any means (graphic, electronic, digital or mechanical, including photocopying, recording, taping, information retrieval systems, or otherwise), in whole or in part, without the written permission of Landcare Research or Tasman District Council.

Landcare Research Page iii

Contents

Summary .................................................................................................................................... v

1 Introduction ...................................................................................................................... 1

2 Background ....................................................................................................................... 1

3 Objectives ......................................................................................................................... 3

4 Methods ........................................................................................................................... 4

5 Erosion and sediment control practices used in New Zealand ........................................ 5

5.1 Urban earthworks and infrastructure ................................................................................ 5

5.2 Forestry............................................................................................................................. 12

5.3 Horticulture and arable cropping ..................................................................................... 12

5.4 Pastoral farming ............................................................................................................... 13

6 Assessment of erosion and sediment control performance for urban earthworks and infrastructure ........................................................................................................................... 24

6.1 Erosion control practices .................................................................................................. 24

6.2 Silt fences and other temporary sediment control practices........................................... 28

6.3 Sediment retention ponds................................................................................................ 32

6.4 Chemical treatment .......................................................................................................... 37

6.5 Summary........................................................................................................................... 46

7 Assessment of erosion and sediment control performance for horticulture and arable cropping ................................................................................................................................... 48

7.1 Wheel track ripping .......................................................................................................... 49

7.2 Wheel track diking ............................................................................................................ 50

7.3 Cover crops ....................................................................................................................... 50

7.4 Grassed riparian buffer strips ........................................................................................... 53

7.5 Sediment retention ponds................................................................................................ 55

7.6 Management of wind erosion .......................................................................................... 56

7.7 Summary........................................................................................................................... 57

8 Assessment of erosion and sediment control performance for pastoral farming ......... 58

8.1 Performance of space-planted trees ................................................................................ 58

8.2 Erosion control using closed canopy vegetation .............................................................. 70


Page iv Landcare Research

8.3 Management of riparian margins and bank erosion ........................................................ 80

8.4 Management of surface erosion ...................................................................................... 83

8.5 Summary........................................................................................................................... 84

9 Assessment of erosion and sediment control performance for forestry....................... 86

9.1 Impacts of forest harvesting on erosion and sediment yield ........................................... 86

9.2 Information on effectiveness of ESC practices used in forestry ....................................... 88

9.3 Summary........................................................................................................................... 91

10 Modelling the effect of erosion and sediment control practices .................................. 92

10.1 Urban development and road construction projects ....................................................... 92

10.2 Pastoral farming ............................................................................................................... 96

11 Current and proposed research projects ....................................................................... 98

12 Conclusions ................................................................................................................... 100

13 Acknowledgements ...................................................................................................... 101

14 References .................................................................................................................... 101

Landcare Research Page v

Summary

Project and Client

Tasman District Council (TDC) received funding from the Ministry for the Environment Community Environment Fund for a project to understand the scope and applicability of the science that underpins present best practice guidance for erosion and sediment control (ESC) in both rural and the urban areas of New Zealand. TDC engaged Landcare Research to undertake this review and NIWA and Southern Skies Environmental Ltd were subcontracted to assist with the review

Objectives

Review the range of ESC practices used in New Zealand

Review the science underpinning choice of ESC practice and performance of each ESC practice

Provide a list of current and proposed future research on ESC in New Zealand

Methods

Information on erosion and sediment control practices used in New Zealand was primarily derived from published sources including regional council and industry ESC guidelines, and the Soil Conservation Technical Handbook.

For each ESC practice a literature review was undertaken, to identify scientific studies of the practice and its performance efficiency, and to identify relevant modelling literature. The review focused on New Zealand-based information, but also reviewed relevant international literature. The review focused on quantitative assessments of performance and covered both the on-site performance and mitigation of off-site effects.

Science databases were used to access published information on ESC treatment performance, and Crown Research Institutes, universities, other research groups, regional councils, central government agencies, industry sector groups and individuals were contacted to identify any relevant unpublished literature.

Results

A wide variety of ESC practices are used in New Zealand, depending on the land use and the type of erosion process(es) that are active. ESC practices for runoff-generated erosion (sheet, rill, gully) can be broadly categorised as: runoff control to reduce water velocity and sediment generation, and to separate clean water and dirty water; erosion control to reduce sediment generation; and sediment control to trap sediment before it moves offsite and into water ways. Mass movement erosion is controlled by practices that influence slope hydrology and/or soil strength. Streambank erosion is controlled by practices that reduce hydraulic scour, or increase bank strength and


Page vi Landcare Research

resistance to erosion. Wind erosion is controlled by practices that reduce soil erodibility, increase soil moisture content or reduce wind erosivity.

Since the late 1990s there have been a number of studies in the Auckland region to establish the performance of ESC practices for earthworks that have backed up overseas performance studies.

ESC practices can be highly effective in reducing the generation of sediment and its discharge from earthworks construction sites (often by an order of magnitude).

Erosion control is best achieved by the use of: fibrous, interwoven materials rather than loose mulches; material with a high percentage cover; relatively thick materials with a high water-holding capacity; flexible, relatively heavy materials that follow the ground contour; combining a number of treatments; re-establishment of vegetation cover as soon as possible; and mulches in combination with topsoil rather than subsoil.

The performance of silt fences and other temporary sediment control practices is improved: where they are sited in locations that promote extensive upstream ponding; by using materials with pore sizes that detain flows and promote clogging of the filter fabric; by regular maintenance to remove the build-up of captured sediments and clogging of filter fabric; by vegetated filter strips are installed downstream of filter fences; and the use of floating decants in decanting earth bunds.

Sediment retention ponds perform best where: design allows extended detention time to promote the settlement of sediments and infiltration through the pond bed; there is a relatively high length:width ratio to maximise the distance between the pond inlet and outlet; there is permanent water in the pond rather than full drainage; design promotes mixing by avoiding dead zones and sheet flow; approach channels, inlets, and pond side walls are protected from erosion; forebays and baffles, which are not readily overtopped, are used to reduce velocity and promote sediment settling; outlets only discharge effluent from the pond water surface rather than the entire water column; and outlet filters (gravel or expanded polystyrene envelopes) are fitted to outlet risers.

ESC practices have generally been found to be less effective for retaining clays and silts than sand, and chemical treatment is used to enhance binding of sediments at source or to promote settling of fine particles in sediment retention devices.

Chemical treatment applied to erosion control practices does not markedly improve the performance of practices such as mulching and grass seeding and its effectiveness reduces with time since application.

Studies of chemical treatment of sediment retention devices show: chemical treatment can markedly improve performance and is more effective as detention time increases, allowing more time for flocculation and settling; well-managed, liquid polyaluminium chloride treatment used in New Zealand performs well, especially during larger storm events and during the winter; careful management is required to avoid over-dosing of ponds and potential discharge of elevated concentrations of dissolved aluminium; solid forms of chemical treatments (e.g. Magnasol® Floc Blocks) can be subject to a number of problems, such as burial and degradation, meaning that runoff control and pre-treatment by sediment forebays may be required.


Landcare Research Page vii

ESC practices to control water and wind erosion on cropland have been little studied in New Zealand.

In row crops compacted wheel tracks are recognised as major sources of runoff and erosion. Ripping of wheel tracks reduced erosion by 95% on strongly structured clay soils. Overseas literature suggests this practice would be most effective on silty and clayey textured soils and less effective on sandy soils.

Most research on erosion on pastoral farmland has been in highly erodible North Island hill country with soft or crushed rock geology where space planted trees (mainly willows and poplars) and afforestation are the most commonly used practices for controlling erosion shallow landsliding, gullying, earthflows, and bank erosion.

Published reductions in landsliding using space-planted trees from quantitative studies range from 70 to 95%, but measured or assessed reductions are often far less than this because plantings are inadequate and poorly maintained. Individual trees influence the amount of landsliding within a radius of c. 10 m.

Similarly, afforestation is often used to control widespread and severe erosion in the most erodible hill country. Comparisons of mature, closed-canopy, indigenous or exotic forest (and scrub) typically show reductions in landsliding (compared with pasture) of c. 90%, and similar reductions in sediment yield at small-catchment scale. Afforestation has been used to control severe gully erosion and reduce rates of earthflow movement.

Bank erosion can be an important source of sediment into waterways but there has been very little quantitative research on rates of bank erosion or mitigation of bank erosion in New Zealand. A combination of ‘soft’ biological erosion control and ‘hard’ engineering works are used to control bank erosion, along with stock exclusion. Research suggests livestock removal from riparian areas improves bank stability, but the effects of riparian planting are more equivocal and are only likely to be observable in the long-term. Recent research suggests that for bank erosion mitigation to be effective an understanding of bank erosion processes is needed to guide which riparian intervention measures may be most effective in different parts of a catchment.

Riparian buffer strips are commonly used to reduce sediment input from surface erosion to streams and have been shown to reduce sediment input by >50% in overseas studies. Factors controlling their effectiveness include width, vegetation composition and structure, contributing slope length and steepness, sediment size, and soil infiltration rate. Most sediment is deposited within the first few metres of a buffer strip, and increasing the width beyond 3–5 m often has little effect on trapping efficiency.

Maintaining a persistent, complete pasture sward reduces the prevalence and severity of surface erosion processes. Grazing management to maintain adequate cover and canopy height is important in minimising soil loss by surface erosion in New Zealand hill country pastures.

Earthworks and clearfelled areas of forests have the potential to generate large amounts of sediment by both surface erosion processes and mass movement. Landslides can mobilise logging slash in debris flows and cause severe off-site effects.


Page viii Landcare Research

The effects of forest harvesting on increasing sediment yield, and the consequences of poor road and landing construction have been well characterised.

In recent years there has been a major emphasis by the forest industry to better manage the environmental impacts of forestry, with a strong emphasis on infrastructure engineering for water and sediment control, and careful siting of roads and landings to reduce erosion hazard. Currently, the relative contribution of sediment from infrastructure and clear-cuts is poorly characterised.

ESC practices promoted for forestry listed in regional council guidelines are largely derived from those used for urban earthworks and infrastructure, with the addition of some practices that are forestry specific and aimed at direct harvesting effects. The latter tend to be general guidelines (e.g. haul away from watercourses, safely dispose of slash). There appear to have been no forestry-specific New Zealand studies on whether ESC design criteria in council guidelines are appropriate. Rather, these criteria are based on the experience of practitioners.

Riparian buffers can contribute to reductions in sediment input to streams, but there is a lack of certainty about what sized buffer or set-back is required to be effective in plantation forestry, especially to mitigate the effects of landslides.

Reviews of the effectiveness of forestry best management practices in the USA mostly demonstrate the effect of implementing multiple best management practices, including silvicultural options, road and track management, and stream crossing management. These reviews show that best management practices can minimize erosion and sedimentation (quoted sediment reduction efficiencies of >50%), and riparian buffers are effective in trapping sediment mobilised in surface runoff. None of the reviews specifically mention management practices aimed at minimising landslides or debris flows.

A number of models have been developed and used in New Zealand for addressing the effects of ESC practices on erosion and sediment yield, ranging from simple empirical models with limited data input requirements to detailed process-based models with high data input requirements.

The USLE (Universal Soil Loss Equation) and GLEAMS (Groundwater Loading Effects of Agricultural Management Systems) have been the most commonly used models to estimate sediment loads discharged from earthworks sites. The USLE provides average sediment losses while GLEAMS provides sediment losses on a daily time step.

In the USLE erosion control practices are accounted for through the C (cover management) and P (supporting practices) factors. The C factor is used to represent the performance of mulching, grass cover and other forms of erosion control and values lie in the range 0.02–0.2 (i.e. 80–98% reduction in sediment losses). The P factor is used to represent the influence of surface roughness in reducing sediment generation and can also be used to take account of soil compaction. The C and P factors can have values >1 where scraping, compaction, and sealing of the upper soil surface reduce infiltration and increase sediment runoff. The USLE does not explicitly take account of sediment control practices (e.g. sediment retention ponds), rather they are accounted for by load reduction factors (LRFs) reflecting the performance of different sediment control practices.


Landcare Research Page ix

In GLEAMS ESC practices are modelled as a number of different ‘bare earth’ cover classes reflecting vegetation cover and erosion control practices. Like the USLE, GLEAMS does not explicitly take account of sediment control practices. In New Zealand uses of GLEAMS they are accounted for by LRFs (e.g. mulching 0.85, chemically treated sediment retention pond 0.65–0.95), or by using a post-processing module to allow explicit simulation of a sediment retention pond and by taking account of differences in particle settling with particle size. GLEAMS is also more capable of representing the staging of construction and ESC works, earthworks scenario analysis, and providing input to receiving environment models.

The impact of ESC control practices in a pastoral farming context has been modelled using NZeem® or SedNetNZ to calculate current erosion rates, then applying effectiveness factors to represent the reduction resulting from applying ESC practices (e.g. afforestation 90%, space-planted trees 70%), along with a maturity factor to represent the age of the trees. This approach has been used to assess the impact of implementing Whole Farm Plans primarily using space-planted trees and afforestation for erosion control.

Conclusions

ESC practices used in New Zealand are based on the following set of principles to control different erosion processes:

Runoff-generated erosion is managed by runoff control to reduce water velocity and to separate clean water and dirty water; erosion control to reduce sediment generation; and sediment control to manage sediment movement offsite

Mass movement erosion is controlled by practices that influence slope hydrology and/or soil strength

Streambank erosion is controlled by practices that reduce hydraulic scour, or increase bank strength and resistance to erosion

Wind erosion is controlled by practices that reduce soil erodibility, increase soil moisture content or reduce wind erosivity.

Based both on local studies in Auckland, and overseas studies, ESC practices can be highly effective in reducing the generation of sediment and its discharge from earthworks construction sites (often by an order of magnitude). These have established both the performance of a range of ESC practices and the factors that determine performance. Particular attention has been directed at methods to retain clay size sediment using chemical treatment.

ESC practices to control water and wind erosion on cropland have been little studied in New Zealand but there is an extensive body of overseas literature. In New Zealand there has been a focus on the importance of compacted areas (especially wheel tracks in row crops) generating runoff and sediment.

Research on erosion on pastoral farmland has focused on the performance of space-planted trees and afforestation in reducing landsliding, gully erosion, and earthflow movement. This has both established performance effectiveness and developed


Page x Landcare Research

recommended treatment options for biological erosion control. There has been limited research on bank erosion control or the performance of riparian buffer strips.

Earthworks and clear-felled areas of plantation forests can generate large amounts of sediment by both surface erosion processes and mass movement. While regional council and industry guidelines for ESC focus on earthworks using similar practices to those employed on other construction sites, there are no forestry-specific New Zealand studies to establish performance of these practices, or to determine the relative contribution of sediment from infrastructure (primarily by runoff driven processes) and from the clear-cuts (primarily by mass movement processes). Recommended ESC practices are based on the experience of practitioners.

Landcare Research Page 1

1 Introduction

Tasman District Council (TDC) has received funding from the Ministry for the Environment Community Environment Fund for a project to understand the scope and applicability of the science that underpins present best practice guidance for erosion and sediment control (ESC) in both rural and the urban areas of New Zealand. The project aimed to establish what relevant research has been conducted both in New Zealand and internationally and what research is presently in progress or planned for the future. TDC sought to understand how applicable this research is to different terrains, climates and land disturbance activities in New Zealand, and also what gaps in present and planned research warrant future investigation.

TDC engaged Landcare Research to:

provide a list of ESC practices used in New Zealand

review the science underpinning the choice of ESC practice and performance for each ESC practice

provide a list of current and proposed future research on ESC in New Zealand

identify information gaps to guide improvements in ESC practice and performance.

Landcare Research subcontracted NIWA and Southern Skies Environmental Ltd to assist with the review. This report provides a list of ESC practices used in New Zealand, and reviews the science available to assess the performance of these different ESC practices.

2 Background

New Zealand has a natural environment and history of land management that predisposes the country to soil erosion (Basher 2013). Erosion processes are naturally very active as a result of a dominance of steep slopes, weak rocks, high rainfall, and common high-intensity rainstorms (e.g. McSaveney 1978; Soons & Selby 1992; Hicks et al. 2011; Basher 2013). Deforestation of much of the country has been relatively recent and extensive, while the introduction of large numbers of grazing animals, and intensive land use in some areas has also accelerated rates of erosion (e.g. Page et al. 2000; Basher & Ross 2002; Glade 2003). Land development associated with urban and roading development has also become a major source of sediment generation (Hicks 1994; Auckland Regional Council 1996). Regional patterns of soil erosion are distinctive, reflecting both the natural environmental variation and land management practices (e.g. Cumberland 1944; Eyles 1983).

Awareness of soil erosion and the need for soil conservation became a matter of national concern by the 1940s following storm events in North Island hill country, and the apparent man-induced degradation of large tracts of the South Island mountainlands (Committee of Enquiry 1939; Gibbs & Raeside 1945; McCaskill 1973; Roche 1994). This resulted in the passing of the 1941 Soil Conservation and Rivers Control Act, the establishment of the Soil Conservation and Rivers Control Council (SCRCC), and catchment boards to manage erosion and sedimentation problems and the development of a wide variety of techniques for


Page 2 Landcare Research

controlling erosion. The initial focus was on controlling extensive gully and earthflow erosion, revegetating extensive areas of bare ground, and river control. Techniques were developed through trial and error and experimentation, and included spaced tree planting, graded banks and terraces, contour cultivation, conservation fencing, contour drains, debris dams, drop structures, farm plans, and identification of land for retirement. The SCRCC undertook research and surveys to underpin the development of soil conservation practice, published bulletins describing the techniques and their application, ran training courses to facilitate practical application, and established experimental farms to experiment and demonstrate soil conservation techniques (McCaskill 1973; Roche 1994).

For rural New Zealand there has been strong emphasis on biological erosion control (either through space-planted trees or blanket afforestation) because of its relative low cost and its effectiveness (Douglas et al. 2013; Phillips & Marden 2005). Since the 1940s, erosion and sediment control techniques have been refined, experimental work has provided better information on treatment performance, and there has been better documentation of the application of ESC techniques, and increasing emphasis on ESC for earthworks in urban environments, infrastructure projects, and forestry. The growth of ESC for earthworks in urban environments has been driven by extensive land development required for population growth in New Zealand, and the potential for severe effects, particularly in receiving environments adjacent to developing cities such as Auckland (Hicks 1994; Auckland Regional Council 1996). In the 1970s and 1980s large areas of erodible hill country were converted from pasture to Pinus radiata forests to control erosion. Subsequently, as a result of government policy, many of these forests have become production forests. Many are currently being harvested and the erosion problems that were evident under pasture are reoccurring (Marden 2004; Phillips et al. 2012).

There is a wide range of information sources on management options for ESC in New Zealand. The Plant Materials Handbook for Soil Conservation (Van Kraayenoord & Hathaway 1986a, b; Pollock 1986) summarised the state of knowledge up to the 1980s on vegetation options for managing soil erosion. Control of Soil Erosion on Farmland (Hicks 1995) was published by MAF and summarised a large amount of information on agricultural techniques for managing soil erosion throughout New Zealand. The Soil Conservation Technical Handbook (Hicks & Anthony 2001) describes a range of biological and engineering techniques for treating all types of erosion. Most regional councils use these sources and have developed locally relevant and practical resources to provide advice to farmers (fact sheets, newsletters, website information, etc.).

In the urban environment, Auckland Regional Council published a set of ESC guidelines for earthworks in 1995 that was significantly revised and published as TP90 in 1999 (Auckland Regional Council 1999). TP90 has formed the basis of ESC guidelines prepared by many other regional councils around New Zealand including Environment Bay of Plenty (Environment Bay of Plenty 2010), Environment Waikato (Environment Waikato 2009), Hawkes Bay Regional Council (Shaver 2009a), Taranaki Regional Council (Taranaki Regional Council 2006), Greater Wellington Regional Council (Greater Wellington Regional Council 2006), and Environment Canterbury (Environment Canterbury 2007). TP90 had limited revision in 2007 and is currently being substantially revised. The New Zealand Transport Agency also recently produced a set of ESC guidelines specifically aimed at the state



highway infrastructure (New Zealand Transport Agency 2014). Practical advice on ESC for building sites is contained in the recently published Builders Pocket Guide (Ecan, undated).

The forest industry has developed a Code of Practice for plantation forestry (NZ Forest Owners Association 2007), providing practical advice for managing ESC, as well as a Road Engineering Manual that provides a comprehensive guide to planning and constructing forest roads and associated infrastructure to manage ESC problems (Gilmore et al. 2011). Many regional councils, including Auckland Council (Bryant et al. 2007), Environment Bay of Plenty (Environment Bay of Plenty 2012), Hawkes Bay Regional Council (Shaver 2009b), Northland Regional Council (Northland Regional Council 2012), Greater Wellington Regional Council (Greater Wellington Regional Council 2006), and Marlborough District Council (Williams & Spencer 2013), have produced ESC guidelines for forestry largely based on adapting TP90 with a focus on managing the ESC effects of earthworks.

Specific ESC guidance has also been produced for the horticulture industry (Franklin Sustainability Project 2000; Barber & Wharfe 2010; Barber 2014). These guidelines draw on research carried out in the Franklin Sustainability Project and from Auckland Regional Councils guidelines for earthworks and forestry (Auckland Regional Council 1999; Bryant et al. 2007).

Alongside the development of guidelines for ESC there has been substantial growth of a private sector ESC industry that undertakes much of the planning and implementation of ESC practices on development sites, and is also responsible for training contractors and other council staff on ESC methodology. The industry also carries out research and experimentation to improve ESC practices.

While there is abundant guidance about available ESC techniques for different erosion processes and land uses, most guidelines provide limited information on treatment performance. When TDC developed a draft ESC guideline they concluded there was uncertainty about the applicability of ESC methods and their specific design used elsewhere in New Zealand for the environment of Tasman region with its large variation in both soils and rainfall. Existing ESC guidelines do provide some guidance for design criteria for application to different types of sites or environmental conditions and the aim of the present project is to provide information which may allow these to be established more precisely.

3 Objectives

Review the range of ESC practices used in New Zealand.

Review the science underpinning choice of ESC practice and performance of each ESC practice.

Provide a list of current and proposed future research on ESC in New Zealand.



4 Methods

Information on erosion and sediment control practices used in New Zealand was derived from published sources including:

Control of Soil Erosion on Farmland (Hicks 1995)

The Soil Conservation Technical Handbook (Hicks & Anthony 2001)

Regional council ESC guidelines (Auckland Regional Council 1999; Taranaki Regional Council 2006; Greater Wellington Regional Council 2006; Environment Canterbury 2007; Environment Waikato 2009; Shaver 2009a; Environment Bay of Plenty 2010)

New Zealand Transport Agency ESC guidelines (New Zealand Transport Agency 2014)

Regional council ESC guidelines for forestry (Greater Wellington Regional Council 2006; Bryant et al. 2007; Shaver 2009b; Environment Bay of Plenty 2012; Northland Regional Council 2012; Williams & Spencer 2013)

Forest industry Environmental Code of Practice and Road Engineering Manual (NZ Forest Owners Association 2007; Gilmore et al. 2011)

ESC guidelines for horticulture (Franklin Sustainability Project 2000; Barber & Wharfe 2010; Barber 2014)

Information was also compiled from Web sites of erosion control companies in New Zealand.

A literature review was undertaken for each ESC practice, with a focus on the most commonly used practices, to identify scientific studies of the practice and its performance efficiency, and relevant modelling literature. While the focus was on New Zealand-based information, relevant international literature was also reviewed. Previously published reviews (e.g. Hicks 1994; Phillips et al. 2000, 2008; Parkyn et al. 2000; Parkyn 2004; Basher et al. 2008a, b; Basher 2013) were used and updated where new information was available. The review focused on quantitative assessments of performance and covered both the on-site performance and mitigation of off-site effects.

Science databases including NZ Science, the Networked Digital Library of Theses and Dissertations, NZResearch, CAB Abstracts, Science Direct and Web of Science, were used to access published information on erosion and sediment control treatment performance. Crown Research Institutes (including NIWA, GNS Science, AgResearch, Plant and Food Research), universities and other research groups (Cawthron Institute), regional councils, central government agencies (including NZTA, MPI, MfE), industry sector groups were also surveyed to identify any relevant unpublished literature.



5 Erosion and sediment control practices used in New Zealand

A wide variety of ESC practices are used in New Zealand, depending on the land use and the type of erosion process(es) that are active. ESC practices for runoff-generated erosion (sheet, rill, gully) can be broadly categorised as:

water management: control of runoff to reduce water velocity and sediment generation, and to separate clean water and dirty water

erosion control: to reduce sediment generation

sediment control: trapping sediment before it moves offsite and into water ways.

Mass movement erosion is controlled by practices that influence slope hydrology (e.g. reduction of soil water content, increase drainage) and/or soil strength. Streambank erosion is controlled by practices that reduce hydraulic scour, or increase bank strength and resistance to erosion. Wind erosion is controlled by practices that reduce soil erodibility, increase soil moisture content or reduce wind erosivity.

In most cases, an integral component is an ESC plan or a farm plan that is not an ESC practice per se but a framework within which to plan ESC (e.g. selection of practices, design of individual practices, location, etc.). ESC practices are listed below (Tables 1–4) for the major applications in New Zealand: urban earthworks and infrastructure, forestry, horticulture and arable cropping, and pastoral farming. In these tables the design criteria are not exhaustively listed (see the references for more detail) but include some factors that are relevant to thinking about how to design for different environmental conditions and assessing performance of ESC practices.

5.1 Urban earthworks and infrastructure

Design of ESC practices for urban earthworks and infrastructure is underpinned by a set of principles (see ARC 1999; NZTA 2014) that include:

controlling runoff from beyond, within, and from a construction site (by using perimeter controls, separating clean and dirty water, designing measures to safely discharge runoff)

controlling sediment generation at source (by minimising the area and intensity of disturbance, stabilising bare areas rapidly, protecting steep slopes and watercourses)

minimising sediment discharge from a site (by using detention devices)

Urban earthworks and infrastructure development involve extensive modification of soils, vegetation cover and hydrology and require a wide range of ESC practices to mitigate the effects of these activities (Table 1).

Table 1 List of erosion and sediment control practices used for urban earthworks and infrastructure

Description of method Design criteria variables Reference

Erosion and sediment control plan

Not an ESC practice per se, but a framework within which to plan ESC management

Auckland Regional Council (1999)

Runoff control

Check dams Small dams constructed across a swale or channel to act as grade control structures and reduce velocity of runoff

Contributing catchment size, slope of catchment, spacing between dams, height of dam, ephemeral watercourses only, construction materials (rock rip-rap, filter socks, sandbags, other non-erodible material)

New Zealand Transport Agency (2014)

Contour drains and cutoffs

Temporary excavated channels or ridges constructed slightly off the slope contour to reduce slope length and runoff velocity

Contributing catchment size, slope of catchment, spacing, bank height, channel depth, gradient, shape, stable outlet


Diversion channels and bunds

Non-erodible channels and/or bunds for the conveyance of runoff (either clean or dirty water) that are constructed for a specific design storm to intercept and convey runoff to stable outlets or sediment retention ponds at non-erosive velocities

Location, flow capacity, shape, gradient, stable walls and floor, stable outlet


Pipe drop structure and flume

Temporary pipe structures or constructed flumes placed from the top of a slope to the bottom of a slope to convey clean or dirty runoff without causing erosion

Gradient, stable entry and exit, construction materials (geotextiles, pipes, rock, sandbags, etc.), pipe size, contributing catchment area, catchment slope


Level spreader A non-erosive outlet for concentrated runoff constructed to disperse flows uniformly across a stabilised slope. Often used in combination with sediment retention ponds

Flow capacity, location (to allow flow to spread not concentrate), size (length, width, depth), stable inlet and outlet, grade of spreader is 0%, construction of spreader lip


Hay bale barriers Temporary barriers of hay bales used to intercept and direct surface runoff from small areas

Location, size and slope of contributing catchment Auckland Regional Council (1999)

Water table drains and culverts

A channel excavated parallel to a road or track to provide permanent drainage of the carriageway and/or to provide a conveyance channel for stormwater. Culvert connects the drain to a stable outfall

Design flow, shape, slope, drain armour, spacing of check dams, size and spacing of culverts, stable outfall

Environment Waikato (2009)


Erosion control

Stabilised entranceway

Stabilised pad of aggregate on a woven geotextile base located at any entry or exit point of a construction site to reduce erosion in heavily trafficked area. Can include shaker ramp and vehicle wash

Location, size, shape, construction materials, depth and size of aggregate


Surface roughening Roughening an unstabilised bare surface with horizontal grooves across the slope or by tracking with construction equipment to increase infiltration, surface roughness, detention storage and entrapment of sediment

Divert run-off from above, soil type and texture, rainfall intensity, machinery type, degree of compaction


Benched slopes Grading of sloped areas to form reverse sloping benches with diversion channels on a slope to minimise erosion by limiting volume and velocity of runoff

Slope length, slope steepness, spacing of benches, bench design (width, slope, flow length, diversion channel design), stable outlets, slope face management (grassing, filter socks, etc), diversion of run-off from above


Topsoiling and grass seeding

Planting and establishment of quick growing and/or perennial grass to provide temporary and/or permanent stabilisation on exposed areas, often undertaken in conjunction with the placement of topsoil. Reduces raindrop impact, runoff volume and velocity

Site preparation (installation of other ESC practices), seedbed preparation, fertiliser requirements, seed application (mixture, rate, application method, irrigation), timing


Hydroseeding Application of seed, fertiliser and paper or wood pulp in a slurry sprayed over an area to provide rapid re-vegetation. Reduces raindrop impact, runoff volume and velocity. Applied to critical or difficult areas

Location, site slope, soil conditions, seed mixture and amendments/binders, fertiliser requirements


Mulching Application of a protective layer of straw or other material (bark, wood residue, wood pulp) to the soil surface to stabilise soil surface and reduce raindrop impact and runoff, prevent soil crusting, and conserve moisture. Can be used in combination with regrassing and may need crimping or binders

Location, site slope, type of mulch, rate of mulch application, site conditions (e.g. windiness)



Turfing Establishment and permanent stabilisation of disturbed areas with a continuous cover of grass turf to provide rapid stabilisation. Reduces raindrop impact, runoff volume and velocity

Surface preparation, site conditions (e.g. temperature, gravel content, compaction), need for irrigation, turf application


Geotextiles, plastic covers, erosion control blankets, geo binders

Placement of a variety of erosion control products to stabilise disturbed soil areas and protect soils from erosion by wind or water. Applied to critical or difficult areas or other areas where there is inadequate space to install sediment controls. Includes temporary biodegradable geotextiles (jute, straw blanket, wood fibre blanket, coconut fire blanket or mesh), permanent non-degradable geotextiles (plastic netting or mesh, synthetic fibre with netting, bonded synthetic fibres) and combination synthetic and biodegradable rolled erosion control products

Type of material and product specifications, method of anchoring on slope, location of installation, site preparation


Soil binders and chemical treatment

Organic or chemical soil-stabilising agents that penetrate the soil and bind particles together to form protective crust which reduces windblown dust generation and raindrop impact

Type of binder, application rate and method, divert run-off from above, avoid trafficking, soil conditions

Environment Canterbury (2007)

Sediment control

Sediment retention pond (including flocculation systems)

Temporary pond formed by excavation into natural ground or by the construction of an embankment, with a decanting device to dewater the pond at a rate that will allow the majority of suspended sediment to settle out

Location, size and slope of contributing catchment, soil conditions, size and shape of pond (volume, length, width, depth, volume of dead and live storage, forebay size), decanting device (type, design and position), inlet and outlet design (including level spreader and emergency spillway), baffle location and type, chemical treatment (type, dose rate), emergency spillway


Decanting earth bunds Temporary bund or ridge of compacted earth to intercept sediment-laden runoff and reduce the amount of sediment leaving the site with a decanting device to dewater the decanting earth bund at a rate that will allow suspended sediment to settle out. Used on

Similar to above New Zealand Transport Agency (2014)


smaller areas or where a sediment retention pond cannot be installed

Silt fences Temporary barrier of woven geotextile fabric used to capture sediments carried in sheet flow

Type of fabric, location, contributing catchment size, slope steepness and length, spacing of returns, maximum, length, height, support type and spacing, soil type and texture


Super silt fences Temporary barrier of woven geotextile fabric over a chain link fence used to capture predominantly coarse sediments carried in sheet flow

Type of fabric, location, contributing catchment size, slope steepness and length, spacing of returns, maximum length, height, support type and spacing, soil type and texture


Filter socks A mesh tube filled with a filter material (e.g. compost, sawdust, straw) used to intercept and filter runoff and reduce the velocity of runoff

Filter material, size of sock, slope steepness and length, spacing of returns, location, support type and spacing


Flocculation including FlocSocks

Added to sediment retention pond inflows via a rainfall-activated system to accelerate coagulation and settlement of fine colloidal particles

Flocculant type and dose rate, dosing system, location of dosing point


Dewatering Removal of water from excavations, trenches and sediment control devices by pumping

Volume of water and the levels of sediment, disposal of water


Stormwater inlet protection

Barrier across or around a stormwater inlet to intercept and filter sediment-laden runoff before it enters a reticulated stormwater system (includes silt fence, geotextile fabric, filter sock, check dam, proprietary products)

Type of barrier, runoff management to and away from device


Sediment sump Temporary pit constructed to trap and filter water before it is pumped to a suitable discharge area

Location, number, size/volume, fill type, stable discharge area


Vegetative buffer zones and turf filter strips

Areas of existing grass cover which are retained at appropriate locations to remove small volumes of sediment from shallow sheet flows.

Location, contributing catchment area and slope, slope, width, spacing of stable returns



Soakage system Temporary soak pits to dispose of clean run-on water and sediment-laden site runoff into the ground where infiltration rates and groundwater levels allow

Fill type and size, groundwater levels, permeability, inlet protection, design of forebays


Sediment curtain Temporary floating geotextile fabric barriers suspended vertically within a water body (stream) to separate contaminated and uncontaminated water to isolate the work area and allow sediments to settle out of suspension

Stream width, velocity, water depth, fabric type, flotation and weighting devices, length and height of curtain


Streamworks

Temporary watercourse crossings

A bridge, ford or temporary structure installed across a watercourse for short term use by construction vehicles to cross watercourses without moving sediment into the watercourse, or damaging the bed or channel

Location, timing of construction, fish migration, loading, design storm flow, culvert size, inlet and outlet protection


Permanent watercourse crossings

Bridge, culvert or ford installed across a watercourse where permanent access is required across a small watercourse

Location, design storm flow, loading, culvert size, inlet and outlet protection

Environment Waikato (2009)

Dam (with pumping or diverting)

Temporary practices used to convey surface water from above a construction activity to downstream of that activity

Dam materials, design flow, pump size and installation, stable outlet


Temporary waterway diversions

A short-term watercourse diversion that allows work to occur within the main watercourse channel under dry conditions. Diverts all flow via a stabilised system around the area of works and discharge it back into the channel below the works to avoid scour of the channel bed and banks

Location, design flow, diversion channel design, diversion dam design


Instream and near stream works

Temporary structures built (from rock, sand bags, wood or a filled geotextile material) within the banks or channel of a waterway to enclose a construction area and reduce sediment delivery from work in or immediately adjacent to the waterway

Many and varied New Zealand Transport Agency (2014)


Rock outlet protection Rock (rip-rap or gabion baskets) placed at the outfall of channels or culverts

Location, slope, rock size, base protection Environment Waikato (2009)



5.2 Forestry

Plantation forests have been established throughout New Zealand with large areas established on steep slopes, erodible rock types and soils, and in areas where high intensity storms are a regular occurrence. Mass movement (landslides, earthflows, slumps), gully erosion and surface erosion can all be active within plantation forests, with many steepland forests having originally been planted for erosion control. While mature trees provide a valuable erosion control function there is a period immediately post-harvest (2–8 years) where slopes are especially vulnerable to erosion (Phillips et al. 2012). Soil disturbance associated with clear-felling, earthworks for landing and road construction, and activities around stream channels all have the potential to cause erosion. In the past earthworks for landing and road construction have been major sediment sources, but improved planning, engineering design and construction have reduced this problem and much of the sediment generated from forestry is generated from the clear-cuts by landsliding and debris flows (Phillips et al. 2012). The principles of urban ESC are now being applied to forestry practices but focus mainly on earthworks and less on how to manage erosion from the clear-cuts (Table 2). The principles for erosion control in forestry include (Amishev et al. 2013):

Keep disturbed areas small and time of exposure short

Control erosion at source

Install perimeter controls

Retain sediment on site

Protect critical areas

Inspect and maintain control measures

Establish the new crop as soon as possible.

Plantation forests located on steeplands are more prone to shallow landsliding for several years following harvesting than at any other time in the forest growth cycle. Removal of trees causes soil moisture conditions to be wetter for longer (because of loss of interception capacity of canopy and reduced evapotranspiration). Once trees are removed, roots slowly decay and soil reinforcement is reduced and not fully compensated for by the replanted trees for several years following planting. The result is a period often referred to as the “window of vulnerability” in which landslides are highly likely to occur if there is a severe storm. Commonly this can also result in highly destructive debris flows. Developing strategies for managing post-harvest landslides and debris flows is particularly problematical.

5.3 Horticulture and arable cropping

Horticulture and arable cropping can be affected by both water erosion and wind erosion. The activities associated with this land use involve intensive cultivation and periods with poor ground cover when the soil is not protected from erosion. A wide range of ESC practices are used (Table 3) with control of water erosion based on runoff control, erosion



control, and sediment control (Barber 2014) while control of wind erosion involves techniques that reduce soil erodibility and wind erosivity (Ross et al. 2000).

5.4 Pastoral farming

Large areas of erodible hill country are used for pastoral farming and these areas are affected by a wide range of erosion processes including shallow landslides, earthflows, gully erosion, streambank and surface erosion (Hicks 1995; Basher 2013). The intensity of these processes is partly controlled by natural factors including rock type, slope, soils, and rainfall, but there is also a strong effect of land use or vegetation cover (Glade 2003; Basher 2013). Large storms can result in regionally extensive areas of landsliding (e.g. Cyclone Bola March 1988, Manawatū-Wanganui February 2004) or more localised areas of intense landsliding (e.g. Wairarapa 1977, Dunedin 1994, Marlborough 2010) within pastoral farmland (Crozier 2005). In the flat lowlands streambank erosion is the most important form of erosion and can be exacerbated by stock access to streams.

Biological methods of erosion control are the most widely used with a large range of vegetation types and species used to control erosion throughout New Zealand (Hicks & Anthony 2001; Basher et al. 2008). Space-planted poplars and willows are widely used ESC plants in New Zealand since they can be established as poles in the presence of grazing animals, and are appropriate for the control of landslide, earthflow, gully and streambank erosion. Afforestation also remains an important ESC method for the areas most susceptible to erosion (Phillips & Marden 2005). Structural measures, control of drainage have also been used to control some forms of erosion. Table 4 lists the range of ESC methods used for erosion control for pastoral farming.

Table 2 List of erosion and sediment control practices used for forestry


Harvest plan Not an ESC practice per se, but outlines the requirements for erosion and sediment control

Bryant et al. (2007)

Runoff control

Diversion channels and bunds

Permanent non-erodible channels and/or bunds to convey clean runoff to stable outlet.

Location, flow capacity, shape, gradient, stable walls and floor, stable outlet


Contour drains and cutoffs

Temporary (usually) excavated channels or ridges constructed slightly off the slope contour to reduce slope length and runoff velocity and deliver runoff to stable outlet

Contributing catchment size, slope of catchment, spacing, bank height, channel depth, gradient, shape, stable outlet


Broad-based dips A dip and reverse slope in a road surface with an out-slope in the dip for natural cross drainage, to provide cross-drainage on in-slope roads and prevent build-up of runoff and erosion

Contributing catchment size, road/track slope, spacing, surface protection


Rolling dip A dip and reverse slope in a road surface with an out-slope in the dip for natural cross drainage to provide cross drainage on in-slope roads and prevent build-up of runoff and erosion; used on roads that are too steep for broad-based dips

Road gradient, spacing, slope Bryant et al. (2007)

Flumes and outfalls Mechanical conveyance system that transports water from one area to another via a stable outlet without causing erosion. Usually associated with culverts

Catchment area, design flow, construction material Bryant et al. (2007)

Check dams Small dams constructed across a swale or channel to act as grade control structures and reduce velocity of runoff

Contributing catchment size, slope, spacing between dams, height of dam, ephemeral watercourses only, construction materials (rock rip-rap, filter socks, sandbags, other non-erodible material), channel protection


Water table drains, culverts and sumps

A channel excavated parallel to a road or track to provide permanent drainage and control runoff and/or to provide a conveyance channel for stormwater. Culvert connects drain to a stable outfall and sump at upstream end of culvert can be included to trap coarse sediment

Design flow, shape, slope, drain armour, spacing of check dams within drain, size and spacing of culverts, stable outfall

Williams & Spenser (2013)


Erosion control

Surface roughening Roughening of a bare surface to create horizontal grooves that will reduce the concentration of runoff, aid infiltration, trap sediment and aid vegetation establishment

Contributing catchment size, soil type and texture, rainfall intensity, machinery type, degree of compaction


Log corduroying Placement of logs to provide a solid working platform, usually in wet processing areas or on access roads to minimise sediment generation

Location, log placement Bryant et al. (2007)

Slash and mulch placement

Application of a protective layer of hay/straw mulch or slash to the soil surface to reduce raindrop impact and prevent sheet erosion

Location, depth Bryant et al. (2007)

Grassing and hydroseeding

Sowing of seed to establish a vegetative cover over exposed soil and reduce raindrop impact and sheet/rill erosion. Hydroseeding allows revegetation of steep or critical areas that cannot be stabilised by conventional sowing methods.

Location, timing, catchment area, site slope, soil conditions, seed mixture, application rate, fertiliser requirements


Rock lining of channels

Protection of bare drains and roadside water tables in erosion prone soils against erosion

Catchment area, drain gradient, shape, construction materials, design flow


Geotextiles Fabrics used to protect soil surfaces against raindrop impact and sheet/rill erosion particularly in spillways and diversion channels

Location, fabric type, method of anchoring on slope, site preparation


Benched slopes Benches constructed on the outside of roads/tracks to place stable fill

Location, size, slope Williams & Spenser (2013)

Slash management Placement of slash to avoid mobilisation in water bodies and off landings

Storm frequency-magnitude, topography, soils, catchment size, proximity of trees to watercourses, watercourse values, benching, storage space, water control, slash placement

Northland Regional Council (2012)

Sediment control

Haybale barriers Temporary sediment retention devices to intercept and divert runoff for very small catchments

Catchment area, location, spacing, anchoring to slope Bryant et al. (2007)


Earth bund Ridge of compacted earth (preferably compacted subsoil) built on the contour to detain runoff and trap sediment

Catchment area, soil materials, height, length, stable outlet


Slash bund Temporary bunds of slash for very small catchments to trap the initial ‘pulse’ of coarse sediment

Catchment area, location, shape, size, amount of slash Bryant et al. (2007)

Earth bund Temporary bund or ridge of compacted earth to detain runoff long enough to allow sediment to drop out of suspension prior to discharge from catchments <0.1.ha. Typically a continuous bund constructed on the contour (e.g. around the toe of a landing) or a ‘horseshoe’ shape incorporating a natural depression

Catchment area, length, height, batter slope, area, compaction


Silt fence Temporary barrier of woven geotextile fabric used to capture sediment carried in sheet flow from small areas

Catchment area, slope steepness, location, slope length, spacing, anchoring to slope, fabric type


Super silt fence (debris dam)

Temporary barrier of woven geotextile fabric over a chain link fence used to capture predominantly coarse sediments carried in sheet flow often constructed in areas of active erosion

Catchment area, location, type of fabric, contributing catchment size, height, spacing, support type and spacing


Silt trap Temporary small sediment retention pond system Catchment area, location, size, stable inlet and outlet Bryant et al. (2007)

Sediment retention pond (including flocculation systems)

Temporary pond formed by excavation into natural ground or by the construction of an embankment, with a decanting device to dewater the pond at a rate that will allow the majority of suspended sediment to settle out

Location, size and slope of contributing catchment, soil conditions, size and shape of pond (volume, length, width, depth, volume of dead and live storage, forebay size), decanting device (type, design and position), inlet and outlet design (including level spreader and emergency spillway), baffle location and type, chemical treatment (type, dose rate)


Sediment trap/soak hole/sump

Constructed hole in porous soils used to control runoff from roads/tracks and trap sediment

Location, spacing, size/volume, soil conditions, stable inlet, use of silt fence

Environment Bay of Plenty (2012)

Streamworks

Harvesting operations Planning of harvesting operations to minimise impacts on stream channels

Fell away from streams if possible, remove slash from streams, don’t haul through streams, stabilise margins post-harvest



Dry stream crossings Temporary crossings of ephemeral channels protected by log corduroying

Location, catchment area Bryant et al. (2007)

Permanent watercourse crossings

Bridge, culvert or ford installed across a watercourse where permanent access is required across a small watercourse

Location, catchment area, design storm flow, culvert size, inlet and outlet protection, road runoff diversion, stabilised approach


Dam (with pumping or diverting)

Temporary practices used to convey surface water from above a construction activity (e.g. culvert installation) to downstream of that activity

Dam materials, design flow, pump size and installation, stable outlet


Temporary waterway diversion

A short-term watercourse diversion that allows work to occur within the main watercourse channel under dry conditions. Diverts all flow via a stabilised system around the area of works and discharges it back into the channel below the works to avoid scour of the channel bed and banks

Location, design flow, diversion channel design, diversion dam design


Table 3 List of erosion and sediment control practices used for horticulture and arable cropping


Erosion management plan

Not an ESC practice per se, but a framework within which to plan ESC management

Barber (2014)

Water erosion

Runoff control

Interception drains Drains to intercept and control runoff from above. If gradient steep then requires check dams

Catchment area and slope, design flow, gradient, soil materials

Barber (2014)

Culverts In drains to pass paddock entranceways Catchment area, design flow, culvert size Barber (2014)

Benched headlands Used to direct runoff to paddock edge or drain (stable outlet). May be grassed to trap sediment

Paddock size, slope length, runoff volume, soil materials Barber (2014)

Diversion bund Earth bund used to divert runoff away from vulnerable paddock or to prevent water discharging directly from a paddock

Location, flow capacity, shape, gradient, stable walls and floor, stable outlet, connection to other ESC measures

Barber (2014)

Contour drains Temporary excavated channels or ridges constructed slightly off the slope contour to reduce slope length and runoff velocity and deliver runoff to stable outlet

Contributing catchment size, slope of catchment, spacing, slope of drain, length, soil materials, depth

Barber (2014)

Grassed swale (within-paddock)

Grass-covered surface drain formed used to direct clean water runoff along the swale, following its natural course, to a stable outlet

Catchment area, swale width, slope length, design flow, gradient, soil materials

Barber (2014)

Stabilised (raised) access ways and discharge points

Metalled access point used to control runoff and direct to a stable outlet or other ESC measure

Location, connection to other ESC measures, culvert size Barber (2014)

Erosion control

Cover crops Crop planted to protect the soil from raindrop impact and sheet/rill/wind erosion between rotations, and ploughed into the soil before planting of a new crop

Type of crop, rate of growth Barber (2014)


Wheel track ripping Shallow cultivation of compacted wheel tracks in row crops to increase infiltration and reduce erosion

Slope length, soil materials, type of implement Barber (2014)

Wheel track diking Use of an implement to create series of closely-spaced soil dams in compacted wheel tracks

Slope length, soil materials, type of implement Barber (2014)

Paddock length Used to break up long paddocks, control runoff and erosion Slope length, soil materials Barber (2014)

Cultivation practices Used to manage soil structure and organic matter, increase infiltration and reduce runoff and erosion. Includes minimum tillage, no-tillage and stubble retention

Type of implements, number of cultivation passes, surface roughness, moisture content, cultivation direction, slope

Hicks & Anthony (2001)

Strip cropping Strips of permanent vegetation retained between crops to break up slope length and reduce water and wind erosion

Spacing, width, vegetation type Hicks & Anthony (2001)

Sediment control

Vegetated buffers and riparian margins

Grass or hedge areas adjacent to waterways or at paddock boundaries to reduce runoff velocity and filter sediment

Contributing catchment area, width, species composition Barber (2014)

Silt/Super Silt fences Temporary barrier of woven geotextile fabric (incorporating a chain link fence – Super Silt fence) used to capture sediments carried in sheet flow from small catchments

Contributing catchment area, slope, spacing, fabric type Barber (2014)

Decanting earth bund Shallow bund or ridge of compacted earth installed at bottom of paddock to pond runoff, with a decanting device to dewater the bund at a rate that will allow suspended sediment to settle out. Used on smaller areas or where a sediment retention pond cannot be installed

Contributing catchment area, location, design flow, volume of dead and live storage, decant type and rate, emergency spillway

Barber (2014)

Silt trap Sediment retention pond formed by excavation into natural ground or by the construction of an embankment, with a decanting device to dewater the pond at a rate that will allow the majority of suspended sediment to settle out

Location, size and slope of contributing catchment, soil conditions, size and shape of pond (volume, length, width, depth, volume of dead and live storage, forebay size), decanting device (type, design and position), baffle location and type, stable outlet

Barber (2014)

Wind erosion


Cultivation management

Used to manage soil structure, organic matter, surface roughness, reduce soil erodibility and erosion. Includes minimum tillage, no-tillage and stubble retention

Type of implements, number of cultivation passes, surface roughness, aggregate size, moisture content, cultivation direction, time soil is bare, field width, soil materials

Ross et al. (2000)

Windbreaks Used to reduce windspeed at ground level and wind erosion Width of shelterbelt, tree species Ross et al. (2000)

Strip cropping Strips of permanent vegetation retained between crops to break up paddock length and reduce wind erosion

Spacing, width, vegetation type Hicks & Anthony (2001)

Table 4 List of erosion and sediment control practices used for pastoral farming


Farm plan Not an ESC practice per se, but a framework within which to plan ESC management

Surface erosion

Pasture management Maintenance of high level of ground cover to reduce sheet/rill/wind erosion

Stocking level, stock type, timing and duration of grazing, species composition, fertiliser management, fencing


Contour furrows Furrow constructed with slight gradient to break up slope to control runoff

Slope, spacing, contributing area, soil type Hicks & Anthony (2001)

Mass movement (shallow landslides, slumps, earthflows)

Spaced planting Planting of spaced poles to reduce soil water content, increase soil strength and reduce erosion

Location of planting, tree species, spacing, extent of planting, pole protection, stock management


Afforestation Blanket planting of closely spaced trees to reduce soil water content, increase soil strength and reduce erosion

Location of planting, extent of planting, spacing, tree species


Reversion Removing stock and fencing erosion-prone areas to encourage reversion to woody vegetation to reduce erosion

Location, seed source, species composition, rate of reversion

Hicks (1995)

Surface drainage Use of surface ditches, cutoff drains and graded banks to reduce infiltration and dewater ponding areas on slumps and earthflows

Location, depth, stable outlet Hicks & Anthony (2001)

Sub-surface drainage Horizontal boring to reduce subsurface water content of earthflows and slumps

Location, depth below surface, number of drains, capacity of drains,


Surface recontouring Smoothing the land surface to enhance runoff, reduce ponding and soil water content

Location, topography, soil materials Hicks & Anthony (2001)

Gully erosion

Spaced planting Planting of spaced poles to stabilise the sides and floors of gullies. Tree species, spacing, extent of planting, pole protection



Afforestation Blanket planting of closely spaced trees to reduce soil water content, increase soil strength and reduce erosion

Planting pattern, tree spacing, species, location (extent) of planting, timing of planting of different parts of gullies

Hicks and Anthony (2001)

Graded banks Series of earth banks formed on long slopes to control surface runoff and divert to a stable outlet

Location, gradient, spacing, stable outlet Hicks and Anthony (2001)

Flumes and chutes Structures to discharge water across/away from gully heads or sidewalls to a stable outlet further down the gully. Mainly used to control migration of gully headcuts

Location, flow capacity, construction material and design


Pipe drop structures Pipes used to discharge water across from gully heads or sidewalls to the gully floor. Often used where flow is small

Location, flow capacity, construction material and design


Sink holes Constructed hole in porous soils used to control runoff and trap sediment. Typically used in highly porous volcanic soils

Location, spacing, size/volume, soil conditions, stable inlet, use of silt fence

Eyles (1993)

Diversion banks Earth bank used to divert runoff away from gully head to stable outlet

Catchment area and slope, design flow, gradient, soil materials


Grassed waterway Grassed waterway used to divert runoff away from gully head to stable outlet

Catchment area and slope, design flow, gradient, soil materials, shape, vegetation type


Drop structures Spillway constructed of concrete, geotextiles, rock, sheet piling used to safely convey runoff over gully head

Location, catchment area and slope, design flow, gradient, constructio9n material


Debris dams Structures constructed of a variety of materials (e.g. timber, pole and netting, brush, logs, iron) to control the grade, reduce channel slope and water velocity, trap debris and stabilise the gully floor

Location, catchment area and slope, gully activity, gradient, construction material, anchoring, height


Streambank erosion

Tree planting Planting of spaced poles or native vegetation to stabilise streambanks. Can include tying together of the vegetation to enhance survival

Location, tree species, spacing, extent of planting, pole protection, fencing


Vegetation lopping and layering

Felling of existing vegetation and layering to stabilise stream banks Location, extent, density, anchoring Gibbs (2007)


Engineering works (rip rap, groynes, gabion baskets, etc)

Rock and netting structures used to control severe bank erosion. Can be used in combination with biological control

Structure type, location, extent, shape Gibbs (2007)

Debris traps Low dams on the bed of small streams, constructed from netting and posts, to stabilise channels, reduce bank erosion and trap sediment

Location, spacing, height, construction materials Gibbs (2007)

Gravel extraction Removal of gravel to take pressure off outside of bends and reduce bank erosion

Amount of gravel removed Gibbs (2007)

Bank shaping Battering of streambanks to reduce potential for bank erosion Location, height of bank, shape of bank Gibbs (2007)

Channel diversion/realignment

Realignment of channel away from actively eroding banks to reduce bank erosion

Location, disturbance, construction method Gibbs (2007)

Riparian fencing Permanent fencing of streambanks to exclude grazing and reduce damage to stream banks by stock

Width of setback, riparian vegetation, type of fence

Hicks (1995)

Controlled grazing Temporary fencing of streambanks to allow infrequent grazing and reduce damage to stream banks by stock

Width of setback, riparian vegetation, frequency of grazing, type of stock

Hicks (1995)

6 Assessment of erosion and sediment control performance for urban earthworks and infrastructure

This section reviews studies that have assessed the performance of erosion and sediment control practices used in urban development and road construction projects, including:

erosion control practices for managing sediment generation at source, for instance, the use of mulches and geotextiles to protect areas of bare earth and stabilization by hydroseeding, and

sediment control practices for managing sediment discharges from a site, for instance, the use of detention ponds and temporary measures such as silt fences.

The international literature contains a range of studies investigating the performance of both groups of practices, most of which have been conducted in the USA. The research effort into erosion and sediment control in the US dates from the 1970s and includes many experimental studies as well as field evaluations conducted on operational construction sites. Since the late 1990s there has been a growing interest in establishing the performance of erosion and sediment control practices in New Zealand. In particular, a number of studies conducted in the Auckland region have provided a local source of information to complement the findings of overseas studies.

The following sections (6.1 to 6.4) review studies of the types mentioned above by grouping them as follows:

1. Erosion control practices

2. Silt fences and other temporary sediment control practices

3. Sediment retention ponds

4. Chemical treatment.

Each section begins by providing a brief description of examples of overseas studies, presented in tabular form, and a summary of the factors identified in these studies as being important in influencing the performance of the types of erosion and sediment control devices investigated. This is followed, in each section, by a more detailed description of relevant New Zealand studies. A summary of the key findings emerging from the review is presented in Section 6.5.

6.1 Erosion control practices

6.1.1 Overseas studies

Table 5 summarises examples of overseas studies that have assessed the performance of erosion control practices. The majority of these studies involved sampling simulated rainfall-runoff from experimental plots to assess the comparative performance of a range of mulching materials and erosion control geotextiles or blankets.

Table 5 Summary of findings of overseas studies to assess performance of erosion control practices used in urban development and road construction projects. Studies are listed in chronological order

Reference Location and Soil Type (if specified)

Scope and Methods Key Findings

Lemly (1982) North Carolina, USA: Red Clay

Comparison of single treatments of asphalt-tacked straw, jute netting, mulch blanket, wood chips and excelsior blanket with multiple treatments (chemical soil binder/erosion checks/ tacked straw/jute netting). Measurement of sediment concentrations in runoff from grass-seeded experimental plots of varying slopes (6–27°) during natural rainfall events.

Multiple treatments found to be more effective than single treatments, with reductions in sediment concentrations in the ranges of 88-90% and 23–78% respectively, relative to bare soil. Single treatments found to be relatively ineffective for removing finer sediment particles (diameter <0.04 mm) even after the establishment of grass cover, especially on slopes >9°. The authors noted that highway embankments can approach 23°, indicating single treatment methods are unsuitable for these situations. Best and worst performing single treatments were excelsior blanket and asphalt-tacked straw, respectively.

Jennings & Jarrett (1985)

Pennsylvania, USA: silty clay loam

Comparison of 11 mulch treatments including straw, bark, jute net, burlap, Hold/Gro® and rocks. Focus on inter-rill erosion of flat slopes, with simulated rainfall events applied to experimental plots of 1° slope.

Porous mulches (straw, bark, burlap and jute) which were capable of holding water resulted in lower runoff depths, suspended sediment concentrations and erosion rates than mulches which could not absorb water. The best performing material was straw (erosion rate 2% of that of untreated soil) and worst performing was rocks (erosion rate 37% of that of untreated soil).

Armstrong & Wall (1990)

Ontario, Canada: loam

Comparison of a mulch and erosion control blanket. Simulated rainfall applied to plots of 27° slope on highway embankment under construction.

The erosion control blanket was more effective than the mulch, with soil loss reductions in the ranges 98–99% and 71–95%, respectively, relative to bare soil. The performance of the mulch deteriorated over time, reflecting its displacement.

Harding (1990) Indiana, USA: silt loam

Comparison of 14 treatments including straw, coconut fibre, nylon monofilamental materials, curled wood fibres, jute net and interwoven paper and thread. Simulated rainfall events applied to experimental plots of 5° slope.

The majority of treatments reduced soil loss by >90%, relative to bare soil. In general, natural organic materials found to be more effective than nylon monofilamental materials, possibly due to lack of moisture holding capacity of the latter. Fibrous, interwoven materials (including synthetics) found to be more effective than conventional mulches such as straw which can be displaced.

Ziegler et al. (1997)

Hawaii, USA: clay Laboratory comparison of performance of 13 Rolled Erosion Control Systems (RECS) for reducing rainsplash detachment of soil particles. Included eight natural RECS (products made of coconut

Natural products tended to be more effective than synthetic products at reducing rainsplash (most achieving >97% reduction relative to bare soil). While some synthetic products were highly effective (best product achieving a 99% reduction), others were not (worst product achieving only



fibres, coir yarn, aspen excelsior, netted wood fibre and jute) and five synthetic RECS (products made of polypropylene fibres and PVC monofilaments).

a 22% reduction). Performance was positively related to the % cover and thickness of products. The ‘best’ group of products also yielded the least mass of fine sediments (<63 µm). The performance of all products decreased with time since onset of rainfall.

Benik et al. (2003)

Minnesota, USA Comparison of five erosion control treatments: bare soil, a disk–anchored straw mulch, a wood–fibre blanket, a straw/coconut blanket, and a bonded–fibre matrix product. Simulated rainfall applied to plots of 20° slope on highway embankment under construction.

Under conditions with little vegetation, the blanket and bonded-fibre treatments achieved an approximate 99% reduction in soil loss relative to bare earth, while straw–mulch achieved an approximate 90% reduction. However, erosion from bare and straw–mulch treatments was greatly reduced by vegetative growth that occurred between the spring and autumn sampling.

Lipscomb et al. (2006)

USA: sand, clay and loam soils

Comparison of blown straw and a netted erosion control blanket using application of simulated rainfall.

The erosion control blanket reduced soil loss by 98% on sand and loam soils and 80% on clay. Blown straw found to be ineffective for erosion control on steep slopes and clay soils.

Rickson (2006) Bedfordshire, UK: clay loam and sandy loam

Laboratory investigation of geotextile performance in relation to (a) rainsplash and (b) overland flow. Comparison of seven geotextiles including products made of jute, coir, wood shavings, polypropylene and nylon.

Rainsplash losses of soil found to be lowest with high % cover, thickness, ability to pond surface water, high water-holding capacity and rough texture. Overland flow erosion found to be reduced by high water-holding capacity (which influenced the degree of contact between geotextile and surface), and roughness (which influenced flow velocity).



The results indicate that erosion control can be highly effective, with some treatments found to reduce sediment loss from bare earth by as much as 99% (Armstrong & Wall 1990; Zielger et al. 1997; Benik et al. 2003). However, not all treatments have been found to be equally effective, with variations in performance found to be a function of the following characteristics:

Potential for displacement – fibrous, interwoven materials have been found to be more effective than loose mulches such as straw, which can become displaced by rainfall and runoff (Armstrong & Wall 1990; Harding 1990; Zielger et al. 1997; Benik et al. 2003; Lipscomb et al. 2006)

Percentage cover – materials with a higher percentage cover have been found to be more effective at reducing soil disturbance by rainsplash (Zielger et al. 1997; Rickson, 2006)

Thickness and associated water-holding capacity – thicker treatments with higher water-holding capacity perform better at reducing overland flow and associated soil loss (Jennings & Jarrett 1985; Harding 1990; Rickson 2006)

Flexibility and weight – flexible, heavier materials that have better contact with the underlying soil surface are better at ponding water, also at reducing overland flow and associated soil loss (Zielger et al. 1997; Rickson 2006)

Number of treatments – the use of a combination of treatments has been found to be more effective than single treatments, particularly in relation to the control of fine sediments (Lemly 1982)

Establishment of vegetation – variations in the performance of different materials have been found to become less marked following the establishment and growth of vegetation (Benik et al. 2003).

6.1.2 New Zealand studies

Auckland Regional Council, Landcare Research, and the University of Waikato collaborated on a study to assess the performance of straw mulch at an earthworks site in Albany, north of Auckland (ARC 2000). Eighteen experimental plots were established, each of which had one of the following five land covers: established grass, mulched topsoil, bare topsoil, mulched subsoil, and bare subsoil. Runoff and sediment discharge was measured during a series of storm events.

Over the duration of the study the sediment loads discharged from mulched topsoil and mulched subsoil plots were approximately 94% and 85% lower than those from bare topsoil and bare subsoil plots, respectively.1 The grassed plots also generated relatively low sediment loads, around 87% lower than bare topsoil plots. The highest loads were

1 Sediment load reductions estimated from graph presented in ARC (2000).



generated from the bare subsoil plots, being around double the load from the bare topsoil plots.

The study also found that, during high intensity rainfall events, applying mulch to bare subsoil was less effective than applying it to topsoil. In particular, mulching subsoils was found to make no significant difference to the discharge of clay and silt-sized particles during these types of storm event. The effect of topsoil in helping to lower sediment discharge (enhanced with the application of mulch) was linked to its higher organic matter content, allowing higher rates of infiltration when compared with subsoils.

6.2 Silt fences and other temporary sediment control practices


Table 6 summarises examples of overseas studies that have assessed the performance of silt fences and other temporary sediment control practices (excluding sediment retention ponds which are covered in Section 6.3). These studies include both field evaluations of performance on operational construction sites and experimental studies involving the simulation of rainfall-runoff on plots or in laboratory installations.

The results indicate that silt fences can be highly effective, with sediment removal efficiencies of up to 99% reported. Performance has been found to be predominantly a function of the settling of sediments in ponded water upstream of a fence rather than a result of filtering by the fence fabric (Kouwen 1980; Wyant 1981; Barrett et al. 1995a, 1998). The following characteristics have been found to be important in influencing the performance of silt fences:

Extent of upstream ponding – performance is likely to be higher where the geometry and slope of the site promote upstream ponding (Kouwen 1980; Barrett et al. 1995a, 1998)

Permeability – the pore size of the filter fabric influences the extent to which runoff is detained upstream of the fence and the extent to which sediment particles are trapped in the fabric. The trapping of sediment in the fabric matrix further reduces its permeability, contributing to the extended detention of incoming runoff (Kouwen 1980; Wyant 1981; Barrett et al. 1995a, 1998)

Soil particle size characteristics – finer soil particles tend to be responsible for the clogging of the filter fabric, reducing permeability and extending detention time (Barrett et al. 1995a, 1998). However, where sediment runoff is dominated by finer particles, removal efficiencies are likely to be relatively low because these finer sized particles settle out less readily than the coarser fractions (Line & White 2001); and

Maintenance – silt fences become less effective over time as the build-up of sediments and clogging of the fabric increases the likelihood of overtopping (Kouwen 1980; Wyant 1981; Barrett et al. 1995a, 1998).



The use of vegetated buffers below silt fences has also been found to enhance fence performance, providing an additional filtering mechanism to reduce suspended sediment concentrations (Pitt et al. 2007). In contrast, some rock structures have been found to be relatively ineffective (Barrett et al. 1995a). Better performance has been reported in situations where rock structures provide for detention, with one study reporting removal efficiencies of up to 69% (Line & White 2001).

Table 6 Summary of findings of overseas studies to assess performance of silt fences and other temporary sediment control practices used in urban development and road construction projects. Studies are listed in chronological order



Kouwen (1980) Ontario, Canada: silica sand

Laboratory comparison of performance of straw bales, 12 silt fence fabrics and gravel berms.

Deposition due to ponding, rather than direct filtration, was found to be the main mechanism for reducing sediment concentrations. As a result, less permeable fabrics were found to be more effective, some with sediment trapping efficiencies of 99%. However, the authors noted that materials need to be permeable enough to prevent overtopping. Well clamped straw bales were 95% efficient (subject to any overtopping), but loose bales were less efficient. Trapping efficiency decreases with time as sediment builds up behind the silt fence/barrier. Efficiency is related to slope through its influence on pond length.

Wyant (1981) Virginia, USA: clay, sand and silt soils

Development of standard methods for evaluating filter fabrics. Laboratory comparison of 15 different filter fabrics.

The majority of fabrics were found to have filtering efficiencies in excess of 90%, even for the clay soil. Only clay sized particles were discharged in effluent from the fabric, with the silt and sand sized particles readily settling in the ponded water upstream.

Barrett et al. (1995a, 1998)

Texas, USA: silty clay

Field evaluation of operational silt fences and a rock berm. Laboratory comparison of four filter fabrics and a rock berm.

In the field study silt fences had a median removal efficiency of 0% (range –61% to 54%: negative efficiencies indicating storms where TSS concentrations were higher downstream of the silt fence than upstream). This apparently poor performance reflected that upstream samples were collected from ponded water (post-settlement). Suspended sediments downstream of the silt fences were dominated by silts and clays. The rock berm was ineffective at reducing Total Suspended Solids concentrations. In the laboratory study fabrics had median removal efficiencies in the range 68–93%, with the highest performing having the lowest flow rate (hence a longer detention time). The performance of all fabrics was positively related to detention time. The rock berm was relatively ineffective, with a median removal efficiency of 7%.

Line & White (2001)

North Carolina, USA: sandy loam and clay loam.

Three-year field evaluation of three operational sediment traps: a rock check dam across a drainage ditch and two berms with rock lined outlets.

The sediment traps had removal efficiencies of 58-69%. Performance at one site was lower (49%) during a high intensity storm event which overtopped the trap. Traps were most effective at removing sand sized particles (68–91%) and least effective at removing clay sized particles (21–40%). With stabilization of the upstream catchment area, finer particles made up an increasing proportion of sediment runoff, resulting in a reduction in removal efficiencies over time.



Pitt et al. (2007)

Alabama, USA Field evaluation of the performance of silt fences, with and without downstream vegetated buffers, based on sampling during intense (≥25 mm/hr) storm events.

Silt fences reduced the mean particle count by around 50% compared with control sites with no fences. Runoff samples collected below a 5m vegetated buffer downstream of a silt fence had mean total solids concentrations about 20% lower than samples collected immediately below silt fences.




Auckland Regional Council commissioned a study to assess the performance of decanting earth bunds (DEBs; Babington and Associates 2004). The principal objective of the study was to assess the sediment removal efficiency of DEBs constructed in accordance with the design specified in ARC’s TP90 guidelines (ARC 1999). Initial trials were conducted at a site at the closed Greenmount Landfill in south Auckland and involved the use of a rainfall simulator to assess performance during specified design rainfall events. Subsequent monitoring was conducted of the performance of a DEB at a site in Tamaki Heights, also in south Auckland, during natural storm events. Both trials involved measurement of inflows and outflows and the collection of water samples either manually or by auto-sampler for the analysis of Total Suspended Solids concentrations.

During the simulated rainfall trials the performance of the DEB was found to vary in relation to event duration, antecedent soil moisture conditions and the extent of available storage, with sediment removal efficiencies reported in the range 47–75%. In general, these events had peak flows similar to the 1-year Average Recurrence Interval (ARI) rainfall event and a total storm volume similar to the 20-year ARI event.

Unfortunately, the trial to assess performance during natural rainfall events encountered a number of problems and was able to estimate sediment removal efficiencies from only four of the events monitored (results in the range 23–79%). However, it is important to note that the report on the study describes various causes of significant uncertainty, even in relation to the results from these four storm events (Babington & Associates 2004).

As part of the simulated rainfall trials a comparison was also made of the performance of the DEB with two alternative decant designs: a standard novacoil pipe upstand and a floating decant. The floating decant was reported to achieve a sediment removal efficiency 38% higher than the novacoil decant, based on performance associated with the descending limb of the outflow hydrograph. As a result, the study recommended that floating decants be adopted as standard practice on DEBs.

6.3 Sediment retention ponds


Table 7 summarises examples of overseas studies that have assessed the performance of sediment retention ponds. Note that these exclude studies involving the use of chemical treatments, which are covered in Section 6.4. Most of the studies summarised in Table 7 involved field evaluations of operational ponds, although some were based on experimental installations or computer modelling.

The performance of sediment retention ponds in these studies varied widely, with reported sediment removal efficiencies ranging from virtually zero to 99%. This wide range reflects not only variations in site conditions, pond sizing and design but also the fact that many of these studies involved assessing various experimental modifications aimed at improving



performance. Better performance has been found to be related to the following characteristics:

Extended detention time, promoting the settling of finer suspended sediments and increasing the proportion of influent water which is lost via infiltration through the base of the pond (Bidelspach et al. 2004)

Appropriate sizing (Barrett et al. 1995b), with performance increasing as a function of the pond surface area to peak discharge ratio (McBurnie et al. 1990)

Greater distance between the pond inlet and outlet, with performance increasing in response to a higher length to width ratio (Barrett et al. 1995b; Gharabaghi et al. 2006; Pitt et al. 2007)

The presence of permanent ponding, as opposed to fully-drained sediment traps (Pitt et al. 2007; McCaleb & McLaughlin 2008)

Pond designs which promote mixing and settling by avoiding dead zones and sheet flow through the upper part of the water column (Nighman & Harbor 1997; Pitt et al. 2007)

Protection and stabilization of approach channels, inlets and pond side walls to prevent erosion (Pitt et al. 2007; McCaleb & McLaughlin 2008)

The presence of forebays and baffles which reduce velocity and promote sediment settling, but which are not readily overtopped (Barrett et al. 1995b)

Outlets that discharge effluent from the pond water surface, rather than from the entire water column (Ward et al. 1979)

The use of outlet filters, such as gravel or expanded polystyrene envelopes fitted to outlet risers (Engle & Jarrett 1995).

Table 7 Summary of findings of overseas studies to assess performance of sediment retention ponds used in urban development and road construction projects. Studies are listed in chronological order



Engle & Jarrett (1995)

Pennsylvania, USA: loam

Laboratory evaluation of two types of filters (expanded polystyrene chips and gravel) fitted to sediment retention pond outlet risers.

The filters resulted in sediment removal efficiencies of 78-89%, varying with filter type and outlet configuration, compared with 60-71% with no filter. For a given outlet configuration, filters improved sediment removal by around 25%.

Nighman and Harbor (1997)

Ohio, USA: silty clay loam

Field evaluation of performance of a sediment retention pond with and without inlet baffles. Monitoring of five storm events before the fitting of a silt fence baffle, with four events monitored subsequently.

The pond had a mean sediment removal efficiency of 68% (range 56–79%) during the five pre-baffle storms. Performance was better for silts (90%) than clays (63%). Following fitting of the baffle, sediment removal efficiency dropped to virtually nothing during events in which the baffle was overtopped but improved slightly to 80% in events with no overtopping. The former result may have been due to a lack of vertical mixing, with sediment-laden inflow over the baffle generating sheet flow across the surface of the pond. The authors comment that, to avoid overtopping, baffles should be located so that their top is higher than the elevation of the outlet and should incorporate perforations.

Bidelspach et al. (2004)

Ohio, USA: silt loam

Evaluation of the influence of inflow-outflow delay time on sediment retention pond performance. Monitoring of experimental pond during 12 simulated 2-year 24 hour rainfall events. Delay in outflow controlled by valve on outlet skimmer.

The pond had a sediment removal efficiency of 90% with an unmodified outlet (for particles <45 µm). Efficiency increased to 92%, 94% and 98% with the modified outlet controlled to produce inflow-outflow delays of 0h (valve opened immediately), 12 h and 168h, respectively. The improved performance was largely due to increased infiltration through the bed of the pond.

Thaxton & McLaughlin (2005)

North Carolina, USA: sandy loam

Evaluation of pond baffles made of three materials: jute/coir, tree protection fencing and silt fencing. Velocity, TSS and turbidity measured in experimental pond subject to three different inflow rates.

Compared with the control (no baffles), all baffle materials were found to greatly reduce and diffuse flow but made very little difference to sediment retention. The jute/coir baffle was the most effective in trapping sediment with a removal rate of 98%, although this was less than 1% greater than the ‘no baffle’ situation. The authors noted that soil injection rates were very low compared to sediment inputs observed in the field. Baffles increased the capture of silts but not clay sized particles. Results indicate three regimes: (1) hydraulically unstable



regime, indicative of low baffle permeability that leads to overtopping and turbulent resuspension, (2) optimum diffusion regime, which diffuses incoming energy across the full width and depth of the pond without introducing turbulence, and (3) partial diffusion regime, which is less effective in reducing turbulence than the optimum.

Gharabaghi et al. (2006)

Toronto, Canada: silt/clay dominated

Field evaluation of performance of an operational sediment retention pond and comparison with results of earlier study, focusing on the influence of length to width ratio. Monitoring through 12 storm events. Supporting hydrodynamic/sediment transport modelling also conducted.

Very high influent TSS concentrations resulted in both ponds having sediment removal efficiencies of over 90%, However the pond with an 8:1 length to width ratio discharged effluent with markedly lower TSS concentrations (37 mg/L) than the 2:1 pond (177 mg/L). The latter pond failed to meet the local authority’s sediment control objectives.

McCaleb & McLaughlin (2008)

North Carolina, USA: sandy loam

Field evaluation of five sediment retention devices differing in terms of presence/absence of permanent ponding, outlet characteristics and presence of baffles. Monitoring during 11–35 storm events over periods of 5 to 13 months.

The most effective devices were those with permanent ponding: a sediment retention pond (efficiency of 99%, falling to 76% following blockage of the floating outlet) and a basin with rock outlet (73%). Three fully drained basins had sediment retention efficiencies of less than 40%. There was rapid erosion of the unprotected vertical inlet and side walls of these devices. Discharges, even from the most effective device, had high turbidity (>1000 NTU average) and TSS concentrations (>1000 mg/L average).




Albany field-based study

Winter (1998) conducted a detailed field evaluation of an operational sediment retention pond at Albany, on the northern fringe of Auckland. The catchment of the pond (area 4.6ha) had slopes in the range 3–9° and soils comprising loamy silts and silt loams (approximately 80% fine sand and silt sized particles). The pond was designed in accordance with Auckland Regional Council’s guidelines in place at the time, providing for storage of 200 m3 for each hectare of contributing catchment. Outflows during events up to the 5-year return period were discharged via a floating decant, with high flows discharged via a manhole riser. The average residence time of water in the pond was determined to be 9 hours.

Monitoring was conducted over two periods. Nine storm events were monitored during the closed winter season between May and October 1995, while two events were monitored during the working summer season in November and December 1995. Inflows and outflows were measured using pressure transducers and rated flumes while water samples were collected at flow-weighted intervals using an automatic sampler.

The overall sediment removal efficiency of the pond over the 11 storm events was calculated to be 90%, with 376 and 39 tonnes of sediment entering and leaving the pond, respectively. Efficiency during the individual storms monitored ranged from 70% to 99%, and was inversely related to peak inflows/outflows and event mean concentrations of suspended solids in outflow samples. During two-thirds of all storm events over the period of the study there was insufficient rainfall to generate outflow from the pond, and as a result 100% of influent sediment was retained during these smaller storm events.

The average sediment removal efficiency during the two summer storms (96%) was higher than during the nine winter events (87%). This was attributed to the influence of earthworking activities increasing the erodibility of soils during the open season, as shown by influent suspended solid concentrations and turbidity generally being higher during summer storms than winter storms.

The sediment removal efficiencies reported by Winter (1998) compare favourably with the results of overseas studies, indicating that, where designed and constructed in accordance with relevant guidelines, sediment retention ponds in New Zealand can be an effective practice for reducing loads of sediment discharged to receiving environments. However, in assessing performance, regard should also be given to effluent quality. While the pond resulted in a reduction in mean suspended sediment concentrations from around 10 000 g m–3 in influent samples to around 1000 g m–3 in effluent samples, these latter concentrations were estimated to be around 10 times higher than background concentrations in the receiving stream. The maximum concentration of suspended solids was around 20,000 g m–3, 200 times the background concentration. In general, higher suspended sediment (SS) concentrations and turbidity were associated with periods of events during which the capacity of the floating decant was exceeded, resulting in the overtopping of the manhole riser. There was evidence of reduced residence time during these periods, with concentrated flow paths observed between the pond inlet and outlet.



Two other field-based studies of the performance of sediment retention ponds have been conducted on operational construction sites north of Auckland area in recent years (Moores & Pattinson 2008; Larcombe 2009). However, because both these studies have focused on the performance of chemically treated ponds, discussion of their results is reserved to Section 6.4.

Auckland University experimental and modelling studies

Khan (2012) conducted a laboratory investigation of residence time and short-circuiting in sediment retention ponds using dye tracer studies and Particle Tracking Velocimetry (PTV) techniques. The study involved constructing a 1:10 scale physical model of an operational sediment retention pond located at the ALPURT B2 motorway construction site, north of Auckland. The pond had an approximate length to width ratio of 3:1 with 2:1 sloping sides. A deflector island and a series of baffles were constructed and placed in different arrangements to investigate their effectiveness for reducing short-circuiting and increasing residence time2.

Counter to the author’s expectations, the incorporation of either the deflector island or 1-2 baffles was found to stimulate short-circuiting and reduce residence time. This effect was attributed to the relatively long, narrow shape of the pond and the fact that a substantial proportion of the bed of the pond comprised the sloping side wall. The island and baffles diverted flow to areas of shallow water overlying the sloping sides, resulting in reduced mixing. However, the introduction of additional baffles was found to increase residence time. The author notes that the unexpected findings of the study may not necessarily be transferable to other pond designs, for instance where the sloping side walls occupy a less substantial proportion of the pond bed.

The same physical model was used by Farjood et al. (2015) to investigate a range of hydraulic performance indices for representing residence time in sediment retention ponds (SRP). Based on the results of dye traced studies they found that a new index, τ5, was the best for representing the degree of short circuiting in the SRP. This index represents the normalised time taken for 5% of influent sediment particles (represented by a dye tracer in the study) to exit the SRP. The lower the value of the τ5 index the greater the extent of short-circuiting.

6.4 Chemical treatment

Many of the studies described in the previous sections have found that erosion and sediment control practices can be highly effective in reducing sediment generation and discharge. However, some have also shown that despite removal efficiencies in excess of 90%, turbidity and concentrations of suspended sediments in effluent discharged from

2 The study also investigated the influence of floating treatment wetlands (FTWs) on pond hydraulics, finding that FTWs increase residence time. However, this aspect of the study is of more relevance to the performance of stormwater treatment ponds rather than sediment retention ponds.



construction sites can still be markedly higher than environmental guidelines and/or background concentrations in receiving aquatic environments (Winter 1998; Garabaghi et al. 2006; McCaleb & McLaughlin 2008). In particular, erosion and sediment control practices have generally been found to be less effective for the retention of fine soil particles, especially clays but also silts, than for coarse sand sized particles (Wyant 1981; Nighman & Harbor 1987; Line & White 2001; Lipscomb et al. 2006).

In order to address this issue, in recent years enhanced erosion and sediment control practices that utilise chemical treatments have become increasingly common. These approaches aim to improve the quality of effluent discharged from earthworks sites by either:

binding sediments at source, as an enhanced erosion control practice, or

flocculating suspended fine sediments, so as to enhance their capture and settlement in sediment retention devices.

The following sections summarise the findings of studies conducted to assess the performance of chemical treatments, relative to ‘traditional’ forms of erosion and sediment control.


Table 8 summarises examples of overseas studies that have assessed the performance of chemical treatments. Most of these studies involved investigating the effectiveness of polyacrylamide (PAM) to improve the performance of erosion control practices, although the use of PAM for sediment control has also been investigated (McLaughlin et al. 2009a, b). A study to evaluate the use of one other chemical, calcium sulphate, is also reported (Przepiora et al. 1998).

The findings of these studies include, for erosion control, that:

chemical treatment does not markedly improve on the performance of traditional physical practices such as mulching and grass seeding (Hayes et al. 2005; McLaughlin & Brown 2006; Soupir et al. 2004), and

the effectiveness of chemical treatment reduces with time since application, with treatment unlikely to remain effective over a full construction season (Soupir et al. 2004).

For sediment control:

Chemical treatment can improve on the performance of physical devices, for instance in markedly lowering turbidity (Faucette et al. 2009; McLaughlin et al. 2009a, b), and

Performance is more effective as detention time increases, allowing more time for flocculation and settlement processes to act (Przepiora et al. 1998).



6.4.2 New Zealand Studies

A number of Auckland-based studies have investigated the performance of chemical treatment for improving the effectiveness of sediment control practices.

Initial trials at ALPURT and Greenhithe

The earliest of these studies investigated the performance and practicality of a range of commercially-available flocculants (Auckland Regional Council 2004). Bench testing of PAM, the chemical of choice in the US studies described above, gave encouraging results.

Table 8 Summary of findings of overseas studies to assess performance of chemical treatment of sediment control measures used in urban development and road construction projects. Studies are listed in chronological order



Przepiora et al. (1998)

North Carolina, USA: silt loam and sandy loam

Evaluation of performance of CaSO4 (moulding plaster) as a chemical treatment of SRPs. Two operational ponds dosed with CaSO4 following 14 storm events. Turbidity of discharges monitored for 50-70 hours and compared with control ponds.

Turbidity in treated ponds was <50 NTU, compared with 100-1650 NTU in the untreated control ponds. Turbidity decreased over time following dosing, with time taken for turbidity to fall below 100 NTU found to be inversely proportional to the dosage of CaSO4. Turbidity in treated ponds fell to less than 10% of initial values. Residual SO4 concentrations exceed 250 mg/L (the local water quality standard) but the authors calculated that a dose resulting in SO4 concentrations of less than 250 mg/L would be effective in reducing turbidity below 100 NTU.

Roa et al. (1998)

Wisconsin, USA: silt loam.

Evaluation of performance of PAM for erosion control. Varying doses of dry and liquid applications applied to dry and wet soil, with and without mulch/seeding. Simulated rainfall applied to experimental plots with 6° slopes.

Plots treated with PAM had sediment yields in the range 93% (liquid PAM with mulch) to 77% (liquid PAM applied to wet soil) lower than a control plot.

Soupir et al. (2004)

Virginia, USA: clay and silt dominated soils

Evaluation of performance of PAM for erosion control. Varying doses of dry and liquid applications compared with hydroseeding and straw mulch. Two groups of simulated rainfall events a month apart applied to experimental plots with 3° slopes.

Overall TSS loads from the PAM treatments were 19-33% lower than from the control plot. Load reductions were higher in the first simulation (28-82%) than in the second simulation (11-25%), one month after application. Dry PAM performed better than the liquid applications. However, all forms of the PAM treatment were less effective than straw mulch and hydroseeding, especially in the second simulation. By this time it was observed that there was very little PAM remaining on the plots.

Hayes et al. (2005)

North Carolina, USA: sandy loam and sandy clay loam

Evaluation of performance of PAM for erosion control. Liquid PAM applied at the recommended rate and at half the recommended rate, with and without grass seeding and mulching on 11° and 27° slope plots. Monitoring during 6-8 storm events at each of three sites.

On the 27° slope the application of PAM on its own was ineffective, while the seed/mulch treatment on its own reduced sediment losses by up to 83%. The addition of PAM to the seed/mulch treatment made no significant difference. On the 11° slope the application of PAM resulted in some improvement to average turbidity over all storms monitored but a statistically significant improvement only occurred during one storm event. The findings suggest the use of PAM is of limited benefit for erosion control on the types of soils and slopes present in the study region.



McLaughlin & Brown (2006)

North Carolina, USA: silt loam

Evaluation of performance of PAM for vegetation establishment and erosion control on its own and in combination with four types of erosion control: straw, straw erosion control blanket, wood fibre and mechanically bonded fibre matrix. Performance monitored on a 2° slope during 5 storm events over 36 days and 6° and 11° slopes with simulated rainfall.

Straw, straw erosion control blanket and mechanically bonded fibre matrix significantly reduced sediment loss and turbidity. The addition of PAM to these covers did not consistently improve runoff quality. PAM did reduce turbidity in some instances but this was assessed as reflecting flocculation in the runoff collection system, rather than as a result of reduced erosion.

Faucette et al. (2009)

Georgia, USA: sandy clay loam.

Field evaluation of temporary perimeter controls used in high intensity/duration single storm events. Comparison of straw bales, mulch filter berms and compost filter with and without the addition of PAM. Synthetic 24 hour, 5-year return period rainfall applied to experimental plots of 6° slope.

The compost filter socks (with and without PAM) were the most effective treatments, achieving a TSS load reduction in the range 84-89%. In comparison, straw bales and mulch filters achieved reductions of 51–54%. These differences reflected the lower flow rates and hence greater detention and infiltration achieved by the compost filter socks. Turbidity from the PAM treated filter sock was significantly lower than from the untreated filter sock.

McLaughlin et al. (2009a)

North Carolina, USA: sandy loam, sandy-clay loam

Field evaluation of three types of sediment retention devices on an operational highway construction site: (1) standard sediment traps with rock outlets; (2) sediment traps with the addition of baffles, a forebay and PAM blocks and powder; and (3) sediment retention ponds with floating outlets, baffles, a forebay and PAM blocks and powder. Monitoring of effluent discharges during up to 31 storm events over a three year period. Limited sampling of influent because high sediment loads often clogged or buried sampler intakes.

Discharges from devices with PAM, forebays, porous baffles and stabilization had average turbidity and TSS concentrations approximately four times lower than from those without these modifications. Differences in peak turbidity and TSS concentrations were more marked. Discharges from the devices with surface outlets were generally poorer than from the modified sediment traps, but this was attributed to differences in storm event characteristics. However, when one of the sediment retention ponds was performing optimally it reduced turbidity by up to 99%.

McLaughlin et al. (2009b)

North Carolina, USA:

Comparison of performance of standard sediment traps and fibre check dams (FCDs, straw wattles and coir logs), the latter with and without the addition of dry PAM. Monitoring of sites on highway construction projects during 20-23 rainfall events over 9 months.

Sediment losses from the FCD sites were around 99% lower than from the standard sediment traps and were lower with PAM than without PAM. Turbidity at FCD sites with PAM were an order of magnitude lower than at FCD sites without PAM and two orders of magnitude lower than at standard sediment trap sites. The addition of PAM reduced turbidity to below 10 NTU (the local regulatory guideline) during some storm events.



However, a number of practical problems were identified in relation to the need to pre-dilute liquid PAM and it was excluded from further trials. In contrast, a solid form of PAM, Magnasol® Floc Blocks, also performed well in bench testing and this product was selected for field testing. In addition, the bench tests found that two aluminium-based liquid chemicals, alum and Polyaluminium Chloride (PAC) had potential, and these were also taken forward into the field testing phase of the study.

Field testing was undertaken at two locations: first, at a number of ponds located on the ALPURT motorway construction project north of Auckland and, second, at a residential subdivision at Greenhithe on the North Shore. While initial trials at ALPURT using alum indicated it was effective for reducing SS concentrations, greater emphasis was placed on the performance of PAC because this was found to result in less marked reductions in pH. Reductions in the pH of runoff discharged to receiving streams are potentially problematic because the toxicity of residual dissolved aluminium increases with acidity.

The performance of liquid PAC was investigated by comparing the characteristics of influent and effluent samples collected at 21 different ponds over summer 1998/99 and winter 1999. Differences in suspended sediment concentrations in the range of 90-99% were reported for well-designed ponds receiving PAC treatment3. Less marked differences in influent and effluent concentrations were associated with ponds that had poorly performing decanting devices, multiple inflow points, high inflow energy or poor separation of inlets and outlets. However, even the smallest difference in SS concentrations (62%) was well in excess of the results reported for an untreated pond (4–12%). While dissolved aluminium concentrations were generally found to be below acute USEPA criterion values, the occasional elevated Al concentration and low pH value indicated situations in which overdosing with PAC had occurred.

Preliminary field trials of Magnasol® Floc Blocks were also conducted at ALPURT, followed by more intensive testing at the Greenhithe site, to develop practical guidelines for its use on sites where liquid dosing systems may be unfeasible. The initial trials found that Floc Blocks became readily buried by the high bed load of untreated runoff. The investigation therefore moved on to assess the performance of Floc Blocks to treat runoff flowing through a flume following initial treatment by a forebay. Floc Blocks were found to be effective in this context, subject to limits on inflow rates and SS concentrations. Notably, the study reported that treatment with Floc Blocks resulted in the more rapid settlement of suspended solids than aluminium coagulant treatments because the flocs that form are large, dense, and fast settling. As a consequence, the study concluded that sediment retention ponds utilising a Floc Block system do not need to be as large as those for an aluminium coagulant treatment system.

3 It should be noted that the sampling methodology adopted in these trials is not described, but from the data tables included in the report it appears performance was assessed on the basis of a single influent and single effluent sample collected during each storm event (ARC 2004). In view of the likely variations in influent and effluent quality during any given storm event, the reported percentage differences in SS concentrations are best treated as being broadly indicative of performance.



However, the further investigation at the Greenhithe site again encountered the problem of burial of the Floc Blocks during high intensity storms, despite pre-treatment of runoff. Problems were also reported in relation to the degradation of the Blocks during periods of dry weather and conversely, their softening when submerged for prolonged periods. Despite these problems, the results of sampling did show that, during some storms, effluent discharged from a treated pond had substantially lower SS concentrations than that discharged from untreated ponds on the same construction site. Based on the results of the Greenhithe trial, the study reported that the ideal operating condition for Floc Block involves exposing them to water only during the storm flow period, with the Block being out of the water when there is no storm flow. It also recommended that Floc Block treatment systems should be designed to treat a specified maximum flow, and to use a flow balancing system ahead of the Floc Block treatment system to ensure that the maximum design flow is not exceeded and to remove bedload prior to treatment of the runoff.

ALPURT B2 study

Reflecting the apparent effectiveness and growing operational use of the PAC dosing of sediment retention ponds, ARC commissioned a further, more intensive field evaluation of this form of chemical treatment (Moores & Pattinson 2008). A field programme comprising hydrological monitoring and the collection of water samples was implemented at the ALPURT B2 motorway construction site near Orewa, north of Auckland. A rainfall gauge, weirs, water level recorders and automatic water samplers were installed at a pair of identically-designed TP90 compliant (ARC 1999) ponds that each received approximately half of the run-off from an earthworks area of 4.4 hectares. The inflow to one pond was treated with PAC by a rainfall activated dosing system whilst the inflow to the other pond was not treated.

Water samples were obtained from seven storm events over the period March to December 2007. The results indicated that the addition of PAC was an effective method of improving the sediment removal efficiency of sediment retention ponds. The estimated total sediment load discharged from the treated pond was a third of that from the untreated pond. The treated pond was found to achieve an estimated sediment removal efficiency of at least 68%, while the untreated pond performed well below this level at around 30%.

However, there were substantial variations in the effectiveness of PAC treatment both during and between storm events. Efficiencies of over 90 per cent were achieved during relatively small events (characterised by rainfall totals in the range 10.5 to 28 mm) during which the efficiency of the untreated pond was also relatively high. The improvement in pond performance as a result of PAC treatment was most marked during events, or periods of events, with relatively high rainfall depths and intensities. While sediment removal efficiency in the treated pond during the larger events (characterised by rainfall totals in the range 48 to 195 mm) was lower than during the smaller events, the additional sediment load retained as a result of PAC treatment during these events was substantial. During a single large event in March 2007 the sediment load discharged to the receiving environment from the treated pond was over four tonnes less than that from the untreated pond. These results indicated that the greatest gains from PAC treatment are achieved through dosing of



ponds during relatively large storm events when the performance of sediment retention ponds without PAC treatment is relatively poor.

The results also indicated that PAC treatment could be of benefit during the winter, when the performance of untreated SRPs tends to be relatively poor because of high pond and soil water levels reducing residence time. The total sediment load retained in the treated pond during three winter events was reported to be over a tonne (82%) greater than that in the untreated pond despite this being the closed season for earthworks.

Despite the generally positive results of the study, there were also occasions during which the PAC treatment appeared to significantly underperform. This was attributed to possible deficiencies in the PAC dosing system resulting in under-dosing or delay in the delivery of PAC. During one major storm (assessed as a 1 in 13 year ARI rainfall event), a sudden deterioration in the quality of effluent from the treated pond appeared to indicate the cessation of treatment: in other words, the dosing system had run out of chemicals. Recognising, however, that when well-managed, PAC treatment was found to markedly improve the performance of the SRP, the study recommended regular inspection be undertaken of the dosing system.

Effluent water samples were also analysed for dissolved aluminium concentrations in order to further investigate the extent to which residual aluminium from PAC treatment may represent a problem for receiving environments. Median dissolved aluminium concentrations in samples from the two pond outlets were found to be similar at 0.047 g m–

3 and 0.044 g m–3 and less than the ANZECC (2000) trigger value for a 95 per cent level of protection.

However, the ANZECC (2000) 95 per cent trigger value was exceeded by maximum concentrations in 13 samples collected at the treated pond outlet and in nine samples from the untreated pond outlet. The level of exceedance was greater in the treated samples than the untreated samples and generally coincided with periods of increasing flows during the early- to mid-part of each event.

Silverdale North monitoring

Two ponds treated with PAC were monitored as part of the conditions of resource consents associated with the Silverdale North development, again near Orewa, north of Auckland (Larcombe 2009). Automatic water samples of pond inflows and outflows were collected during three and four storm events, respectively, at the two ponds over the 2008/9 earthworks season. Average SS concentrations in effluent samples were in the range 36–180 g m–3, compared with 899–42 691 g m–3 in influent samples.

The overall sediment removal efficiency of both ponds was approximately 99%, with efficiencies in all but one storm event in the range 94–99.8%. The sediment removal efficiency during the remaining event was relatively low at 80%, which was attributed to this being a low intensity rainfall event. Runoff from the site continued to enter the pond well after the rainfall (and hence PAC dosing of the pond) had ceased.



Long Bay development monitoring

The Long Bay development lies just inside Auckland’s northern urban boundary and is located adjacent to a regional park and a sensitive east coast receiving environment. Consent for the development, which is currently in progress, was granted subject to a range of monitoring conditions relating to the performance of sediment control measures and the state of freshwater and marine ecosystems downstream of the development (Ridley & De Luca 2015). This includes the collection of water samples from sediment control device inflows and outflows and their analysis for TSS concentrations. Samples of outflows from selected ponds are collected by automatic water samplers while samples of inflows and outflows from additional sediment ponds and decanting earth bunds are collected manually. All the monitored devices are subject to chemical treatment by liquid PAC.

The results of the monitoring are submitted quarterly to Auckland Council while, in addition, separate reports of the results of sampling are submitted for each rainfall event in which more than 25 mm falls over 24 hours. The reporting of results focuses primarily on the estimation of the total load of sediment discharged from monitored sediment control devices and the consequent effects (if any) of discharges on receiving environments, rather than the estimation of sediment removal efficiencies. However, given that some sampling of inflows is undertaken, the monitoring results do provide for an assessment of performance based on comparison of influent and effluent TSS concentrations.

Table 9 summarises the results of monitoring as provided by the project’s sediment control consultant for the period September 2013 to May 20154. The table presents summary statistics with and without the inclusion of a small number of sampling results from an event on 14 December 2014 during which elevated TSS concentrations were measured and, as a result, mitigation actions were taken to address underperformance of the monitored sediment control devices.

Table 9 Summary of TSS concentrations in samples collected from inflows to and outflows from sediment control devices at Long Bay Development, September 2013 to May 2015. Values in brackets reflect the exclusion of data relating to the event of 14 December 2014

Manual Automatic

Inflow Outflow Outflow

Number of samples 24 41 (38)

272 (264)

TSS Concentrations (g m-3)

Range 70–30 000 4–3600

(4–140) 1–4100 (1–195)

Median 2100 28

(27) 21

(19)

95th %ile 26 200 470

(115) 76

(59)

4 G. Ridley, pers. comm., 13 January 2016.



Based on the results of manual sampling, median and 95th percentile TSS concentrations in outflow samples were 98–99% lower than concentrations in inflow samples over the period of monitoring. The quality of treated effluent indicated by the manual sampling programme is supported by the results for the much larger number of outflow samples collected by automatic samplers, with the median and 95th percentile TSS concentrations in automatically collected samples being slightly lower than in the manually collected samples.

The results of manually collected water samples from the Vaughan Stream, upstream of the development, are also relevant for an assessment of the performance of sediment control measures. While based on a relatively small number of samples (n=11), the median and 95th percentile TSS concentrations in upstream samples collected from the Vaughan Stream during the storm events monitored over the September 2013 to May 2015 period were 18 g m–3 and 79 g m–3, respectively. This compares with median and 95th percentile TSS concentrations of 19 g m–3 and 59 g m–3, respectively, in pond outflow auto-samples (excluding the samples collected during the event of 14 December 2014). This comparison suggests that, with the exception of the event of 14 December 2014, sediment concentrations in effluent discharged from the monitored sediment control devices is generally consistent with that in receiving streams, providing further evidence of the extent to which chemical treatment can enhance the performance of sediment control devices.

6.5 Summary

The studies reviewed in the preceding sections indicate that erosion and sediment control practices can be highly effective in reducing the generation of sediment and its discharge from construction sites. Many of these studies have reported order of magnitude (or even two orders of magnitude) reductions in sediment loads and concentrations. Key factors that have been identified as promoting the effectiveness of erosion and sediment control include:

In relation to erosion control practices:

the use of fibrous, interwoven materials as opposed to loose mulches

the use of material with a high percentage cover

the use of relatively thick materials with a high water-holding capacity

the use of flexible, relatively heavy materials which follow the ground contour

using a number of treatments in combination

the re-establishment of vegetation, and

the use of mulches in combination with topsoil, as opposed to subsoil.

In relation to silt fences and other temporary sediment control practices:

siting measures in locations which promote extensive upstream ponding

the use of materials of a pore size which detains flows and promotes clogging of the filter fabric



regular maintenance to address the build-up of captured sediments and clogging of filter fabric

the use of vegetated filter strips downstream of filter fences, and

the use of floating decants rather than novacoil upstands in decanting earth bunds.

In relation to sediment retention ponds:

designs which provide for extended detention, not only to promote the settlement of sediments but also infiltration through the pond bed

a relatively high length to width ratio to maximise the distance between the pond inlet and outlet

the presence of permanent ponding, as opposed to full drainage

designs which promote mixing by avoiding dead zones and sheet flow

protection and stabilization of approach channels, inlets and pond side walls to prevent erosion

the presence of forebays and baffles which reduce velocity and promote sediment settling, but which are not readily overtopped

outlets which discharge effluent from the pond water surface, rather than from the entire water column, and

the use of outlet filters, such as gravel or expanded polystyrene envelopes fitted to outlet risers.

Some of the studies reviewed have also shown that despite removal efficiencies in excess of 90%, turbidity and concentrations of suspended sediments in effluent discharged from construction sites can still be markedly higher than environmental guidelines and/or background concentrations in receiving aquatic environments. In particular, erosion and sediment control practices have generally been found to be less effective for the retention of fine soil particles, especially clays but also silts, than for coarse sand sized particles.

The use of chemical treatments aimed at addressing these issues has become increasingly common in recent years. A number of studies have investigated the performance of these approaches, finding that:

In relation to erosion control practices:

chemical treatment does not markedly improve on the performance of traditional physical practices such as mulching and grass seeding, and

the effectiveness of chemical treatment reduces with time since application.

In relation to sediment control practices:

Chemical treatment can improve on the performance of physical devices, for instance markedly lowering turbidity and TSS concentrations



Performance is more effective as detention time increases, allowing more time for flocculation and settlement processes to act

The use of solid forms of chemical treatments can be subject to a number of problems, such as burial and degradation, meaning that runoff control and pre-treatment by sediment forebays may be required

Where well-managed, liquid PAC treatment used in New Zealand performs well

PAC treatment makes a greater difference to effluent quality discharged during larger events, when the performance of non-treated ponds is relatively poor

PAC treatment can also be of benefit during the winter, when the performance of untreated ponds tends to be relatively poor

Concentrations of residual aluminium from PAC treatment are generally, but not always, below relevant water quality guidelines. Careful management is required to avoid the over-dosing of ponds and potential discharge of elevated concentrations of dissolved aluminium.

7 Assessment of erosion and sediment control performance for horticulture and arable cropping

ESC measures to mitigate sheet and rill erosion from cropland follow similar principles to ESC for urban and earthworks construction sites, namely:

runoff control to reduce water entering paddocks (interception drains, culverts) and limit runoff generation within paddocks

erosion control measures to reduce sediment generation (wheel track ripping and diking, cover crops, cultivation management, strip cropping)

sediment control measures to reduce sediment discharge from paddocks (sediment retention ponds, silt fences).

A number of these practices were trialled during the Franklin Sustainability Project (Franklin Sustainability Project 2000) and the “Holding it together” (HIT) project (Johnstone et al. 2011), although limited quantitative data were produced. Compacted areas in fields (wheel tracks and headlands) are major sources of runoff and erosion and a number of practices are targeted at reducing runoff and erosion from these areas.

Practices used to control wind erosion include limiting the time the soil is bare (maintain a vegetative cover or surface residue) and has a dry surface (irrigate, use of mulches, terracing), and reducing wind velocity through increased surface roughness (stubble mulching, ridge-till, coarse seedbeds) or windbreaks and strip cropping.

Hicks (1995) included the following estimates of the performance of a variety of ESC practices on cropland mostly derived from North American literature:

minimum tillage reduces soil loss by 26–52% compared with conventional tillage

stubble mulching reduces soil loss by 50–80% compared with bare soil, depending on the amount of stubble cover



contour cultivation reduces soil loss by 30–90+% compared with cultivation up-and-down slope

grassed waterways reduce soil loss by 50+% compared with bare soil

7.1 Wheel track ripping

Compacted wheel tracks are recognised as major sources of runoff and erosion within crops that are grown in beds, often oriented up- and down-slope (Basher & Ross 2001). Many vegetable crops, including onions, cabbages, lettuces, broccoli, squash, leeks, and carrots, are grown in beds, and the wheel tracks between the beds have reduced infiltration and hydraulic conductivity and are the key zones for initiation of surface runoff and erosion. Ripping of wheel tracks with a tined implement down to a depth of 30–40 cm can break up compacted surface and subsurface soil and hence reduce runoff and erosion. Ripping leaves a semi-continuous opening in the centre of the wheel track and a disrupted, uneven surface in the remainder of the wheel track that enhances infiltration and hydraulic conductivity.

Basher and Ross (2001) report the results of trials carried out at Pukekohe in which rates of erosion and infiltration rates were compared between ripped and unripped wheel tracks in onion crops. The experiments were on strongly structured clay loam soils, in paddocks that

were gently sloping (2–9) with long slope lengths (120–170 m), over the winter period. In the first trial rainfall was below average with few significant rainfall events and no erosion occurred. In the second trial the erosion rate from the unripped treatment was 21 t ha–1 compared with1 t ha–1 for the ripped treatment. Soil was mobilised along the edge and base of the wheel tracks, with no evidence of erosion of the onion beds. Measurement of infiltration rates showed strong contrasts between the ripped and unripped wheel tracks, and the onion beds (0.5, 60 312, 411 mm hr–1 respectively at the start of the trial, and 77, 8582, 907 mm hr-1 respectively at the start of the trial). The infiltration rates measured in the unripped compacted wheel tracks during the early part of the trial would commonly be exceeded by short-term rainfall intensities. The increase of infiltration rate in unripped wheel tracks occurred after the soil surface began to dry out, and frequent wetting and drying cycles caused the compacted surface soil to crack. Most erosion occurred in the winter/early spring period when storm frequency and rainfall intensity was highest, and infiltration rates in the uncultivated wheel tracks lowest. The trials showed the cultivation of wheel tracks is a simple and effective practice to increase infiltration of rainfall and reduce erosion rates on clay-rich, strongly structured soils.

Basher et al. (2004) found similar results on young volcanic soils at Ohakune under intensive vegetable cropping where compacted wheel tracks had low infiltration rates (4 mm hr–1) compared with carrot beds (853 mm hr–1). However, it was suggested ripping might exacerbate soil loss from wheel tracks on light weakly structured volcanic soils.

In Britain and Europe compacted wheel tracks with reduced infiltration rate have also been recognised as key factors controlling runoff and sediment generation and can account for up to 80% of the runoff and erosion from cropped fields (e.g. Fullen 1985; Reed, 1986; Robinson & Nagizadeh 1992; Ludwig et al., 1995). However, few studies have reported on the effectiveness of wheel track management for reducing erosion. Deasy et al. (2010) and



Bailey et al. (2013) report results from a trial on clayey, silty and sandy soils in England under cereal crops where ‘tramline disruption’ with a single tine implement was used to manage compaction in wheel tracks. They found losses of suspended sediment were 4–230 times higher from wheel tracks than from vegetated beds, with the highest losses on sandy and silty soils and smallest losses on clayey soils. Tramline disruption reduced suspended sediment generation (compared to untreated tramlines) by 98–99% on silty soils and 75–96% on sandy soils.

7.2 Wheel track diking

Diking creates small indents in wheel tracks, which increase water retention time and improve infiltration. It is a practice that has been used widely in Australia and the USA to increase efficiency of water use during rainfall, especially in dry areas, since the practise increases surface detention time, but also has benefits for reducing soil loss. Most literature reports its benefits for enhancing moisture retention and reducing runoff rather than impacts on sediment generation. There are a wide variety of styles of diking that can be implemented using different implements; some of these are illustrated in Johnstone et al. (2011).

Demonstrations in the HIT project showed the potential for wheel track diking to reduce surface runoff and sediment loss. Thirteen demonstration trials were installed at Pukekohe and Hawke’s Bay and vegetable growers observed that diking reduced surface runoff and in-field ponding on a range of soil types, crops and slope steepness. In a Hawke’s Bay trial the volume of winter runoff from a trial bed that had been diked was reduced by over 90% compared with an un-diked control. Johnstone et al. (2011) state there were clear effects on the amount of sediment lost in the runoff but present no data other than a visual comparison of runoff.

A number of studies in the international literature report results of studies relevant to the performance of diking (variously referred to as micro-basin tillage, furrow diking or basin tillage), and all show that it is very effective in reducing runoff and soil loss. In China Xiao et al. (2012) using rainfall simulation on 11-m long plots on loessial black soil found diking reduced runoff and soil loss by 70% and 62% respectively. Similarly, Sui et al. (2016), in a field study with 25-m long plots, found diking reduced runoff and soil loss by 63% and 96% respectively. On sandy soils in the United States using rainfall simulation on 3-m-long plots diking reduced runoff by 67% and erosion by 71% (Truman & Nuti 2009). In Israel, on a crusting loess soil, Rawitz et al. (1983) report results of studies of 25-m-long field-based plots on diking in the fallow period before cotton was planted. They found that erosion was reduced by 67–95% where diking was used.

7.3 Cover crops

Cover crops, which are grown to be incorporated back into the soil, are used widely in the New Zealand vegetable industry but there is not much data on their performance in reducing sediment loss. Cover crops can be used between successive crops to avoid leaving a bare soil surface, or within wheel tracks. Cover cropping is particularly appropriate for



vegetable production systems, which typically leave few surface residues for erosion protection, return little organic matter to soils, and are often associated with soil structural problems that can limit productivity (e.g. Haynes & Tregurtha 1999; Basher & Ross 2002).

The performance of cover crops (oats, sorghum, mustard, phacelia, annual ryegrass) planted between successive crops at Pukekohe as part of the Franklin Sustainability project is summarised in Hicks (2006). The trials demonstrated that cover crops improved aggregate stability (i.e. on average aggregates were larger) compared with values before they were planted and compared to leaving the soil fallow. By inference this should reduce soil erodibility and erosion.

In the HIT project several cover crop trials were reported (Johnstone et al. 2011):

Comparison of the performance of a small number of cover crops (annual ryegrass, two types of mustard) with fallow soil over winter in the Horowhenua on a clay loam soil. The cover crops improved aggregate stability, proportion of larger aggregates and infiltration rates compared to fallow.

Comparison of the performance of 14 different cover crops (including annual ryegrass, oats, lupin, mustard, vetch, peas, phacelia, clover) with fallow soil over winter in the Horowhenua on a silt loam soil. Results were more variable in this trial but in general terms it again suggested benefits from cover crops of improved aggregate stability and proportion of larger aggregates.

A trial at Pukekohe comparing soil loss from a wheat cover crop with bare fallow soil shaped to form potato beds. Broadcasting wheat across the fallow soil reduced soil loss by approximately 38% between May and June, and by approximately 26% between June and July. The difference was rather low and not statistically significant because of a high level of spatial variability in runoff patterns but visually there was a clear benefit from the wheat treatment in most rows. The smallest losses were where wheat had successfully established in the wheel tracks as well as on top of the beds.

A trial at Pukekohe to assess the potential of wheel track-planted oats to reduce soil erosion in a potato crop. A silt fence was installed to trap sediment at the bottom of the rows and compare tracks with and without oats. No soil loss results were reported, which presumably means there was no soil movement over the course of the trial.

There have been no detailed studies of the effects of cover crops on erosion rates in New Zealand. However, plot studies at Pukekohe illustrate the mitigating effect of vegetation on erosion rates. At a trial site on clay loam, strongly structured soils, Basher et al. (1997) provide erosion rate data from plots of bare, cultivated soil and grass plots (Table 10). When the plots were bare soil loss from the bare plots was two orders of magnitude higher (5700–6900 t km–2 yr–1) than soil loss from grass plots (30–50 t km–2 yr–1).



Table 10 Plot soil losses at Pukekohe from bare soil and pasture (from Basher et al. 1997). Plots 2 and 3 were converted to grass in March 1972

Soil loss (t km-2 yr-1)

Period Plot 1 Plot 2 Plot 3 Plot 4

26/2/71 – 14/3/72 6944 6679 5723 5878

26/2/71 – 14/3/72 4820 48 27 4062

26/2/71 – 11/9/73 5690 4806

After a severe storm at Pukekohe in 1999 (>50 yr recurrence interval) Basher and Thompson (1999) noted the influence of cover crops on the extent of erosion semi-quantitatively. They classified the severity of erosion (as a percentage of field area – see Table 11) in each paddock into 5 classes and compared the extent of erosion for different crops and surface conditions

Table 11 Relationship between severity of erosion and surface condition/land use during a storm at Pukekohe in 1999 (from Basher & Thompson 1999)

Erosion extent*

No. of fields Negligible Very slight Slight Moderate Severe

Cover crop 11 100

Bare, cultivated

134 34 2 30 18 16

Greens 35 88 5 8

Onions 46 75 0 13 8 3

* Negligible = <1, very slight = 1–2%, slight = 3–5%, moderate = 6–10%, severe = 11–50%

There have been a large number of studies internationally on the influence of cover crops on erosion rates, both in understanding the function of cover crops (providing resistance to soil particle detachment by water and wind, adding organic matter and increasing soil carbon and structural stability) and defining the extent of their impact in reducing soil loss. Soil conservation technology developed in the 1930s and 1940s in the USA led to the derivation of the Universal Soil Loss Equation in which one of the key parameters is the cover and management factor (C). This expresses the average soil loss for a particular crop system and management relative to bare soil (C = 1) and values vary widely from <0.05 to >0.8 for different cropping systems (Wischmeier & Smith 1978). Values of C have not been published for New Zealand cropping systems, but for use on earthworks in New Zealand the following values are used (Auckland Regional Council undated): pasture 0.02, establishing grass 0.1, mulch on topsoil 0.05. Gabriels et al. (2003) compiled C factors for 40 different crop rotation systems in Belgium that accounted for crop growth stage (fallow, seedbed, establishment, development, maturity and residue/stubble), type of rotation and the length of crop rotation. The main crops considered of relevance to New Zealand were wheat, barley, potato, maize, beans, peas, leek, carrot, and for most crop rotations involving these crops the C factor was between 0.28 and 0.38.



There have been many studies of the impact of cover crops in reducing erosion. For examples:

Use of winter cover crops to reduce erosion of a silty soil in Belgium from a continuous silage maize system is described by Laloy and Bielders (2010). During the intercropping period, cover crops reduced runoff and erosion by more than 94% compared with untilled, post- maize harvest plots. Rough tillage after maize harvest proved equally effective as a late sown cover crop. There was no effect of cover crop destruction and burial dates on runoff and erosion during the intercropping period, probably because rough tillage for cover crop burial compensates for the lack of soil cover.

Similarly Kaspar et al. (2001) found rye and oat cover crops grown between no-till soybean crops on silty soils in the USA reduced erosion by c. 40%. The rye cover crop reduced erosion significantly more than oats.

In Uruguay Alliaume et al. (2014) characterised the effect of oat cover crops on silty soil used for outdoor tomato production in raised beds. Soil cover reduced soil loss from raised beds by more than 98%. In this study, reduced tillage combined with mulching also had a larger impact on reducing erosion than the cover crop.

7.4 Grassed riparian buffer strips

There have been no studies of the effectiveness of grass riparian buffer strips (also called vegetative filter strips) in New Zealand. However numerous overseas studies have shown that they can be highly effective in reducing sediment delivery to streams by decreasing the velocity of runoff and allowing particles to settle (see reviews in Dosskey 2001; Yuan et al. 2009).

Kronvang et al. (2000) found that a 29-m wide buffer was able to trap 100% of the sediment from rill erosion on sandy and loamy soil in arable cropland in Denmark, and even a 0.5 m wide buffer strip trapped 62% of the eroded soil.

Dillaha et al. (1987) used a field plot based rainfall simulator study to evaluate the effectiveness of grass buffer strips (0, 4.6, or 9.1 m width) for the removal of sediment from cropland runoff on a silt loam soil. The 9.1 and 4.6 m grass buffer strips with shallow uniform flow removed an average of 84 and 70% of the incoming sediment. They also suggested from observation of existing grass buffer strips that on-farm vegetative buffers were not likely to be as effective as experimental buffer strips because of problems with flow concentration. Similarly, Magette et al. (1989) examined the effectiveness of grass buffer strips 4.6 and 9.2 m long in removing sediment from cropland runoff. Mass losses of total suspended solids were reduced by 66% by 4.6-m-wide buffer strips and 100% by 9.2-m-wide buffer strips.

Gharabaghi et al. (2000) studied the effect of buffer width on sediment trapping efficiency in plot-based experiments. Their results on a high clay content soil showed average sediment-removal efficiency varied from 50 to 98% as buffer width increased from 2.44 to 19.52 m. Almost all the easily removable aggregates (>40 µm diameter) were removed within the first five metres of the buffer strip but the small-size aggregates were very



difficult to remove by filtering flow through grass. The only effective mechanism for removal of small-size sediments was infiltration. Experiments with appreciable infiltration showed removal efficiencies of >90%. The sediment-removal efficiency of the buffer strip did not increase much by increasing the width of the filter strip beyond 10 m.

Field experiments on a silt loam soil with simulated runoff were used by Abu-Zreig et al. (2004) to study the efficiency of grass buffer strips for sediment removal from cropland runoff. Twenty buffer strips with varying width (2, 5, 10 and 15 m), slope, and vegetated cover were used. The results showed that the average sediment trapping efficiency of all buffers was 84%, and ranged from 68% in a 2-m-wide buffer strip to as high as 98% in a 15-m-wide buffer strip. The width of buffer was the predominant factor affecting sediment deposition in buffer strips up to 10 m; increasing buffer length to 15 m did not improve sediment trapping efficiency.

The effect of a 6-m-wide buffer strip in reducing runoff and suspended solids from a loamy soil growing maize, winter wheat and soybean in North-East Italy is described by Borin et al. (2005). The buffer strip was composed of two rows of regularly alternating trees and shrubs, with grass in the inter-rows. The buffer strip reduced runoff by 78%, total suspended solid concentrations by 78%, and losses by 94%.

Dosskey (2001) and Yuan et al. (2009) review a large number of studies of the performance of grass buffers and make the following observations:

Published studies show buffers typically retain 40–100% of the sediment mass that enters them but the effectiveness of buffers in removing sediment varied widely among the studies analysed

sediment trapping performance depends strongly on buffer width, buffer type (vegetation type, density and spacing), particle size of the sediment, the ability of the vegetation to retard flow, soil infiltration rate, the amount of runoff, slope gradient and length of contributing slope

the first 3–6 m of buffer plays a dominant role in sediment trapping

Dosskey (2001) suggests it remains difficult to extrapolate current knowledge of buffer functions into accurate estimates of sediment reduction in response to buffer installation on crop land at any given site.

Yuan et al. (2009) provide some guidelines to performance:

under conditions of relatively shallow flow not concentrated in channels, gently sloping, densely vegetated 3-m buffers are likely to limit transport of sediment from fields to streams, whereas moderately steep, less densely vegetated buffers of 3 m may be vulnerable to much higher rates of sediment delivery

buffers greater than 6 m are effective and reliable in removing sediment from any situation

the sediment trapping to buffer width relationship is best fitted with a logarithmic model but, based on the published literature, there is a large amount of scatter in this relationship (Fig. 1).



Figure 1 Relationship between buffer width and sediment trapping efficiency (from Yuan et al. 2009).

7.5 Sediment retention ponds

There is little information on the performance of sediment retention ponds used for sediment control from arable cropping, either in New Zealand or internationally. Pellow and Barber (unpublished Franklin Sustainability Project Newsletter, June 2004) provide some very crude estimates of soil loss (based on a very limited amount of water sampling and measurement of total suspended solids) from different sized sediment retention ponds at Pukekohe (Table 12). They compared estimated annual and soil losses from flat and sloping fields, with and without sediment retention ponds, and adequately sized with undersized sediment retention ponds. Where a properly designed sediment retention pond was installed, estimated annual soil loss was reduced to one third of that where no sediment retention pond was used. The undersized pond produced 50% more sediment than a correctly designed pond. They suggest that while the study was very limited in scope, the relative soil losses provide a good indication of the performance of sediment retention ponds, and that the effect of ponds is greater in medium-high intensity rainfall events than low intensity events.

Table 12 Estimated annual soil loss at Pukekohe (Pellow & Barber, unpublished)

Paddock description TSS1 (g m–3)

Annual soil loss (t ha–1 yr–1)

Event soil loss (kg ha–1 yr–1)

Flat, no SRP 370 2 55

Sloping, no SRP 2040 12 68

Sloping, design guideline SRP2 690 4 25

Sloping, substandard SRP3 1070 6 79

1 Total Suspended Solids 2 SRP constructed to design specifications (based on 50 or 100 m3 ha-1 of catchment depending on catchment area, slope and row length; see http://www.aucklandcouncil.govt.nz/EN/environmentwaste/sustainabilityconservation/landandsoilconservation/Documents/franklinsustainabilityprojectguide.pdf) 3 undersized SRP

http://www.aucklandcouncil.govt.nz/EN/environmentwaste/sustainabilityconservation/landandsoilconservation/Documents/franklinsustainabilityprojectguide.pdf

http://www.aucklandcouncil.govt.nz/EN/environmentwaste/sustainabilityconservation/landandsoilconservation/Documents/franklinsustainabilityprojectguide.pdf



There are few studies in the international literature where the performance of sediment retention ponds within cropland has been characterised. Brown et al. (1981) describe a 5-year study of a pond with a capacity of about 3400 m3 receiving surface water runoff from 4050 ha of irrigated land (0.8 m3/ha) mostly under grain crops. The pond was kidney shaped, 150 m long and 1.2 m deep, constructed on an irrigation canal. It removed 65–76% of the sediment entering the pond with efficiency depending on the flow rate and the sediment concentration of flow entering the pond. The measured efficiency was higher than the design efficiency (50%) despite high flow rates overloading the pond. The pond removed sand and silt sized particles but was unable to settle clay sized particles without flocculation.

Fiener et al. (2005) described the performance of small sediment retention ponds established at field borders in areas growing potatoes and grain. The field borders were raised by earth embankments to create small detention ponds and these were drained by underground tile outlets with perforated risers. Four detention ponds with volumes of 30–260 m3 ha–1 were monitored over 8 years. The ponds trapped 54–85% of the incoming sediment amounting to up to 15 t ha–1 yr–1. While these were very low cost ponds, they required regular sediment removal.

7.6 Management of wind erosion

Basher and Painter (1997) and Ross et al. (2000) reviewed practices for controlling wind erosion in New Zealand. These include limiting the time the soil is bare (by maintaining vegetative cover or surface residue) and has a dry surface (using irrigation, mulches), and reducing wind velocity through increased surface roughness (stubble mulching, ridge-till, preparation of coarse seedbeds), shelter belts, and strip cropping. There has been limited research in New Zealand on these mitigation practices but there is an extensive overseas literature (see Chepil & Woodruff 1963; Skidmore 1986) that shows the key principles for controlling wind erosion are: stabilizing erodible surfaces with surface residue; producing a rough, cloddy surface; reducing field width or the distance wind travels in crossing an unprotected field with barriers and strip crops; and establishing and maintaining sufficient vegetative cover.

Basher and Painter (1997) identify the key factors affecting potential for wind erosion in New Zealand as:

wind erosivity. Patterns of erosivity are described in Painter (1978) and values are relatively high due to the windy climate of New Zealand

evaporative energy supply which reduces near-surface soil water content

soil erodibility (often estimated by the proportion of soil aggregates >0.84 mm, Woodruff and Siddoway, 1965) is high owing to low clay content and weak soil structure in many soils

surface soil water content. In the temperate climate of New Zealand, near-surface soil water content is a more important influence on wind erosion rates than in arid climates.



Field shelter by planting windbreaks has historically been widely practised to protect cropland in eastern parts of New Zealand (Stringer 1977; Sturrock, 1984). Considerable research has been directed at windbreak design and performance, suitable species, the advantages of shelter, its costs and benefits (see Salter 1984; Gilchrist 1984). Windbreaks reduce wind speed at ground level for a distance of about 20–30 times the height of the wind break on the leeward side and 5 times on the windward side. The length of the shelter and distance between windbreaks are also important for determining the area protected. Recommended windbreak design and composition is described by Hicks and Anthony (2001).

Little research has been undertaken on other aspects of wind erosion mitigation except for analysis of the influence of surface soil water content at time of cultivation on soil erodibility. Cresswell et al. (1991) examined the effects of pre-tillage soil water content and multiple-pass tillage on soil physical condition on a silt-loam soil susceptible to wind erosion. They found excessive tillage of this soil should be avoided, especially at low soil-water content, because it produced large quantities of wind-erodible aggregates, and a smoother soil surface that increases susceptibility to wind erosion.

7.7 Summary

ESC practices to control water and wind erosion on cropland have not been much studied in New Zealand.

In row crops, compacted wheel tracks are recognised as major sources of runoff and erosion. Ripping of wheel tracks reduced erosion by 95% on strongly structured clay soils. Overseas literature suggests this practice would be most effective on silty and clayey textured soils and less effective on sandy soils.

Wheel track diking has been shown in New Zealand trials to reduce runoff, but the impact on soil loss has not been characterised. Overseas studies have shown this practice can reduce soil loss by 60 to >92%.

Cover crop trials in New Zealand have demonstrated an improvement in aggregate size and stability which by inference should reduce soil erodibility and erosion. At Pukekohe a cover crop trial produced a relatively small reduction in soil loss (26–38%). In a severe storm in this area in 1999 there was little erosion where cover crops were present. There have been a large number of studies internationally on the influence of cover crops on erosion rates that show they can reduce erosion by more than 90%.

There have been no studies of the effectiveness of grass riparian buffer in cropland New Zealand. Overseas studies show that they can be highly effective in reducing sediment delivery to streams by decreasing the velocity of runoff and allowing particles to settle and infiltrate. Buffers typically retain 40–100% of the sediment mass that enters them but the effectiveness of buffers in removing sediment varies widely depending on buffer width, buffer type, particle size of the sediment, the ability of the vegetation to retard flow, soil infiltration rate, the amount of runoff, slope gradient, and length of contributing slope. The first 3–6 m of buffer plays a dominant role in sediment trapping.



There is little information on the performance of sediment retention ponds used for sediment control from arable cropping, either in New Zealand or internationally. One Pukekohe study found annual soil loss was reduced to one third of that where no sediment retention pond was used. One overseas study showed ponds can be highly efficient (>65%) in removing sand and silt but was unable to remove clay particles in runoff.

Field shelter provided by windbreaks has historically been widely practised to protect cropland from wind erosion in eastern parts of New Zealand. Most research has been directed at windbreak design and performance (on wind speed rather than erosion mitigation), suitable species, the advantages of shelter, and its costs and benefits. Little research has been undertaken on other aspects of wind erosion mitigation in New Zealand except for analysis of the influence of surface soil water content at time of cultivation on soil erodibility.

8 Assessment of erosion and sediment control performance for pastoral farming

Large tracts of pastoral farm land in New Zealand are located on erodible hill country. In these areas naturally high rates of erosion (as a result of steep slopes, highly erodible rocks, generally high rainfall and common high-intensity rainstorms) have been exacerbated by recent deforestation and conversion to pastoral farmland (Basher 2013). Erosion in the hill country has been a concern since the 1930s and there is a long history of planting trees to control landslide, gully earthflow, and bank erosion (primarily wide-spaced plantings, but also blanket afforestation) and the use of farm plans with a narrow soil conservation focus. More recently, Whole Farm (WFP) and Environment Plans have been undertaken by regional councils and other industry groups to reduce soil erosion but also to take a broader view of land management and integrate soil conservation strategies with farming operations. Surface erosion (sheet and rill), often associated with poor vegetation cover and heavy stocking that reduces infiltration rates, can also occur but is generally less of an issue than mass movement and gully erosion. In the lowlands bank erosion can be a significant issue, often associated with stock access to river banks.

8.1 Performance of space-planted trees

The most commonly used practice for reducing erosion on hilly pastoral farmland is wide spaced planting of trees, usually Poplars (Populus) and willows (Salix) (van Kraayenoord & Hathaway 1986a, b) but other species such as Acacia or Eucalyptus are sometimes used. Their use was reviewed by Basher et al. (2008a) and Phillips et al. (2000, 2008). Despite the widespread use of space-planted trees for erosion control in New Zealand there has been relatively little experimental or quantitative work to establish their effectiveness in reducing erosion in relation to factors such as tree size and planting density, and there are no published studies on their measured effect on sediment yield (Douglas et al. 2009). Nor is there any information on their effectiveness over a range of different storm size (recurrence interval). The published studies emphasise the importance of both initial establishment of the trees and subsequent maintenance to ensure their survival and effectiveness. Results of



studies that have directly examined the effectiveness of space-planted trees for erosion control on pastoral farmland are summarised in Table 13.

In early work Hawley and Dymond (1988) analysed the effects of space-planted trees on landsliding on a small area underlain by mudstone at Ngatapa, Gisborne. The trees were originally planted at 20-m spacing (25 stems per hectare, sph) but there was some mortality. They used computer processing of digital aerial photographs to determine the relationship between landsliding and proximity to space-planted trees by characterising the fraction of ground eroded with distance from trees. Measurements were made on individual trees and converted to the average fraction of ground eroded which increased with distance from trees (Fig. 2).

Figure 2 Relationship between average fraction of ground eroded and distance from trees (from Hawley & Dymond 1988).

They found the influence of trees was negligible beyond 11 m. From the relationship between average fraction of ground eroded and distance from trees they calculated that if trees had been planted at 10-m spacing with 100% establishment and survival there would have been a reduction in landslide damage of 70%. On the hillslope examined, where the spacing of 14-year-old trees was 20 m and 66% of the planted trees had survived, the actual reduction in landslide damage due to space-planted trees was only 14%. This work was extended by Hawley (1988) to determine the influence of root encroachment from neighbouring trees on slope stability. At a spacing of 11.5 m (75 sph), he estimated that landsliding around one tree, including the contribution from a neighbouring tree, reduced from 8.2% to 1.4%.

Table 13 Studies of effects of space-planted trees under pastoral farming

Reference Location, terrain, storm history

Erosion type(s) Scope and Methods Key Findings

McIvor et al. (2015)

Coastal Hawkes Bay; moderately steep young sediments; storm rainfall estimated 700 mm

Shallow landslides

Aerial photo assessment to identify comparable sites with (86 sites) and without trees (25 sites), field measurement of landslide distribution, site and tree characteristics (number, species, DBH, tree spacing, distance of landslides from nearest tree). Used to calculate area of protection from trees (total area protected minus area of slip scars) and the effectiveness of trees linked with tree size, tree species, and tree spacing.

Tree spacing varied from 9 to 11 m. Landsliding reduced by 78% on sites with trees compared to pasture sites. Mature poplars and willows reduced landsliding within a zone of ~10 m of the trees to almost zero. Where plantings had a mean DBH of <20 cm, their effectiveness was reduced dependent on spacing. For trees with a mean DBH of ~10 cm effectiveness was negligible regardless of spacing.

Douglas et al. (2009, 2013)

Storms in 2004 (Manawatū – daily rainfall >200 mm) and 2006 (Wairarapa – daily rainfall >100 mm); steep young sediments

Shallow landslides

Compared occurrence of landslides on slopes with space planted trees (poplars, willows, Eucalytpus) and pasture at 65 sites. Collected tree attributes (height, DBH, canopy radius, tree density/spacing (range 32–65 sph, 12–18 m). Assumed radius of influence of 10 m and calculated area of landslides within 10 m of a tree. The percentage of shallow landslide scar area at each tree (TSA) and paired pasture (PSA) site was calculated by expressing scar area as a percentage of total site (polygon) area.

Trees reduced landslide occurrence by 95 per cent compared to paired pasture control sites (0.4% vs. 7.9% scar area, respectively), and scars occurred on fewer sites with trees than pasture (10 vs 45). There were no significant differences between species in their effectiveness in reducing landslide occurrence. On the tree sites where landslides occurred, per cent scar area was <3.5 per cent except at one site where it was 11.3 per cent. Greatest extent of landsliding occurred where trees had a DBH of < 30 cm. Suggest that if all poles survive to produce trees, trees could be thinned to increase understorey pasture production without compromising slope stability, providing retained trees have DBH > 30 cm and are no further apart than 18 m.

Varvaliu (1997)

Pakihikura valley, Rangitikei, storms August/September 1992, steep young sediments

Shallow landslides

362 slopes, aerial photo interpretation and field checking, measurement of area and % of slope eroded.

Average % eroded of unstable slopes was 11.7% under pasture, and 7–8% under space-planted trees, pines and indigenous forest. Soil conservation planting reduced landsliding by 33 to 37%.



Lough (1993) Pohangina valley, storms August/September 1992, steep young sediments

Shallow landslides

Aerial photo interpretation and field checking, measurement of area and % of slope eroded.

Average % erosion on unstable slopes in pasture was 6.7% compared to 2% with space-planted poplars, manuka scrub (0.4%, native forest (0.5%), pines (1.5%)

Cameron (1991); Hicks (1991a)

Whareama catchment, Wairarapa; 305 sites on argillite and young Tertiary siltstones storm 8–11 April 1991, storm rainfall 200–300 mm

Shallow landslides, gullying, streambank erosion

Field-based subjective assessment of storm damage (stability, type of erosion, damage to fences, tracks, pasture and drains, feasibility of treatment, type of soil conservation measures) along a road-based transect and relationship to landform, LUC unit and vegetation type (soil conservation planting, native scrub and forest). Damage mapped in field, office analysis to assign sites as stable, unstable/untreatable, unstable/unplanted, unstable/inadequately planted, unstable/adequately planted and calculated several ratios (need – proportion of sites needing treatment, extent – proportion of sites with treatment installed, adequacy – proportion of sites with adequate treatment) and frequency distribution of indices.

Adequately installed measures reduced gullying by 50%, bank erosion by 24%, landsliding by c. 70% relative to unstable, unplanted slopes. Soil conservation measures installed on

67% of unstable channels, but only 55% of these adequately installed

on 52% of unstable footslopes, but only of these 39% adequately installed

on 43% of unstable hillslopes, but only 43% of these adequately installed

Soil conservation measures reduced catchment sediment supply by

35% from gullying

21% from bank erosion

22% from landsliding

Soil conservation measures were only installed 2/3rds of unstable channels and half the unstable slopes. Only 50% of the treated channels and 40% of hillslopes had adequately installed and maintained soil conservation measures.

Pain & Stephens (1990)

Eltham, Taranaki; steep young siltstone and mudstone; storm rainfall 202 mm over 3 days

Digital aerial photographic assessment compared with field-based visual assessment, single storm.

8% of c. 10 000 ha of pasture affected by landslides (visual) cf. 9.6% (digital assessment of 5 sample areas). Extent of landsliding affected by vegetation type – pasture 9.6%, native forest 0.27%, pines 0%, regrowth 0.97%.



Hawley (1988); Hawley & Dymond (1988)

Ngatapa, Gisborne; steep slopes underlain by ash over mudstone

Shallow landslides

Assessed location of landslides relative to space-planted trees to calculate average fraction of ground eroded vs distance to trees. Tree density 25 sph.

From the relationship between average fraction of ground eroded vs distance to trees calculated that if trees had been planted at 10-m spacing (100 sph) with 100% establishment and survival there would have been a reduction in landslide damage of 70%. However the spacing of 14-year-old trees was 20 m and 66% of the planted trees had survived, the actual reduction in landslide damage due to space-planted trees was only 14%.

Hicks (1988, 1989a, 1992a)

Waihora catchment, Gisborne; steep young siltstone and mudstone; post Cyclone Bola, storm rainfall of 300-600 mm

Shallow landslides, gullying, streambank erosion

Field and aerial photo subjective assessment of storm damage to hillslopes and channels along transects (generally adjacent to roads) to record extent of erosion, nature of soil conservation planting or other vegetation, stability/treatment status. Damage mapped in field, office analysis to assign sites as stable, unstable/untreated, unstable/inadequately treated, unstable/adequately treated Analysis of relationship of damage to landform, LUC unit and vegetation type (soil conservation planting, native scrub and forest). Mapped, area affected by mass movement, length of water courses affected, length of fences and tracks damaged using grid square sampling.

Describes general relationship between LUC units and storm damage (most hillslope and stream damage on class 7 with lesser amount on class 6). Criteria for adequacy were: appropriate planting pattern for type of instability, all unstable areas planted at adequate density (<10 m spacing), trees were mature and healthy. About 7% of transect hillslopes damaged by fresh mass movement. Farm conservation measures reduced damage on 34% of hillslopes where they were installed and maintained, but did not reduce damage where they were inadequately installed or maintained (66% of hillslopes). Mass movement was 22% less than it would have been in the absence of conservation planting. Stream bank plantings reduced bank erosion compared with untreated streams. Damage repair costs 20% lower than they would have been in the absence of conservation planting, and could have been 63% less had measures been installed to an adequate standard wherever they were required.



Hicks (1992b) Waihora catchment, Gisborne; steep young siltstone and mudstone; post Cyclone Bola, storm rainfall of 300–600 mm. Whareama catchment, Wairarapa. Waipa, Waikato

Streambank. Same as Hicks (1988, 1989a, 1992a) In both areas two thirds of sites were planted but half of these were untreatable. In Waihora, 51% of banks eroded under grass compared with 2–3% under poplars and willows; erosion reduced by 50% where planting adequate and maintained, 40% of sites not planted, and 27% inadequately treated. In Whareama, number of bank failures/km lower under willows (3) and poplars (12) than grass (16), only 33% of banks adequately treated. In Waipa, % of bank eroded reduced from 23% under grass to <10% under willows and poplars), only 33% of banks adequately treated.

Pain (1986) 3 sites near Mangaweka, Manawatū; steep young mudstone with some inter-bedded sandstone

Shallow landslides

Digital aerial photographic assessment before-and-after space planting, comparison of treated and untreated hillslopes, field assessment of landslide occurrence; 1952–85/86 period.

Early attempt at assessing effectiveness of trees in reducing landsliding using digital image analysis. Most landslides occurred before the earliest imagery (1952). Limited effects of space planted trees but mostly due to experimental design – many landslides occurred under trees but trees were preferentially planted in more susceptible (concave) sites, or only a short time since trees planted

Phillips et al. (2008)

Gisborne, steep Tertiary soft sedimentary

Earthflow, gully, slump

Field assessment based on Thompson and Luckman (1993). 17 earthflow sites, erosion, 13 gully sites, 1 slump site. Trees >12–15yrs old and average diameter at breast height of 23–79 cm.

Successful treatment for both earthflows and gullies occurred with willows and poplars at final spacings of 4–6 m, though this was highly variable. Treatment unsuccessful where tree cover was sparse (because of losses) or where trees were planted wider than 10 m apart. Recommend initial plant spacings of 4 m for gully erosion control and 5–6 m for earthflows and 8-m + for landslides.



A somewhat similar approach has been used in several recent studies that have quantitatively determined the effect of space-planted trees (including the species, their size and density) in reducing landsliding. Douglas et al. (2009, 2013) examined the effects of small groups (5–10) of mature space-planted trees (dominantly poplar with some willow and Eucalyptus) at 40 sites in the Manawatū and 25 sites in the Wairarapa following large storms in 2004 and 2006. Trees were planted at 32–65 sph, height was 8–43 m, and slope averaged 27°. Measurements were made of the area of individual shallow landslide scars within a 10-m radius of each tree in the measured group, assuming this was the radius of influence of individual trees (based on Hawley & Dymond 1988). The effect of the space-planted trees was then compared with landslide occurrence in comparable pasture sites without trees to assess the influence of the trees.

Trees reduced landslide occurrence by 95 per cent compared with paired pasture control sites (0.4 per cent vs 7.9 per cent scar area, respectively), and scars occurred on fewer sites with trees than pasture (10 vs 45). For the 10 tree sites with scars, the area of scars was <3.5 per cent, except at one site where it was 11.3 per cent – the greatest extent of landsliding occurred where trees had a DBH of < 30 cm. Mature trees reduced landsliding by 95% when planted at spacings closer than 13–18 m. Younger trees (DBH <30 cm) were less effective, and willows were more effective than poplars.

A similar study is described in McIvor et al. (2015). They characterised the performance of space-planted trees in a severe storm that hit coastal Hawke’s Bay in 2011. Using the same approach as Douglas et al. (2009), data were collected from 86 sites with mostly 20–30-year-old poplar and willow trees and 25 nearby pasture sites. The data set included both mature trees and smaller diameter (mostly willow) trees. The poplar sites had a mean DBH of 31.1 cm compared to the mean DBH of 18.0 cm at willow only sites

Landsliding was reduced by 78% on sites with trees compared with pasture sites. Mature plantings of groups of both poplar and willow reduced landsliding within a zone of ~10 m of the trees to almost zero. There was a moderately strong relationship between DBH and area of protection for poplars, whereas it was weak for willows. Where plantings had a mean DBH of <20 cm, their effectiveness was reduced dependent on spacing. For trees with a mean DBH of ~10 cm, effectiveness was negligible, regardless of spacing. Despite this, trees with DBH of ≥10 cm were generally able to withstand shallow slipping without being totally dislodged. Where trees were planted in gullies the same criteria applied for their effectiveness in halting the advancement of the gully upslope.

Varvaliu (1997) assessed the effectiveness of soil conservation measures in reducing landsliding for a series of storms in 1992 in an area of Manawatū Tertiary mudstone hill country. He found the area of landslides on unstable slopes was 11.73% under pasture, 7.45% under space-planted poplars, 7.74% under pines, and 7.84% under indigenous forest. This study found a much lower reduction in landsliding (33%) compared with pasture, and little difference with vegetation type. Lough (1993) analysed the effect of the same storms in part of the Pohangina valley and found that the average % erosion on unstable slopes in pasture was 6.7% compared with 2% with space-planted poplars, manuka scrub (0.4%), native forest (0.5%), pines (1.5%), equivalent to reductions of 70%, 94%, 92%, and 77%. Hicks et al. (1993) analysed the extent of erosion across a wider area of the Manawatū-Wanganui hill country for the same storm, using a grid-based approach, and found landslide



erosion was c. 85% less under scrub and pine forest than under pasture. No erosion was observed under indigenous forest. In summarising their observations they suggest that, compared with pasture and on unstable slopes, erosion reduces by 10% with sporadic space-planting, 60% with extensive space-planting, and 90% under mature pines or indigenous forest.

Earlier studies used less quantitative methods of assessing performance of space-planted trees. Cameron (1991) and Hicks (1991a) assessed the effect of space-planted trees on landsliding in a storm in the Whareama catchment, Wairarapa. They found adequately installed measures reduced gullying by 50%, streambank erosion by 24%, mass movement of colluvial footslopes by 67%, and steep hills by 71% compared with unstable, unplanted slopes. Only about half the soil conservation measures were adequately installed. He also suggests catchment sediment supply was 23% less than could have been expected in the absence of soil conservation measures, and that it could have been halved by upgrading inadequately installed plantings.

The performance of soil conservation measures in the Waihora catchment during Cyclone Bola is described by Hicks (1988, 1989a, 1992a). Based on assessments of the extent of erosion (landslides, earthflows, gullies) and performance of soil conservation measures on a transect through the catchment, he suggests that erosion was 22% lower (measured as area of damage) than it would have been in the absence of soil conservation measures. He also suggests it could have been reduced by 74% had soil conservation measures been installed everywhere they were needed, and to an adequate standard. Of the soil conservation measures that had been used, only 35% were assessed as adequate.

The effectiveness of soil conservation treatments for controlling gully and earthflow erosion in the Gisborne – East Coast region following Cyclone Bola was evaluated by Thompson and Luckman (1993). A comprehensive study of sites affected by gully erosion (136 sites) and earthflow erosion (142 sites) collected a standard set of data describing each site and the soil conservation treatment applied (see section 6.2). The treatments were based mainly on the planting of poplar and willow trees (aged 10–30 years at the time of the assessment). For gully erosion afforestation, gully wall planting, channel (pair) planting and debris dams were evaluated. Earthflow erosion was treated by afforestation, space-planted trees, localised close tree planting, gully control at the toe of the earthflow, graded diversion banks, surface smoothing and drainage. The site types were grouped into a series of 12 classes for assessment of effectiveness (Fig. 3), since different site types require different soil conservation treatment. From the extensive data collected inference-based procedures for interpreting this information were developed and documented (Luckman & Thompson 1990; Thompson & Luckman 1993) to ensure assessment of effectiveness was consistently applied. Computer-based knowledge systems were developed to integrate the field observations and available historical information, which were also compared with direct subjective field-based assessment of effectiveness.



Figure 3 Site classifications for earthflow and gully erosion (from Luckman & Thompson 1993).

Treatment of erosion was successful at 42% of gully sites and 63% of earthflow sites. The study enabled minimum tree configurations (tree spacing and coverage of erosion landforms) to be specified. Feasible treatments existed for 8 of the 12 site classes studied. For gullies less than 2-m deep: tree spacing of 2–4 m over the entire length of the gully was successful in 9 of 10 sites, whereas a spacing of 4–8 m was successful in 11 of 32 sites. Deeper gullies (2–5 m) require pair channel planting at spacings of 4–8 m or less, in combination with planting 40–60% of the gully walls at similar spacing. An integrated approach combining debris dams, gully wall, and channel planting was also successful, as was afforestation. For gullies >5 m deep, or gullies in bentonite or argillite no suitable treatment was identified. Earthflow sites with shallow, untreated toe gullying were successful where tree spacing was 5–8 m or less covering >60% of the earthflow area. If the gully was treated, wider spacing (8–10 m) combined with planting 40% of the earthflow area was successful. Where gully depth at the toe of an earthflow was >2 m, an appropriate gully treatment combined with planting of >20% of the earthflow area at spacing up to 5–8 m was needed. If wider tree spacing was used (8–10 m) >60% of the earthflow area need to be



planted. Reforestation was also successful in treating earthflows. The results suggested that the interaction between roots of neighbouring trees at close tree spacing was a major factor in conferring treatment success.

Phillips et al. (2008) adopted similar field-based criteria for assessment of effectiveness of space-planted trees in controlling earthflow, gully and landslide erosion at 30 sites within the East Coast Forestry project with all treatments at least 12–15 years old. Evaluation of treatment success (classed as successful or unsuccessful) was based on current erosion activity including the surrounding area (using an LUC-based assessment of erosion severity), tree survival and tree condition to classify current land condition into 5 classes (very poor to very good). This data showed that successful treatment for both earthflows and gullies occurred with willows and poplars at final spacings of 4–6 m, though this was highly variable. Unsuccessful treatments were where remaining tree cover was sparse (because of losses) or where trees were planted wider than 10 m apart. They concluded that to obtain effective erosion control for active gullies, initial plant spacings needed to be at 4 × 4 m (625 stems ha–1) or less and for earthflows, spacings of 5–6 m (400–280 stems ha–1) was recommended. Effective erosion control for shallow landslides could be achieved at wider spacings of 8 m or wider. They also provided a set of recommended erosion control treatments.

Root and canopy development are the primary drivers for the effectiveness of space-plantings on slopes for reducing the incidence and severity of landsliding and other mass movement erosion types. This has been the driver for a number of studies of above- and below-ground plant growth characteristics, including canopy growth rates, canopy closure modelling, root strength and decline, root growth rates, root morphology, root biomass, root occupancy (e.g. Phillips et al. 2000, 2014; McIvor et al. 2008, 2009, 2011; Douglas et al. 2010; Schwarz et al. 2016). This knowledge has been used for recommending appropriate species, for planting stock, and for tree spacings to minimise erosion. For example, Phillips et al. (2000) review plant performance for the possible erosion control treatments funded under the East Coast Forestry project. They argue that the first 8 years of any treatment are the most important for sediment reduction and develop a set of parameters related to above- and below-ground plant characteristics with which to evaluate different treatments. These were ranked in order of importance as:

canopy occupancy = lateral root site occupancy> root cross-sectional area per shear area >= root depth

>>root biomass

Some of these parameters can be derived from other growth parameters (e.g. canopy occupancy from diameter-at-breast-height, tree height, crown width). Using this approach they ranked the performance of potential treatments

mature reversion > plantation forestry >= supplementary planting >> within-gully treatment

They use this approach to develop a method for assessing the sediment reduction effectiveness of treatments allowed as part of the East Coast Forestry project based on individual plant performance criteria and their hydrological and mechanical influence on slope stability. This was further extended by Phillips et al. (2008), who provided guidelines for choosing effective erosion control treatments for a range of erosion processes aimed primarily at the Gisborne – East Coast region (Table 14).



Table 14 Recommended plant treatments for control of erosion features (from Phillips et al. 2008)

Feature Severity or risk Current activity status

Area (ha)

Depth (m)

Treatment1 Plant treatment, spacings and minimum plant densities (stems/ha)2

Earthflow

Severe Active > 1 ha +/- > 2 m gully Feature + Reversion OR Afforestation OR 5 m on feature (400 spha) AND 250 spha buffer

Moderate Active < 1 ha +/- > 2 m gully Feature + 6 m on feature (250 spha) AND 200 spha buffer

Active or inactive < 1 ha +/- < 2 m gully Feature 6–7 m (200–250 spha)

Gully

Severe to extreme Active > 2 ha > 2 m Feature + Reversion OR Afforestation OR 4 m (600 spha) (but probably not practical) AND 250 spha buffer

Moderate Active < 2 ha > 2 m Feature + 5–6 m on feature (250–400 spha) AND 200 spha buffer

Active or inactive < 2 ha < 2 m Feature 5–6 m (250–400 spha)

Slump Severe Active > 1 ha N/A Feature + 5 m on feature (400 spha) AND 250 spha buffer

Moderate Active or inactive < 1ha N/A Feature 6–7 m (200–250 spha)

Landslide only High N/A N/A N/A Coverage3 6–7 m (200–250 spha)

Moderate N/A N/A N/A Coverage 8–10 m (100–150 spha)

Combination4

High/severe Active N/A N/A Coverage Reversion OR Afforestation OR 5 m on active parts (400 spha) and 250 spha elsewhere

Moderate Active or inactive N/A N/A Coverage 6–7 m (200–250 spha)5

1 Feature – requires just the erosion feature to be treated; Feature +/– requires the erosion feature plus a proportion of contributing/surrounding catchment area to be treated (to a natural catchment boundary or limit of signs of instability. 2 Final stocking at age 8–10 years. 3 Coverage = general planting across slopes to achieve required stems per hectare. Preference is for equal spacing but mixed densities or species depending on topography and severity of individual features acceptable.

4 Combination, e.g. a slope that has a mix of erosion types but generally while individual flows, gullies etc. are moderate to severe they occupy a small proportion of the total area to be treated.



Recently, Schwarz et al. (2016) combined field data and a modelling approach to show that root distribution data are important inputs for quantifying root reinforcement at the hillslope scale, and that root distribution strongly depends on local environmental conditions and on tree planting density. The combination of soil mechanical properties (soil angle of internal friction and cohesion) and topographic conditions (slope inclination) are the major parameters needed to define how much root reinforcement is needed to stabilise a specific slope, and thus the spacing of the trees to achieve this at different tree size. Schwarz et al. concluded that planting density between 330 and 160 stems per hectare (spacing of 5.5 and 8 m, respectively) would assure significant root reinforcement for slope stabilisation and reduce the volume of triggered shallow landslides by up to 100 %.

Key issues for the use of space-planted trees for erosion control are selection of areas to be treated, initial tree spacing and tree survival. One approach to selecting areas for treatment in the Manawatū-Wanganui region is use of the Highly Erodible Land model (Dymond et al. 2006) which was developed to identify land that requires protection by woody vegetation cover. It uses a combination of slope thresholds for different geological terrain (ranging from 24° for mudstone and crushed mudstone/argilllite to 28° for hard greywacke/argillite and volcanic rocks) to identify land prone to landsliding, as well as mapping of earthflow and gully erosion from the New Zealand Land Resource Inventory, to identify land needing treatment by either space-planting or afforestation.

The effectiveness space-planting is highly dependent on successful establishment of the trees at the required spacing (see Table 14) and subsequent maintenance to ensure their survival. This was emphasised by early research that found performance was strongly limited by inadequate spacing and poor survival (e.g. Hawley & Dymond 1988; Hicks 1988, 1989a, 1992a; Thompson & Luckman 1993) and this may still be an issue. For example, a recent study by Marden and Phillips (2013) examined survival of poplar and willow poles planted in the Gisborne – East Coast area and found 24% of poles had died within 24 months, and 40% had died within 45 months. They attributed this to a combination of poor pre-treatment of poles, poor planting technique, site factors and stock damage.

8.2 Erosion control using closed canopy vegetation

For widespread and severe erosion afforestation is often used to control erosion, typically using conifers such as Pinus radiata or Douglas-fir (Pseudotsuga menziesii), or scrub and native forest reversion. There is far more data available for the effect of closed canopy (forest and scrub) vegetation, including both afforestation and reversion, on erosion and it has been extensively reviewed by O’Loughlin (1995, 2005), Glade (2003), Basher et al. (2008a), Marden (2004, 2012), Blaschke et al. (2008), and Phillips et al. (2012). This includes both process-based studies documenting the hydrological and mechanical mechanisms underlying the impact of trees on slope stability, and data comparing erosion rates under different vegetation communities. Most data on vegetation effects on erosion come from observations of storm damage (particularly after Cyclone Bola), generally comparing pasture with tree cover (either exotic pine plantations or mature native forest) or with scattered regenerating scrub, and mostly providing data on landsliding rates. There are also data on



the effect of forest cover on suspended sediment yield and the effects of afforestation, especially in controlling severe gully erosion.

8.2.1 Influence of closed canopy vegetation on erosion rates

Many studies have characterised the effect of young pines, mature closed canopy pines and native vegetation (forest and scrub) on erosion rates during large storm events such as Cyclone Bola.

Marden and Rowan (1993) described differences in landslide density under six different vegetation types before and after Cyclone Bola (Table 15). The incidence of landsliding was c. 80% lower under indigenous forest, pines >8 years old or scrub than pasture prior to Cyclone Bola, and increased to c. 90% lower during Cyclone Bola (except for scrub). Landslide density under young pine trees (< 6 years old) was little different to pasture. In a similar study Phillips et al. (1990) compared landslide densities and volumes under different vegetation for a series of storms in the Gisborne – East Coast area (Table 16). The impact of landsliding was greatest on pasture and pines < 6 years old and least for pines > 8 years old. Before Cyclone Bola landslide densities were 74% lower under pines > 8 years old than under pasture, and after Bola this increased to 91%. These data also show little difference in the percent reduction whether landslide density or volume is used as the metric for assessing vegetation effects.

Table 15 Landslide density (landslides ha–1) before and after Cyclone Bola (Marden & Rowan 1993)

Landslide density % reduction1

Vegetation type Pre-Bola Post-Bola Pre-Bola Post-Bola

Pasture 0.139 0.564

Indigenous forest 0.031 0.066 78 88

Pines > 8 yr 0.028 0.048 80 92

Pines 6–8 yr 0.07 0.162 50 71

Pines < 6 yr 0.135 0.496 3 12

Regenerating scrub 0.029 0.12 79 79

1 Compared with pasture

Table 16 Landslide density (landslides ha–1) and volume (m3 ha–1) before and after Cyclone Bola (Phillips et al. 1990)

% reduction

Vegetation type Pre-Bola landslide density

Post-Bola landslide density

Post-Bola landslide volume

Pre-Bola landslide density

Post-Bola landslide density

Post-Bola landslide volume

Pasture 0.23 0.68 916

Pines > 8 yr 0.06 0.06 48 74 91 95

Pines 6–8 yr 0.20 0.21 370 12 69 60

Pines < 6 yr 0.18 0.62 790 22 9 14



Marden et al. (1991) compared volumetric landslide rates in the Uawa catchment during Cyclone Bola for pasture and trees of different age to demonstrate how tree age affected the amount of landsliding (Table 17). The rate of landsliding was 87% lower under pines > 8 years old than under pasture, 40% lower for trees between 2 and 8 years old, while trees < 1 year old produced 24% more sediment than did pasture.

Table 17 Landslide volume (m3 ha–1) during Cyclone Bola in the Uawa catchment (Marden et al. 1991)

Landslide volume % reduction

Pasture 1631

Pines > 8 yr 217 87

Pines 2–8 yr 1002 39

Pines < 1 yr 2019 –24

Hicks (1989b, c, 1991b) compared the incidence of mass movement in Cyclone Bola (as percent of hillslope eroded) in part of the Waipaoa catchment under pasture, pine, indigenous forest, and scrub. The level of landslide damage was much less under plantation or indigenous forest compared with pasture or reverting scrub. The proportion of uneroded hillslopes increased from 6% under pasture to 16% under pine forest, 27% under scrub, and 33% under indigenous forest. These data were compiled as frequency distribution plots of landslide damage in classes (no mean values), making direct comparison with other studies difficult but clearly illustrating the effect of vegetation in altering the frequency distribution (Fig. 4). Similar results were observed in Taranaki hill country during Cyclone Hilda in 1990 (Hicks 1990).

Figure 4 Frequency distribution of proportion of hillslope eroded under different vegetation types in Cyclone Bola (Hicks 1989a, 1991).

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

Pro

po

rtio

n

% hillslope eroded

pasture

pine forest

native forest

scrub



Fransen and Brownlie (1995) compared landslide density in adjacent catchments in northern Hawke’s Bay hill country before and after afforestation of one of the catchments (Table 18). Tamingimingi was in pasture at all times, while Pakuratahi was afforested in 1971/72. At three different times after afforestation in 1971/72 (1981, 1988, 1994) landslide density under pine trees was >80% lower than on pasture, and this difference increased as the trees grew. These differences were similar for two metrics (percent landslide area, landslide density). In this study the greatest differences did not occur after Cyclone Bola. Before afforestation (1943, 1970) there were differences in landslide density between the two catchments that related at least in part to the amount of scrub in the two catchments (Fransen & Brownlie 1995).

Table 18 Differences in landslide density (landslides km–2) and area of landslides (percent of total catchment area) in Pakuratahi and Tamingimingi catchments (from Fransen & Brownlie 1995)

Pakuratahi Tamingimingi % difference

Tamingimingi/Pakuratahi

% landslides Landslide density % landslides Landslide density % landslides Landslide density

1943 1.37 296 1.18 232 –16 –28

1970 0.07 16 0.16 45 56 64

1981* 0.01 2 0.06 17 83 88

1988 0.14 22 0.91 130 85 83

1994 0.02 7 0.34 75 94 91

*Pakuratahi was afforested in 1971/72

Scrub reversion can also be effective in reducing landsliding. Bergin et al. (1993) report that after Cyclone Bola areas of scrub had 74% less landsliding (measured as area affected by landslides) than pasture, and that the age and density of scrub affected the amount of landsliding – in 8-year-old scrub the amount of landsliding was reduced by 54% compared with pasture, in 16-year-old scrub by 91%, and in 30-year-old scrub there was little landsliding. Similarly, Bergin et al. (1995) found that landslide damage reduced by 65% in 10-year-old scrub and by 90% in 20-year-old scrub.

In the Manawatū-Wanganui February 2004 storm topography, geology and vegetation all influenced the extent of landsliding (Hancox & Wright 2005a, b; Dymond et al. 2006). Despite very high rainfall (>600 mm) in the Tararua and Ruahine ranges there were very few landslides on forested or tussock-covered slopes underlain by greywacke (Hancox & Wright 2005a). The greatest numbers of landslides were in the Tertiary hill country, with forested and scrub-covered slopes much less affected than areas under pasture. Hancox and Wright (2005b) undertook detailed analysis of four areas in the Manawatū-Wanganui hill country with the worst landslide damage. Using a grid-based assessment reflecting density of landsliding, 39.8% of pasture hillslopes were affected by landsliding compared with c. 7.5% for forest (pine, indigenous and scrub) and 17.2% for poplar/willow (Table 19). On average, the percentage area affected by landsliding was 80% lower under forest than under pasture, and 57% lower under poplars and willows. There were substantial spatial variations in the relative effect of vegetation cover, particularly for poplars and willows.



For the same event, Dymond et al. (2006) mapped landslide scars throughout the area affected by the storm, from satellite imagery, as a test of the Highly Erodible Land (HEL) model (Page et al. 2005). This showed that the effect of woody vegetation in reducing the probability of landsliding tended to increase as slope steepness increased (similar results were found in Taranaki; see DeRose 1996; DeRose et al. 1996). Although no average data are presented, Dymond et al. (2006) suggest forest generally reduced landsliding by 90% and scrub by 80% with only minor differences in response to vegetation cover between different Tertiary-aged rock types (e.g. soft mudstone, consolidated and unconsolidated sandstone). The location of landsliding was only moderately well predicted by the HEL model (with a 58% accuracy). Many landslides occurred on slopes below the slope thresholds used for identification of highly erodible land. Hicks and Crippen (2004) compared mapping from the satellite imagery with higher resolution aerial photos and found the satellite imagery assessment slightly underestimated the amount of bare ground. They also compiled estimates of the percent bare ground from mass movement for a range of vegetation types (Table 20). Damage (as percent area of landsliding assessed from 20 randomly chosen sites) to forest (pines or indigenous) was c. 70% less than to pasture, c. 30–40% less where extensive trees were present, and little different where only scattered trees were present.

Table 19 Differences in percent area affected by landslides under different vegetation cover in four study sites of the Manawatū-Wanganui region during the February 2004 storm (from Hancox & Wright 2005b)

% area affected % reduction

Area Pasture Bush/Scrub Pine Poplar/Willo

w Bush/Scrub

Pine

Poplar/Willow

Mangawhero 49.6 4.7 6.5 12.5 91 87 75

Whangaehu 42.5 6.6 9.2 30.3 85 78 29

Turakina 31 6.5 7.1 7.4 79 77 76

Pohangina 35.9 11.6 7.9 18.4 68 78 49

Average 39.8 7.5 7.7 17.2 81 81 57

Table 20 Differences in percent area affected by landslides under different vegetation cover in 20 randomly chosen sites of the Manawatū-Wanganui region during the February 2004 storm (Hicks & Crippen 2004)

Vegetation Mass movement % area % reduction

Pasture 4.9

Pasture with scattered scrub 4.8 2

Pasture with scattered indigenous trees 3.4 31

Pasture with scattered exotic trees 4.9 0

Pasture with extensive scrub 5.1 -4

Pasture with extensive native trees 3.3 33

Pasture with extensive exotic trees 3 39

Conifers 1.6 67

Indigenous scrub 2.6 47



Indigenous forest 1.5 69

There is a very limited amount of published data, all from a single set of studies in the Gisborne area, on the influence of afforestation in reducing earthflow movement. Over 4 years, surface movement rates on forested earthflows were 2–3 orders of magnitude lower than on grassed earthflows (Zhang et al. 1993). However, subsequent monitoring showed recurrent displacements of forested (indigenous and exotic) earthflows can occur under a closed canopy forest with critical failure thresholds leading to the initiation of movement determined by the duration of antecedent soil moisture surplus and elevated pore water pressure (Marden et al. 2008a).

Thompson and Luckman (1993) also comment on the performance of biological erosion control on earthflows (which included both space planting and afforestation), suggesting treatment was ‘successful’ at 63% of sites they assessed so long as trees were closely (< 5–8 m) and extensively planted (>60% of earthflow surface). Phillips et al. (2000) list two other unpublished earthflow studies (at Kaitangata Station and the Waimata valley) where treatments included dense (<5 m spacing) poplar, willow, and pine planting, surface smoothing, graded banks on the earthflows as well as treatment of associated gully erosion (by willow planting and debris dams. Both sites showed little movement during Cyclone Bola although no quantitative data was presented.

There is limited information on the influence of forest vegetation on gully erosion, other than in the Gisborne – East Coast region. Gully erosion in this area is closely associated with deforestation, and afforestation has been extensively used to control gully erosion (Marden et al. 2005, 2008b, 2011, 2012). In 1957, 95% of the total area of gullies was under pasture and <1% under indigenous forest (Marden et al. 2012) – see Table 21. By 1997 the number of gullies had reduced (from 3360 to 2150) but the total area had increased (from 5600 to 7710 ha) and only 50% of the total gully area was under pasture. Reforestation had stabilised 2367 ha of gullies but 1720 ha (mostly formerly pasture) of gullies remained active under pine forest (22% of total gully area in 1997). Reforestation has been focused on the most erodible terrain where gully stabilisation through reforestation is most difficult.

Table 21 Changes in the number and area of gullies in the Gisborne – East Coast region between 1957 and 1997 under different vegetation types (from Marden et al. 2012)

Number of gullies Area of gullies

(ha) %

1957 1997 1957 1997 1957 1997

Pasture 3160 1350 5319 3850 95 50

Indigenous forest 25 75 28 660 0.5 9

Shrubland 175 340 253 1480 4.5 19

Exotic forest 0 385 0 1720 0 22

Total 3360 2150 5600 7710

The ability to stabilise gullies with trees is highly dependent on gully size and shape at the time of planting, with an 80% chance of success (i.e. stabilisation over one forest rotation)



for gullies <1 ha and little chance of success once gullies exceed 10 ha. Thompson and Luckman (1993) also found that treatment of gully erosion was successful at only 42% of sites they examined and it required very closely spaced trees to be highly effective. Where gullies were > 5 m deep space-planting was ineffective.

8.2.2 Influence of closed canopy vegetation on suspended sediment yield

The effects of vegetation cover on sediment yield have been assessed directly by comparing small catchments that have uniform mature vegetation cover, and where other environmental characteristics that affect sediment yield are similar (e.g. geology, rainfall, topography), so that the differences in sediment yield are primarily due to vegetation. Studies that document links between vegetation and sediment yield have typically been undertaken in small catchments where vegetation cover is uniform, and mostly have focused on differences between pasture and mature forest, with more limited data on scrub. There are no measured data on the effect of space-planted trees for soil conservation on sediment yield, and only two studies that have directly characterised the effect of afforestation (Eyles & Fahey 2006; Hughes at al. 2012). It is also noteworthy that most of these studies have relatively short periods of data collection (< 5 years) and therefore have some limitations for defining the long-term differences between different vegetation covers. Table 22 summarises results of studies comparing sediment yield from pasture and mature pine trees, and Table 23 summarises comparisons of pasture and indigenous forest. Blaschke et al. (2008) have previously summarised data on vegetation effects on suspended sediment yield.

Table 22 Comparison of suspended sediment yields (SSY) for pasture and mature pine forest catchments

Area Site Vegetation SSY (t km–2 yr–1)

% reduction

Area (km2)

Geology Basis of comparison Source

Northland/ Waikato

Glenbervie Mature pine forest 46 51 0.63 Greywacke

Mean annual yield calculated from short-term (1–3 yr) record

DM Hicks (1990)

Scotsmans Pasture 95 0.16

Northland Topuni Mature pine forest 27 59 0.88

Crushed mudstone Kokopu Pasture 68 3.08

Rotorua Puruki Mature pine forest 2 93 0.34

Pumice Purutaka Pasture 23 0.23

Nelson Moutere 14 Mature pine forest 4 95 0.07

Moutere Gravel Moutere 5 Pasture 79 0.04

East Otago

Jura Mature pine forest 10 −901

1.92

Schist Storm Mature pine forest 10 1.14

Kintore Pasture 4 2.92

Vollveillerburn Pasture 6 1.63

Hawke’s Bay Pakuratahi Mature pine forest 14 67 3.45 Mudstone and

sandstone Pre-harvest (2.4 yr) Eyles & Fahey

(2006) Tamingimingi Pasture 43 7.95

Central North Island

Puruki Mature pine forest 4 82 0.34 Ash over volcanics Average yield over 3 yr (1982–84)

Dons (1987)

Purutaka Pasture 22 0.1

Nelson

Moutere C2 Pasture 21 0.07 Moutere gravel Average of 2 yr (1986–87)

Smith (1992)

Moutere C3 Pasture with riparian pine 67 −219 0.03

Moutere C4 Pasture with riparian pine 32 −52 0.03

1 Average of two pasture and two pine catchments

Table 23 Comparison of suspended sediment yields (SSY) for pasture and indigenous forest catchments

Area Site Vegetation SSY (t km–2 yr–1)

% reduction

Area (km2)

Geology Basis of comparison Source

Whatawhata

Whakakai Indigenous forest 32 68 3 Greywacke 1 yr (dry) of data (1996–97)

Quinn & Stroud (2002)

Kiripaka Mixed land use 263 −166 2.66

Mangaotama Pasture 99 2.59

Whatawhata

Whakakai Indigenous forest 44 80 3 Greywacke 2 yr of data (Aug 1995 – July 1997)

Dodd et al. (2008a, b)

Kiripaka Mixed land use 803 −274 2.66

Mangaotama Pasture 215 2.59

Whatawhata

Whakakai Indigenous forest 60 38 3 Greywacke 12 yr of data (1999–2011)

Hughes et al. (2012)

Mangaotama Pasture 97 2.59 Tree planting began 2000

Central North Island

Puruwai Indigenous forest 27 −23 0.28 Ash over volcanics 3 yr of data (1982–1984)

Dons (1987)

Purutaka Pasture 22 0.1

Tararua

Ballance Indigenous forest 12 91 0.1 Siltstone, sandstone

1 yr of data Bargh (1977, 1978)

Tuapaka Pasture 140 1.8 Greywacke, alluvium



The strongest influence of forest cover on sediment yield is typically found in small-catchment studies. DM Hicks (1990) reports comparisons of sediment yields from small (0.04–3.08 km2) pasture and mature pine forest catchments from Northland to Otago on a variety of rock types. For a given storm magnitude, forested catchments yield on average 63% less (range 40–78%) sediment than pasture catchments. In terms of mean annual sediment yield, forested catchments typically yielded 50–95% less than pasture catchments, although there was a group of catchments in East Otago where the forested catchments yielded 90% more sediment than pasture catchments. While most studies have found that sediment yield from forest is less than from pasture there are a number of studies where the reverse was found with higher sediment supply in forested catchments (e.g. Dons (1987) native forest catchment; Quinn & Stroud 2002).

Only one study has directly followed the effect of afforestation on sediment yield. Sediment yields have been monitored since 1995 from small (2–3 km2) pasture, indigenous forest and mixed vegetation catchments on weathered greywacke and argillite in the Waikato hill country at Whatawhata. In 2001/02 land use in the pasture catchment was drastically modified with 1.53 km2 afforested with Pinus radiata, 0.07 km2 planted in indigenous trees and shrubs, 1000 poplar poles planted in steep, erosion-prone areas, and riparian fencing within the remaining pasture area. In the most comprehensive analysis of the pasture and indigenous forest catchments Hughes et al. (2012) found that the indigenous forest catchment yielded 38% less sediment than the pasture catchment over a 12-year period (including both before and after land use change) but the differences were greater for the largest storm events (yields were c. 70% lower for the indigenous forest catchment during storm events with >5 year ARI). These results characterise the first 9 years following vegetation conversion, mostly before the trees have reached canopy closure, and Hughes et al. (2012) suggest results may reflect the removal of stock from steep slopes and riparian areas rather than the direct effect of trees. During that time the annual sediment yields have been highly variable in both catchments (coefficients of variation in annual yield are c. 40%) and there is no evidence of a progressive reduction in yield in the treated catchment or any temporal trend in the ratio of suspended sediment yield of the two catchments (Fig. 5). They also suggest that channel widening associated with afforestation may have contributed to increased sediment load from bank erosion and that sediment yield may not decline until a new state of channel morphology is achieved.

A study in nearby Waitetuna catchment in which estimates of storm-based sediment yield were calculated for small (1–3 km2) subcatchments with different vegetation covers (pasture, indigenous forest, pine forest) provided contrasting results (McKergow et al. 2010). During the storm, sediment yield was c. 300% higher in small forested (pine or indigenous) catchments than in a pasture catchment. McKergow et al. (2010) suggest the primary source of sediment was bank erosion and that the low yield from the pasture catchment in part reflected trapping of sediment within ponds and wetlands in that catchment and very active bank erosion in both forested catchments.



Figure 5 Comparison of annual export of suspended sediment from an indigenous forest catchment and a pasture catchment following land use change (from Hughes et al. 2012). Solid line is the ratio of pasture yield to indigenous forest yield.

8.3 Management of riparian margins and bank erosion

Bank erosion can be an important source of sediment because it delivers sediment directly into stream channels. It is likely to be an important source of sediment in New Zealand although there has been very little quantitative research on rates of bank erosion or mitigation of bank erosion (Hughes 2016). In New Zealand a combination of ‘soft’ biological erosion control and ‘hard’ engineering works are used to control bank erosion, and stock exclusion is also used to improve bank stability (see Davies-Colley & Parkyn 2001). The influence of riparian management on bank erosion, and sediment delivery by overland flow, has been reviewed by Parkyn (2004) drawing on the international literature and Hughes (2016) has reviewed local studies of the effect of riparian management on bank erosion.

Hughes (2016) identified only 9 studies in New Zealand that assess the effectiveness of riparian management interventions for reducing stream bank erosion (Table 24). Most used qualitative or semi-quantitative analysis methods and typically compared stream banks in pasture catchments (with unlimited livestock access) with stream banks where livestock were excluded and riparian shrubs/trees were present. He concluded that in the studies reviewed although riparian management appeared to have been mostly effective (in terms of an observable or inferred reduction in bank erosion or decreased suspended sediment concentration/yield) this had only been established semi-quantitatively at best and noted that in the one study where extensive riparian management (stock exclusion) had been carried out in headwater streams there was no observable change in sediment yield over a 12-year period. He also noted that while the studies reviewed demonstrated the benefits of livestock removal, the effects of riparian planting were more equivocal and are only likely to be observable in the long term. For example, Smith (1992) found riparian afforestation with

0.0

0.5

1.0

1.5

2.0

2.5

3.0

0

50

100

150

200

rati

o p

as

ture

/In

dig

en

ou

s f

ore

st

An

nu

al yie

ld (

t km

-2)

Indigenous forest

PastureLand use change



pine trees increased sediment yield (by a factor of 2). She attributed this increase to a lack of riparian ground cover in the afforested catchment allowing ready sediment availability.

Table 24 Summary of results of bank erosion studies in New Zealand (from Hughes 2016)

Study Measurement method Intervention measure/Monitored effectiveness

Smith (1989) Runoff plots Fenced grass riparian buffers. Space for time study. Bank erosion not measured but TSS concentrations in hillslope runoff were c. 90% lower at treated sites.

Smith (1992) Suspended sediment yield

Riparian afforestation with pine trees increased sediment yield due to lack of riparian ground cover.

Williamson et al. (1992)

Semi-quantitative assessment

Riparian zone retired from grazing. Space-for-time substitution study. No evidence that grazed banks >2m wide were more susceptible to erosion than retired banks. For streams >2m wide grazing on wet riparian soils resulted in increased erosion.

Hicks (1992b) Visual assessment Visual comparison of bank erosion at planted sites with nonplanted sites, generally after large flow events. Found that i) tree species used was important; ii) interlocking roots formed by several years growth of dense tree plantings provided superior protection from bank erosion; and iii) removal of dead trees or trees that had fallen into the river reduced the occurrence of bank erosion. Tree planting can reduce bank erosion so long as: appropriate species are used; a sufficient length of the stream is treated; and the plantings are maintained. Where plantings were adequate channel damage was reduced substantially (by >50% in the Waihora), but 40–60% of the plantings were rated as inadequate.

Williamson et al. (1996)

Visual assessment Planted and fenced riparian buffers. Before and after study indicated a decrease of actively eroding banks from 30% to c. 4%.

Boothroyd et al. (2004)

Channel width measurements and visual assessment

Retention of riparian buffers during pine harvesting. Stream channels from clearcut sites were significantly wider than pine forest sites with indigenous vegetation buffers (for both pre- and post-harvest sites).

Parkyn et al. (2003) Qualitative assessment (Pfankuch method)

Planted (with trees) and fenced riparian buffers. Space-for-time substitution study indicated that 3 out of 9 assessment sites (where riparian buffers were established) scored better.

Hughes et al. (2012)

Suspended sediment load estimates

Planted and fenced riparian buffers.* Before and after study. No measured change in sediment loads. No bank measurements made.

Wilcock et al. (2013)

Monthly water quality sampling

Planted and fenced riparian buffers.* Before and after study. Reductions in TSS concentration of between 4 and 11%. Improvements attributed to livestock exclusion. No bank measurements made.

Collins et al. (2013) Turbidity measurements

Planted and fenced riparian buffers. Space for time study. Marginal difference (1.6 NTU) in mean nephelometric turbidity at treated sites. No bank measurements made.

*Other catchment rehabilitation measures also implemented



Hughes (2016) also suggested that for riparian management to be effective, an understanding of bank erosion processes is needed (i.e. the relative role of mass failure, fluvial entrainment and preparatory processes such as wetting/drying and stock trampling). There is likely to be scale-dependence of these processes, with preparatory processes dominating in headwater streams, fluvial entrainment in mid-reaches, and mass failure in lower reaches of catchment. He suggests this can be used to identify the riparian intervention measures that may be most effective in different parts of a catchment. He illustrates this concept using the Williamson et al. (1992) study that found livestock grazing of the riparian areas had a greater effect on narrow (<2 m), low order channels than on wider, higher-order channels where banks were higher and fluvial entrainment was a more important contributor to bank erosion.

Parkyn (2004) reviews research on the efficiency and management of riparian buffer zones in reducing sediment input to streams (including grass filter strips, native and introduced trees, grazing management), including consideration of the effects of riparian buffer width, vegetation type. The removal of stock from streams and riparian areas has clear benefits for reducing sediment input to streams by reducing direct damage to stream banks and reducing soil compaction from stock (e.g. Belsky et al. 1999). Much of the research into the effectiveness of buffer zones for removing contaminants from surface runoff has focussed on effectiveness of vegetated filter strips, usually consisting of rank grass, in trapping sediment using laboratory or field experiments (e.g. Young et al. 1980; Dillaha et al. 1989; Magette et al. 1989; Daniels & Gilliam 1996). These studies typically report trapping efficiencies exceeding 50% for sediment (Table 25) and sediment removal rates increasing non-linearly with buffer width (see Fig. 1). Grass filter strips (5–10 m width) are particularly effective at removing sediment from overland flow (Gharabaghi et al. 2002). Parkyn (2004) notes that the effectiveness of grass buffer strips as filters for sediment is less in steep hilly terrain than in rolling land, as overland flow is concentrated in channelised natural drainage-ways, giving rise to high flow velocities, and buffers therefore need to be wider. She suggests that optimal widths for sediment removal can be highly variable but recommends a minimum of 10–20 m.

Table 25 Sediment removal efficiencies for grass buffer strips (from Parkyn 2004)

Buffer width (m)

Removal (%) Reference

30.5 90 Wong & McCuen (1982)

61 95 Wong & McCuen (1982)

24.4 92 Young et al. 1980

22.9 33 Schellinger & Clausen (1992)

61 80 Horner & Mar (1982)

30 75-80 Lynch et al. (1985)

9.1 85 Ghaffarzadeh et al. (1992)

9.1 84 Dillaha et al. (1989)

4.6 70 Dillaha et al. (1989)



DOC and NIWA prepared a set of guidelines (Collier et al. 1995) that provide practical measures to improve the design of riparian buffer zones to manage bank stability (as well as light climate, water temperature, carbon supply, habitat diversity, flood flows, and contaminants). For contaminants (including sediment), the guidelines can be used to calculate the optimal filter strip width for attenuating overland flow based on the modified CREAMS model (Chemical, Runoff, and Erosion from Agricultural Management Systems). This requires information on topography, slope, soil types for drainage and clay content categories, and hillslope length. Generally, buffer widths need to widen as the slope length, slope angle and clay content of the adjacent land increase and as soil drainage decreases.

Numerous studies have shown that the width of the buffer zone is important in determining trapping efficiency. Duzant et al. (2011) quote studies that observed trapping efficiencies varied between 68% and 98% as the width increased from 2 to 10 m, a 15-m-wide buffer strip was three times more efficient than a 1-m-wide buffer, and trapping efficiency ranged from 53 to 86% with 4.6 m wide buffers and from 70 to 98% with 9.1 m wide buffers. However, they also observe that since most sediment is deposited within the first few metres of the strip, increasing the width beyond 3–5 m has little effect on trapping efficiency. Duzant et al. (2011) also note that the effectiveness of riparian buffers depends on sediment characteristics, with wider buffers required for clayey than sandy sediment, and the architecture of the vegetation making up the buffer. Sturdy, tall, perennial species are considered the most suitable, and short, flexible grasses less so, while an open structure to the buffer strip can enhance rather than protect against erosion by allowing concentration of runoff through gaps in the buffer.

8.4 Management of surface erosion

Maintaining a persistent, complete pasture sward reduces the prevalence and severity of surface erosion processes of wind, sheet wash, and rilling (Hicks 1995). This can be achieved through strategic grazing management in spring to maintain a short, leafy pasture, reducing grazing pressure during drought, cold, or wet conditions to avoid loss of plant cover, use of fertiliser to maintain sward vigour and growth, and establishing improved pastures using seed mixes comprising new cultivars of grass and legume species such as perennial ryegrass (Lolium perenne), cocksfoot (Dactylis glomerata), tall fescue (Festuca arundinacea), white clover (Trifolium repens), and subterranean clover (Trifolium subterraneum) (Basher et al. 2008a). New cultivars are better adapted to the harsh conditions in hill country than earlier available material (Kemp et al. 1999). In a review of previous work, Hicks (1995) concluded that improving pasture could reduce surface erosion by 50–80% compared with levels occurring on land with unimproved pasture.

Grazing management is important in minimising soil loss by surface erosion. Animal treading by cattle has been shown to reduce infiltration rates and increase soil loss, especially where only a short grass canopy remains after grazing. Russell et al. (2001) used rainfall simulation to characterise the effects of variable treading intensity and variable canopy height after grazing on soil loss in hill country pasture. Although they were unable to directly demonstrate that increased treading caused increased soil loss (because of very high variability between plots) they suggested a 20-mm canopy height is needed to minimize the effects of a short-term treading event on soil water infiltration rate and sediment loss and



that management of canopy height is more important than grazing intensity. Sheep grazing can also decrease infiltration rates and increase soil loss. Soil losses were measured in runoff produced by rainfall simulation by Elliott and Carlson (2004). After intensive winter grazing rainfall at high intensity increased suspended sediment concentrations by 13–16 times, and they reduced to background levels within 6 weeks. The effect of grazing was smaller during summer because grazing removed less grass cover. They suggest intensive grazing leading to exposure of bare ground should be avoided in winter to reduce soil loss.

8.5 Summary

Most research on erosion on pastoral farmland has been in highly erodible North Island hill country with soft or crushed rock geology.

Space-planted trees (mainly willows and poplars) and afforestation are the most commonly used practices for controlling erosion on pastoral farmland. They can be highly effective in reducing erosion at hillslope scale, especially by shallow landsliding but also for gullying and earthflows.

The on-site performance of space-planted trees in reducing erosion, mainly by landsliding, has been examined in a small number of studies using quantitative and semi-quantitative methods. There are no published studies on their effect on sediment yield.

A small number of quantitative studies have measured the effectiveness of individual trees or small groups of trees. Key findings include:

Published reductions in landsliding using space-planted trees from quantitative studies can range from 70 to 95%, but often measured or assessed reductions are far less than this because plantings are inadequate.

Individual trees influence the amount of landsliding within a radius of c. 10 m.

Semi-quantitative approaches tend to be based on site or transect and have a broader perspective but limited direct measurement and require an assessment of trends in erosion status. There have been more of these types of studies. Key findings include:

Different combinations of erosion type and extent require different erosion control treatments.

There are feasible treatments for most erosion problems but they sometimes require a combination of biological and structural erosion mitigation. Recommended erosion control treatments are based on type(s) of erosion, risk of erosion, current activity of erosion, size and depth of feature, extent of treatment required, plant spacing required.

The effectiveness of space-planted trees depends on correctly identifying the area needing treatment, the size area treated, establishment of trees at sufficient density, and subsequent maintenance to ensure the survival of trees. Poor survival of trees has been identified as a major constraint to performance of space-planted trees (due to poor pre-treatment of poles, poor planting technique, site factors and stock damage)



There have been a number of studies of above- and below-ground plant growth characteristics (including canopy growth rates, canopy closure modelling, root strength and decline, root growth rates, root morphology, root biomass, root occupancy) since these determine effectiveness of space-planted trees for reducing erosion. This knowledge has been used for recommending appropriate species, planting stock and tree spacings to minimise erosion.

Afforestation is often used to control widespread and severe erosion. Mature, closed-canopy, indigenous or exotic forest (and scrub) typically reduces landsliding by 90%, and has been used to control severe gully erosion and reduce rates of earthflow movement. Trees younger than about 8 years, before canopy closure, are far less effective in reducing erosion. In the Gisborne – East Coast region stabilisation of severe gully erosion by afforestation is highly dependent on gully size and shape at the time of planting, with an 80% chance of success for gullies <1 ha and little chance of success once gullies exceed 10 ha.

Mature, closed-canopy, indigenous or exotic forest also typically reduces sediment yield compared to pasture catchments (typically by 50–90%). However, sediment yields following afforestation are as much influenced by storm event magnitude and other factors (e.g. channel widening) as by land use change.

Bank erosion can be an important source of sediment because it delivers sediment directly into stream channels and is likely to be an important source of sediment although there has been very little quantitative research on rates of bank erosion or mitigation of bank erosion in New Zealand. A combination of ‘soft’ biological erosion control and ‘hard’ engineering works are used to control bank erosion, along with stock exclusion. Research suggests livestock removal from riparian areas improves bank stability, but the effects of riparian planting are more equivocal and are only likely to be observable in the long-term.

Recent research suggests for bank erosion mitigation to be effective an understanding of bank erosion processes is needed to guide which riparian intervention measures may be most effective in different parts of a catchment.

Riparian buffer strips are commonly used to reduce sediment input from surface erosion to streams and have been shown to reduce sediment input by >50%. Factors controlling their effectiveness include width, vegetation composition and structure, contributing slope length and steepness, sediment size, and soil infiltration rate. Most sediment is deposited within the first few metres of a buffer strip, and increasing the width beyond 3–5 m often has little effect on trapping efficiency.

Maintaining a persistent, complete pasture sward reduces the prevalence and severity of surface erosion processes. Grazing management to maintain adequate cover and canopy height is important in minimising soil loss by surface erosion.



9 Assessment of erosion and sediment control performance for forestry

ESC for forestry is focused on the pre-harvest, harvest and post-harvest period when substantial earthworks are carried out and trees are typically clearfelled over large areas. These activities have the potential to generate large amounts of sediment. The impacts of forest harvesting on erosion and sediment generation are well characterised (see reviews in O’Loughlin 1995, 2005; Phillips et al. 2012; Amishev et al. 2014) but rather less is known about the performance of ESC practices in a forestry context.

9.1 Impacts of forest harvesting on erosion and sediment yield

Forest harvesting has a wide range of potential effects on erosion processes and slope stability. These range from direct effects of disturbance (e.g. by soil scraping, road and landing construction, disruption of natural slope hydrological flow paths) to indirect effects such as the change in water balance caused by removal of trees. The major erosion response to forest removal is an increase in landsliding, but there is also an increase in earthflow movement and surface erosion either on the clear-cuts or associated with roads and forest infrastructure (Phillips et al. 2012). As a consequence, sediment yields also increase in the few years following harvesting but then drop to pre-harvest levels, typically within 2–5 years. Post-harvest sediment yields appear to be as much influenced by climatic conditions in the post-harvest period as by infrastructure development and the removal of trees (e.g. Eyles & Fahey 2006). Landslides can also mobilise logging slash in debris flows which can cause severe off-site effects (e.g. Douglas et al. 2011).

The effects of forest harvesting on increasing sediment yield have been well described in a number of studies (summarised in Amishev et al. 2014). In Northland, Hicks and Harmsworth (1989) found sediment yield increases at storm-event scale up to 100 times and that the harvesting period produced 70% of the total sediment load through the entire forest cycle. At Pakuratahi near Napier, sediment yield increased by an order of magnitude during and following forest harvesting but the increase only persisted for 2 years (Fahey & Marden 2000; Fahey et al. 2003; Eyles & Fahey 2006). In three small to moderate-sized (6–24 km2) catchments on weathered granite in Nelson, sediment yields over a 7-year period were on average about five times higher in catchments being harvested than from mature pine catchments (Basher et al. 2011). At storm-event scale, yields were up to 15 times higher in the harvested catchments. The extent of this increase in erosion and sediment yield is dependent on many factors, but it generally increases with storminess (storm frequency, amount and intensity of rainfall), the erodibility of the underlying rock types, and with slope steepness.

The reasons for the decrease in slope stability are related largely to the hydrological and mechanical effects of the removal of trees. This causes soil moisture levels to be higher for longer mainly due to loss of interception capacity of the tree canopy. Also, once trees are removed, the roots slowly decay and the reinforcement they give to soil is reduced and is not fully compensated for by the replanted trees for several years following planting (Fig. 8). This phenomenon is particularly pronounced in ‘softwood’ tree species like Pinus radiata and results in a ‘window of vulnerability’, in which slope stability is reduced and if a significant rainstorm coincides with a recently harvested area, then mass movements are



highly likely to result. The window of vulnerability between rotations for New Zealand plantation forests is estimated to be from 2 to 3 years after harvesting until canopy closure of the next rotation (~year-8) but is species and density dependent.

Figure 6 Changes in forest-vegetation root strength or root reinforcement after timber harvesting (from Phillips et al. 2012).

Forest harvesting involves a considerable amount of earthworks as roads, tracks and landings are constructed, bridges and culverts are installed, and large numbers of trucks and other equipment use the road system. This gives rise to considerable soil disturbance and alterations to natural slope hydrology, and has the potential to cause severe erosion both at the time of construction and post-harvest as roads and landings are decommissioned. The main concerns are the length of roads in steep terrain, the cutting of roads at mid-slope locations, water control, recognition of highly unstable landscape features, overall road design, layout and construction considerations, maintenance during and post-harvest, and life of the road (Phillips et al. 2012).

Many studies worldwide have shown that roads can be the major source of sediment from plantation forests. Most New Zealand studies on harvest-related impacts recognise the significance of roads, tracks and landings as sediment sources (Mosley 1980; Pearce & Hodgkiss 1987’ Fahey & Coker 1989, 1992; Smith & Fenton 1993). In New Zealand roads can increase landslide erosion by about two orders of magnitude compared with undisturbed forest land (Coker & Fahey 1993; Phillips et al. 2012) and can generate significant surface erosion (Fahey & Coker 1989, 1992; Coker et al. 1993). Throughout the 1980s and 1990s a series of studies of forest-road and landing-related erosion were carried out in the Nelson, Marlborough Sounds and Coromandel (reviewed in Fransen et al. 2001). It was concluded:

Infrequent road-related mass movements are major sources of sediment within forests.



Mass-movement erosion rates decline with road age, but may increase to earlier levels when roads are upgraded for harvesting activities.

Road mass-erosion rates are up to three orders of magnitude greater than surface erosion rates.

Poorly constructed landings can also be a major source of sediment (Pearce & Hodgkiss 1987). Coker et al. (1990) concluded that landings could be prone to mass movement in high rainfall areas of New Zealand and suggested a series of measures to reduce the potential for failure. These included measures at the time of landing construction (full benching of fill slopes, end hauling of spoil if necessary, removal of large woody debris before fill is deposited) and post-harvest (managing drainage, reduction of woody over-burden, and replanting of sidecast fill slopes).

Since that time there has been a major effort by the forest industry to better manage the environmental impacts of forestry, with a strong emphasis on infrastructure, which has resulted in an Environmental Code of Practice (New Zealand Forest Owners Association 2007), a much improved forest road engineering manual (Gilmore et al. 2011), and implementation of environmental management systems by large forestry companies. There have been few recent quantitative studies of erosion as a result of roads and landings, but recent storm damage assessments have generally concluded that little of the damage was caused by poorly constructed roads and landings (e.g. Douglas et al. 2011; Philips & Marden 2011; Ngapo 2012) and that most of the erosion occurs on the clear-cuts or by channel scour. In a Coromandel study of immediate post-harvest sediment sources that examined sediment contribution by soil scraping, slopewash and landsliding in a 36-ha catchment, Marden et al. (2006) concluded that sediment generation was dominated by landsliding (72% of total sediment generated) although they did not directly quantify sediment production from road and landing failures.

9.2 Information on effectiveness of ESC practices used in forestry

ESC practices used in forestry listed in regional council guidelines (see Table 2) are largely derived from those used for urban earthworks and infrastructure, with the addition of some practices that are forestry specific (e.g. log corduroying, slash management) and are primarily aimed at managing the impacts of earthworks. Most are derived substantially from the guidelines prepared by Auckland Council (Bryant et al. 2007). The science basis for many of these practices is covered in section 6, but there appear to have been no New Zealand studies that are forestry specific to test that the design criteria are appropriate. Rather they are based on experience of practitioners. Similarly there is a lack of literature internationally on these practices in a forestry context. The ESC guidelines for forestry focus mostly on management of roads and landings with limited guidance on ESC for the clear cuts other than to extract logs away from watercourses and manage slash carefully.

Since early studies of erosion that was caused by roads and landings, and recognition of the importance of good engineering practice to minimise the environmental effects of soil disturbance, there have been considerable improvements in construction of forest infrastructure. Because mechanical slope stabilisation is costly along low-traffic-volume



gravel roads and tracks, erosion prevention can be partly achieved by road location and construction methods that recognise erosion hazard (Phillips et al. 2012).This has been assisted by the preparation of an environmental code of practice for forestry operations (New Zealand Forest Owners Association 2007) and a road engineering manual specifically written for forest roads (Gilmore et al. 2011). The latter covers both roads and landings and provides guidance on methods to reduce erosion that include many of the same techniques listed in regional council guidelines, but with less information on design. There is also a strong focus on water control and avoiding erosion.

Sediment generation and delivery associated with forest road stream crossings has recently been analysed by Brown and Visser (2016) from a field survey of 39 haul road-stream crossings across six regions of New Zealand (Southland, Otago, Canterbury, Nelson, Marlborough, Gisborne, and Waikato). The crossing types included open fords, single culverts, drift decks and battery culverts and the Universal Soil Loss Equation modified for forest land (Dissmeyer & Foster 1984) was used to calculate potential erosion rates. They found that excessive road drainage lengths and slopes were being avoided streams but suggested water control could be enhanced by adding a cross-drain culvert, cut-out or broad-based dip to limit the approach length to 20 m (compared with an observed average of 40 m). Recently constructed road-stream crossings had higher sediment delivery potential due to greater soil disturbance. Calculated potential erosion ranged from 0.01 to 8.7 t ha–1 yr–1, with a median of 0.7 t ha–1 yr–1 and was strongly related to aggregate surfacing on the road surface.

Riparian management is one ESC practice that has received some attention in the New Zealand literature. Boothroyd and Langer (1999) reviewed the international literature on forest harvesting and riparian management guidelines and suggested they had little relevance in New Zealand. Boothroyd et al. (2004) characterised the effect of remnant native riparian buffers in harvested pine plantations in Coromandel streams, with buffer widths ranging from 8 to >25 m. Bank erosion was measured along both banks of the study reach by recording the length of bank slumping and fresh erosion scars, and was expressed as a percentage of total study reach length. In harvested reaches with no riparian buffer bank erosion averaged 31% of stream length compared to 9% where riparian buffers had been left. No published work has looked at the trapping efficiency of riparian buffers for managing sediment associated with landslides and debris flows in steep erodible forested hill country. However, while there is a general acceptance that riparian buffers or streamside management zones have value and can contribute to reductions in sediment input to streams, there is a lack of certainty about the exact benefits of riparian buffers and what sized buffer or set-back was required to be effective (Amishev et al. 2014).

Management of woody residue (slash) is regarded as a key issue because it is the combination of coarse slash left behind after harvesting with soil and regolith generated from landsliding that is a major contributor to the development of the debris flows that cause the most severe off-site and downstream damage (Baillie 1999; Douglas et al. 2011; Ngapo 2012; Amishev et al. 2014). The slash is produced from breakages during felling and extraction, and from trimming and processing operations. Slash has both benefits (surface cover, return of nutrients, improved stream habitat where not excessive) and risks (mobilisation in debris flows, degraded stream habitat where excessive). A balance is needed between retaining slash for its beneficial effects (see Baillie 2011) and avoiding the



adverse effects, especially in large storm events. Concentrations of slash typically occur around streams (as a result of difficulties with directional felling away from streams and hauling across streams), and around landings (where trees are trimmed and processed for loading out). Slash management plans are often required as part of the consent process for forest harvesting. Douglas et al. (2011) provided specific recommendations for the Bay of Plenty region for identifying areas at risk of slash mobilisation (based on geology, soils and slope) and slash management on clear-cuts and downstream.

Little work has been done on the source areas for slash (i.e. the relative contribution of ‘bird’s nests’ associated with landings, slash deposited in streams, and slash that is widely dispersed over clear-cuts) that causes downstream problems (Douglas et al. 2011). In the 80 events analysed by Baillie (1999) it was estimated 48% of the debris was sourced from landslides and 38% from in-stream log jams and debris dams, with the remainder from landing failures and road collapses.

There have been several reviews of effectiveness of forestry best management practices (BMP) in the USA largely based on paired catchment studies. In most cases multiple BMPs were implemented (involving silvicultural options (such as size of clear cuts, size of logs extracted, direction of hauling, burning, spray application), road and track management and stream crossing management) so the results show the cumulative effect of implementing BMPs:

Cristan et al. (2016) review 81 studies of the implementation of forestry best management practices from across the United States, with a focus on timber harvesting, skid trails, roads construction and maintenance, stream crossing, riparian management, and site preparation. Some general conclusions were:

Assessment of BMPs is often region specific. Those from the western USA are probably most relevant to New Zealand

BMPs can minimize erosion and sedimentation but implementation rates and quality are critical to BMP effectiveness for reduction of erosion and sediment yield

Forested riparian buffers are effective in trapping sediment and reducing stream suspended sediment concentrations

Critically important BMP practices for forest roads include proper drainage structures, surfacing, erosion control of cut and fill slopes, traffic control, and closure

Sediment control structures applied to stream crossing approaches can significantly reduce runoff and sediment delivery

BMPs need to be applied during forest operations, not only as a closure measure

Effective skid trail closure practices can include installing waterbars and/or applying slash, mulch, or a combination of mulching and seeding

Improved stream crossings such as portable skidder bridges and temporary culverts can decrease TSS concentrations and turbidity compared with unimproved stream crossing structures.



Anderson and Lockaby (2011) review 17 studies of implementation of forestry best management practices in the south-eastern USA. These include paired catchment studies as well as experimental treatments for road and stream crossing management, riparian zone management, clearcutting and drainage. Results reported include:

Riparian management (buffer strips) was 71–99% effective in trapping sediment transported in concentrated flow from ephemeral swales

No relationship between buffer width (between 7.6 and 30.5 m) and sediment trapping efficiency (surface roughness of the understorey vegetation was more important than width in trapping sediment)

Trapping efficiency in buffer strips was strongly influenced by sediment size. Larger aggregates (>20 μm) were retained in buffers strips as narrow as 10 m, but for smaller aggregates infiltration is important and wider buffer strips are needed (up to 16 m)

Logging roads and skid trails were major sources of sediment but there were variable results from studies comparing road surface treatments and runoff management (ranging from no effect to a 99% reduction in sediment loss).

Edwards and Williard (2010) review 3 paired catchment studies in eastern USA with and without BMPs. BMP efficiencies ranged from 53 to 94% (calculated as a ratio of sediment yield with and without BMPs) during harvest and up to 1 year after harvesting with the reduction in runoff and sediment loss associated with reduced surface erosion.

None of these reviews specifically mention management practices aimed at minimising landslides or debris flows.

9.3 Summary

Earthworks and clearfelled areas have the potential to generate large amounts of sediment by both surface erosion processes and mass movement. Landslides can mobilise logging slash in debris flows and cause severe off-site effects. The effects of forest harvesting on increasing sediment yield, and the consequences of poor road and landing construction and maintenance have been well characterised.

In recent years there has been a major effort by the forest industry to better manage the environmental impacts of forestry, with a strong emphasis on infrastructure engineering for water and sediment control, and careful siting of roads and landings to reduce erosion hazard.

ESC practices promoted for forestry listed in regional council guidelines are largely derived from those used for urban earthworks and infrastructure, with the addition of some practices that are forestry specific and aimed at direct harvesting effects. The latter tend to be general guidelines (e.g. haul away from watercourses, safely dispose of slash).



There appear to have been no New Zealand studies that are forestry specific to test that the ESC design criteria in council guidelines are appropriate. Rather they are based on experience of practitioners.

Riparian buffers can contribute to reductions in sediment input to streams, but there is a lack of certainty about what sized buffer or set-back is required to be effective in plantation forestry, especially to mitigate the effects of landslides.

Reviews of the effectiveness of forestry best management practices in the USA mostly demonstrate the effect of implementing multiple best management practices including silvicultural options, road and track management and stream crossing management. They show that best management practices can minimize erosion and sedimentation (quoted sediment reduction efficiencies of >50%) but implementation rates and quality are critical. Riparian buffers are effective in trapping sediment (efficiencies of 71–99% for sediment in surface runoff) and trapping efficiency is influenced by sediment size and roughness of understorey vegetation. None of the reviews specifically mention management practices aimed at minimising landslides or debris flows.

10 Modelling the effect of erosion and sediment control practices

A number of models have been developed and used in New Zealand for addressing the effects of erosion and sediment control practices on erosion and sediment yield. Modelling approaches included simple empirical models with limited data input requirements (e.g. Universal Soil Loss Equation, New Zealand Empirical Erosion Model), detailed process-based models with high data input requirements (e.g. GLEAMS), and hybrid empirical–process-based models (e.g. SedNetNZ). Elliott and Basher (2011) review the use of erosion models in New Zealand.

10.1 Urban development and road construction projects

This section reviews two approaches adopted in New Zealand for modelling sediment loads associated with urban development and roading projects: the USLE and GLEAMS. In particular, it focuses on the ways in which each method takes account of erosion and sediment control practices in the estimation of sediment loads discharged from construction sites.

10.1.1 USLE

The Universal Soil Loss Equation (USLE) is a widely used method in New Zealand for estimating sediment losses associated with urban development and road construction projects. Developed in the USA from the results of extensive plot-scale experiments (Wischmeier & Smith 1978), the equation estimates the average annual soil loss per unit area (A) as the product of five factors:

A = R.K.LS.C.P



where: R is the erosivity factor, a function of rainfall intensity; K is the soil erodibility factor; LS is the slope length and steepness factor; C is the cover management factor; and P is the supporting practices factor.

Values for each factor are calculated from formulae or obtained from look-up tables. The subsequent Revised Universal Soil Loss Equation (RUSLE) retains these same factors but incorporates the results of additional research in updating the way that each of the factors is determined (Pitt et al. 2007).

In applying the USLE or RUSLE for the estimation of sediment losses from earthworks projects, erosion and sediment control practices are partly taken account of through the C and P factors. For bare earth in the absence of erosion and sediment control measures, both C and P take a value of 1 (or higher where the surface of the soil has been modified, see below). As the effectiveness of measures increases, the values of C and/or P reduce, resulting in a reduction in the estimate of sediment loss.

The C factor is used to represent the performance of mulching, grass cover and other forms of erosion control. Pitt et al. (2007) present values of C developed for the original USLE. These lie in the range 0.02 – 0.2 for a range of unconsolidated mulches, indicating an expected 80–98% reduction in sediment losses associated with these practices. The authors note that these values do not reflect advances in erosion control associated with alternative mulching practices and the development of erosion control blankets. Pitt et al. (2007) also present a range values for the C factor associated with the timing and extent of revegetation of bare earth surfaces. As vegetation passes through various stages from seed preparation to maturity, the value of the C factor drops from 1 (no erosion control) to 0.06 (94% sediment retention).

While the P factor was designed as a way of reflecting the influence of various conservation cropping practices on soil loss from agriculture (Haan et al. 1994), in the context of construction earthworks it can be used to represent the influence of surface roughness on sediment generation. P values of less than 1 are associated with rough, irregular sites and the use of disking to improve soil permeability (Tasman District Council 2014).

C and P can also take values higher than 1 to reflect situations in which management practices may increase soil loss. In relation to construction earthworks, this reflects the scraping, compaction, and sealing of the upper soil surface which reduces infiltration, leading to greater potential to generate sediment runoff. Different authors describe contrasting approaches for representing the influence of these forms of soil modification on sediment generation. Haan et al. (1994) take account of soil compaction through the C factor, while New Zealand sources do it through the P factor (Winter 1998; Tasman District Council 2014)5.

5 Compacted, sealed soils are represented in Haan et al. (1994) by the adoption of C factor values of up to 1.4 and in Winter (1998) and Tasman District Council (2014) by the adoption of a P factor value of 1.2 – 1.32.



While the C and P factors are used to reflect the influence of erosion control and soil modification on earthworks sites, the USLE does not explicitly take account of sediment control practices. Instead, load reduction factors (LRFs) reflecting the performance of different sediment control practices can be applied once the untreated sediment load generated on a site has been calculated. Remembering that the USLE allows the calculation of soil loss per unit area (A), this involves first multiplying the value of A by the site area and a factor representing the sediment delivery ratio (SDR). The SDR represents the proportion of sediment generated on a site that will be transported to sediment control devices and increases with slope steepness and proximity to waterways (Tasman District Council 2014). LRFs are then applied to estimate the sediment load discharged to receiving environments following treatment by sediment control devices.6 Tasman District Council (2014) suggests the adoption of LRFs in the following ranges:

Silt fences and super-silt fences – 0.4 – 0.75

Decanting earth bunds – 0.6 or 0.8 if chemically treated; and

Sediment retention ponds – 0.5 – 0.8 or 0.75 – 0.95 if chemically treated.

Software for running RUSLE calculations is currently available as a second version (RUSLE2) from the US Natural Resources Conservation Service (NRCS)7. This provides for the estimation of both annual average and design storm yields of sediment. C and P factor values are drawn from databases linked to the model based on user inputs on management practices (Pitt et al. 2007).

In examples of applications of the USLE and its derivatives, Parker et al. (1995) adopted a GIS-based approach which took into account variations in the C factor, while Wachal et al. (2009) used RUSLE2 to assess the performance of erosion and sediment controls associated with the development of natural gas wells in Texas. Annual average and design storm yields were modelled for a range of cover management and support practices in association with three soil types and three slopes. The most effective practice modelled was sediment retention ponds, with annual average sediment yields 75–93% lower than from unprotected sites, depending on slope and soil type. Seeding was the least effective practice, with sediment yields 50–71% lower than from unprotected sites, again depending on slope and soil type. Sediment removal efficiencies were generally estimated to decline as the return period of storm events increased from 1 to 10 years, but the reductions were relatively minor. The efficiency of ponds, for instance, was modelled to fall from 78% to 73% (Wachal 2009).

10.1.2 GLEAMS

A number of studies to assess sediment generation associated with major New Zealand construction projects have used the Groundwater Loading Effects of Agricultural Management Systems (GLEAMS) model. Recent examples include the Waterview

6 Treated load = untreated load x (1-LRF)

7 http://fargo.nserl.purdue.edu/rusle2_dataweb/ and USDA ARS (2008)

http://fargo.nserl.purdue.edu/rusle2_dataweb/



Connection and Puhoi-to-Warkworth Road of National Significance (RNS) state highway projects, both in the Auckland region (Harper & Semadeni-Davies 2010; Harper et al. 2013)

GLEAMS is a physically based mathematical model developed for continuous simulation (at a daily time step) of surface runoff and sediment losses from the land on a field scale (Knisel 1993). The procedure for deriving sediment loads using GLEAMS field-scale predictions involves dividing a given study area (usually a catchment) into a number of land “cells”, each assumed to be of uniform land-cover, slope and soil type. The GLEAMS model uses a long-term climate record (rainfall, temperature and solar radiation) together with parameter values reflecting land-cover, slope and soil type to calculate a daily series of surface runoff and sediment yields for each cell. The results are then aggregated for the study area as a whole.

Earthworks sites are modelled as one of a number of ‘bare earth’ land-cover classes. These classes have parameter values representing the absence of vegetation or other cover protection, resulting in the generation of markedly higher sediment yields than vegetated (or impervious) covers, holding all else equal. Stabilization of areas of bare earth is represented in GLEAMS by the selection of cover classes reflecting the post-construction land cover, such as grassland or impervious covers.

Two approaches for modelling the influence of erosion and sediment control practices in GLEAMS studies have been adopted. The first of these involves the calculation of a treated sediment load by applying an LRF to the untreated load calculated by GLEAMS, in the same way as described above in relation to the USLE. In the Puhoi-to-Warkworth RNS study, for instance, erosion and sediment control practices were represented by the adoption of the following LRFs (Harper et al. 2013):

Mulching – 0.85

Super-silt fences – 0.5 – 0.8

Decanting earth bunds – 0.6 – 0.9

Chemically-treated sediment retention ponds – 0.65 – 0.95.

These LRFs were specified by the project’s erosion and control specialists, taking account of the results of New Zealand monitoring described in Section 1 of this report. The ranges of values for the three sediment control devices reflected variations in performance with storm events with average return intervals in the range 2 to 50-years, based on evidence of deterioration in performance as storm magnitude increases.

The second approach is the application of a GLEAMS post-processing module simulating a sediment retention pond. The untreated sediment loads estimated by GLEAMS are partitioned into ten particle size classes according to soil type. The model calculates the proportion of the influent sediments that are removed by the pond according to their distribution among the different size classes and the respective settling speeds. Holding all else equal, a pond is modelled to remove a higher proportion of the load of a relatively coarse-grained soil than a relatively fine-grained soil. Where a sediment pond is chemically treated, the settling speeds of the smaller sediment particles entering the pond can be adjusted (increased), to reflect their aggregation as a result of flocculation. This approach



was adopted in the estimation of sediment loads associated with the Waterview Connection project, with the application of the pond module resulting in a mean sediment removal efficiency of 94% (Harper & Semadeni-Davies 2010).

Compared with the relative simplicity of the USLE, the use of GLEAMS allows more sophistication in the assessment of sediment losses but involves more effort and requires access to appropriate data sources, including long-term time series of climate data inputs. However, this use of a long-term series allows the model to capture a suitable range of climate conditions that enable estimation of runoff and sediment loads associated with storm events of specified return periods, as well as mean annual loads. It also results in the generation of a time-series of model outputs (runoff and sediment load) at a daily time step than can be used as inputs to receiving environment models, as has been the case in a number of studies to predict the effects of urban development on harbour sedimentation and sediment quality (e.g. Parshotam 2008; Parshotam & Wadwha 2008).

The use of GLEAMS also allows the staging of works to be represented through changing the land cover type applying to each cell in a given study area as a project progresses. Similarly, the model allows investigation of alternative scenarios with differing extents of open earthworks areas and/or variations in the potential sediment contribution of high risk areas (e.g. steeper zones). Based on this versatility, its foundation in physical processes and its track record in New Zealand studies, the use of GLEAMS has proven to be a well-suited method for estimating sediment losses from earthworks. The model is probably best suited to applications on projects that are large in scale and/or are located in the catchments of sensitive receiving environments, as has been the case to date.

10.1.3 WEPP

The process-based WEPP model has been used to predict sediment yield during development of urban subdivisions by Morton (1996) and Winter (1998). In both cases the emphasis was on comparing predicted sediment discharge using WEPP and the USLE rather than assessing the impact of erosion and sediment control practices.

10.2 Pastoral farming

The New Zealand Empirical Erosion Model (NZeem®) was developed to address the effects of soil conservation and land use scenarios on erosion and sediment yield by Dymond et al. (2010). Erosion is modelled as:

E = aCRb

where E = long-term average annual erosion rate (t km-2 year-1) R = mean annual rainfall (mm year–1) C = 1 for tall, closed canopy woody vegetation, 10 for non-woody vegetation (herbaceous vegetation and bare ground) a = the erosion terrain coefficient b = 2



The model also incorporates a sediment delivery term defined by connection to a stream network derived from DEM analysis. NZeem® was calibrated using estimates of suspended sediment yield for New Zealand rivers. NZeem® assumes a factor-of-10 reduction in erosion rates for land areas covered in trees. This model has been used to assess:

the effects of different land-use scenarios in the Motueka catchment (pre-human vegetation, present land use, intensive land use with all production forestry converted to pastoral agriculture). NZeem® predicted catchment sediment yields of 150 000, 320 000, and 750 000 t yr–1 respectively.

national trends in erosion with trends in land cover between 1975 (c. 95 m t yr–1) and 2002 (c. 80 m t yr–1).

strategies for implementing on-farm sediment control measures on sediment loads in the Manawatū River catchment. Dymond et al. (2010) assumed that a fully implemented Whole Farm Plan (WFP) would reduce erosion by 70%. They analysed the impact of the current WFP implementation plan to predict its effect on mean sediment discharge (t yr–1) of the Manawatū, as well as water clarity (Ausseil & Dymond 2008). The model predicted that after maturation of the soil conservation plantings the mean sediment discharge of the Manawatū River would reduce by 48% (from 3.1 to 1.6 million t yr–1). This reduction would approximately double water clarity and would move the median water clarity of the middle Manawatū from 0.9 to approximately 1.8 m.

Douglas et al. (2008) and McIvor et al. (2011) proposed an approach based on NZeem® to estimating the effects of conservation works on sediment export from a farm. Underpinning this approach is the need to collect data (using aerial photography) during implementation of a works programme on vegetation type and canopy cover, the area of each works (at the individual site level) and the age of the vegetation. NZeem® is used to calculate the reduction in sediment export resulting from WFP implementation assuming:

Afforestation (pine or indigenous) reduces erosion by 90% once trees are mature (>20 years). For trees <20 years, a maturity factor (Mf) is applied using the age of the trees (Agef) and assuming a linear reduction in erosion through time:

𝑀𝑓 = 𝐴𝑔𝑒𝑓

20

Mature (>15 years) space-planted trees reduce (landslide and gully) erosion by 70%. For trees <15 years, a maturity factor (Mp) is applied assuming a linear reduction in erosion through time, and is weighted by the fraction of WFP works that have been implemented (f):

𝑀𝑝 = 𝑓 × 𝐴𝑔𝑒𝑝

15

Vegetation on retired land reverting to indigenous forest is assumed to be at one of five phases with maturity factor Mr: reverting pasture Mr = 0.0; incomplete scrub canopy closure (early stage) Mr = 0.1; incomplete scrub canopy closure (intermediate



stage 3 years) Mr = 0.5; complete scrub canopy closure Mr = 0.9 (after 5 years); and indigenous forest Mr = 1.0

Sediment export (S) from a farm where a WFP is being implemented is calculated as

𝑆 = 𝑒𝑓(1 − 0.9�̅�𝑓)𝐴𝑓 + 𝑒𝑝(1 − 0.7�̅�𝑝)𝐴𝑝 + 𝑒𝑟(1 − 0.9�̅�𝑟)𝐴𝑟

where �̅� is the mean maturity factor for exotic forestry (�̅�𝑓), spaced planted tree (�̅�𝑝)

and indigenous forestry sites (�̅�𝑟)

𝐴 is the area of exotic forestry (𝐴𝑓), spaced planted trees (𝐴𝑝) and indigenous forestry

sites (𝐴𝑟)

𝑒 is the mean erosion rate under pasture (from NZeem®) of the area of exotic forestry (𝑒𝑓), spaced planted tree (𝑒𝑝) and indigenous forestry sites (𝑒𝑟)

The same approach has also been implemented using the hybrid empirical-process model SedNetNZ (Dymond et al. 2014, 2016) to assess the effect of implementation of WFPs in the Manawatū catchment. Initially, it was applied by Schierlitz et al. (2006) using the assumption that a fully implemented WFP would reduce erosion by 70%. The model was used to test the effects of different scenarios for implementation of WFPs (numbers of WFPs and different potential scenarios for selection of areas needing WFPs) and demonstrated the effect of targeting WFPs to the most erosion prone terrain. More recently, Dymond et al. (2016) used factors for effectiveness that varied with the type of work implemented (Table 26). This analysis provided information on the effect of WFPs on different erosion processes and different subcatchments of the Manawatū. This analysis predicted that current WFP initiatives would reduce the sediment load of the Manawatū by 34% (from 2.1 million t yr–1 to 1.4 million t yr–1).

Table 26 Effectiveness of soil conservation works and the required time to reach maturity (from Dymond et al. 2016)

Soil conservation work Maturity (yrs) Effectiveness (%)

Afforestation 10 90

Bush retirement 10 90

Riparian retirement 2 80

Space-planted trees 15 70

Gully tree planting 15 70

Sediment traps 1 70

Drains 1 70

11 Current and proposed research projects

There is a limited amount of current research on the effectiveness of erosion and sediment control practices including:



A proposed holistic review of sediment retention pond performance in the Auckland region including research into water temperature (both in the main pond body and in the receiving environment), flocculant types and dosing rates, the hydraulic properties of the SRP’s, pH levels for the SRP’s and soils in the catchments, receiving environment effects of sediment and pH, and the cultural aspects of sediment in relation to traditional Mana Whenua values and mahinga kai. The potential of high sediment transportation during low frequency, high magnitude events will be investigated with attention to be paid to how improvements to SRP design and operation may increase retention capability. Field trials and monitoring will be undertaken to test hypotheses and the effect different soil structures may have on SRP effectiveness. Led by ClaytonFordham Ltd and Auckland Council.

Monitoring undertaken as part of the Transmission Gully Motorway project (TGP). The consents issued for the TGP project included a requirement to monitor the effects that the construction may have on streams that cross the construction area. Seven stream sites have been instrumented for monitoring of flow and turbidity and a further 3 for turbidity only. In addition, the performance of erosion and sediment control practices will be monitored, including sediment retention ponds and decanting earth bunds. Undertaken by NIWA on behalf of the Wellington Gateway Partnership with oversight from a Sediment Peer Review Panel.

A 3-year Sustainable Farming Fund project “Don’t muddy the water” aimed at quantifying the effectiveness of sediment control on cultivated land. At present measurements are being made of SRP performance in croplands at Pukekohe (aimed at verifying the efficiency of SRPs based on 0.5% storage volume and use of decanting devices) and efficiency of grass buffer strips at Levin. Further ESC practices are likely to be assessed in the next 2 years (cover crops, wheel track diking). Undertaken by NIWA and Landcare Research on behalf of Horticulture NZ.

Measurement of the above-ground (DBH, root collar diameter, plant height, canopy spread, plant survival) and below-ground (root bole, root diameter-size classes and total root length, root depth, root spread) characteristics of mānuka growing on erosion prone sites in the East Coast and Hawke’s Bay regions. The primary aim is to collect plant growth metrics for modelling the influence of planting density on the erosion control effectiveness of space-planted mānuka during their formative years following establishment. A secondary aim is to use the growth metrics to calculate carbon sequestration rates. Undertaken by Landcare Research for MPI.

Root reinforcement and slope stability modelling as part of both the Growing Confidence in the Future of Forestry and Sustainable Land Use Research Initiative (SLURI) programmes. This work aims to quantify the effectiveness of soil conservation plants through measurements of root distribution, quantify the effects of tree management (e.g. pollarding) on root growth, and measure how soil texture affects root growth. Measurements of root architecture and growth, and root-soil shear strength will be incorporated into an empirical model that predicts the effectiveness of different soil conservation plants in mitigating landslides. Undertaken by AgResearch, Plant and Food Research and Landcare Research.

A current project to identify causes of concentrated sediment runoff to streams associated with forest harvesting operations in New Zealand. The focus is on



identifying where stream crossings and drainage structures associated with permanent and temporary roads, hillslope and landing failures, soil disturbance by machine tracking or scarification of hillslopes from hauler extraction cause sediment runoff to streams. The goal is to assess the location and importance of these sites to develop targeted best management practices to reduce sediment runoff. Field surveys are being undertaken in Otago, Southland, Nelson-Marlborough, Gisborne, Auckland, Wairarapa, and Northland. Undertaken by University of Canterbury with funding from Forest Growers Levy Trust.

A Sustainable Farming Fund project to mitigate the costs to farmers of severe rainstorms and to provide recommendations for tree planting on farms and a decision support tool based on LUC classes. Undertaken by Plant and Food Research.

As part of the Whatawhata Research Station ICM programme, one sub-catchment was substantially afforested (c. 50%) and the effect on suspended sediment yield was monitored. Monitoring of suspended sediment by NIWA continues and a current research proposal includes monitoring of sediment export from ephemeral channels on steep land and an assessment of the trapping performance of vegetative barriers.

A research proposal has been submitted to MBIE by Landcare Research (in collaboration with Massey University and iwi) titled “Next-generation erosion and sediment management”. It aims to continue development of the SedNetNZ model and provide improved information on erosion mitigation performance.

12 Conclusions

ESC practices used in New Zealand are based on a set of principles for control of different erosion processes:

Runoff-generated erosion is managed by runoff control to reduce water velocity and to separate clean water and dirty water; erosion control to reduce sediment generation; and sediment control to manage sediment movement offsite

Mass movement erosion is controlled by practices that influence slope hydrology and/or soil strength

Streambank erosion is controlled by practices that reduce hydraulic scour, or increase bank strength and resistance to erosion

Wind erosion is controlled by practices that reduce soil erodibility, increase soil moisture content or reduce wind erosivity.

Based on both local studies, in Auckland, and overseas studies, ESC practices can be highly effective in reducing the generation of sediment and its discharge from earthworks construction sites (often by an order of magnitude). These have established both the performance of a range of ESC practices and the factors that determine performance. Particular attention has been directed at methods to retain clay size sediment using chemical treatment.

ESC practices to control water and wind erosion on cropland have been little studied in New Zealand but there is an extensive body of overseas literature. In New Zealand



there has been a focus on the importance of compacted areas (especially wheel tracks in row crops) generating runoff and sediment.

Research on erosion on pastoral farmland has focused on the performance of space-planted trees and afforestation in reducing landsliding, gully erosion, and earthflow movement. This has established both performance effectiveness and developed recommended treatment options for biological erosion control. There has been limited research on bank erosion control or the performance of riparian buffer strips.

Earthworks and clearfelled areas of plantation forests can generate large amounts of sediment by both surface erosion processes and mass movement. Both regional council and industry guidelines for ESC focus on earthworks using similar practices to those employed on other construction sites but there are no forestry-specific New Zealand studies to establish performance of these practices, or to determine the relative contribution of sediment from infrastructure (primarily by runoff driven processes) and from the clear-cuts (primarily by mass movement processes). Recommended ESC practices are largely based on experience of practitioners.

13 Acknowledgements

The Ministry for the Environment provided funding to Tasman District Council from the Community Environment Fund Round 6 for this work. We are grateful to all those individuals and organisations who took the time to discuss the use of ESC with us. The authors wish to thank Graeme Ridley, Earl Shaver, and Todd Property Ltd for providing copies of reports and monitoring data relating to a number of the New Zealand studies included in this review. Ian Lynn and Anne Austin commented on an earlier draft of this report.

14 References

Abu-Zreig M, Rudra RP, Lalonde MN, Whiteley HR, Kaushik NK 2004. Experimental investigation of runoff reduction and sediment removal by vegetated filter strips. Hydrological Processes 18: 2029–2037.

Alliaume F, Rissing WAH, Tittonell P, Jorge G, Dogliotti S 2014. Reduced tillage and cover crops improve water capture and reduce erosion of fine textured soils in raised bed tomato systems. Agriculture, Ecosystems and Environment 183: 127–137.

Amishev D, Basher L, Phillips C, Hill S, Marden M, Bloomberg M, Moore J 2014. New forest management approaches to steep hills. MPI Technical Paper 2014/39. Prepared for MPI by Scion, Landcare Research, and University of Canterbury.

Anderson CJ, Lockaby BG 2011.The effectiveness of forestry best management practices for sediment control in the Southeastern United States: a literature review. Southern Journal of Applied Forestry 35: 170–177.

Armstrong JJ, Wall GJ 1990. Quantitative evaluation of the effectiveness of erosion control material. Ontario Ministry of Transportation Report MAT-90-10.

http://www.ingentaconnect.com/content/saf/sjaf

http://www.ingentaconnect.com/content/saf/sjaf



Auckland Regional Council 1996. The environmental impacts of accelerated erosion and sedimentation. ARC Technical Publication No. 69.

Auckland Regional Council 1999. Erosion and sediment control guidelines for land disturbing activities in the Auckland Region. ARC Technical Publication No. 90.

Auckland Regional Council (ARC) 2000. The effectiveness of mulching earthworked surfaces. ARC Landfacts S04.

Auckland Regional Council (ARC) 2004. the use of flocculants and coagulants to aid the settlement of suspended sediment in earthworks runoff: trials, methodology and design. Auckland Council Technical Publication 227 (draft).

Auckland Regional Council undated. Estimating sediment yield using the Universal Soil Loss Equation (USLE). Landfacts S05.

Ausseil A-G.E, Dymond JR 2008. Estimating the spatial distribution of sediment concentration in the Manawatu river, New Zealand, under different land-use scenarios. IAHS Publication 325: 502–509.

Australian and New Zealand Environment and Conservation Council 2000. Australian and New Zealand guidelines for fresh and marine water quality. Canberra, Australian and New Zealand Environment and Conservation Council and Agriculture and Resource Management Council of Australia and New Zealand.

Babington and Associates (2004) Ltd 2008. Decanting earth bund efficiency trial, April 2008 with amendments September 2008. Prepared for Auckland Regional Council.

Bailey A, Deasy C, Quinton J, Silgram M, Jackson B, Stevens C 2013. Determining the cost of in-field mitigation options to reduce sediment and phosphorus loss. Land Use Policy 30: 234–242.

Baillie BR 1999. Management of logging slash in streams – results of a survey. Project Report PR85, LIRO, Rotorua.

Baillie BR 2011.The physical and biological function of wood in New Zealand’s forested stream ecosystems. PhD thesis, University of Waikato, Hamilton.

Barber A 2014. Erosion and sediment control guidelines for vegetable production: good management practices, version 1.1. Prepared for Horticulture New Zealand by Agrilink, Kumeu.

Barber A, Wharfe L 2010. Code of Practice for commercial vegetable growing in the Horizons Region: best management practices for nutrient management and minimising erosion on cultivated land, version 2010/1. Prepared for Horticulture New Zealand by Agrilink, Kumeu.

Bargh BJ 1977. Output of water, suspended sediment and phosphorus and nitrogen from a small forested catchment. New Zealand Journal of Forestry Science 7: 162–171.



Bargh BJ 1978. Output of water, suspended sediment and phosphorus and nitrogen from a small agricultural catchment. New Zealand Journal of Agricultural Research 21: 29–38.

Barrett ME, Kearney JE, McCoy TG, Malina JF 1995a. An evaluation of the use and effectiveness of temporary sediment controls, CRWR Online Report 95-6. Austin, TX, The University of Texas at Austin, Center for Research in Water Resources,. https://www.crwr.utexas.edu/reports/pdf/1995/rpt95-6.pdf. (accessed 4 Dec 2015).

Barrett ME, Zuber RD, Collins ER, Malina JF, Charbeneau RJ, Ward GH 1995b. A review and evaluation of literature pertaining to the quantity and control of pollution from highway runoff and construction. CRWR Online Report 95-5. Austin TX, The University of Texas at Austin, Center for Research in Water Resources,. https://www.crwr.utexas.edu/reports/pdf/1995/rpt95-5.pdf (accessed 9 Dec 2015).

Barrett ME, Malina JF, Charbeneau RJ 1998. An evaluation of geotextiles for temporary sediment control. Water Environment Research 70: 283–290.

Basher LR 2013. Erosion processes and their control in New Zealand. In: Dymond J ed. Ecosystem services in New Zealand – conditions and trends. Lincoln, Manaaki Whenua Press. Pp. 363–374.

Basher LR, Botha N, Dodd MB, Douglas GB, Lynn I, Marden M, McIvor IR, Smith W 2008a. Hill country erosion: a review of knowledge on erosion processes, mitigation options, social learning and their long-term effectiveness in the management of hill country erosion. Landcare Research Contract Report LC0708/081 for the Ministry of Agriculture and Forestry.

Basher LR, Botha N, Douglas GB, Kingsley-Smith H, Phillips C, Smith W 2008b. Hill country erosion: outcome of a workshop to prioritise research requirements. Landcare Research Contract Report LC0708/112.

Basher LR, Hicks DM, Clapp B, Hewitt T 2011. Sediment yield responses to forest harvesting and large storm events, Motueka River, New Zealand. New Zealand Journal of Marine and Freshwater Research 45: 333–356.

Basher LR, Hicks DM, Ross CW, Handyside B 1997. Erosion and sediment transport from the market gardening lands at Pukekohe, Auckland, New Zealand. Journal of Hydrology (NZ) 36: 73–95.

Basher LR, Painter DJ 1997. Wind erosion in New Zealand. Proceedings of the International Symposium on Wind Erosion, Manhattan, Kansas, 3–5 June 1997, USDA-ARS, Manhattan (http://www.weru.ksu.edu/symposium/proceed.htm).

Basher LR, Ross CW 2001. Role of wheel tracks in runoff and sediment generation under vegetable production on clay loam, strongly structured soils at Pukekohe, New Zealand. Soil and Tillage Research 62: 117–130.

https://www.crwr.utexas.edu/reports/pdf/1995/rpt95-6.pdf

https://www.crwr.utexas.edu/reports/pdf/1995/rpt95-5.pdf



Basher LR, Ross CW 2002. Soil erosion rates under intensive vegetable production on clay loam, strongly structured soils at Pukekohe, New Zealand. Australian Journal of Soil Research 40: 947–961.

Basher LR, Ross CW, Dando J. 2004: Impacts of carrot growing on volcanic ash soils in the Ohakune area, New Zealand. Australian Journal of Soil Research 42: 259–72.

Basher LR, Thompson T 1999. Erosion at Pukekohe during the storm of 21 January 1999. Contract report for the Franklin Sustainability Project, Landcare Research Contract Report LC9899/96.

Belsky AJ, Matzke A, Uselman S 1999. Survey of livestock influences on stream and riparian ecosystems in the western United States. Journal of Soil and Water Conservation 54: 419–431.

Benik SR, Wilson BN, Biesboer DD, Hansen B, Stenlund D 2003. Performance of erosion control products on a highway embankment. Transactions of the American Society of Agricultural Engineers 46: 1113–1119.

Bergin DO, Kimberley MO, Marden M 1993. How soon does regenerating scrub control erosion? New Zealand Forestry, August 1993: 38–40.

Bergin DR, Kimberley MO, Marden M 1995. Protective value of regenerating tea-tree stands on erosion-prone hill country, East Coast, North Island, New Zealand. New Zealand Journal of Forestry Science 25: 3–19.

Bidelspach DA, Jarrett AR, Vaughan BT 2004. Influence of increasing the delay time between the inflow and outflow hydrographs of a sediment basin. Transactions of the American Society of Agricultural Engineers 47: 439–444.

Blaschke P, Hicks D, Meister A 2008. Quantification of the flood and erosion reduction co-benefits, and co-costs, of climate change mitigation measures in New Zealand. Wellington, Ministry for the Environment.

Boothroyd IKG, Langer ER 1999. Forest harvesting and riparian management guidelines. NIWA Technical Report 56.

Boothroyd IKG, Quinn JM, Langer ER, Costley KJ, Steward G 2004. Riparian buffers mitigate effects of pine plantation logging on New Zealand streams: 1. Riparian vegetation structure, stream geomorphology and periphyton. Forest Ecology and Management 194: 199–213.

Borin M, Vianello M, Morari F, Zanin G 2005. Effectiveness of buffer strips in removing pollutants in runoff from a cultivated field in North-East Italy. Agriculture, Ecosystems and Environment 105: 101–114.

Brown K, Visser R 2016. Evaluating the potential for sediment delivery at forest road-stream crossings in New Zealand. New Zealand Journal of Forestry 61: 31–38.

http://www.sciencedirect.com/science/journal/01678809

http://www.sciencedirect.com/science/journal/01678809



Brown MJ, Bondurant JA, Brockway 1981. Ponding surface drainage water for sediment and phosphorus removal. Transactions of the American Society of Agricultural Engineers 24: 1478–1481.

Bryant S, Handyside B, Dunphy M 2007. Forestry operations in the Auckland Region: a guideline for erosion and sediment control. Technical Publication 223. Auckland, Auckland Regional Council.

Cameron D 1991. Effect of soil conservation tree plantings on damage sustained by the Whareama catchment during the storm of 8–11 April 1991. Broadsheet 9: 25–29.

Chepil WS, Woodruff NP 1963. The physics of wind erosion and its control. Advances in Agronomy 15: 211–302.

Coker RJ, Fahey BD 1993. Road-related mass movement in weathered granite, Golden Downs and Motueka forests, New Zealand: a note. Journal of Hydrology (NZ) 31: 65–70.

Coker RJ, Fahey BD, Payne JJ 1993. Fine sediment production from truck traffic, Queen Charlotte Forest, Marlborough Sounds, New Zealand. Journal of Hydrology (NZ) 31: 56–64.

Coker RJ, Pearce AJ, Fahey BD 1990. Prediction and prevention of forest landing failures in high-intensity areas of northern New Zealand. International Symposium on Research Needs and Applications to reduce Erosion and Sedimentation in Tropical Steeplands, Fiji, June 11–15, 1991. IAHS Publication 192: 311–317.

Collier KJ, Cooper AB, Davies-Colley RJ, Rutherford JC, Smith CM, Williamson RB 1995. Managing riparian zones: a contribution to protecting New Zealand's rivers and streams. Volume 2: Guidelines. Wellington, New Zealand, Department of Conservation

Collins KE, Doscher C, Rennie HG, Ross JG. 2013. The effectiveness of riparian ‘restoration’ on water quality – a case study of lowland streams in Canterbury, New Zealand. Restoration Ecology 21: 40–48.

Committee of Inquiry 1939. Maintenance of vegetative cover in New Zealand, with special reference to land erosion. Department of Scientific and Industrial Research Bulletin No. 77.

Cresswell HP, Painter DJ, Cameron KC 1991. Tillage and water content effects on surface soil physical properties. Soil and Tillage Research 21: 67–83.

Cristan R, Aust WM, Bolding MC, Barrett SM, Munsell JF, Schilling E 2016. Effectiveness of forestry best management practices in the United States: literature review. Forest Ecology and Management 360: 133–151.

Crozier MJ 2005. Multiple-occurrence regional landslide events in New Zealand: hazard management issues. Landslides 2: 247–256.



Cumberland KB 1944. Soil erosion in New Zealand: a geographic reconnaissance. Christchurch, Whitcombe and Tombs.

Daniels RB, Gilliam JW 1996. Sediment and chemical load reduction by grass and riparian filters. Soil Science Society of America Journal 60: 246–251.

Davies-Colley R. and Parkyn S 2001. Effects of livestock on streams and potential benefits of riparian management. Issues and Options in the Auckland Region. Prepared by NIWA for Auckland Regional Council. ARC Technical Publication 351.

Deasy C, Quinton J, Silgram M, Staotae C, Jackson R, Stevens CJ, Bailey AP 2010. Mitigation options for phosphorus and sediment (MOPS): reducing pollution in run-off from arable fields. The Environmentalist 108: 12–17.

DeRose RC 1996. Relationship between slope morphology, regolith depth, and the incidence of shallow landslides in eastern Taranaki hill country. Zeitschrift für Geomorphologie NF Supplement 105: 49–60.

DeRose RC, Trustrum NA, Blaschke PM 1996. Post-deforestation soil loss from steepland hillslopes in Taranaki, New Zealand. Earth Surface Processes and Landforms 18: 131–144.

Dillaha TA, Reneau RB, Mostaghimi S, Lee D 1989. Vegetative filter strips for agricultural non-point source pollution control. Transactions of the American Society of Agricultural Engineers 32: 513–519.

Dissmeyer, G.E. and Foster, G.R. 1984. A guide for predicting sheet and rill erosion on forest land. Atlanta, GA, United States Department of Agriculture, Forest Service, Southern Region Technical Publication R8-TP 6..

Dodd MB, Quinn JM, Thorrold BS, Parminter TG, Wedderburn ME 2008a. Improving the economic and environmental performance of a New Zealand hill country farm catchment: 3. Short-term outcomes of land-use change. New Zealand Journal of Agricultural Research 51: 155–169.

Dodd MB, Thorrold BS, Quinn JM, Parminter TG, Wedderburn ME 2008b. Improving the economic and environmental performance of a New Zealand hill country farm catchment: 1. Goal development and assessment of current performance. New Zealand Journal of Agricultural Research 51: 155–169.

Dons A 1987. Hydrology and sediment regime of a pasture, native forest, and pine forest catchment in the central North Island, New Zealand. New Zealand Journal of Forestry Science 17: 161–178.

Dosskey MG 2001. Toward quantifying water pollution abatement in response to installing buffers on crop land. Environmental Management 28: 577–598.

http://www.tandfonline.com/action/doSearch?action=runSearch&type=advanced&result=true&prevSearch=%2Bauthorsfield%3A(Dodd%2C+M.+B.)

http://www.tandfonline.com/action/doSearch?action=runSearch&type=advanced&result=true&prevSearch=%2Bauthorsfield%3A(Quinn%2C+J.+M.)

http://www.tandfonline.com/action/doSearch?action=runSearch&type=advanced&result=true&prevSearch=%2Bauthorsfield%3A(Thorrold%2C+B.+S.)

http://www.tandfonline.com/action/doSearch?action=runSearch&type=advanced&result=true&prevSearch=%2Bauthorsfield%3A(Parminter%2C+T.+G.)

http://www.tandfonline.com/action/doSearch?action=runSearch&type=advanced&result=true&prevSearch=%2Bauthorsfield%3A(Wedderburn%2C+M.+E.)

http://www.tandfonline.com/action/doSearch?action=runSearch&type=advanced&result=true&prevSearch=%2Bauthorsfield%3A(Dodd%2C+M.+B.)

http://www.tandfonline.com/action/doSearch?action=runSearch&type=advanced&result=true&prevSearch=%2Bauthorsfield%3A(Thorrold%2C+B.+S.)

http://www.tandfonline.com/action/doSearch?action=runSearch&type=advanced&result=true&prevSearch=%2Bauthorsfield%3A(Quinn%2C+J.+M.)

http://www.tandfonline.com/action/doSearch?action=runSearch&type=advanced&result=true&prevSearch=%2Bauthorsfield%3A(Parminter%2C+T.+G.)

http://www.tandfonline.com/action/doSearch?action=runSearch&type=advanced&result=true&prevSearch=%2Bauthorsfield%3A(Wedderburn%2C+M.+E.)



Douglas G, Dymond J, McIvor I 2008. Monitoring and reporting of whole farm plans as a tool for affecting land use change. Report for Horizons Regional Council, Palmerston North, AgResearch.

Douglas GB, McIvor IR, Manderson AK, Koolaard JP, Todd M, Braaksma S, Gray RAJ 2013. Reducing shallow landslide occurrence in pastoral hill country using wide spaced trees. Land Degradation and Development 24: 103–114.

Douglas GB, McIvor IR, Manderson AK, Todd M, Braaksma S, Gray RAJ 2009. Effectiveness of space-planted trees for controlling soil slippage on pastoral hill country. In: Currie LD, Lindsay CL eds Nutrient management in a rapidly changing world. Occasional Report No. 22. Palmerston North, Fertilizer and Lime Research Centre, Massey University

Douglas GB, McIvor I R, Potter JF, Foote AG 2010. Root distribution of poplar at varying densities on pastoral hill country. Plant and Soil 333: 147–161.

Douglas J, Stokes S, Waiora Soil Conservation 2011. Report on exotic forest debris management related to storm events in the Bay of Plenty. Operations Publication 2011/03. Whakatane, Bay of Plenty Regional Council.

Duzant JH, Morgan RPC, Wood GA, Deeks LK 2011. Modelling the role of vegetated buffer strips in reducing transfer of sediment from land to watercourses. In: Morgan RPC, Nearing MA eds Handbook of erosion modelling. Chichester, Wiley-Blackwell. Pp. 249–262.

Dymond JR, Ausseil AG, Shepherd JD, Buettner L 2006. Validation of a region-wide model of landslide susceptibility in the Manawatu–Wanganui region of New Zealand. Geomorphology 74: 70–79.

Dymond JR, Betts HD, Schierlitz CS 2010. An erosion model for evaluating regional land-use scenarios. Environmental Modelling and Software 25: 289–298.

Dymond JR, Herzig A, Ausseil A-G, 2014. Using SedNetNZ to assess the impact of the Sustainable Land Use Initiative in the Manawatu-Wanganui region on river sediment loads. Landcare Research Contract Report LC1895 for Horizons Regional Council.

Dymond JR, Herzig A, Basher L, Betts HD, Marden M, Phillips CJ, Ausseil A-G, Palmer DJ, Clark M, Roygard J 2016. Development of a New Zealand SedNet model for assessment of catchment-wide soil-conservation works. Geomorphology 257: 85–93.

Edwards PJ, Williard WJ 2010. Efficiencies of forestry best management practices for reducing sediment and nutrient losses in the Eastern United States. Journal of Forestry 108: 245–249.

Elliott AH, Basher LR 2011. Modelling sediment flux: a review of New Zealand catchment-scale approaches. Journal of Hydrology (NZ) 50: 143–160.

Elliott AH, Carlson WT 2004. Effects of sheep grazing episodes on sediment and nutrient loss in overland flow. Australian Journal of Soil Research 42: 213–220.



Engle BW, Jarrett AR 1995. Sediment retention effectiveness of sedimentation basin filtered outlets. Transactions of the American Society of Agricultural Engineers 3: 435–439.

Environment Bay of Plenty 2010. Erosion and sediment control guidelines land disturbing activities. Guideline 2010/01. Whakatane, Environment Bay of Plenty.

Environment Bay of Plenty 2012. Erosion and sediment control guidelines for forestry operations. Guideline 2012/04. Whakatane, Environment Bay of Plenty.

Environment Canterbury 2007. Erosion and sediment control guidelines for the Canterbury region. Ecan Report No. R06/23.

Environment Canterbury undated. Builders Pocket Guide: practical advice on managing worksites and the environment.Christchurch, Environment Canterbury (available at http://www.bpg.co.nz/pdf/builders-pocket-guide-new.pdf).

Environment Waikato 2009. Erosion and sediment control guidelines for soil disturbing activities. Environment Waikato Technical Report No.2009/02.

Eyles GO 1983. Severity of present erosion in New Zealand. New Zealand Geographer 39: 12–28.

Eyles GO 1993. Gully erosion control techniques for pumice lands. Lincoln, Landcare Research.

Eyles G, Fahey B 2006. The Pakuratahi Land Use Study: a 12 year paired catchment study of the environmental effects of Pinus radiata forestry. HBRC Plan No. 3868. Napier, Hawke’s Bay Regional Council.

Fahey BD, Coker RJ 1989. Forest road erosion in the granite terrain of south-west Nelson, New Zealand. Journal of Hydrology (NZ) 28: 123–141.

Fahey BD, Coker RJ 1992. Sediment production from forest roads in Queen Charlotte Forest and potential impact on marine water quality, Marlborough Sounds, New Zealand. New Zealand Journal of Marine and Freshwater Research 26: 187–195.

Fahey BD, Marden M 2000. Sediment yields from a forested and a pasture catchment, coastal Hawke’s Bay, North Island, New Zealand. Journal of Hydrology (NZ) 39: 49–63.

Fahey B, Marden M, Phillips C 2003. Sediment yields from plantation forestry and pastoral farming, coastal Hawke's Bay, North Island, New Zealand. Journal of Hydrology (NZ) 42: 27–38.

Farjood A, Melville BW, Shamseldin AY, Adams KN, Khan S 2015. Evaluation of hydraulic performance indices for retention ponds. Water Science and Technology 72: 10–21.

Faucette LB, Governo J, Tyler R, Gigley G, Jordan CF, Lockaby BG 2009. Performance of compost filter socks and conventional sediment control barriers used for perimeter control on construction sites. Journal of Soil and Water Conservation 64: 81–88.



Fiener P, Auerswald K, Weigand S 2005. Managing erosion and water quality in agricultural watersheds by small detention ponds. Agriculture Ecosystems and Environment 110: 132–142.

Franklin Sustainability Project 2000. Doing It Right, Franklin Sustainability Project Guide to Sustainable Land Management. Pukekohe, Pukekohe Vegetable Growers Association.

Fransen PJB, Brownlie RK 1995. Historical slip erosion in catchments under pasture and radiata pine forest, Hawke’s Bay hill country. New Zealand Forestry 40(4): 29–33.

Fransen PJB, Phillips CJ, Fahey BD 2001. Forest road erosion in New Zealand: overview. Earth Surface Processes and Landforms 26: 165–174.

Fullen MA 1985. Compaction, hydrological processes and soil erosion on loamy sands in East Shropshire, England. Soil and Tillage Research 6: 17–29.

Gabriels D, Ghekiere G, Schiettecatte W, Rottiers I 2003. Assessment of USLE cover-management C-factors for 40 crop rotation systems on arable farms in the Kemmelbeek watershed, Belgium. Soil and Tillage Research 74: 47–53.

Ghaffarzadeh M, Robinson CA, CruseRM 1992. Vegetative filter strip effects on sediment deposition from overland flow Agronomy abstracts. Madison, WI, American Society of Agronomy. P. 324.

Gharabaghi B, Fata A, Van Seters T, Rudra RP, MacMillan G, Smith D, Li JY, Bradford A, Tesa G 2006. Evaluation of sediment control pond performance at construction sites in the Greater Toronto Area. Canadian Journal of Civil Engineering 33: 1335–1344.

Gharabaghi B, Rudra R, Whiteley HR, DickinsonWT 2000. Sediment removal efficiency of vegetative filter strips. Annual Research Report of Guelph Turfgrass Institute (http://131.104.104.3/00anrep/vegfilter.pdf)

Gharabaghi B, Rudra R, Whiteley HR, DickinsonWT 2002. Development of a management tool for vegetative filter strips. Journal of Water Management Modeling R208-18: doi: 10.14796/JWMM.R208-18.

Gibbs HS, Raeside JD 1945. Review of soil erosion in the high country of the South Island. DSIR Bulletin No. 92.

Gibbs M 2007. Best practice guidelines for vegetation management and in stream works. Environment Waikato Technical Report No.2007/41. Hamilton, Environment Waikato.

Gilchrist AN 1984. Wind erosion hazard and its mitigation by shelter: a review. Publication No. 6, Soil Conservation Centre, Ministry of Works and Development, Palmerston North, New Zealand.

Gilmore B, Mackie G, Meredith K 2011. New Zealand forest oad engineering manual. Wellington, NZ Forest Owners Association.



Glade T 2003. Landslide occurrence as a response to land use change: a review of evidence from New Zealand. Catena 51: 297–314.

Greater Wellington Regional Council 2006. Erosion and Sediment Control Guidelines for the Wellington Region. Wellington, Greater Wellington Regional Council.

Haan CT, Barfield BJ, Hayes JC 1994. Design hydrology and sedimentology for small catchments. San Diego, CA, Academic Press. 588 p.

Hancox GT, Wright K 2005a. Landslides caused by the February 2004 rainstorms and floods in southern North Island, New Zealand. Institute of Geological and Nuclear Sciences Science Report 2005/10.

Hancox GT, Wright K 2005b. Analysis of landsliding caused by the February 2004 rainstorms in the Wanganui-Manawatu hill country, southern North Island, New Zealand. Institute of Geological and Nuclear Sciences Science Report 2005/11.

Harding MV 1990. Erosion control effectiveness: comparative studies of alternative mulching techniques. In: Berger JJ ed. Environmental restoration: science and strategies for restoring the earth. Washington, DC, Island Press. Pp 149–156.

Harper S, Semadeni-Davies A 2010. Western Ring Route – Waterview Connection Report G30: associated sediment and contaminant generation report. NIWA Client Report No. AKL-2010-022 prepared for New Zealand Transport Agency.

Harper S, Moores J, Elliott A 2013. Puhoi–Warkworth Road of National Significance: estimates of construction sediment loads. NIWA Client Report No. AKL2013-008 prepared for Further North Alliance.

Hawley JG 1988. Influence of trees on landslip. Proceedings New Zealand Association of Soil Conservators conference, Havelock North, August 1988.

Hawley JG, Dymond JR 1988. How much do trees reduce landsliding? Journal of Water and Soil Conservation 43: 495–498.

Hayes SA, McLaughlin RA, Osmond DL. 2005. Polyacrylamide use for erosion and turbidity control on construction sites. Journal of Soil and Water Conservation 60: 193–199.

Haynes RJ, Tregurtha R 1999. Effects of increasing periods under intensive vegetable production on biological, chemical and physical indices of soil quality. Biology and Fertility of Soils 28: 259–266.

Hicks DL 1988. An assessment of soil conservation in the Waihora catchment, East Coast. Draft Report for Soil Conservation Centre, Aokautere and East Cape Catchment Board, Palmerston North, Water and Soil Division, Ministry of Works and Development.

Hicks DL 1989a. Farm conservation measures effect on hill country erosion: an assessment in the wake of Cyclone Bola. Division of Land and Soil Sciences Technical Record PN3.



Hicks DL 1989b. Upstream versus downstream damage during large storms. DSIR Land Resources Technical Record PN1.

Hicks DL 1989c. Storm damage to bush, pasture and forest: some evidence from Cyclone Bola. DSIR Land Resources Technical Record PN2.

Hicks DL 1990. Landslide damage to pasture, pine plantations, scrub and bush in Taranaki. DSIR Land Resources Technical Record 31.

Hicks DL 1991a. Effect of soil conservation tree plantings on damage sustained by the Whareama catchment during the storm of 8–11 April 1991. DSIR Land Resources Contract Report 91/106 for Wellington Regional Council.

Hicks DL 1991b. Erosion under pasture, pine plantations, scrub and indigenous forest: a comparison from Cyclone Bola. New Zealand Forestry 36(3): 21–22.

Hicks DL 1992a. Impact of soil conservation on storm-damaged hill grazing lands in New Zealand. Australian Journal of Soil and Water Conservation 5: 34–40.

Hicks DL 1992b. Effect of soil conservation tree plantings on stream bank stability. DSIR Land Resources Technical Record 118.

Hicks DL 1995. Control of soil erosion on farmland: a summary of erosions’ impact on New Zealand agriculture, and farm management practices which counteract it. MAF Policy Technical Paper 95/4.

Hicks DL 2006. A review of scientific information relating to the sustainability of current land use practices on cultivated land in the Franklin District of Auckland. Auckland Regional Council Technical Publication No. 319.

Hicks DL, Anthony T 2001. Soil conservation technical handbook. Prepared for the Ministry for the Environment by the New Zealand Resource Management Association.

Hicks D, Crippen T 2004. Erosion of Manawatu-Wanganui hill country during the storm on 15–16 February 2004. Contract Report for Horizons Regional Council.

Hicks DL, Fletcher JR, Eyles GO, McPhail CR, Watson M 1993. Erosion of hill country in the Manawatu-Wanganui region 1992: impacts and options for sustainable land use. Landcare Research Contract Report LC9394/051 for Federated Farmers.

Hicks DM 1990. Suspended sediment yields from pasture and exotic forest basins. Abstracts New Zealand Hydrological Society Symposium, 1990, Taupo.

Hicks DM 1994. Storm sediment yields from basins with various land uses in Auckland area. ARC Environment Division Technical Publication 51.

Hicks DM, Harmsworth GR 1989. Changes in sediment yield regime during logging at Glenbervie Forest, Northland, New Zealand. Comparisons in Austral Hydrology, Hydrology and Water Resources Symposium, 28–30 November 1989, , University of Canterbury, Christchurch, New Zealand. Pp. 424–428.



Hicks DM, Shankar U, McKerchar AI, Basher L, Jessen M, Lynn I, Page M 2011. Suspended sediment yields from New Zealand rivers. Journal of Hydrology (NZ) 50: 81–142.

Horner RR, Mar BW 1982. Guide for water quality impact assessment of highway operations and maintenance. Washington Department of Transport Report WA-RD39/14.

Hughes AO 2016. Riparian management and stream bank erosion in New Zealand. New Zealand Journal of Marine and Freshwater Research 50: DOI:10.1080/00288330.2015.1116449

Hughes AO, Quinn JM, McKergow LA 2012. Land use influences on suspended sediment yields and event sediment dynamics within two headwater catchments, Waikato, New Zealand. New Zealand Journal of Marine and Freshwater Research 46: 315–333.

Jennings GD, Jarrett AR 1985. Laboratory evaluation of mulches in reducing erosion. Transactions of the American Society of Agricultural Engineers 28: 1466–1470.

Johnstone PR, Wallace DF, Arnold N, Bloomer D 2011: Holding it together – soils for sustainable vegetable production. Report for Horticulture New Zealand prepared by Plant and Food Research and Landwise.

Kaspar TC, Radke JK, LaflenJM 2001. Small grain cover crops and wheel traffic effects on infiltration, runoff, and erosion. Journal of Soil and Water Conservation 56: 160–164.

Kemp PD, Matthew C, Lucas RJ 1999. Pasture species and cultivars. In: White J, Hodgson J eds New Zealand pasture and crop science. Auckland, Oxford University Press. Pp. 83–99.

Khan S 2012. Hydrodynamics of sediment retention ponds. Unpublished PhD thesis, University of Auckland. 436 p,

Knisel WG ed. 1993. GLEAMS. Groundwater Loading Effects of Agricultural Management Systems, version 2.10. Biological and Agricultural Engineering Department Publication No. 5. Tifton, GI, University of Georgia Coastal Plain Experiment Station. 260 p.

Kronvang B, Laubel AR, Larsen SE, Iversen HL 2000. Soil erosion and sediment delivery through buffer zones in Danish slope units. “The Role of Sediment Transport in Nutrient and Contaminant Transfer”, IAHS Publication No. 263: 67–73.

Kouwen N 1980. Silt fences to control sediment movement on construction sites. Ontario, Canada, Ministry of Transportation Report MAT-90-03.

Laloy E, Bielders CL 2010. Effect of intercropping period management on runoff and erosion in a maize cropping system. Journal of Environmental Quality 39: 1001–1008.

Larcombe M. 2009. Summary of sediment retention pond performance, mass discharges of sediment to the Orewa Estuary receiving environment and effects of sediment discharge to the Orewa Estuary for the 2008-2009 earthworks season. Consent compliance report submitted to ARC.

http://scholar.google.co.nz/citations?view_op=view_citation&hl=en&user=2VcBbMcAAAAJ&sortby=pubdate&citation_for_view=2VcBbMcAAAAJ:2P1L_qKh6hAC



Lemly AD 1982. Erosion control at construction sites on red clay soils. Environmental Management 6: 343–352.

Line DE, White NM 2001. Efficiencies of temporary sediment traps on two North Carolina construction sites. Transactions of the American Society of Agricultural Engineers 44: 1207–1215.

Lipscomb CM, Johnson T, Nelsen R, Lancaster T 2006. Comparison of erosion control technologies: “Blown straw vs. erosion control blankets, quantification of performance and value”. Land and Water 50(4): 48.

Lough CJ 1993. Effect of soil conservation measures on 1992 erosion of hill country farmland in the Pohangina district. Research report for B. Hort., Massey University, Palmerston North.

Ludwig B, Boiffin J, Chadœuf J, Auzet AV 1995. Hydrological structure and erosion damage caused by concentrated flow in cultivated catchments. Catena 25: 227–252.

Lynch JA, Corbett ES, Mussallem K 1985. Best management practices for controlling nonpoint-source pollution on forested watersheds. Journal of Soil and Water Conservation 40: 164–167.

Magette WL, Brinsfield RB, Palmer RE, Wood JD 1989. Nutrient and sediment removal by vegetated filter strips. Transactions of the American Society of Agricultural Engineers 32: 663–667.

Marden M 2004. Future-proofing erosion-prone hill country against soil degradation and loss during large storm events: have past lessons been heeded? New Zealand Journal of Forestry 49: 11–16.

Marden M 2012. Effectiveness of reforestation in erosion mitigation and implications for future sediment yields, East Coast catchments, New Zealand: a review. New Zealand Geographer 68: 24–35.

Marden M, Arnold G, Gomez B, Rowan D 2005. Pre- and post-reforestation gully development in Mangatu Forest, East Coast, North Island. New Zealand. River Research and Applications 21: 757–771.

Marden M, Arnold G, Seymour A, Hambling R 2012. History, distribution and stabilisation of steepland gullies in response to revegetation, East Coast region, North Island, New Zealand. Geomorphology 153–154: 81–90.

Marden M, Betts H, Arnold G, Hambling R. 2008b. Gully erosion and sediment load: Waipaoa, Waiapu and Uawa rivers, eastern North Island, New Zealand. In: Sediment dynamics in changing environments (Proceedings of a symposium held in Christchurch, New Zealand, December 2008). IAHS Publication 325: 330–350.



Marden M, Herzig A, Arnold G. 2011. Gully degradation, stabilisation and effectiveness of reforestation in reducing gully-derived sediment, East Coast region, North Island, New Zealand. Journal of Hydrology (NZ) 50: 19–36.

Marden M, Phillips CJ 2013. Survival and growth of poplar and willow pole plantings on East Coast hill country. Landcare Research Contract Report LC1622 for Plant and Food Research.

Marden M, Phillips CJ, Rowan D 1991. Declining soil loss with increasing age of plantation forests in the Uawa catchment, East Coast region. In: Henriques PR ed. Proceedings, International Conference on Sustainable Land Management, Napier, Hawke’s Bay, New Zealand, 17–23 November 1991. Pp. 358–361.

Marden M, Phillips C, Rowan D 2008a. Recurrent displacement of a forested earthflow and implications for forest management, East Coast Region, New Zealand. IAHS Publication 325: 491–501.

Marden M, Rowan D 1993. Protective value of vegetation on Tertiary terrain before and during Cyclone Bola, East Coast, North Island, New Zealand. New Zealand Journal of Forestry Science 23: 255–263.

Marden M, Rowan D, Phillips C 2006. Sediment sources and delivery following plantation harvesting in a weathered volcanic terrain, Coromandel Peninsula, North Island, New Zealand. Australian Journal of Soil Research 44: 219–232.

McBurnie JC, Barfield BJ, Clar ML, Shaver E 1990. Maryland sediment detention pond design criteria and performance. Applied Engineering in Agriculture 6: 167–173.

McCaleb MM, McLaughlin RA 2008. Sediment trapping by five different sediment detention devices on construction sites. Transactions of the American Society of Agricultural and Biological Engineers 51: 1613–1621.

McCaskill L 1973. Hold this land: a history of soil and water conservation in New Zealand. Wellington, AH and AW Reed.

McIvor I, Clarke K, Douglas G 2015. Effectiveness of conservation trees in reducing erosion following a storm event. In: Currie LD, Burkitt LL eds Proceedings, 28th Annual Fertiliser and Lime Research Centre Workshop ‘Moving farm systems to improved attenuation’, 8–9 February 2007. Occasional Report 28. Palmerston North, Fertiliser and Lime Research Centre.

McIvor IR, Douglas GB, Hurst SE, Hussain Z, Foote AG 2008. Structural root growth of young Veronese poplars on erodible slopes in the southern North Island, New Zealand. Agroforestry Systems 72: 75–86.

McIvor IR, Douglas GB, Benavides R 2009. Coarse root growth of Veronese poplar trees varies with position on an erodible slope in New Zealand. Agroforestry Systems 76: 251–264.



McIvor I, Douglas G, Dymond J, Eyles G, Marden M 2011. Pastoral hillslope erosion in New Zealand and the role of poplar and willow trees in its reduction. In: Soil erosion issues in agriculture. Pp. 257–278. http://www.intechopen.com/articles/show/title/ pastoral-hill-slope-erosion-in-new-zealand-and-the-role-of-poplar-and-willow-trees-in-its-reduction.

McKergow LA, Pritchard M, Elliott AH, Duncan MJ, Senior AK 2010. Storm fine sediment flux from catchment to estuary, Waitetuna-Raglan Harbour, New Zealand. New Zealand Journal of Marine and Freshwater Research 44: 53–76.

McLaughlin RA, Brown TT 2006. Evaluation of erosion control products with and without added polyacrylamide. Journal of the American Water Resources Association 42: 675–684.

McLaughlin RA, Hayes SA, Clinton DL, McCaleb MM, Jennings GD 2009a. Water quality improvements using modified sediment control systems on construction sites. Transactions of the American Society of Agricultural and Biological Engineers 52: 1859–1867.

McLaughlin RA, King SE, Jennings GD 2009b. Improving construction site runoff quality with fiber check dams and polyacrylamide. Journal of Soil and Water Conservation 64: 144–154.

McSaveney MJ 1978. Magnitude of erosion across the Southern Alps. Proceedings of the Conference on Erosion Assessment and Control in New Zealand, Christchurch, August 1978. Christchurch, New Zealand Association of Soil Conservators. Pp. 7–24.

Moores J, Pattinson P 2008. Performance of a Sediment Retention Pond Receiving Chemical Treatment. Auckland Regional Council Technical Report 2008/021.

Morton D 1996.Prediction of sediment discharge during development of a residential subdivision, Auckland, New Zealand. Unpublished MSc thesis, University of Waikato, Hamilton, New Zealand.

Mosley MP 1980. The impact of forest road erosion in the Dart Valley, Nelson. New Zealand Journal of Forestry 25(2): 184–198.

New Zealand Forest Owners Association 2007. New Zealand environmental code of practice for plantation forestry. Wellington, NZ Forest Owners Association.

New Zealand Transport Agency 2014. Erosion and sediment control guidelines for state highway infrastructure: construction stormwater management. Wellington, New Zealand Transport Agency.

Ngapo N 2012. Addressing harvest related sedimentation. Paper to Environmental Defence Society Conference ‘Growing Green’, Auckland on 6–7 August 2012. http://www.edsconference.com/content/docs/2012_papers/Ngapo,%20Norm.pdf.



Nighman DM, Harbor JM 1997. Trap efficiency of a stormwater basin with and without baffles. Proceedings of the International Erosion Control Association 28: 469–483.

Northland Regional Council 2012. Forestry earthworks and harvesting guidelines for Northland. Northland Regional Council in association with Hancock Forest Management Ltd, Rayonier/Matariki Forests, Chandler Fraser Keating Ltd, Northland Forest Managers (1995) Ltd, P F Olsen Ltd, Juken New Zealand Ltd.

O’Loughlin CL 1995. The sustainable paradox – an examination of the plantation effect – a review of the environmental effects of plantation forestry in New Zealand. New Zealand Forestry 39: 3–8.

O’Loughlin CL 2005. The protective role of trees in soil conservation. New Zealand Journal of Forestry 49: 9–15.

Page MJ, Shepherd JD, Dymond JR, Jessen M 2005. Defining highly erodible land for Horizons Regional Council. Landcare Research Contract Report LC0506/049.

Page MJ, Trustrum NA, Gomez B 2000. Implications of a century of anthropogenic erosion for future land use in the Gisborne-East Coast region of New Zealand. New Zealand Geographer 56: 13–24.

Pain CF 1986. The effectiveness of tree planting in the control of landslides, North Island, New Zealand: development of methodologies. Soil Conservation Centre Internal Report No. 165. Aokautere, Ministry of Works and Development.

Pain CF, Stephens PR 1990. Storm damage assessment using digitised aerial photographs: Eltham, New Zealand, 24–25 February 1986. New Zealand Geographer 46: 21–25.

Parker DG, Parker SC, Stader TN 1995. Use of GIS to predict erosion in construction. Proceedings of the 1st International Conference on Water Resources: Part 1 (of 2), 14–18 Aug 1995, San Antonio, Texas. New York, NY, ASCE. Pp. 839–843.

Parkyn S 2004. Review of riparian buffer zone effectiveness. MAF Technical Paper No: 2004/05.

Parkyn SM, Davies-Colley RJ, Halliday NJ, Costley KJ, Croker GF 2003. Planted riparian buffer zones in New Zealand: do they live up to expectations? Restoration Ecology 11: 436–447.

Parkyn S, Shaw W, Eades P 2000. Review of information on riparian buffer widths necessary to support sustainable vegetation and aquatic functions. NIWA Client Report ARC00262.

Parshotam A 2008. Southeastern Manukau Harbour / Pahurehure Inlet contaminant study: sediment load model results. Auckland Regional Council Technical Report 2008/052.

Parshotam A, Wadwha S 2008. Central Waitemata Harbour Contaminant Study. GLEAMS model results for rural and earthwork sediment loads from the catchment. Auckland Regional Council Technical Report 2008/041.



Pearce AJ, Hodgkiss PD 1987. Erosion and sediment yield from a landing failure after a moderate rainstorm, Tairua Forest. New Zealand Forestry 32: 19–22.

Phillips C, Marden M 2005. Reforestation schemes to manage regional landslide risk. In: Glade T, Anderson MG, Crozier MJ eds, Landslide hazard and risk. Chichester, UK, Wiley. Pp. 517–548.

Phillips CJ, Marden M 2011. Houpotu Forest storm damage assessment and review. Landcare Research Contract report LC777.

Phillips C, Marden M, Basher L 2012. Plantation forest harvesting and landscape response – what we know and what we need to know. New Zealand Journal of Forestry 56: 4–12.

Phillips C, Marden M, Douglas G, McIvor I, Ekanayake J 2008. Decision support for sustainable land management: effectiveness of wide-spaced trees. Landcare Research Contract Report LC0708/126.

Phillips CJ, Marden M, Lambie S 2014. Observations of root growth of young poplar and willow planting types. New Zealand Journal of Forestry Science 44: DOI: 10.1186/s40490-014-0015-6

Phillips CJ, Marden M, Miller D 2000. Review of plant performance for erosion control in the East Coast region. Landcare Research Contract Report LC9900/111 for MAF Policy.

Phillips CJ, Marden M, Pearce A 1990. Effectiveness of reforestation in prevention and control of landsliding during large cyclonic storms. Proceedings of the 19th World IUFRO Congress (Division 1, Vol. 1), Montreal, Canada, August 1990. Pp. 340–350.

Pitt RE, Clark SE, Lake DW 2007. Construction site erosion and sediment controls. Planning, Design and Performance. Lancaster, PA, DEStech Publications, Inc.. 381 p.

Pollock KM 1986. Plant materials handbook for soil conservation, Part 3 Native plants. Water & Soil Miscellaneous Publication 95.

Przepiora A, Hesterberg D, Parsons JE, Gilliam JW, Cassel DK, Faircloth W 1998. Field evaluation of calcium sulfate as a chemical flocculant for sedimentation basins. Journal of Environmental Quality 27: 669–678.

Quinn JM, Stroud MJ 2002. Water quality and sediment and nutrient export from New Zealand hill-land catchments of contrasting land use. New Zealand Journal of Marine and Freshwater Research 36: 409–429.

Rawitz E, Morin J, Hoogmoed WB, Margolin M, Etkin H 1983. Tillage practices for soil and water conservation in the semi-arid zone. I: management of fallow during the rainy season preceding cotton. Soil and Tillage Research 3: 211–231.

Reed AH 1986. Soil loss from tractor wheelings. Soil and Water 14: 12–14.

Rickson RJ 2006. Controlling sediment at source: an evaluation of erosion control geotextiles. Earth Surface Processes and Landforms 31: 550–560.



Ridley G, De Luca S 2015. Vaughan’s catchment adaptive environmental monitoring and management response plan 2015 (earthworks three, four and five Long Bay). Report by Ridley Dunphy Ltd, prepared for Long Bay Communities Ltd.

Roa A, Bubenzer GD, Miyashita ES 1998. Determination of PAM Use Potential in Erosion Control In; Pereira et al. eds Proceedings of the First European Conference of Water and the Environment: Innovation Issues in Irrigation and Drainage. Portuguese National Committee of ICID (International Commission on Irrigation and Drainage). Pp. 49-57.

Robinson DR, Nagizadeh R 1992. The impact of cultivation practice and wheelings on runoff generation and soil erosion on the South Downs: some experimental results using simulated rainfall. Soil Use and Management 8: 151–156.

Roche M 1994. Land and Water: water and soil conservation and central government in New Zealand 1941–1988. Wellington, Historical Branch, Department of Internal Affairs.

Ross CW, Basher LR, Baker CJ 2000. Management of erosion risk on arable soils: a review. Landcare Research Contract Report LC9900/090 for Environment Canterbury.

Russell JR, Betteridge K, Costall DA, Mackay AD 2001. Cattle treading effects on sediment loss and water infiltration. Journal of Range Management 54: 184–190.

Salter RT 1984. Wind erosion. In: Speden I, Crozier MJ comps Natural hazards in New Zealand. Wellington, New Zealand National Commission for Unesco. Pp. 206–248.

Schellinger GR, Clausen JC 1992. Vegetative filter treatment of dairy barnyard runoff in cold regions. Journal of Environmental Quality 21: 40–45.

Schierlitz C, Dymond J, Shepherd J 2006. Erosion/sedimentation in the Manawatu associated with scenarios of whole farm plans. Landcare Research Contract Report LC0607/028.

Schwarz M, Phillips C, Marden M, McIvor IR, Douglas GB, Watson A 2016. Modelling of root reinforcement and erosion control by ‘Veronese’ poplar on pastoral hill country in New Zealand. New Zealand Journal of Forestry Science 46: DOI: 10.1186/s40490-016-0060-4.

Shaver E 2009a. Hawke’s Bay waterway guidelines: erosion and sediment control. Draft for Hawke’s Bay Regional Council April 2009, HBRC Plan Number 4109.

Shaver E 2009b. Hawkes Bay waterway guidelines: forestry erosion and sediment control. draft for Hawkes Bay Regional Council April 2009, HBRC Plan Number 4108.

Skidmore E 1986. Wind erosion control. Climatic Change 9: 209–218.

Smith CM 1989. Riparian pasture retirement effects on sediment, phosphorus, and nitrogen in channelised surface run-off from pastures. New Zealand Journal of Marine and Freshwater Research 23: 139–146.



Smith CM 1992. Riparian afforestation effects on water yields and water quality in pasture catchments. Journal of Environmental Quality 21: 237–245.

Smith M, Fenton T 1993. Sediment yields from logging tracks in Kaingaroa Forest. New Zealand Logging Industry Research Organisation Report 18.

Soons JM, Selby MJ eds 1992. Landforms of New Zealand (2nd ed.). Hong Kong, Longman Paul

Soupir ML, Mostaghimi S, Masters A, Flahive KA, Vaughan DH, Mendez A, McClellan PW 2004. Effectiveness of polyacrylamide (PAM) in improving runoff water quality from construction sites. Journal of the American Water Resources Association 40: 53–66.

Stringer, D. J. 1977: A guide to the establishment of windbreaks. South Canterbury Catchment Board Publication 16.

Sturrock JW ed. 1984. Shelter research needs in relation to primary production: report of the National Shelter Working Party. Water and Soil Miscellaneous Publication 59.

Sui Y, Ou Y, Yan B, Xu X, Rousseau AN, Zhang Y 2016. Assessment of micro-basin tillage as a soil and water conservation practice in the black soil region of northeast China. PLOS One 11(3) doi: 10.371/journal.pone.0142313.

Taranaki Regional Council 2006. Guidelines for earthworks in the Taranaki region. New Plymouth, Taranaki Regional Council.

Tasman District Council 2014. Erosion and sediment control guidelines. Final Draft for Discussion, 12 March 2014.

Thaxton CS, McLaughlin RA 2005. Sediment capture effectiveness of various baffle types in a sediment retention pond. Transactions of the American Society of Agricultural Engineers 48: 1795–1802.

Thompson RC, Luckman PG 1993. Performance of biological erosion control in New Zealand soft rock hill terrain. Agroforestry Systems 21: 191–211.

Truman CC, Nuti RC 2009. Improved water capture and erosion reduction through furrow diking. Agricultural Water Management 96: 1071–1077.

United States Department of Agriculture – Agricultural Research Service (USDA – ARS) 2008. User’s Reference Guide (DRAFT) revised universal soil loss equation version 2 (RUSLE2). http://fargo.nserl.purdue.edu/rusle2_dataweb/userguide/RUSLE2_User_Ref_Guide_2008.pdf Accessed 21 December 2012.

Van Kraayenoord CWS, Hathaway RL eds 1986a. Plant materials handbook for soil conservation. Volume 1: Principles and practices. Water and Soil Miscellaneous Publication 93.

http://fargo.nserl.purdue.edu/rusle2_dataweb/userguide/RUSLE2_User_Ref_Guide_2008.pdf

http://fargo.nserl.purdue.edu/rusle2_dataweb/userguide/RUSLE2_User_Ref_Guide_2008.pdf



Van Kraayenoord CWS, Hathaway RL (Eds) 1986b. Plant materials handbook for soil conservation, Volume 2: Introduced Plants. Water and Soil Miscellaneous Publication 94. Wellington, National Water and Soil Conservation Authority.

Varvaliu A 1997. Effect of soil conservation measures on 1992 soil slip erosion of the Pakihikura valley, Rangitikei district, North Island, New Zealand. Unpublished MAgrSc thesis, Massey University, Palmerston North, New Zealand.

Wachal DJ, Banks KE, Hudak P, Harmel RD 2009. Modeling erosion and sediment control practices with RUSLE 2.0: a management approach for natural gas well sites in Denton County, TX, USA. Environmental Geology 56: 1615–1627.

Ward AD, Haan CR, Barfield BJ 1979. Prediction of sediment basin performance. Transactions of the American Society of Agricultural Engineers 22: 126–136.

Wilcock RJ, Monaghan RM, Quinn JM, Srinivasan MS, Houlbrooke DJ, Duncan MJ, Wright-Stow AE, Scarsbrook MR 2013. Trends in water quality of five dairy farming streams in response to adoption of best practice and benefits of long-term monitoring at the catchment scale. Marine and Freshwater Research 64: 401–412.

Williams P, Spencer M 2013. Environmental guidelines forest harvesting. Blenheim, Marlborough District Council.

Williamson R, Smith C, Cooper A. 1996. Watershed riparian management and its benefits to a eutrophic lake. Journal of Water Resources Planning and Management 122: 24–32.

Williamson RB, Smith RK, Quinn JM 1992. Effects of riparian grazing and channelisation on streams in Southland, New Zealand. 1: channel form and stability. New Zealand Journal of Marine and Freshwater Research 26: 241–258.

Winter R 1998. Predicting sediment yield during the earthworks development stage of a subdivision, Auckland, and assessment of the efficiency of a sediment retention pond. Unpublished MSc Thesis, University of Waikato.

Wischmeier WH, Smith DD 1978. Predicting rainfall erosion losses – a guide to conservation planning. USDA AH537. Washington, DC.

Wong SL, McCuen RH 1982. The design of vegetative buffer strips for runoff and sediment control. Unpublished report to Maryland Coastal Zone Management Program, University of Maryland.

Wyant DC 1981. Evaluation of filter fabrics for use in silt fences. Transportation Research Record 832: 6–12.

Xiao B, Wang Q, Wang H, Wu J, Yu D 2012. The effects of grass hedges and micro-basins on reducing soil and water loss in temperate regions: a case study of Northern China. Soil and Tillage Research 122: 22–35.

Young RA, Huntrods T, Anderson W 1980. Effectiveness of vegetated buffer strips in controlling pollution from feedlot runoff. Journal of Environmental Quality 9: 483–487.



Yuan Y, Bingner RL, Locke MA 2009. A review of effectiveness of vegetative buffers on sediment trapping in agricultural areas. Ecohydrology 2: 321–336.

Zhang X, Phillips CJ, Marden M 1993. A comparison of earthflow movement rates on forested and grassed slopes, Raukumara Peninsula, North Island, New Zealand. Geomorphology 6: 175–187.

Ziegler AD, Sutherland RA, Tran LT 1997. Influence of rolled erosion control systems on temporal rainsplash response – a laboratory rainfall simulation experiment. Land Degradation & Development 8: 139–157.

Scientific basis for erosion and sediment control ... · Scientific basis for erosion and sediment...

Documents

Transcript of Scientific basis for erosion and sediment control ... · Scientific basis for erosion and sediment...