Characterizing geomorphological change to support sustainable …€¦ · Advanced Review...

Advanced Review

Characterizing geomorphologicalchange to support sustainable riverrestoration and managementRobert C. Grabowski,1,2∗ Nicola Surian3 and Angela M. Gurnell1

The hydrology and geomorphology of most rivers has been fundamentally alteredthrough a long history of human interventions including modification of river chan-nels, floodplains, and wider changes in the landscape that affect water and sed-iment delivery to the river. Resultant alterations in fluvial forms and processeshave negatively impacted river ecology via the loss of physical habitat, disrup-tion to the longitudinal continuity of the river, and lateral disconnection betweenaquatic, wetland, and terrestrial ecosystems. Through a characterization of geo-morphological change, it is possible to peel back the layers of time to investigatehow and why a river has changed. Process rates can be assessed, the historicalcondition of rivers can be determined, the trajectories of past changes can bereconstructed, and the role of specific human interventions in these geomorpho-logical changes can be assessed. To achieve this, hydrological, geomorphological,and riparian vegetation characteristics are investigated within a hierarchy of spa-tial scales using a range of data sources. A temporal analysis of fluvial geomor-phology supports process-based management that targets underlying problems.In this way, effective, sustainable management and restoration solutions can bedeveloped that recognize the underlying drivers of geomorphological change, theconstraints imposed on current fluvial processes, and the possible evolutionary tra-jectories and timelines of change under different future management scenarios.Catchment/river basin planning, natural flood risk management, the identificationand appraisal of pressures, and the assessment of restoration needs and objectiveswould all benefit from a thorough temporal analysis of fluvial geomorphology. ©2014 Wiley Periodicals, Inc.

How to cite this article:WIREs Water 2014. doi: 10.1002/wat2.1037

INTRODUCTION

In many countries, regulatory objectives nowrequire rivers to be managed in a holistic manner

that balances human use and modification with thepreservation and improvement of aquatic and riparian

∗Correspondence to: [email protected] School of Geography, Queen Mary, University of London,London, UK2Cranfield Water Science Institute, Cranfield University, Cranfield,UK3Department of Geosciences, University of Padova, Padova, ItalyConflict of interest: The authors have declared no conflicts ofinterest for this article.

ecosystems. Implicit in these objectives is the acknowl-edgment that the hydrology and geomorphology ofrivers has been fundamentally altered through a longhistory of direct intervention to river form and waterflow as well as wider changes in the landscape thatimpact water and sediment delivery to the river.1,2

These alterations in fluvial forms and processes havenegatively impacted river ecology through the lossof physical habitat, disruption to the longitudinalcontinuity of the river, and a disconnection betweenaquatic, wetland, and terrestrial ecosystems.3,4 How-ever, the numerous demands on rivers (e.g., freshwatersupply, navigation, flood protection) and dwindlingfinancial resources to maintain current infrastructureand maintenance operations have placed an emphasis

© 2014 Wiley Per iodica ls, Inc.

Raj Bridgelall

Highlight

Raj Bridgelall

Highlight

Advanced Review wires.wiley.com/water

on the identification of effective managementand restoration approaches that yield sustainablesolutions.5,6 Working with the river’s natural hydro-logical and geomorphological processes, as opposedto imposing form and behavior, offers the best oppor-tunity to do so.4,7,8 A process-based managementapproach requires an understanding of not only thecurrent geomorphological condition a river, but alsohow this has changed over time. While from anecological perspective, river restoration and broadermanagement should target the measures and riverreaches that would provide the greatest and mostcost-effective impact in terms of reinstating andsustaining natural processes, it is critical that directand indirect human interventions are factored intosuch an approach. Furthermore, although this paperemphasizes the understanding of processes and theirdirect relevance to restoration, this is only a smallpart of the many factors that are incorporated intoriver restoration in practice. Thus good science cancontribute to the design of effective and sustain-able restoration and management schemes, but thelocation(s), detailed design, and financing of suchschemes are often more dependent upon a host ofhuman-related issues that are beyond the scope of thisreview.9

Rivers change over time. This is an inherentproperty of rivers and floodplains, and is driven byforces operating within the channel (i.e., intrinsic) andas a result of changes in the wider catchment (i.e.,extrinsic). Temporal changes in fluvial geomorphologycan be expressed in a river in many different ways: thespatial location of the channel (e.g., lateral migrationand avulsions); riverbed and floodplain levels (e.g.,channel incision and floodplain sedimentation); chan-nel planform and dimensions; bed sediment character-istics; and the frequency and diversity of geomorphicunits in the channel and floodplain. Some of thesechanges may be natural for the river type, while oth-ers are induced by changes that have occurred else-where in the catchment. By recognizing that rivers aredynamic, a temporal analysis of fluvial geomorphol-ogy can support river management and restoration byproviding information on:

• Rates of geomorphological, hydrological, andecological processes (e.g., water flow, sedimenttransport, and riparian and aquatic plant growthand succession),

• The previous condition of the catchment, flood-plain, and channel,

• Rates and trajectories of past change in channeland floodplain characteristics,

• Identification of human pressures and how theyhave changed over time,

• Channel response to past natural disturbancesand human pressure.

While there is a growing recognition of the roleof process-based management and restoration of rivers(e.g., the concept of ‘Making space for water’ pro-moted by the Department for Environment Food andRural Affairs, in the UK), the identification, planning,and implementation of individual measures is oftenthe best outcome of the complex interaction betweenvarious legal frameworks, regulatory drivers, policyinitiatives, and stakeholder engagement which is com-pounded by the opportunistic nature of restorationprojects (i.e., a willing landowner). Frequently, theresult is piecemeal management that treats the symp-toms of alterations to geomorphological processesrather than the causes, with the consequence that mea-sures may not meet their intended objectives. Theoutcomes of a thorough temporal analysis providemanagers with the information needed to develop aholistic understanding of their rivers and floodplains.It allows them to identify the nature, magnitude,and underlying causes of geomorphological changein a reach, the human constraints imposed on futurerestoration and management, and the possible evo-lutionary trajectories and timelines of change underdifferent future management scenarios. This informa-tion can be used to develop effective and sustainablesolutions with process-based objectives, regardless ofwhether they are integrated catchment-scale measuresto tackle multiple pressures and improve ecologicalstatus, or reach-scale projects to improve physicalstream habitat or local amenity.

The aim of this review is to provide guide-lines and suggestions for the temporal analysis ofgeomorphological change in rivers which can informprocess-based river management and restoration. Thearticle is structured around a spatial hierarchicalframework that nests the reach and its distinctivegeomorphic forms and processes into a wider spatialcontext. A brief outline of the spatial scales is given,followed by an introduction to the types of approachesused in a temporal analysis and the timescales overwhich they are relevant. Then at each spatial scale,characteristics are identified that control critical flu-vial processes or are indicative of channel adjustment,alteration, or artificiality. Recommendations are pro-vided on the approaches for gaining information oneach characteristic, the range of data that can be col-lected using those approaches, suitable analytical tech-niques and methods to assess data accuracy. Finally,the role of a temporal analysis of geomorphology in


WIREs Water Characterizing geomorphological change in rivers

the development of sustainable river restoration andmanagement strategies is discussed.

SPATIAL AND TEMPORAL SCALESOF ANALYSIS

The geomorphological character of river reachesdepends not only upon interventions and processeswithin the reach but also within the upstream (andsometimes the downstream) catchment. In addition,the character of river reaches responds in a delayedway to processes and interventions within the catch-ment. As a result, understanding geomorphology atthe reach scale requires an understanding of currentand past processes and interventions at larger spatialscales. Without such a multiscale understanding,management strategies are not fully informed andmay not provide sustainable solutions.

Spatial hierarchical frameworks have been pro-posed in many forms in the literature, each developedwith a particular application or set of applications inmind.10–17 Addition of a formal temporal analysis tosuch frameworks is rare, although Ref 15 providesan excellent description of how this may be achieved.Nevertheless, many researchers acknowledge spaceand timescales over which processes may be influen-tial and forms may persist8,10,13,16; while others con-sider scenarios of process dynamics and change.11,12

This article complements these existing frameworksby providing guidance on the types of characteristicsthat should be investigated, the various data sourcesthat can be assembled to investigate each characteris-tic, and data analysis techniques that can be used tosupport a scientifically rigorous interpretation of tem-poral change.

Spatial Scales of AnalysisFor the present application, a hierarchy composed offour levels of spatial units is used, which is based onand coherent with earlier delineations.10,11,15 Hydro-logical, geomorphological, and riparian vegetationproperties are investigated within this hierarchy todevelop a comprehensive picture of geomorphologicalprocess–form interactions and their changes over time(Figure 1; Table 1).

The catchment is an area of land that is drainedby a river and its tributaries, and, for the purposesof this approach, can be delineated based on thetopographic divide (watershed).

Landscape units, i.e., physiographicregion/province, are portions of the catchment withsimilar geomorphological characteristics. The catch-ment is divided into landscape units that are broadly

Catchment

102–105 km2

103–104 years

Landscape unit

102–103 km2

103–104 years

Segment

101–102 km

102–103 years

Reach

100–101 km

101–102 years

FIGURE 1 | Hierarchy of spatial scales for the assessment of rivergeomorphology with indicative spatial and time scales.

TABLE 1 Temporal Change Is Investigated at Different SpatialScales

Spatial Scale Characteristics

Catchment Land cover/land use (LCLU)

Land topography

Landscape unit LCLU and sediment production

Land topography and sediment production

Rainfall and groundwater

Segment Valley setting

Channel gradient

River flows and levels

Sediment delivery

Riparian corridor and wood production

Reach Channel planform, migration, and features

Channel geometry

Sediment transport

Riparian vegetation, aquatic vegetation, wood



consistent in terms of their topography, geology, andland cover, as these factors determine the hydrologicalresponsiveness of a catchment and the sources anddelivery pathways of sediment to the river system.

River segments are sections of the river networkthat are subjected to similar valley-scale influencesand flow energy conditions. Delineation is based onmajor changes in valley gradient, major tributaryconfluences, and valley confinement.

Geomorphologically speaking, a reach is asection of river along which boundary conditions aresufficiently uniform that the river maintains a nearconsistent set of process–form interactions, resultingin characteristic planform patterns and landforms inthe channel and floodplain, such as meanders, gravelbars, and oxbow lakes.

The reach is arguably the most important scale.It is the key spatial scale at which the mosaic offeatures found within river channels and floodplains(1) responds to the cascade of influences from largerspatial scales and (2) is influenced by interactions andfeedbacks between geomorphic and hydraulic unitsand smaller elements such as plants, large wood, andsediment particles within the reach. The reach is alsothe scale at which people view and interact with ariver, and the scale at which most management andrestoration work is directed.

Approaches and Timescales of AnalysisA diverse array of techniques can be applied to inves-tigate temporal changes in geomorphological formsand processes from the catchment down to the reachscale. These techniques can be broadly categorizedaccording to the disciplines within which they havebeen developed, the data sources they utilize, andthe temporal scale at which they can be applied(Table 2; Figure 2). For the present review, techniquesare divided into four major approaches: field sur-vey, remote sensing, historical, and palaeo approaches.Table 2 lists the methods and data sources includedwithin each approach, the timescales over which theyare typically applied, and their strengths and weak-ness for the characterization of geomorphologicalchange.

The choice of approach for an analysis of tem-poral change is dependent on the data sources thatare available for an area, the history of pressures inthe catchment, and the responsiveness of the riverto pressures. Some data sources are preferred, typ-ically those that are scientifically derived, unbiasedand are supported by metadata detailing methods anduncertainties/errors. However alternative data sourcescan be used when and where the preferred data are

unavailable, but this may impact on the level of detailor confidence of the resulting interpretation. A riversituated in a region with a long history of humanmodifications may require a longer timescale of anal-ysis if causal linkages are to be identified betweenpressures and channel change. Likewise, a river thatresponds slowly to external forces (e.g., a lowland, lowenergy river with cohesive banks) may require a longertimescale of analysis to fully capture the trajectory ofchange that is occurring.

Accuracy, error, and uncertainty are discussed inmore detail in a later section, but it is important to bearin mind that the reliability of data to faithfully rep-resent geomorphological forms, processes, and eventsvaries considerably within and between data sources.All data sources should be checked to determine theoriginal purpose of the data, the person or author-ity that recorded the data, when it was recordedand subsequently published, the methods or instru-ments used, and reported levels of accuracy (spatial,temporal, attribute) to determine its suitability for aparticular analysis.

Integrating Data from Different Sourcesand ScalesOne of the main challenges of a temporal analysis isto integrate data from a wide range of sources withvarying levels of reliability in order to detect gen-uine changes in the catchment, floodplain, and riverchannel. This is where a geographical informationsystem (GIS) becomes particularly useful. Once thedatasets are correctly loaded into a GIS, they canbe queried and analyzed using a veritable toolbox oftechniques.

A chronology to visualize the changes that haveoccurred in the catchment, riparian corridor, andchannel over time provides a useful way of synthesiz-ing changes and their potential causes18,19 (Figure 3).The chronology pulls together information on thecharacteristics that influence geomorphological pro-cesses and those that respond to changes in those pro-cesses. This allows changes in characteristics to betracked over time (e.g., land cover, riparian vegeta-tion, human interventions, river flow regime includingmajor flood or drought events, channel planform pat-tern, channel width, etc.) and also allows the causallinkages between them to be explored.

CHARACTERISTICS INVESTIGATEDAT EACH SPATIAL SCALE

This section outlines what characteristics should beexamined at each spatial scale, which approaches and



TABLE 2 Four Approaches to Investigate Temporal Change in Fluvial Geomorphology

Approach (Timescale) Methods/Data Sources Strengths Weaknesses

Field survey (n/a) River reconnaissanceMorphological quality index (MQI)River styles framework

Quick and relativelyinexpensive to conduct

Detailed information on currentchannel/floodplain formsand processes

Essential for reach scaleanalysis if data areunavailable from otherapproaches

Only applicable at the reach scaleCan only indicate possible

changeCannot estimate rates of changeRequires an experienced

geomorphologist

Remote sensing (Decades) Platforms: satellite, airplane,remotely operated vehicles(kites, drones)

Data: photography,multi/hyperspectral, altimetry(radar, light detection andranging—LiDAR, terrestrial laserscanning, TLS)

Large variety of data types thatare suitable for mostcharacteristics at all spatialscales

Aerial photography archivesextend back up to 100 years

Satellite data extend back up to30 years, and have hightemporal frequency

Most freely availablemultispectral satellite datahave low spatial resolution, soonly suitable for large spatialscales or for large, wide riversat the reach scale

High resolution photography andmultispectral data good forsegment and reach scale, butare expensive topurchase/commission.

Data processing andinterpretation requiresspecialist knowledge

Historical (Centuries) MapsLand/tax surveysAgricultural censusesRiver topographic surveysMonitoring station recordsDocumentary evidence (diaries,

deeds, estate records, etc)Photography, paintings, etc.

Historical maps can extend thetimescale of analysis tocenturies, and be used tostudy many characteristics

Topographic surveys andgauging station records areoften the only data sourcesfor bed level changes.

Documentary evidence cancorroborate evidence fromother data sources

Information is captured andinterpreted through the ‘lens’of the observer

Availability and reliability ofsources is highly variable, andboth generally decrease as theanalysis is extended furtherback in time

Scale and original purpose of amap limits its application

Palaeo (Millennia) SedimentologyStratigraphyDating: radiocarbon, OSL, tree

rings

Insight into the underlyingprocesses

Provides accurate dating. Wellconstrained layers can bedated to decadal or evenannual resolution

Requires specialist knowledgeDating using OSL and

radiocarbon is expensive

data sources are recommended to investigate them,and how they can be analyzed and interpreted toquantify temporal change. While we identify preferredapproaches and data sources, we recognize that thesemay not be available for every location or time periodbeing investigated. Therefore, a range of alternativedata sources is presented for all characteristics tomaximize the likelihood of finding information tosupport the characterization. Finally, some discussionof the limitations of data sources is presented, butreaders are referred to Table 2 for a general overview

of the timescales of analysis, strengths, and weaknessof different approaches.

Catchment/Landscape Unit ScaleThe geomorphological characteristics investigated atthe catchment and landscape unit scales relate tothe underlying drivers of river change: water andsediment. This section explores temporal variationsin land cover/land use (LCLU), land topography,and precipitation and groundwater. Some important



FIGURE 2 | Temporal scales over which differentapproaches may yield useful information (solid linesare the core temporal scales; dashed lines illustratethe potential range of temporal scales).

0.01 0.1 1 10

Field survey

Remote sensing

Historical

Palaeo

Timescale (years)

100 1000 10 000 100 000

0.8 W / Wmax

ΔZ

0.6

0.4

W / W

max

0.2

0

Gravel extraction

Groynes, bankprotections structures

Land use change inthe catchment

Torrent-control worksin the catchment

Dams

Alteration of formativedischarge regime

Intensity of single controlling factor Relative effect on channel dynamic in the study reach

Time

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10

min

1800

1810

1820

1830

1840

1850

1860

1870

1880

1890

1900

1910

1920

1930

1940

1950

1960

1970

1980

1990

2000

2010

min

max

max

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

1

–0.2

–0.4

ΔZ (

m)

–0.6

–0.8

–1

0

FIGURE 3 | A chronology is a valuable tool to integrate data sources, track changes in hydrological and geomorphological characteristics overtime, and explore causal linkages. An example from the Tagliamento River that explores the impact of pressures on channel width (dimensionless,W/Wmax) and bed level. (Reprinted with permission from Ref 20. Copyright 2001 Elsevier)

characteristics, most notably geology, are not includedin this analysis, as they do not change substantiallyover the timescales under consideration. To facilitatepresentation, the catchment and landscape unit spatialscales are combined here because characteristics andkey processes that are subject to temporal change aresimilar at both scales (Table 1); however a higherlevel of detail would be expected for characteristicsevaluated at the landscape unit rather than catchmentscale.

Land Cover/Land UseLCLU is a significant controlling factor on catch-ment hydrology and sediment production. Large-scale

changes in LCLU can alter surface run-off and sedi-ment production21,22 and in severe cases even influenceregional climate and precipitation patterns.23 An anal-ysis of temporal changes in LCLU relies principallyon remote sensing and historical approaches, utiliz-ing satellite imagery, aerial photography, and land/taxsurveys.

Satellite imagery is now the most commonly useddata source for quantifying changes in LCLU over timeat the catchment and landscape unit scales. A largerange of datasets is currently available for this pur-pose, varying in the type of sensor used, spectral res-olution and range, and the spatial resolution of theresulting data, and their applicability depends on the



TABLE 3 The Minimum Spatial and Spectral Requirements for Satellite Data and Minimum Photographic Scale for Aerial Photograph forIdentification of Land Cover and Attributes24,26

Land Cover/Use

Attributes

(USGS Levels)

Minimum Spatial Resolution

Required for Identification from

Satellite Data

Spectral

Requirements1 Data Sources

Minimum Scale Required

If Aerial Photos Are Used as

the Main Data Source

Land cover (I) 20 m–1 km VIS, IR, Radar MODIS 1:40,000

Orbview-1

NOAA AVHRR

Landsat MSS

EnviSat-1 (MERIS)

Cover types (II) 10–100 m VIS, IR, Radar Landsat TM 4-7 1:20,000

Landsat ETM 7

IRS (XS)

ASTER

RADARSAT

Species dominance (III) 1–30 m VIS, IR, Panchromatic IKONOS 1:10,000

Spot 5

Quickbird

Species identification (IV) 0.1–2 m Panchromatic GeoEye-1 1:2400-1:1200

WorldView-1

OrbView-3

LiDAR

1Spectral bandwidths: VIS, visible (red, green, blue); IR, near- and middle-infrared; Radar, microwave; Panchromatic, grayscale images sensitive to the visibleand ultraviolet spectra.

spatial scale and level of detailed needed24,25 (Table 3).Aerial photography can be used to extend the tempo-ral analysis of LCLU further back in time, in manycountries to at least the mid-20th century. Other typesof data from airborne sensors (e.g., LiDAR—lightdetecting and ranging, and hyperspectral) can be usedto investigate land cover but the high spatial resolu-tion of the data and the correspondingly low spatialcoverage make them more suited to characterizationat the segment scale.

Classification of LCLU from satellite data andaerial photographs can be done manually based onimage characteristics (e.g., tone, color, texture, shapesize, context), but is now more commonly achievedusing image analysis software and semi-automated(i.e., supervised) or automated pixel-based or objectoriented approaches.27,26 Temporal change can be rep-resented simply with catchment/landscape unit sum-maries of the areal cover of the land cover types orusing a spatially explicit approach that detects changein the attributes of individual pixels. However cautionmust be exercised, particularly in relation to the latteroption, to minimize errors associated with the positionor classification of pixels. Many countries or regions

have their own land survey data sets, often based onclassification of satellite imagery, which are invaluableto LCLU change analysis and have the added benefitof harmonized LCLU classes (e.g., Corine Land Coverdata for Europe28).

Finally, some countries have long histories ofdetailed land and tax surveying (e.g., cadastral sur-veys) that can provide an excellent source of infor-mation for the analysis of LCLU. Recent work fromGermany29 and Sweden30,31 are good examples of thisapproach. The records should be checked prior to useto ensure they are spatially complete for the studyregion, and that LCLU classes are harmonized overtime. Where maps were produced as a part of theland/tax surveying process, they were typically at alarge-scale and can often be analyzed quantitativelyin a GIS, following standard processing and georefer-encing steps. For example, cadastral maps date back tothe 17th century in Sweden and have been used to doc-ument transitions in LCLU over time.30 Where recordsare in written format, additional map data, such asparcel locations on a more recent cadastral map,are needed to conduct a spatial analysis of changein a GIS.



Land Topography (Tectonics, Seismic Activity,and Mass Movements)Changes in land topography over time will impacton both catchment hydrology and sediment produc-tion. However, over the timescales of interest to riverrestoration and management, they are primarily linkedto changes in sediment production.32 Tectonic move-ment, seismic activity, and mass movements triggeredby a variety of processes (land cover change, climatevariation, deglaciation, etc.) are major producers ofcoarse and fine sediment that can be delivered to theriver channel. In this section, approaches are presentedto assess changes in sediment production over timeacross the catchment or landscape units. The deliveryof sediment to the river channel (i.e., hillslope-channelconnectivity or coupling) is addressed at the segmentscale and sediment transport at the reach scale.

Remote sensing is the preferred approach toassess changes in land topography and sediment pro-duction over time at the catchment and landscape unitscales, but historical maps and geomorphological sur-veys supported by stratigraphic and sedimentologicaldata can help to lengthen the timescale of analysis andverify results from remote sensing.

The identification and quantification of massmovements has traditionally involved geomorpholog-ical field mapping and the manual interpretation ofaerial photographs.33 While these are valuable andtime-tested methods, other remotely sensed datasetshave the potential to reduce analytical cost and time,improve feature identification, and extend spatial andtemporal coverage.34,35 For example, recent stud-ies have highlighted the possibility of automatic orsemi-automatic extraction of mass movement featuresusing high resolution LiDAR DEMs.36,37 Likewise,the volumetric analysis of sediment mobilized dur-ing mass movements can now be easily calculatedusing remotely sensed data; a DEM of Difference(DoD) can be produced by comparing DEMs of thelandscape prior to and following the event.38 Whilemany DEMs are now freely available online (e.g.,SRTM, ASTER-G-DEM), the high spatial resolutionof laser-derived DEMS (LiDAR and Terrestrial LaserScanning—TLS) has expanded the types of processesthat can be investigated and has markedly increasedthe precision of volumetric measurements. While onlylarge mass movement events could have been realis-tically quantified in the past, aerial LiDAR and ter-restrial laser scanning can resolve small changes inlandscapes that yield detailed information on coarseand fine sediment production.39

Where they exist, historical topographic andlandslide inventory maps can help to identify thelocation and extent of landslides in a region. An

individual landslide map can indicate the level oflandslide activity, but maps from different periodsin time allow the calculation of landslide frequencyand, if elevation is included on the maps, a roughestimation of sediment produced.35,40 Documentaryand photographic evidence can be used to supportgeomorphological and stratigraphic interpretations.41

Landslide susceptibility datasets derived from an anal-ysis of climate, slope, lithology, and land cover arealso excellent resources to explore the likelihood oflandslide activity and any spatial variations within acatchment42 (e.g., European Landslide SusceptibilityMap, Joint Research Centre, European Commission).

To lengthen the timeframe of the temporalanalysis, palaeo-seismic and palaeo-landslide activitycan be estimated from topographic, stratigraphic,and sedimentological evidence.33,43,44 For example,palaeo-landslide work based on stratigraphy andradiocarbon dating has demonstrated links betweenlandslide frequency and climate change45 that arerelated to glacial erosion and debutressing fol-lowing glacial retreat,46 anthropogenic land coverchanges41,47 and fluctuations in temperature and thetiming, frequency, and magnitude of rainfall.48

Precipitation and GroundwaterWater drives rivers. Thus data on precipitation,surface hydrology, and groundwater are essentialto studies of temporal change in geomorphology.The primary source of information is hydrologicalmonitoring records, which are the focus of this shortsection, though remote-sensing is increasingly beingused to characterize surface hydrology and detectchange over time.49

Hydrological monitoring records are crucialto the investigation of temporal changes in precip-itation or groundwater levels. A simple analysis oftrends in average, maximum, and minimum annualand monthly precipitation or historical intensity–duration–frequency analyzes can be extracted fromprecipitation gauge records to examine generalchanges in the input of water to the catchment.50,51

Similarly, spatial and temporal variations in ground-water levels from monitored boreholes can also beinvestigated.52 Because of the complex patterns intime series data as well as the interactions betweenglobal climate oscillations and precipitation andgroundwater levels, time series data may be bet-ter analyzed using a standardized procedure, suchas the Standardized Precipitation Index53 (SPI) orStandardized Groundwater level Index54 (SGI) thatwere designed to identify periods of drought, or theycan be investigated using non-stationary approacheslike Fourier and wavelet analysis.55 Where borehole



or piezometer data are unavailable, information ongroundwater can be obtained from age dating, chemi-cal proxies, or various hydrogeophysical techniques56

(e.g., electrical/electromagnetic methods or land-basedgravity surveying). If there is evidence of significantchanges in climate, land use, or groundwater levelsand the necessary data are available, a water budgetcan be assembled from current and historical data toexplore changes in the amount of water delivered tothe channel.57,58

Additional information on groundwater abstrac-tion or inter-basin water transfers can be obtainedfrom national scientific agencies, municipal water sup-pliers, or private water companies.

SegmentGeomorphological characteristics at the segment levelrelate to the boundary conditions that dictate chan-nel form and processes: valley setting (gradient andwidth); river channel gradient; river flows and levels;sediment delivery to the channel; and natural riparianvegetation.

Valley Setting (Gradient and Width)The valley setting is influenced by forces operatingat vastly different timescales, from tectonic upliftacting over millennia to valley blockage by landslidesand glacial surges inducing very rapid geomorphicresponse. These forces can alter the valley gradient,impacting upon river energy and sediment transport,and the valley width, which in turn impacts theplanform and lateral mobility of the river as well asthe extent of the active floodplain.

Methods from all four of the approaches(Table 2) are typically used in combination to iden-tify, confirm, and date topographic features in thelandscape that are indicative of changes in valleysetting.59–61 These features, such as river terracesand palaeo-landslides, are identified using geomor-phological surveys and remote sensing techniquesand may be depicted on historical topographic maps.Stratigraphic, sedimentological, and dating techniquesare used to confirm the origin of the features and con-strain the dates for their formation. Indicators ofchanges in valley setting, such as inset river terraces,can also be associated with rapid channel narrowingand incision caused by anthropogenic interventions.62

These changes are discussed in more detail in the fol-lowing sections, but it is important to point out herethat in addition to the changes that occur in channelgeometry and bed level, the floodplain width maybe severely diminished, which can have significantimplications for the conveyance of high flows and thedistribution of riparian vegetation.

Anthropogenic structures that influence the val-ley gradient and effective valley width should also bestudied. Large dams that span the width of the flood-plain have a profound and immediate impact on thewater surface slope, and cause significant changes inupstream bed elevation over time due to sedimentdeposition as well as profoundly influencing the flowregime and sediment delivery downstream. Extensiveartificial levée networks associated with flood con-trol structures or infrastructure (e.g., rail and roadembankments) constrict the valley width, limiting thespatial extent of flood inundation and restricting thelateral mobility of the channel. Information on engi-neering structures can be obtained from maps, gov-ernment records, or can be identified from aerial pho-tographs and remotely sensed data. Semi-automatedapproaches have been developed to identify and clas-sify earthworks in floodplains from DEMs, satellitemultispectral data and aerial photography.63 By link-ing the spatial representation of engineering structureswith a timeline of their constructions and flood levels,it becomes possible to quantify changes in floodplainwidth over time.

Channel Gradient: Changes to LongitudinalProfileChannel gradient is set initially by the valley set-ting, but is further controlled by planform patternand geometry. Channel gradient will naturally adjustover time, in response to normal geological and geo-morphological processes. Significant changes in chan-nel gradient over short timescales, though, are oftencaused by anthropogenic modifications to the channelor catchment, such as changes to channel planform(i.e., channel realignment and meander cut-off), bedlevel (e.g., weirs, dams, and gravel mining), or sedi-ment delivery from the catchment. Channel gradientis one of the fundamental properties that determinethe amount of fluvial energy available to transportsediment within the river channel.

An investigation of changes in channel gradi-ent requires information on two variables at multi-ple points in time: (1) the length of the river in thesegment and (2) the bed elevation at a minimum oftwo locations along the segment. In some situations,this information can be gathered from remote sensingsources, but the most reliable data come from histor-ical sources, particularly systematic river topographicsurveys. Accurate topographic surveying of riversbegan in the mid-19th to early 20th century in Europeand North America due to the development of riversfor navigation, flood control, and water resources,and offers a wealth of data for historical analyzes offluvial geomorphology.64,65 For example, repeated



0.0

15 20 25 30 35 40 45 50

Distance from mouth (km)

55 60 65 70 75 80 85 9010

–2.5

–5.0

2.5

5.0

7.5

Pisa

Bientinaculvert

GonfolinaGorge

Lower valdarno

1844–1847

1952–1955

1978–1980

10.0

Ele

vatio

n (

m a

.s.l.)

12.5

15.0

17.5

20.0

22.5

25.0

FIGURE 4 | Changes in bed level over time for the Arno River, Italy. (Reprinted with permission from Ref 67. Copyright 2007 Elsevier)

topographic surveys have been conducted in manyEuropean rivers and have been successfully used toquantify bed aggradation and incision associated withclimate change and anthropogenic impacts62,66–68

(Figure 4). Care must be exercised when compar-ing historical bed-levels as problems can arise dueto differences in geographical reference systems, sur-vey techniques, or in the attribute measured65 (e.g.,average bed, thalweg, or water surface level).

When systematic surveys are unavailable,channel gradient can be estimated by combiningchannel length and bed level estimates from differentsources, or from gauging station records. Channellength can be derived from plan sources includingmaps, aerial photographs, and other remotely senseddatasets, while bed-level change can be derived fromcross-sectional surveys conducted for other purposes,such as bridge construction and maintenance, floodrisk management, or river restoration.69–71 Changesin bed level can also be inferred from gauging stationrecords in an approach known as specific gauge anal-ysis, in which water surface levels at set discharges arecompared over time using empirical ratings curves foreach year of the analysis to reconstruct average bedelevation.72

If no quantitative information on historicalbed levels is available, then some indication of bedlevel changes can be inferred from a field survey. Forexample, inset floodplain terraces, undercut bridgepiers, and exposed bedrock/former floodplain layersin an alluvial river may all indicate incision.73,74

Conversely, buried engineering structures, largeuncompacted point bars, and thick fine sedimentdeposits overlying a gravel bed may indicate aggra-dation. The occurrence of these properties variesdepending on the catchment characteristics and thelocation of the segment within the catchment, so mustbe assessed by an experienced geomorphologist. Fieldsurveys of bed level change should be conducted atmultiple locations within a segment to ensure a reli-able assessment. Stratigraphic, sedimentological, andbotanical evidence can support conclusions drawnfrom a geomorphological survey and help to constrainthe timing of bed level changes.75,76

River Flows and LevelsInformation on spatial and temporal variations inriver flow and level are vital to any analysis of tempo-ral river change, because they are the primary controlon sediment mobilization, transport, and deposi-tion which in turn induce land form change. Themost accurate and complete records come from rivergauging stations, but some information can also beobtained from remotely sensed data and documentarysources.

Many indicators can be extracted from riverflow records77 (e.g., average and extreme flows andtheir timing) and can be used to estimate hydrologi-cal alteration.78 Gauging station records spanning atleast 20 years79 are required for this type of anal-ysis with a minimum temporal resolution of 1 day,or less if short-term events such as hydropeaking



are significant. The entire time series can be ana-lyzed to investigate temporal trends, in magnitude,frequency, timing, duration, and rate of change inflow; divided into time periods related to significantchanges in the flow regime (e.g., pre- and post-damconstruction); or applied to observed and ‘natural-ized’ flows, where the latter take account of modifica-tions attributable to flow abstractions or additions. Inthe second and third options, indicators are extractedfrom the pre- or naturalized time series and comparedto the post- or observed time series to assess hydrolog-ical alteration.78 Different flow characteristics may besignificant in different climatic regions and morpho-logical settings.77,80 Changes in any of these indicatorsthrough time or in comparison with natural or ‘natu-ralized’ conditions will be accompanied by hydromor-phological changes within the segment and, in mostcases will affect downstream segments as well. Whilesmall shifts may be attributable to climate change,major shifts usually reflect human interventions, withhydropeaking being a distinct indicator of artificialityin the flow regime. Figure 5 illustrates how dam con-struction and operation can impact maximum annualflow, monthly average flows, and daily flows.

Other useful geomorphological indicators aretotal and specific stream power. Total stream power(Ω) is the rate of energy dissipation per unit down-stream length (W m–1) and is calculated as

Ω = 𝜌gQS

where 𝜌 is the density of water (1000 kg m–3), gis acceleration due to gravity (9.8 m s–2), Q is dis-charge (in m3 s–1), and S is slope (in m m–1). Amorphologically meaningful discharge indicative of,for example, bankfull conditions is most informative.Thus the median annual maximum flow (Qpmedian) orthe annual flood with a 2-, 5-, or 10-year return periodhave all been used for this purpose. Specific streampower is stream power per unit channel width (W m–2)and is calculated by dividing total stream power bythe average channel width for the segment. Streampower has been correlated to a wide range of geo-morphological forms and processes, including channelsize, planform pattern, sediment transport, and islandformation.83–88

Where river gauging station records do notexist, modeling, remote sensing, historical records,and palaeo approaches can be used to estimate aspectsof the flow regime. For example, the UK’s ‘Flood Esti-mation Handbook’ presents methods to estimate floodindicators (e.g., Qpmedian) for ungauged sites based onattributes of the catchment, river network, and pre-cipitation in the UK.51 Remote sensing can provideinformation on the spatial extent or elevation of the

water surface that can be used, for example, to esti-mate flood levels and extent. River discharge cannotbe directly quantified from remotely sensed data, butcan be estimated from altimetry data by calibratingriver level with gauging station records or throughthe use of hydraulic relationships.49,89 Observationsof flood levels and extents can also be obtained fromdocumentary sources and combined with hydraulicmodeling to reconstruct flood discharge, which canextend the analysis further back in time.90 Lastly,palaeo-hydrological techniques can be used to esti-mate bankfull flow based on cross-section or planformgeometry of palaeo-channels91–93 and flood recordsbased on fluvial sediment deposits.94,95

Finally, to assess the impacts of human interven-tion on the flow of water in the river, a chronologyof anthropogenic changes in the segment should beconstructed. Of particular interest are the dates ofconstruction and the size of water flow impedancesor storage structures, be they for water diversion,hydropower, flood management, or water consump-tion purposes. Information to complete the chronol-ogy can come from any number of historical sources,including maps, aerial photographs, and water com-pany records.

Sediment DeliverySediment delivery refers to the transfer of sedi-ment from the areas of production identified atthe catchment/landscape scale to the river chan-nel. The importance of coupling (i.e., connectivity)between channels and adjacent hillslopes has beenlong acknowledged.96–98 Evaluation of the degree ofcoupling, and its change through time, is critical todrainage basin sediment dynamics as it controls inwhat proportion hillslope sediment flux contributesto drainage basin sediment storage and fluvial sedi-ment yield, respectively.99,100 Remote-sensing and fieldmapping are the most commonly used approachesfor discrete sediment sources, whereas the palaeoapproach is the preferred method for investigating dif-fuse sediment sources, particularly of fine sediment.

In the remote sensing approach, DEMs areused to track changes in the topography of sedimentsources over time to estimate sediment delivery to thechannel. For coarse sediment, the sources are typicallydiscrete and in close proximity to the river channel(e.g., landslides), whereas for fine sediment they canbe discrete (e.g., earth flows) or diffuse sources (e.g.,soil sheet erosion). The DoD method works bestwith discrete events for which there are DEMs thatcharacterize the topography before and after theevent, preferably with multiple post-event DEMs topermit the calculation of delivery rates. DEMs derived



FIGURE 5 | Alteration of flow regimecaused by dam construction and operation. (a)Annual floods on the Savannah River (USA),pre- and post-construction of the ThurmondDam in 1942 (Modified with permission fromthe author81) (b) Changes to the annualhydrograph caused by construction ofsuccessive dams, and (c) changes to daily flows(i.e., hydropeaking) as a result of damoperation on the Aragón River (Spain).(Reprinted with permission from Ref 82.Copyright 1980 Springer)

10000(a)

(b)

(c)

9000

8000

Dam construction

7000

6000

An

nu

al flo

od

pe

ak (

m3 s

–1)

Daily

ave

rage d

ischarg

e (

m3 s

–1)

Dis

charg

e (

m3 s

–1)

5000

4000

3000

2000

1000

0

200

180

160

140

120

100

80

60

40

20

0

0:00 3:00 6:00 9:00 12:00

Time

15:00 18:00 21:00

5

10

15

20

25

30

35

0Sep Oct Nov Dec Jan Feb Mar

Month

10:15 h

12:45 h8:45 h

Apr May Jun Jul Aug Sep

1885 1890 1900 05

1913–1956

1957–1999

2000–2010

1930 1940 1950 1960 1970 1980 1990 2000

Year



from photogrammetry, field surveys, and LiDAR canall be used, but consideration must be given to theuncertainty in the topographic measurements and theamount of change being detected. For example, uncer-tainties in LiDAR-derived elevation measurements arestill in the centimeter to decimeter range, so care mustbe exercised in interpreting topographic change overshort time spans, or when the amount of change beingdetected is of similar magnitude to the positionalaccuracy.101–103 A process-based geomorphologicalmapping method, which combines field surveyingand remote sensing approaches, can be particularlyuseful for assessing coarse sediment connectivity andtransfer.104

Palaeo approaches are the primary empiricalmethods for estimating the delivery of fine sedi-ment to the river channel. Stratigraphic and sedi-mentological interpretation of sediment deposits fromthe channel bed, overbank deposits, fill depositsin cutoffs and avulsions, and reservoir/lake sedi-ments can determine the amount, timing, and sourceof sediment.94,95,105–108 Additional topographic andhistorical data (e.g., historical maps, diaries, pho-tographs, legislation) can corroborate the evidencegathered from palaeo approaches, and can illustratethe impacts of altered sediment delivery on fluvialforms and processes.109 Cosmogenic approaches areparticularly useful for sediment budgeting,101,110 butmay only be feasible in areas with severe or complexfine sediment delivery problems due to the cost andexpertise involved. Alternatively, soil erosion mod-els can be combined with information on changesin LCLU, precipitation to predict changes in finesediment delivery.111–113

Finally, temporal changes in sediment deliv-ery may be identifiable in field surveys of the riverchannel.18,114 For example, a decrease in coarsesediment supply may result in bed incision, bedarmouring, a reduction in geomorphic features, or achange in river pattern (e.g., from braided to wander-ing). An increase in fine sediment delivery may resultin the clogging or burial of a coarse-grained bed,bed aggradation, and the presence of fine sedimentgeomorphic features (e.g., silt bars and benches).

Riparian Vegetation and WoodThis section covers the analysis of riparian vegetationcharacteristics for both the segment and reach scales.Riparian vegetation is not only important from anecological perspective, but its extent and structurecan indicate past river dynamics and the potentialcharacter size and quantity of wood to the river.Wood delivery in turn has important influences onflow hydraulics, sediment retention, and landform

construction within the river channel and its margins,as does the extent and morphological structure ofaquatic vegetation.115,116 At the segment scale, the keycharacteristics include the size, width, and continuityof the riparian corridor and the potential for woodrecruitment to the river. At the reach scale theyrelate to the structure, spatial distribution, and speciescomposition of the riparian vegetation; the species,abundance, morphology (i.e., submerged/emergent) ofaquatic vegetation; and the presence of large wood inthe channel and its margins. Similar data sources andmethods are used at each scale, but the level of detailrequired is higher for the reach scale. The primarysources of information come from remote sensing andecological field surveys (not discussed here), althoughdetailed land survey maps can contribute to theanalysis.

Remotely sensed data is perhaps the bestsource of information to assess change in ripar-ian vegetation over a decadal timescale, includingaerial photographs; multi- and hyperspectral datafrom airborne or satellite-based platforms; and air-borne LiDAR.117–122 The choice of remotely senseddata for a particular river segment depends upondata availability and the spatial resolution of the datain comparison to the width of the riparian corridorand the amount of change being detected. For riverswith large and continuous riparian vegetation cover,small-scale aerial photography and freely availablesatellite imagery can be used to assess segment scalecharacteristics. For segments with narrow or patchyriparian vegetation and for all reach-level characteris-tics, higher resolution data is needed. For guidance onscale and resolution for vegetation identification andclassification, see Table 3. Classification methods aresimilar to those presented earlier for LCLU, thoughadditional supporting information is often needed(e.g., DEMs and floodplain extents).

Where available, LiDAR data is particularlyuseful for characterizing riparian vegetation structureand spatial distribution. The point cloud data thatis generated by LiDAR provides information on thepresence of vegetation, vegetation height and canopystructure, which can be used to interpret vegetationtype, vegetation age, and ground topography.119,123

LiDAR can also be combined with other remotelysensed data to more thoroughly characterize riparianvegetation structure.119,124 Changes over time canbe investigated using height frequency distributions,DoDs or areal coverage of vegetation classes (e.g.,height or species).

Historical maps can be a valuable resource, par-ticularly large-scale land and tax maps that havedetailed land use information associated with them.


Raj Bridgelall

Highlight


0.20

0.30

0.40

Pro

po

rtio

n o

f la

nd

cove

r ty

pe

are

a

0.50

0.60

0.70

0.10

Floodplain age class

0.00

1870

(>127) (111 to 128) (101 to 110) (93 to 100) (45 to 58) (31 to 44) (21 to 30) (10 to 20) (1 to 9) (age range)(59 to 92)

1887 1896 1904 1938 1952 1966 1978 1987 1997 Year

Cottonwood forests

Grassland

Mixed Riparian forest

Riparian scrub

Gravel and sand bars

FIGURE 6 | Floodplain age and vegetation community in the Sacramento River. Floodplain age was determined from a historical analysis ofplanform changes using historical maps and aerial photographs. Note the shift from gravel bars to cottonwood forest to mixed riparian forest withincreasing floodplain age. (Reprinted with permission from Ref 125. Copyright 2011 Elsevier)

Historical cadastral maps have been used to assesschanges in the extent and composition of riparianvegetation.118 This information can be paired withmodern vegetation survey data to link historical chan-nel change to current vegetation structure and speciescomposition125,126 (Figure 6) or to estimate changes inhabitat type, age, and turnover.127,128

Changes in the distribution and frequency oflarge wood in the channel can be investigated effec-tively using remotely sensed data, including verticaland oblique aerial photography,129 airborne hyper-spectral data,130 and a combination of LiDAR, obliqueground photographs and field surveys.131 Structurefrom Motion (SfM) photogrammetry may be use-ful for this purpose, particularly using ground orlow-altitude aerial photography132,133 (e.g., a cameraon a pole). SfM is a newly developed technique ingeomorphology that can generate DEMs from anyseries of overlapping digital photographs with posi-tional accuracies as good as LiDAR. This opens upthe possibility of tracking volumetric changes in largewood using DoDs from historical photos.

Finally, information on riparian vegetation andlarge wood can come from other historical sourcessuch as travel accounts, ground photographs, andgovernment policy/records.64,134 For example, largewood may have been, and may still be, removed fromchannels by the local population for use as fuel or

to improve drainage and reduce local flooding, andby governments to maintain channels and protectinfrastructure. Any information on how the spatialextent and intensity of these practices has variedover time will help to develop an understanding ofhow large wood has influenced the current and pastgeomorphological condition of the river.

ReachWhile the characteristics investigated at the largerspatial scales were largely associated with controlson geomorphology, those at the reach scale are pri-marily indicators of function, channel adjustment, oralteration/artificiality. Geomorphological characteris-tics are grouped into three categories: planform mor-phology and channel migration; channel geometry;and bed sediment caliber. Riparian vegetation, aquaticvegetation, and wood should also be assessed at thereach scale, but this has already been discussed in thesegment-scale description above.

Planform Morphology and Channel MigrationThis section addresses changes in the two-dimensionalform of rivers over time, and includes river planformand associated characteristics (e.g., channel width andsinuosity, braiding, and anabranching indices); chan-nel migration; and geomorphic units within the chan-nel or floodplain. This encompasses a large variety of


Raj Bridgelall

Highlight


characteristics, but they are united by the data sourcesand analytical techniques used to investigate tempo-ral change.135 Analysis of these characteristics (e.g.,channel pattern, channel width) allows reconstruc-tion of evolutionary trajectory of river morphology.This is crucial in river management for understand-ing present morphology and processes and predictingpossible channel evolution in the near future.

The analysis of temporal change in planformrelies primarily on remotely sensed data and histor-ical maps. In fact, these sources are often used incombination. Aerial photographs or satellite data arefrequently used to characterize recent planform, andhistorical maps to extend the analyzes further back intime. The basic premise of the analysis is to overlayimages from multiple years and check to see if therehas been a change in the position of a feature (e.g.,bankline, Figure 7) or a change in the characteristicsof a feature (e.g., channel width, Figure 3). Becausethis type of analysis is based on a comparison of geo-graphical positions, it is crucial that the data sourcesare properly registered to a common coordinate sys-tem in a GIS and accuracy/uncertainty is estimated foreach source and at each time point.

Maps, aerial photography, and satellite imagerycan all be used to investigate temporal changes inrivers that cover the full range of sizes, patterns, anddynamics. The major consideration is the scale ofthe data sources in relation to the size of the featurebeing detected (e.g., channel width) and the amountof change being detected (e.g., lateral migration).Consequently, studies of temporal change in narrowor slowly adjusting rivers need large-scale maps andaerial photographs (minimum 1:10,000 scale) orhigh resolution satellite imagery.136–139 Large anddynamic rivers can be studied with smaller-scalemaps and aerial photographs120,140,141,20 or withcoarse-resolution satellite data.142–144 Infrared bandsof multispectral satellite data, e.g., MODIS band 2or Landsat Thematic Mapper band 5, can be usedto automatically segregate water and land based ona pixel threshold and so to differentiate banklines,particularly for large rivers with clear water.145 Inaddition, well-tested band ratios can be used todiscriminate vegetated from unvegetated surfaces(Normalized Difference Vegetation Index146), andwetter from drier surfaces (Modified NormalizedDifference Water Index147). Maps, aerial photogra-phy, and satellite data can also be used to identifygeomorphic features within the channel and trackchanges in their size, frequency, and location overtime.148–151

Geomorphological surveys can provide insightsinto channel migration and changes in channel width,

particularly when combined with botanical, sedi-mentological, or stratigraphic evidence.152,153 Forexample, channel narrowing can be identified fromactive accretion of sediment on opposite banks, partic-ularly when such accretion is stabilized by vegetationencroachment. The species composition and age struc-ture of riparian vegetation can also provide clues tothe direction of channel change. For example, lateralbanding in the height and ground cover of riparianvegetation due to vegetation succession can underpinestimates of lateral migration extent and rates154,155

and modes of lateral floodplain construction,156,157

whereas lateral and downstream changes in the speciescomposition or morphological structure of riparianvegetation can be indicative of distinct geomorphicfeatures, subject to contrasting inundation and soilmoisture regimes.158 Thus changes in vegetation struc-ture and composition can reveal channel bed inci-sion or aggradation76 through their influence on mois-ture conditions within the geomorphic features.159 Togo further back in time, the planform configurationof palaeochannels can be investigated based on theirtopographic signature in the floodplain and supportedby sedimentological and stratigraphic evidence.160,161

Finally, the chronology of physical pressuresshould be updated with the dates and extent ofriver realignment and channel bank and bed rein-forcement. This information can come from maps,remote-sensing, and water agency records.

Channel GeometryChannel geometry refers to the cross-sectional formor bed configuration of a channel. Changes in channelgeometry over time can indicate changes in the flow orsediment regime or direct channel interventions suchas sediment removal (mining). These are all impor-tant indicators of instability that need to be takeninto account if any channel management or restora-tion is envisaged. They are also indicators of inducedchanges in other processes, for example, bank hydrol-ogy and flow hydraulics, that may in turn impact onriparian as well as the aquatic ecology. While informa-tion on channel width can be gained from maps andaerial photography, additional data are essential for afull characterization of channel geometry. The recom-mended data source for this analysis is topographicsurveys, although several remote-sensing approachesare applicable in certain situations.

Cross-sectional surveys are the core data sourcesto examine changes in channel geometry over time.They are conducted across the river channel, perpen-dicular to the flow direction, and provide a wealthof morphometric information about the channel(bankfull and low flow channel width, bed-level,



Floodplain(a)

(b)

(c)

Study section

Water bodies at approx. low water (LW+)

Water bodies at approx. low water (LW)

Water bodies at low water (LW)

2006

km

0

N

1

1859

1812

Unvegetated gravel and sand

Active zone (study site)

FIGURE 7 | An analysis of historical maps reveal significant anthropogenic alterations to the Danube River that have impacted its planform andthe presence of geomorphic features within the channel and floodplain. (a) The Danube ‘riverscape’ prior to significant human alteration (1812), afteran intensive channelization period (1859), and after the construction of a hydropower plant and further channelization (2006). (Reprinted withpermission from Ref 128. Copyright 2013 John Wiley & Sons)



56

54

52500 100 150 200

Distance (m)

250 300 350 400 450

58

1932

1966

1979

1984

1997

60

62

Ele

vation (

m a

.s.l.)

64

66

68

FIGURE 8 | Temporal changes in cross-section form and bed level for a reach in the Brenta River, Italy (1932–1997). (Reprinted with permissionfrom Ref 68. Copyright 1997 Wiley)

water level at the time of survey, bank profiles, etc.)as well as indices used in hydraulic modeling (e.g.,bankfull cross-section area and hydraulic radius). Inregions where a network of cross-sections has beenestablished for regular monitoring, cross sectionsfrom different points in time can be easily overlaid toinvestigate changes in channel geometry (Figure 8).However internal checks on the surveys should stillbe conducted to ensure that the same reference pointsand start/end locations have been used and that therehas not been a change in the survey approach whichwould affect the way the survey was conducted, theaccuracy of the measurements or the interpretation oflandforms.

Remote sensing approaches to characterizechannel morphology fall into two categories. The firstuses altimetry data from photogrammetry, LiDAR,or TLS to create 3D models of the channel bed,i.e., DEMs. The DEMs are then used to identifyfeatures, detect changes in the morphology over time,and even calculate volumetric differences over time.This approach is mostly applicable to shallow, widerivers for which a substantial portion of the bed isexposed. Large gravel-bed rivers have been studiedextensively using this method.162–166 However highresolution LiDAR and TLS have been applied to thestudy of bank and cliff erosion in meandering riversin conjunction with aerial imagery,167,168 and recentwork has demonstrated the potential for automatedextraction of channel networks and bank faces from

LiDAR.36,169–171 LiDAR has an additional use inbathymetric data collection. Bathymetric LiDARcan measure the bed topography of water bodiesup to ca. 60-m depth with high vertical accuracy.It does not suffer from problems associated withsun glint, shadows, or surface disturbances like thespectral approach described below, but its applicationis limited to waters with low suspended sedimentconcentrations and is not suitable for application tovery shallow water172,173 (<1.5 m deep). While thefocus of discussion on remote sensing techniquesthroughout this review is on airborne and satelliteapproaches, it is worth pointing out that bathymetricsonar174,175 and other related acoustic devices (e.g.,sub-bed profiler176) can be used to map and detectchanges in riverbed topography.

The second approach estimates water depthsusing the spectral signature of aerial photographsand multi/hyperspectral data.173 This techniqueis well developed and has been used successfullyto study changes in many types of water bodies,particularly coastal areas. It is used increasinglyto characterize river bed topography (e.g., fromaerial photography164,166,177–179; airborne multi- andhyperspectral data130,177,180,181; multispectral satellitedata182). Although analysis of remotely sensed datacan provide good characterization of spatial changesin water depth, the absolute accuracy of the depthestimates depends on calibration using synchronouswater depth measurements. This has limited the use


Raj Bridgelall

Highlight


of spectrally derived depth measurements in histori-cal analyzes, but see Ref 164 for one solution to theproblem of ground-truth data for historical aerial pho-tographs. Furthermore, spectrally based bathymetryis limited not only to shallow water depths (typicallya few meters) but also requires clear water conditions,substrate with bright and reflective surfaces, good illu-mination, and minimal atmospheric interference.183

In some circumstances, a geomorphological fieldsurvey may be the only available option to assesschanges in channel geometry over time. This may betrue for remote, narrow, or slowly adjusting streamswhich may not be represented on maps or may be sub-jected to high levels of uncertainty in spatial positionwhich exceed the amount of change being detected.Channel widening can be evidenced by bank erosionor undercutting on opposite banks, whereas channelnarrowing can be indicated by stabilizing, vegetatedbars or benches on both banks. Field evidence of bedlevel changes was discussed earlier in the channel gra-dient section.

Sediment Transport and Bed Sediment SizeInformation on bed and bank sediment size and sed-iment transport is crucial to understanding the geo-morphological style and likely dynamics of rivers.Changes in the sediment regime, specifically in bed-load transport, can cause channel instability thatresults in changes to channel planform, bed levels,and type, geomorphic features, etc. Therefore infor-mation on sediment transport is key for sustainable,process-based river management and restoration.

When available, long-term monitoring data forsuspended sediment and bedload provide invaluableinformation on sediment transport within a reach.Suspended sediment is more commonly monitoredthan bedload transport, as it is an aspect of waterquality that is typically measured by water companiesand national environmental agencies. Bedload is moredifficult to quantify, and consequently monitoring sta-tions are usually located only in areas where bedloadposes a very significant river management problem.184

These sources can be readily analyzed and combinedwith river flow information to assess changes in sedi-ment delivery and transport over time.185

Unfortunately, sediment transport is not mon-itored as commonly as water discharge, and manyrivers have very limited or no sediment monitoringrecord. In this situation, changes in sediment deliv-ery and transport associated with human disturbanceto the system can be explored by creating a histori-cal inventory of engineering structures that impact thelateral or longitudinal transport of sediment (i.e., sed-iment connectivity). For coarse sediment these struc-tures can include dams, check dams, weirs, and torrent

controls,68,186 while for fine sediment they can alsoinclude drainage ditches in the catchment and artificiallevées.109 Depending on the catchment history, it mayalso be pertinent to acquire data for sediment-relatedactivities within the channel, such as records detail-ing the quantity and location of sediment dredgingor mining from the channel.187,188 This inventory canbe combined with information on land cover, topog-raphy, and sediment delivery collected at the catch-ment/landscape unit and segment scales to formulatean integrated chronology of sediment flux.

Remote sensing has enormous potential for usein sediment transport estimates and sediment budgets.This includes the detection and estimation of volumet-ric change in bed topography (i.e., the morphologicalapproach) from aerial photos189 or high resolutionDEMs,162 as well as monitoring fine sediment concen-trations using aerial photography and multispectralsatellite data.190,191 The morphological approach toestimate bed-load has been successfully used in numer-ous studies,189,192–194 in particular where direct mea-surements using samplers are difficult to carry outor where it is not possible to capture the wide spa-tial and temporal variability of sediment transport(e.g., in large gravel-bed rivers). Besides, it has beenshown that morphological methods provide reason-ably robust estimates of the time- and space-averagedbedload transport.195 These approaches rely on mor-phological changes, requiring comparison of DEMsof river channels,196 or cross-sections.68 Consideringthe increasing availability of LiDAR data, but alsothe possibility of deriving DEMs from archival aerialphotos,164 there will be more and more opportuni-ties to apply morphological approaches for sedimenttransport estimation. Even in the absence of favorableconditions for estimation of bed-load transport (themorphological approach requires that sediment trans-port is known at one cross-section within the studyreach), comparison of DEMs represents the best toolfor calculation of the sediment budget and, therefore,for assessing the evolutionary trend of channel mor-phology in a given reach.

Temporal changes in bed sediment caliber can beinvestigated using remote sensing, field surveying, andpalaeo approaches. Techniques have been developedfor the extraction of bed material size from aerial pho-tography based on image texture.197–201 For shallowrivers with non-turbid water, these techniques offer thepossibility of extracting sediment sizes from archivalaerial photos to assess change over time, particularlyin light of recent analytical developments that allowfor automated sediment size measurement withoutthe need for field calibration.202 If photography-basedmethods are not appropriate, a combination of field



survey, stratigraphy, and sedimentology can be used toidentify morphological forms and structures that areindicative of a change in bed caliber (e.g., bed armour-ing or extensive fine sediment deposits in a gravel-bedriver) and to quantify the timing and magnitude ofchange.

ACCURACY, UNCERTAINTY,AND ERROR

All data is subject to error, and so a careful appraisal oferror is essential to scientific data analysis. Accuracy,uncertainty, and error are related, are frequentlyused interchangeably, and are all associated with thereliability of the data to represent the true formor process in nature.203 The differences are subtle.When errors have been quantified for a particular datasource, they are typically referred to as ‘accuracy’;when they are unknown or not clearly defined, theterm ‘uncertainty’ is used; and the term ‘error’ is usedvariously and often when it is quantified by the user. Inthis section, we use accuracy preferentially, and reserveuncertainty or error for the discussion of estimationmethods when accuracy is not defined in advance fora dataset.

Types of AccuracyAccuracy can by subdivided into three components:position, attribute, and time.204 Positional accuracyrefers to the location of a feature on a graphical repre-sentation (e.g., map, photograph, or remotely senseddataset) in relation to other features (i.e., relativeaccuracy), or its true location in nature (i.e., absoluteaccuracy). It is influenced by the methods employed tocollect, interpret, and display the data. For example,the absolute accuracy of a river drawn on a map isdependent on the accuracy of the original survey or theresolution of the aerial photographs it is derived from;the interpretation of a feature from those sources (e.g.,banklines); the geographical projection used; and thepurpose and scale of the map. Positional accuracyis routinely quoted for national/regional maps andsatellite datasets. For example, a 1:10,000 scale UKOrdnance Survey map represents rivers at their truescale, with two banklines, when the river channel is atleast 5-m wide. Average positional accuracy is quotedat ±4 m (±7 m, 95% confidence level), meaning thatthe channel’s location on the map is on average 4 moff relative to its true position, and most points arewithin 7 m. Larger-scale maps typically have higherpositional accuracy. A UK Ordnance Survey map at1:2500 scale represents rivers to scale when they are2-m wide, and has an absolute accuracy of ±2.8 m.

When comparing maps over time in a diachronicanalysis, a threshold for planform change detectionmust be set that incorporates the positional accuracyof each source.

Attribute accuracy relates to how the identifi-cation of a feature or the characteristics of a pixelcompares to its true characteristics at that location.Some degree of interpretation, simplification, or clas-sification is inherent when data is recorded, analyzed,and displayed graphically, whether this was done bythe original surveyor and mapmaker of a historicalmap or a satellite-based sensor and a GIS technician,so attribute accuracy is always an issue. For example,for satellite-based multispectral data, the spectral sig-nature of a feature is influenced by the spatial res-olution of the data relative to the feature size, aswell as by changes in illumination (e.g., sun angle),atmospheric conditions (e.g., clouds or haze), andviewing geometry.27 The spectral signature is then pro-cessed, interpreted, and classified, all of which canaffect attribute accuracy. If features are small rela-tive to spatial resolution, pixels will represent morethan one feature (i.e., mixed pixels), adding additionaluncertainty to feature identification or classification.Techniques have been developed to help overcome thisproblem, e.g., classification of mixed pixels for landcover using fuzzy logic,27,205 but in general it is bestto consider the spatial scale of a feature a priori whenselecting a data source.

Temporal accuracy relates to the reported datefor the observations or data. This is primarily a con-cern for historical data sources, such as maps and doc-umentary evidence. For example, the time lag betweenthe initial field survey and the publication of a map canvary substantially. Often with historical maps, a singlepublication date is listed for the entire map collection,even though locations were surveyed and map sheetsproduced at different times. An additional problemwith maps is partial resurveying, in which only a por-tion of an earlier map is updated and labeled with thenew date. These resurveys introduce significant tem-poral uncertainty if the extent of the resurvey is notindicated. Temporal accuracy is less of an issue forremotely sensed datasets, which are typically time/datestamped at collection or processing, but can be aproblem for archival aerial photographs.

Assessing Accuracy/UncertaintyA wide range of data sources can be used in the anal-ysis of temporal change in river form and processes.These sources differ substantially in their inherent reli-ability and it is extremely important that sources areassessed prior to inclusion into a study. Assessmentinvolves a series of internal and external checks that



verify the positional, attribute, and temporal accuracyof a source.206 For example, a historical map can bechecked to see if it is a partial resurvey by examin-ing accompanying records, comparing the map againstearlier or later ones from the same source, or com-paring the map to other sources from the same timeperiod (e.g., land survey records, aerial photograph).If the data sources are judged to be sufficiently reliablefor the analysis, the accuracy or uncertainty of the datacan be estimated and integrated with the other sourcesin the temporal analysis to support change detec-tion. In the remainder of this section, further informa-tion is provided on estimating positional and attributeaccuracy/uncertainty.

When not reported for a data source, positionalaccuracy can be estimated by comparing positionson the graphical representation with their true loca-tion (e.g., ground control points) or with locationson a map or digital product with higher accuracy.When using a GIS, this process takes place when thedata source is registered to a geographical projection(i.e., georeferencing). To illustrate this, we provide anexample using historical maps. A similar procedurewould be conducted with aerial photographs, how-ever there are additional steps that should be takento correct for image distortion or perspective, i.e.,orthorectification (for an introduction see a relevanttextbook207). A historical map is typically registeredto a coordinate system by identifying common land-marks on a modern large-scale map.208 Landmarksshould be stable in space and time (e.g., a building),as precise as possible (e.g., the corner of a build-ing), and evenly distributed over the map. Geomet-ric transformations are then used to alter the scale,displacement, and rotation of the historical map.209

For most maps, a first-order transformation should beused unless there is significant evidence of shrinkageand distortion of the paper map.136 The output of thisprocess is an average displacement of positions on thehistorical map, which is typically represented as a rootmean square error (RMSE) and often used to assesspositional accuracy.210 However, methods to estimatepositional error and how it propagates through dataanalysis have advanced significantly, and recent workprovides further details on methods and underlyingassumptions.165,203,209,211

Attribute accuracy/uncertainty is discussed herewith a focus on raster datasets. Numerous tech-niques are available to assess uncertainty and detectchange, and the choice is dependent on the dataand type of change being detected.212,213 For landcover, error mis-classification matrices are commonlyused post-classification to estimate attribute accuracyand detect change.214,215 A fuzzy logic approach is

particularly appropriate when attribute classes arenot standardized over time or between sources,205,216

and a multilayer (GIS-based) approach can be use-ful when multiple data sources are integrated for theclassification.212,213,215 A direct comparison of pixelsbetween years can be used, but this approach is moresensitive to positional and attribute errors.

An attribute that deserves special attention iselevation. DEMs are datasets with elevation as anattribute, and a characterization of uncertainty inthese measurements is essential for detecting changesin topography over time using DoDs. Similar to thediscussion of 2D change detection, volumetric changedetection can use a single threshold of change or amore advanced spatially distributed approach.162,217

APPLYING THE TEMPORAL ANALYSISOF GEOMORPHOLOGICAL CHANGETO RIVER RESTORATION ANDMANAGEMENT

A temporal analysis allows us to peel back the lay-ers of time to explore what a channel and its flood-plain looked like in the past, how they have changedover time, how quickly these changes have occurred,and what the role of human interventions is inthese changes. In other words, it supports holistic,sustainable river restoration and management bypermitting the quantification of hydrological andgeomorphological processes (e.g., water flow, sedi-ment transport, riparian, and aquatic plant growthand succession), the identification of natural andhuman-induced alterations to these processes, and theestimation of the impacts of alteration on geomor-phological process rates and forms within a reach.This information allows managers to identify the rootcauses of geomorphological change in river-floodplainecosystems, identify constraints on restoration poten-tial, and assess the possible trajectories and timelinesof change under different management scenarios141

(Figure 9).Geomorphological degradation of a reach is

caused by changes that have occurred both withinthe reach itself and at larger spatial scales. The lossof physical habitat over time may be related to directphysical alteration of the reach in the past or cur-rent management practices, but equally it may bethe symptom of hydrological and geomorphologicalchanges that have occurred upstream (or downstream)of the reach or in the wider landscape in the past. Forexample, changes in LCLU will alter water and sed-iment production at the catchment/landscape scale,which impacts the delivery of water and sediment to



0.6Recent changes

A

B

C

Re

lative

wid

th c

ha

ng

eR

ela

tive

ele

vatio

n c

ha

ng

e

Future scenarios

0.5

0.4

0.3

Reach + basin scale interventionsReach scale interventions onlyNo interventions

0.21990 2000 2010 2020

Time

2030 2040 2050

S1

S2

S3

S4

FIGURE 9 | Possible evolutionary trajectories for the (line A) upperPiave, (line B) lower Brenta, and (line C) Cellina rivers (Italy) based ondifferent sediment management strategies (no interventions, reachscale interventions, or reach+ basin scale interventions). (Reprintedwith permission from Ref 141. Copyright 2013 Wiley)

the channel and floodplain at the segment scale, andwhich ultimately affects channel planform, dimen-sions, bed levels, bed sediment size and transport,and the creation of the hydraulic and geomorphicfeatures that support aquatic and riparian ecosys-tems at the reach scale. Urbanization of a catchmentis an excellent example of this cascade. Numerousstudies have shown how changes to sediment andwater delivery, flow regimes, and riparian vegetationassociated with urbanization can cause reach-scaleproblems such as channel incision, widening, bedarmouring, and a decrease in the diversity and fre-quency of geomorphic features.21,218 In this example,a temporal analysis of geomorphology would allowpractitioners to quantify changes in land use, chan-nel gradient, channel cross-sectional form/widthand the extent and type of riparian vegetation,and to determine how the key geomorphic pro-cesses have been altered (e.g., runoff generation,sediment delivery, river flows, sediment transport,etc.). By identifying these root causes of temporalchanges at reach scale, restoration and managementstrategies can be developed to target the underlyingprocesses to allow for a better geomorphologicalfunctioning of the channel-floodplain ecosystemor to support a comprehensive restoration plan,rather than simply tackling the symptoms of thedegradation.

Once the underlying causes of geomorpho-logical degradation are identified, the potential forrestoration of those processes can be appraised. Inheavily modified catchments or those that supportlarge human populations, industries, or services, itis unlikely that all of the processes will be restorableand some human constraints on geomorphologicalprocesses will have to persist. In these situations, an

assessment is needed on how the impacts of these con-straints can be minimized. For example, hydropowerdams may be required in the headwaters of a river forthe medium- to long-term to provide electricity forurban or industrial areas further downstream. Whilea disruption to bedload transport may be unavoidablefor a large dam, changes to the dam operation canminimize the impacts on the flow regime by mim-icking natural flow magnitude, timing, duration, andfrequency.81 In this example, the temporal analysisof river flows allows practitioners to identify whatthe natural flows would have been prior to humaninterventions and to establish patterns of flow (daily,seasonal, and annual) that are as close as possibleto the natural ones. Other impediments to coarsesediment delivery downstream of the dam can beidentified, and, if appropriate, removed to reconnectthe coarse sediment supply to the channel. If the majoralterations to geomorphology cannot be remedied,then it becomes necessary to target the reach-scalesymptoms in light of the current altered processesin order to increase geomorphological diversity andmaximize ecological benefits.

Finally, with an understanding of the hydro-logical and geomorphological processes and humanconstraints, we can begin to predict the evolution-ary trajectories of the river and floodplain under dif-ferent management scenarios, set management endgoals, and estimate timescales for change.7,219 Pre-vious conditions of the river and the direction andrates of change that resulted from alterations tohydrological and geomorphological processes in thepast give an indication of how a channel and flood-plain will respond to future changes. Practitionersare referred to a range of approaches that can aidthe development of evolutionary trajectories: assess-ment frameworks15; conceptual and empirical modelsof channel evolution220–222 and channel and flood-plain morphologies84,223; and numerical models ofmorphodynamics and sediment transport.224,225 It isimportant to stress that past condition does not meanreference condition. Human interventions to riversand catchments extend back centuries to millenniadepending on the region, and historical condition isinstead an image of what the river and floodplainlooked like under those boundary conditions andhow it changed when those conditions were altered.Restoration should aim to restore geomorphologicalfunction or work within the current boundary con-ditions (water flows, sediment fluxes, etc.) to developobtainable and sustainable targets. The rates of changein the system provide an indication of the potentialand timescale for natural channel evolution. Dynamicrivers that adjust rapidly and respond rapidly to



extrinsic factors have the best prospects for renatu-ralization. A high energy, gravel-bed river that hasincised, narrowed, or shifted to wandering planformbecause of sediment control and exploitation has agood potential to reach a good geomorphologicalcondition in a short period of time once processes arenaturalized, because rates of hydrological and geo-morphological processes are high. Conversely, a for-merly anastomosing river in a lowland setting that wassimplified, channelized, straightened, and widenedmay take considerably longer to recover a good con-dition because process rates are much slower. In thesesituations, a temporal analysis can provide a guidingprinciple with which to develop restoration measures.

CONCLUSION

The aim of holistic river basin management is to bal-ance the demands of human use and modificationof rivers with the preservation and improvement ofphysical structure and condition to support naturaland diverse ecological communities. A process-basedapproach to holistic river management works withriver processes (ecological, chemical, hydrological,

geomorphological) to facilitate the development ofsustainable management and restoration strategies.An analysis of change in fluvial geomorphology sup-ports this approach through a quantification of thekey processes at multiple scales that structure theriver and floodplain at the reach scale, an identifica-tion of alterations to these processes, and an assess-ment of how past alterations to the processes haveaffected and continue to affect the geomorphologyof the river and floodplain. This information allowsmanagers to identify the underlying causes of geomor-phological change in river-floodplain systems, iden-tify constraints on restoration potential, and assess thepossible trajectories and timelines of change under dif-ferent management scenarios. The recommendationson hydrological and geomorphological characteristics,data sources, and analysis provided in this reviewform a flexible framework with which to conduct atemporal analysis that develops an improved under-standing of how a river functions in response to tem-poral changes in the spatial hierarchy of processes thatinfluence it and so provides a foundation on whichto base holistic and sustainable river restoration andmanagement decisions.

ACKNOWLEDGMENTS

The authors would like to thank Massimo Rinaldi, Marta González Del Tánago, Christian Wolter, Stuart Lane,and two anonymous reviewers for their helpful comments on the manuscript, and Ed Oliver for modifyingand improving the figures. The review is a product of research conducted on the characterization of riverhydromorphology within the REFORM collaborative project funded by the European Union Seventh FrameworkProgramme under grant agreement 282656.

REFERENCES1. Dynesius M, Nilsson C. Fragmentation and flow regu-

lation of river systems in the northern 3rd of the World.Science 1994, 266:753–762.

2. Syvitski JPM, Vorosmarty CJ, Kettner AJ, Green P.Impact of humans on the flux of terrestrial sedi-ment to the global coastal ocean. Science 2005, 308:376–380.

3. Ward JV, Tockner K, Arscott DB, Claret C. River-ine landscape diversity. Freshwater Biol 2002,47:517–539.

4. Kondolf GM, Boulton AJ, O’Daniel S, Poole GC,Rachel FJ, Stanley EH, Wohl E, Bang A, CarlstromJ, Cristoni C, et al. Process-based ecological riverrestoration: visualizing three-dimensional connectivityand dynamic vectors to recover lost linkages. Ecol Soc2006, 11:5.

5. Sear DA. River restoration and geomorphology. AquatConserv: Mar Freshwater Ecosyst 1994, 4:169–177.

6. Palmer MA, Bernhardt ES, Allan JD, Lake PS, Alexan-der G, Brooks S, Carr J, Clayton S, Dahm CN, ShahJF, et al. Standards for ecologically successful riverrestoration. J Appl Ecol 2005, 42:208–217.

7. Dufour S, Piégay H. From the myth of a lost paradiseto targeted river restoration: forget natural referencesand focus on human benefits. River Res Appl 2009,25:568–581.

8. Beechie TJ, Sear DA, Olden JD, Pess GR, BuffingtonJM, Moir H, Roni P, Pollock MM. Process-basedprinciples for restoring river ecosystems. Bioscience2010, 60:209–222.

9. McDonald A, Lane SN, Haycock NE, Chalk EA.Rivers of dreams: on the gulf between theoretical and



practical aspects of an upland river restoration. TransInst Br Geogr 2004, 29:257–281.

10. Frissell CA, Liss WJ, Warren CE, Hurley MD. A hier-archical framework for stream habitat classification:viewing streams in a watershed context. Environ Man-age 1986, 10:199–214.

11. Montgomery DR, Buffington JM. Channel processes,classification and response potential. In: Naiman RJ,Bilby RE, eds. River ecology and Management. NewYork: Springer-Verlag; 1998, 13–42.

12. Montgomery DR. Process domains and the river con-tinuum. J Am Water Resour Assoc 1999, 35:397–410.

13. Habersack HM. The river-scaling concept (RSC): abasis for ecological assessments. Hydrobiologia 2000,422:49–60.

14. González del Tánago M, García de Jalón D. Hier-archical classification of rivers: a proposal foreco-geomorphic characterization of Spanish riverswithin the European Water Frame Directive. In: Gar-cía de Jalón D, Vizcaíno P, eds. Fifth InternationalSymposium on Ecohydraulics. Aquatic Habitats,Analysis and Restoration. Madrid, Spain: IAHRCongress Proceedings; 2004, 205–211.

15. Brierley GJ, Fryirs KA. Geomorphology and RiverManagement: Applications of the River Styles Frame-work. Oxford, UK: Blackwell; 2005.

16. Thorp JH, Thoms MC, Delong MD. The river-ine ecosystem synthesis: biocomplexity in river net-works across space and time. River Res Appl 2006,22:123–147.

17. Rinaldi M, Surian N, Comiti F, Bussettini M. Amethod for the assessment and analysis of the hydro-morphological condition of Italian streams: the Mor-phological Quality Index (MQI). Geomorphology2013, 180:96–108.

18. Sear D, Newson MD, Thorne CR. Guidebook ofApplied Fluvial Geomorphology. London: ThomasTelford; 2010, 262 p.

19. Sear DA, Newson MD, Brookes A. Sediment-relatedriver maintenance—the role of fluvial geomorphology.Earth Surf Proc Land 1995, 20:629–647.

20. Ziliani L, Surian N. Evolutionary trajectory ofchannel morphology and controlling factors ina large gravel-bed river. Geomorphology 2012,173-174:104–117.

21. Paul MJ, Meyer JL. Streams in the urban landscape.Annu Rev Ecol Syst 2001, 32:333–365.

22. Owens PN, Batalla RJ, Collins AJ, Gomez B, HicksDM, Horowitz AJ, Kondolf GM, Marden M, PageMJ, Peacock DH, et al. Fine-grained sediment in riversystems: environmental significance and managementissues. River Res Appl 2005, 21:693–717.

23. Malhi Y, Roberts JT, Betts RA, Killeen TJ, Li WH,Nobre CA. Climate change, deforestation, and the fateof the Amazon. Science 2008, 319:169–172.

24. Rogan J, Chen DM. Remote sensing technology formapping and monitoring land-cover and land-usechange. Prog Plan 2004, 61:301–325.

25. Giri C, Pengra B, Long J, Loveland TR. Next gener-ation of global land cover characterization, mapping,and monitoring. Int J Appl Earth Obs Geoinf 2013,25:30–37.

26. Morgan JL, Gergel SE, Coops NC. Aerial photogra-phy: a rapidly evolving tool for ecological manage-ment. Bioscience 2010, 60:47–59.

27. Lillesand TM, Kiefer RW, Chipman JW. In: LillesandTM, Kiefer RW, Chipman JW, eds. Remote Sensingand Image Interpretation. 5th ed. New York: JohnWiley & Sons; 2004.

28. Feranec J, Jaffrain G, Soukup T, Hazeu G. Deter-mining changes and flows in European landscapes1990–2000 using CORINE land cover data. ApplGeogr 2010, 30:19–35.

29. Bender O, Boehmer HJ, Jens D, Schumacher KP.Using GIS to analyse long-term cultural landscapechange in Southern Germany. Landsc Urban Plan2005, 70:111–125.

30. Cousins SAO. Analysis of land-cover transitions basedon 17th and 18th century cadastral maps and aerialphotographs. Landsc Ecol 2001, 16:41–54.

31. Skanes HM, Bunce RGH. Directions of land-scape change (1741–1993) in Virestad, Sweden—characterised by multivariate analysis. Landsc UrbanPlan 1997, 38:61–75.

32. Hinderer M, Kastowski M, Kamelger A, Bartolini C,Schlunegger F. River loads and modern denudation ofthe Alps—a review. Earth Sci Rev 2013, 118:11–44.

33. Geertsema M, Clague JJ, Schwab JW, Evans SG. Anoverview of recent large catastrophic landslides innorthern British Columbia, Canada. Eng Geol 2006,83:120–143.

34. Metternicht G, Hurni L, Gogu R. Remote sensing oflandslides: an analysis of the potential contribution togeo-spatial systems for hazard assessment in moun-tainous environments. Remote Sens Environ 2005,98:284–303.

35. Guzzetti F, Mondini AC, Cardinali M, Fiorucci F,Santangelo M, Chang KT. Landslide inventory maps:new tools for an old problem. Earth Sci Rev 2012,112:42–66.

36. Tarolli P, Sofia G, Dalla Fontana GD. Geomorphicfeatures extraction from high-resolution topography:landslide crowns and bank erosion. Nat Hazards2012, 61:65–83.

37. Van Den Eeckhaut M, Kerle N, Poesen J, HervasJ. Object-oriented identification of forested landslideswith derivatives of single pulse LiDAR data. Geomor-phology 2012, 173:30–42.

38. Coe JA, Glancy PA, Whitney JW. Volumetric analysisand hydrologic characterization of a modern debris



flow near Yucca Mountain, Nevada. Geomorphology1997, 20:11–28.

39. DeLong SB, Prentice CS, Hilley GE, Ebert Y. Multi-temporal ALSM change detection, sediment delivery,and process mapping at an active earthflow. Earth SurfProc Land 2012, 37:262–272.

40. Galli M, Ardizzone F, Cardinali M, Guzzetti F,Reichenbach P. Comparing landslide inventory maps.Geomorphology 2008, 94:268–289.

41. Glade T. Landslide occurrence as a response to landuse change: a review of evidence from New Zealand.Catena 2003, 51:297–314.

42. Günther A, Reichenbach P, Malet J-P, EeckhautM, Hervás J, Dashwood C, Guzzetti F. Tier-basedapproaches for landslide susceptibility assessment inEurope. Landslides 2013, 10:529–546.

43. Soldati M, Corsini A, Pasuto A. Landslides and climatechange in the Italian Dolomites since the Late glacial.Catena 2004, 55:141–161.

44. McCalpin J. Paleoseismology. 2nd ed. London: Aca-demic Press; 2009, 629 p.

45. Stoffel M, Huggel C. Effects of climate change onmass movements in mountain environments. ProgPhys Geogr 2012, 36:421–439.

46. Holm K, Bovis M, Jakob M. The landslide responseof alpine basins to post-Little Ice Age glacial thinningand retreat in southwestern British Columbia. Geo-morphology 2004, 57:201–216.

47. Dapples F, Lotter AF, van Leeuwen JFN, van derKnaap WO, Dimitriadis S, Oswald D. Paleolimnolog-ical evidence for increased landslide activity due toforest clearing and land-use since 3600 cal BP in thewestern Swiss Alps. J Paleolimnol 2002, 27:239–248.

48. Borgatti L, Soldati M. Landslides as a geomorpho-logical proxy for climate change: a record from theDolomites (northern Italy). Geomorphology 2010,120:56–64.

49. Tang QH, Gao HL, Lu H, Lettenmaier DP.Remote sensing: hydrology. Prog Phys Geogr 2009,33:490–509.

50. Longobardi A, Villani P. Trend analysis of annual andseasonal rainfall time series in the Mediterranean area.Int J Climatol 2009, 30:1538–1546.

51. Shaw EM. Hydrology in Practice. 4th ed. London:Spon Press; 2011, 543 p.

52. Bui D, Kawamura A, Tong T, Amaguchi H, Nak-agawa N. Spatio-temporal analysis of recentgroundwater-level trends in the Red River Delta,Vietnam. Hydrogeol J 2012, 20:1635–1650.

53. WMO. Standardized Precipitation Index: UserGuide. 2012. Available at: http://www.wamis.org/agm/pubs/SPI/WMO_1090_EN.pdf. (Accessed June29, 2014).

54. Bloomfield JP, Marchant BP. Analysis of ground-water drought building on the standardised

precipitation index approach. Hydrol Earth SystSci 2013, 17:4769–4787.

55. Holman IP, Rivas-Casado M, Bloomfield JP, Gur-dak JJ. Identifying non-stationary groundwater levelresponse to North Atlantic ocean-atmosphere telecon-nection patterns using wavelet coherence. Hydrogeol J2011, 19:1269–1278.

56. Green TR, Taniguchi M, Kooi H, Gurdak JJ, AllenDM, Hiscock KM, Treidel H, Aureli A. Beneath thesurface of global change: impacts of climate change ongroundwater. J Hydrol 2011, 405:532–560.

57. Hiscock KM. Hydrogeology: Principles and Practice.Malden, MA/Oxford, UK: Blackwell; 2005.

58. Claessens L, Hopkinson C, Rastetter E, Vallino J.Effect of historical changes in land use and climate onthe water budget of an urbanizing watershed. WaterResour Res 2006, 42:W03426.

59. Lave J, Avouac JP. Fluvial incision and tectonic upliftacross the Himalayas of central Nepal. J Geophys ResSolid Earth 2001, 106:26561–26591.

60. Korup O, Strom AL, Weidinger JT. Fluvial responseto large rock-slope failures: examples from theHimalayas, the Tien shan, and the southern Alps innew Zealand. Geomorphology 2006, 78:3–21.

61. Bridgland DR. The record from British Quaternaryriver systems within the context of global fluvialarchives. J Quat Sci 2010, 25:433–446.

62. Surian N, Rinaldi M. Morphological response to riverengineering and management in alluvial channels inItaly. Geomorphology 2003, 50:307–326.

63. Steinfeld CMM, Kingsford RT, Laffan SW.Semi-automated GIS techniques for detecting flood-plain earthworks. Hydrol Process 2013, 27:579–591.

64. Trimble SW. The use of historical data and artifacts ingeomorphology. Prog Phys Geogr 2008, 32:3–29.

65. Grabowski RC, Gurnell AM. Using historical data influvial geomorphology. In: Kondolf GM, Piegay H,eds. Tools in Fluvial Geomorphology. Chichester, UK:John Wiley & Sons; In press.

66. Piégay H, Peiry JL. Long profile evolution of a moun-tain stream in relation to gravel load management:example of the middle Giffre River (French Alps). Env-iron Manage 1997, 21:909–919.

67. Rinaldi M, Simon A. Bed-level adjustments in the ArnoRiver, Central Italy. Geomorphology 1998, 22:57–71.

68. Surian N, Cisotto A. Channel adjustments, bedloadtransport and sediment sources in a gravel-bed river,Brenta River, Italy. Earth Surf Proc Land 2007,32:1641–1656.

69. Brooks AP, Brierley GJ. Geomorphic responsesof lower Bega River to catchment disturbance,1851–1926. Geomorphology 1997, 18:291–304.

70. Erskine WD. Oscillatory response versus progres-sive degradation of incised channels in SoutheasternAustralia. In: Darby SE, Simon A, eds. Incised River


http://www.wamis.org/agm/pubs/SPI/WMO&uscore;1090&uscore;EN.pdf


Channels: Processes, Forms, Engineering and Man-agement. Chichester, UK: John Wiley & Sons; 1999,67–95.

71. Kondolf GM, Swanson ML. Channel adjustmentsto reservoir construction and gravel extractionalong Stony Creek, California. Environ Geol 1993,21:256–269.

72. Pinter N, Heine RA. Hydrodynamic and morphody-namic response to river engineering documented byfixed-discharge analysis, Lower Missouri River, USA.J Hydrol 2005, 302:70–91.

73. Rinaldi M. Recent channel adjustments in alluvialrivers of Tuscany, Central Italy. Earth Surf Proc Land2003, 28:587–608.

74. Natural Resources Conservation Service. Sedimentimpact assessments, Chapter 13. In: Stream Restora-tion Design National Engineering Handbook. Wash-ington DC: U.S. Department of Agriculture; 2007.

75. Dean DJ, Scott ML, Shafroth PB, Schmidt JC. Strati-graphic, sedimentologic, and dendrogeomorphic anal-yses of rapid floodplain formation along the RioGrande in Big Bend National Park, Texas. Geol SocAm Bull 2011, 123:1908–1925.

76. Hupp CR, Rinaldi M. Riparian vegetation patternsin relation to fluvial landforms and channel evolutionalong selected rivers of Tuscany (Central Italy). AnnAssoc Am Geogr 2007, 97:12–30.

77. Olden JD, Poff NL. Redundancy and the choiceof hydrologic indices for characterizing streamflowregimes. River Res Appl 2003, 19:101–121.

78. Richter BD, Baumgartner JV, Powell J, Braun DP.A method for assessing hydrologic alteration withinecosystems. Conserv Biol 1996, 10:1163–1174.

79. Huh SH, Dickey DA, Meador MR, Ruhl KE. Temporalanalysis of the frequency and duration of low andhigh streamflow: years of record needed to characterizestreamflow variability. J Hydrol 2005, 310:78–94.

80. Poff NL. Managing for variability to sustain fresh-water ecosystems. J Water Resour Plan Manag 2009,135:1–4.

81. Richter BD, Thomas GA. Restoring environmentalflows by modifying dam operations. Ecol Soc 2007,12:1.

82. Ibisate A, Díaz E, Ollero A, Acín V, Granado D. Chan-nel response to multiple damming in a meanderingriver, middle and lower Aragón River (Spain). Hydro-biologia 2013, 712:5–23.

83. Bagnold RA. An approach to the sediment transportproblem from general physics. USGS ProfessionalPaper 422, 1966, 37 p.

84. Nanson GC, Croke JC. A genetic classification offloodplains. Geomorphology 1992, 4:459–486.

85. Bagnold RA. An empirical correlation of bedloadtransport rates in flumes and natural rivers. Proc RoySoc Lond Math Phys Sci 1980, 372:453–473.

86. Gurnell A, Petts G. Trees as riparian engineers: theTagliamento River, Italy. Earth Surf Proc Land 2006,31:1558–1574.

87. Yang CT. Unit stream power and sediment transport.J Hydraul Div 1972, 98:1805–1826.

88. van den berg JH. Prediction of alluvial channelpattern of perennial rivers. Geomorphology 1995,12:259–279.

89. Tarpanelli A, Barbetta S, Brocca L, Moramarco T.River discharge estimation by using altimetry data andsimplified flood routing modeling. Remote Sens 2013,5:4145–4162.

90. Uribelarrea D, Perez-Gonzalez A, Benito G. Chan-nel changes in the Jararna and Tagus rivers (centralSpain) over the past 500 years. Quat Sci Rev 2003,22:2209–2221.

91. Gregory KJ, Lewin J, Thornes JB. Palaeohydrology inPractice: A River Basin Analysis. Chichester, UK: JohnWiley & Sons; 1987, 370 p.

92. Starkel L, Gregory KJ, Thornes JB. Temperate Palaeo-hydrology: Fluvial Processes in the Temperate Zoneduring the Last 15000 Years. Chichester, UK: JohnWiley & Sons; 1991, 548 p.

93. Sidorchuk AY, Panin AV, Borisova OK. Morphologyof river channels and surface runoff in the Volga Riverbasin (East European Plain) during the Late Glacialperiod. Geomorphology 2009, 113:137–157.

94. Macklin MG, Lewin J. River sediments, great floodsand centennial-scale Holocene climate change. J QuatSci 2003, 18:101–105.

95. Macklin MG, Benito G, Gregory KJ, Johnstone E,Lewin J, Michczynska DJ, Soja R, Starkel L, Thomdy-craft VR. Past hydrological events reflected in theHolocene fluvial record of Europe. Catena 2006,66:145–154.

96. Caine N, Swanson FJ. Geomorphic coupling of hill-slope and channel systems in two small mountainbasins. Z Geomorphol 1989, 33:189–203.

97. Fryirs KA, Brierley GJ, Preston NJ, Spencer J.Catchment-scale (dis)connectivity in sediment fluxin the upper Hunter catchment, New South Wales,Australia. Geomorphology 2007, 84:297–316.

98. Fryirs K. (Dis)Connectivity in catchment sediment cas-cades: a fresh look at the sediment delivery problem.Earth Surf Proc Land 2013, 38:30–46.

99. Reid LM, Dunne T. Rapid Evaluation of SedimentBudgets. Reiskirchen, Germany: Catena; 1996, 164 p.

100. Roberts RG, Church M. The sediment budget inseverely disturbed watersheds, Queen CharlotteRanges, British Columbia. Can J Forest Res 1986,16:1092–1106.

101. Brown AG, Carey C, Erkens G, Fuchs M, HoffmannT, Macaire JJ, Moldenhauer KM, Walling DE. Fromsedimentary records to sediment budgets: multipleapproaches to catchment sediment flux. Geomorphol-ogy 2009, 108:35–47.



102. Chartin C, Evrard O, Salvador-Blanes S, HinschbergerF, Van Oost K, Lefevre I, Daroussin J, MacaireJJ. Quantifying and modelling the impact of landconsolidation and field borders on soil redistributionin agricultural landscapes (1954–2009). Catena 2013,110:184–195.

103. Croke J, Todd P, Thompson C, Watson F, Denham R,Khanal G. The use of multi temporal LiDAR to assessbasin-scale erosion and deposition following the catas-trophic January 2011 Lockyer flood, SE Queensland,Australia. Geomorphology 2013:111–126.

104. Theler D, Reynard E, Lambiel C, Bardou E. The con-tribution of geomorphological mapping to sedimenttransfer evaluation in small alpine catchments. Geo-morphology 2010, 124:113–123.

105. Lexartza-Artza I, Wainwright J. Making connec-tions: changing sediment sources and sinks in anupland catchment. Earth Surf Proc Land 2011,36:1090–1104.

106. Hoffmann T, Erkens G, Gerlach R, Klostermann J,Lang A. Trends and controls of Holocene floodplainsedimentation in the Rhine catchment. Catena 2009,77:96–106.

107. Lewin J. Medieval environmental impacts and feed-backs: the lowland floodplains of England and Wales.Geoarchaeology 2010, 25:267–311.

108. Macklin MG, Jones AF, Lewin J. River response torapid Holocene environmental change: evidence andexplanation in British catchments. Quat Sci Rev 2010,29:1555–1576.

109. Walter RC, Merritts DJ. Natural streams andthe legacy of water-powered mills. Science 2008,319:299–304.

110. Brown AG. Colluvial and alluvial response to land usechange in Midland England: an integrated geoarchaeo-logical approach. Geomorphology 2009, 108:92–106.

111. Wischmeier WH, Smith DD. Predicting rainfall-erosion losses. Washington, DC: USDA/Science andEducation Administration, US. Govt. Printing Office;1978 55 p.

112. Renard KG, Foster GR, Weesies GA, Porter JP.RUSLE—Revise universal soil loss equation. J SoilWater Conserv 1991, 46:30–33.

113. de Vente J, Poesen J, Verstraeten G. The applicationof semi-quantitative methods and reservoir sedimenta-tion rates for the prediction of basin sediment yield inSpain. J Hydrol 2005, 305:63–86.

114. Sear DA, Newson MD. Environmental change in riverchannels: a neglected element. Towards geomorpho-logical typologies, standards and monitoring. Sci TotalEnviron 2003, 310:17–23.

115. Gurnell AM. Wood in fluvial systems. In: Shroder J Jr,Wohl E, eds. Treatise on Geomorphology, vol. 9. SanDiego, CA: Academic Press; 2013, 163–188.

116. Gurnell AM. Plants as river system engineers. EarthSurf Proc Land 2014, 39:4–25.

117. Beschta RL, Ripple WJ. River channel dynamics fol-lowing extirpation of wolves in northwestern Yellow-stone National Park, USA. Earth Surf Proc Land 2006,31:1525–1539.

118. Kondolf GM, Piégay H. Changes in the riparian zoneof the lower Eygues River, France, since 1830. LandscEcol 2007, 22:367–384.

119. Bertoldi W, Gurnell AM, Drake NA. The topographicsignature of vegetation development along a braidedriver: results of a combined analysis of airborne lidar,color air photographs, and ground measurements.Water Resour Res 2011, 47:W06525.

120. Comiti F, Da Canal M, Surian N, Mao L, Picco L,Lenzi MA. Channel adjustments and vegetation coverdynamics in a large gravel bed river over the last 200years. Geomorphology 2011, 125:147–159.

121. Dean DJ, Schmidt JC. The role of feedback mecha-nisms in historic channel changes of the lower RioGrande in the Big Bend region. Geomorphology 2011,126:333–349.

122. Henshaw AJ, Gurnell AM, Bertoldi W, Drake NA.An assessment of the degree to which Landsat TMdata can support the assessment of fluvial dynamics, asrevealed by changes in vegetation extent and channelposition, along a large river. Geomorphology 2013,202:74–85.

123. Antonarakis AS, Richards KS, Brasington J.Object-based land cover classification using airborneLiDAR. Remote Sens Environ 2008, 112:2988–2998.

124. Geerling GW, Vreeken-Buijs MJ, Jesse P, RagasAMJ, Smits AJM. Mapping river floodplain ecotopesby segmentation of spectral (Casi) and structural(Lidar) remote sensing data. River Res Appl 2009,25:795–813.

125. Greco SE, Fremier AK, Larsen EW, Plant RE. A toolfor tracking floodplain age land surface patterns on alarge meandering river with applications for ecologicalplanning and restoration design. Landsc Urban Plan2007, 81:354–373.

126. Meitzen KM. Lateral channel migration effects onriparian forest structure and composition, Conga-ree River, South Carolina, USA. Wetlands 2009,29:465–475.

127. Hohensinner S, Habersack H, Jungwirth M, Zauner G.Reconstruction of the characteristics of a natural allu-vial river-floodplain system and hydromorphologicalchanges following human modifications: the DanubeRiver (1812–1991). River Res Appl 2004, 20:25–41.

128. Hohensinner S, Jungwirth M, Muhar S, SchmutzS. Spatio-temporal habitat dynamics in a changingDanube River landscape 1812–2006. River Res Appl2011, 27:939–955.

129. Lassettre NS, Piégay H, Dufour S, Rollet A-J. Decadalchanges in distribution and frequency of wood in afree meandering river, the Ain River, France. Earth SurfProc Land 2008, 33:1098–1112.



130. Marcus WA, Legleiter CJ, Aspinall RJ, Boardman JW,Crabtree RL. High spatial resolution hyperspectralmapping of in-stream habitats, depths, and woodydebris in mountain streams. Geomorphology 2003,55:363–380.

131. Bertoldi W, Gurnell AM, Welber M. Wood recruit-ment and retention: the fate of eroded trees on abraided river explored using a combination of field andremotely-sensed data sources. Geomorphology 2013,180:146–155.

132. Fonstad MA, Dietrich JT, Courville BC, Jensen JL,Carbonneau PE. Topographic structure from motion:a new development in photogrammetric measurement.Earth Surf Proc Land 2013, 38:421–430.

133. Westoby MJ, Brasington J, Glasser NF, Hambrey MJ,Reynolds JM. ‘Structure-from-Motion’ photogramme-try: a low-cost, effective tool for geoscience applica-tions. Geomorphology 2012, 179:300–314.

134. Maser C, Sedell JR. From the Forest to the Sea:The Ecology of Wood in Streams, Rivers, Estuaries,and Oceans. Delray Beach, FL: St. Lucie Press; 1994,200 p.

135. Lawler DM. The measurement of river bank erosionand lateral channel change—a review. Earth SurfProcess Landf 1993, 18:777–821.

136. Gurnell AM, Downward SR, Jones R. Channel plan-form change on the River Dee meanders, 1876–1992.Regul River 1994, 9:187–204.

137. Gurnell AM. Channel change on the River Dee mean-ders, 1946–1992, from the analysis of air photographs.Regul River 1997, 13:13–26.

138. Hooke JM. Cutoffs galore! Occurrence and causes ofmultiple cutoffs on a meandering river. Geomorphol-ogy 2004, 61:225–238.

139. Hooke JM. Spatial variability, mechanisms and prop-agation of change in an active meandering river. Geo-morphology 2007, 84:277–296.

140. Surian N. Channel changes due to river regulation: thecase of the Piave River, Italy. Earth Surf Process Landf1999, 24:1135–1151.

141. Surian N, Ziliani L, Comiti F, Lenzi MA, Mao L.Channel adjustments and alteration of sediment fluxesin gravel-bed rivers of North-eastern Italy: potentialsand limitations for channel recovery. River Res Appl2009, 25:551–567.

142. Yao Z, Ta W, Jia X, Xiao J. Bank erosion andaccretion along the Ningxia-Inner Mongolia reaches ofthe Yellow River from 1958 to 2008. Geomorphology2011, 127:99–106.

143. Gupta N, Atkinson PM, Carling PA. Decadal lengthchanges in the fluvial planform of the River Ganga:bringing a mega-river to life with Landsat archives.Remote Sens Lett 2013, 4:1–9.

144. Mount NJ, Tate NJ, Sarker MH, Thorne CR. Evolu-tionary, multi-scale analysis of river bank line retreat

using continuous wavelet transforms: Jamuna River,Bangladesh. Geomorphology 2013, 183:82–95.

145. Pavelsky TM, Smith LC. RivWidth: a software toolfor the calculation of river widths from remotelysensed imagery. IEEE Geosci Remote Sens Lett 2008,5:70–73.

146. Rouse JC, Hass RH, Schell JA, Deering DW, Har-lan JC. Monitoring the vernal advancement and ret-rogradation (greenwave effect) of natural vegetation.NASA/Texas A&M University Report, 1974.

147. Xu HQ. Modification of normalised difference waterindex (NDWI) to enhance open water features inremotely sensed imagery. Int J Remote Sens 2006,27:3025–3033.

148. Zanoni L, Gurnell A, Drake N, Surian N. Islanddynamics in a braided river from analysis of historicalmaps and air photographs. River Res Appl 2008,24:1141–1159.

149. Latrubesse EM, Amsler ML, de Morais RP, AquinoS. The geomorphologic response of a large pristinealluvial river to tremendous deforestation in the SouthAmerican tropics: the case of the Araguaia River.Geomorphology 2009, 113:239–252.

150. Hooke JM, Yorke L. Channel bar dynamics onmulti-decadal timescales in an active meandering river.Earth Surf Process Landf 2011, 36:1910–1928.

151. White JQ, Pasternack GB, Moir HJ. Valley widthvariation influences riffle-pool location and persistenceon a rapidly incising gravel-bed river. Geomorphology2010, 121:206–221.

152. Gurnell AM. Vegetation along river corridors: hydro-geomorphological interactions. In: Gurnell AM, PettsGE, eds. Changing River Channels. Chichester, UK:John Wiley & Sons; 1995.

153. Hupp CR, Bornette G. Vegetation as a tool in theinterpretation of fluvial geomorphic processes andlandforms in humid temperate areas. In: KondolfGM, Piegay H, eds. Tools in Fluvial Geomorphol-ogy. Chichester, UK: John Wiley & Sons; 2003,269–288.

154. Perucca E, Camporeale C, Ridolfi L. Influence ofriver meandering dynamics on riparian vegetationpattern formation. J Geophys Res Biogeosci 2006,111:G01001.

155. Perucca E, Camporeale C, Ridolfi L. Significanceof the riparian vegetation dynamics on meander-ing river morphodynamics. Water Resour Res 2007,43:W03430.

156. Page K, Nanson G. Concave-bank benches and asso-ciated floodplain formation. Earth Surf Process Landf1982, 7:529–543.

157. Salo J, Kalliola R, Hakkinen I, Makinen Y, NiemelaP, Puhakka M, Coley PD. River dynamics and thediversity of amazon lowland forest. Nature 1986,322:254–258.



158. Hupp CR, Osterkamp WR. Riparian vegetation andfluvial geomorphic processes. Geomorphology 1996,14:277–295.

159. Toledo ZO, Kauffman JB. Root biomass in relationto channel morphology of headwater streams. J AmWater Resour Assoc 2001, 37:1653–1663.

160. Hoyle J, Brooks A, Brierley G, Fryirs K, Lander J.Spatial variability in the timing, nature and extent ofchannel response to typical human disturbance alongthe Upper Hunter River, New South Wales, Australia.Earth Surf Process Landf 2008, 33:868–889.

161. Brown AG, Toms PS, Carey CJ, Howard AJ, Chal-lis K. Late Pleistocene-Holocene river dynamics at theTrent-Soar confluence, England, UK. Earth Surf Pro-cess Landf 2013, 38:237–249.

162. Wheaton JM, Brasington J, Darby SE, Sear DA.Accounting for uncertainty in DEMs from repeat topo-graphic surveys: improved sediment budgets. EarthSurf Process Landf 2010, 35:136–156.

163. Moretto J, Rigon E, Mao L, Picco L, Delai F, LenziMA. Channel adjustments and island dynamics in theBrenta River (Italy) over the last 30 years. River ResAppl 2013. doi: 10.1002/rra.2676.

164. Lane SN, Widdison PE, Thomas RE, Ashworth PJ,Best JL, Lunt IA, Smith GHS, Simpson CJ. Quantifi-cation of braided river channel change using archivaldigital image analysis. Earth Surf Process Landf 2010,35:971–985.

165. Lane SN, Westaway RM, Murray HD. Estimation oferosion and deposition volumes in a large, gravel-bed,braided river using synoptic remote sensing. Earth SurfProcess Landf 2003, 28:249–271.

166. Westaway RM, Lane SN, Hicks DM. Airborne remotesensing of clear water, shallow, gravel-bed rivers usingdigital photogrammetry and image analysis. Pho-togramm Eng Remote Sens 2000, 67:1271–1281.

167. O’Neal MA, Pizzuto JE. The rates and spatial patternsof annual riverbank erosion revealed through terres-trial laser-scanner surveys of the South River, Virginia.Earth Surf Process Landf 2011, 36:695–701.

168. De Rose RC, Basher LR. Measurement of river bankand cliff erosion from sequential LIDAR and his-torical aerial photography. Geomorphology 2011,126:132–147.

169. Passalacqua P, Do Trung T, Foufoula-Georgiou E,Sapiro G, Dietrich WE. A geometric framework forchannel network extraction from lidar: nonlinear dif-fusion and geodesic paths. J Geophys Res 2010,115:F01002.

170. Passalacqua P, Belmont P, Foufoula-Georgiou E.Automatic geomorphic feature extraction from lidarin flat and engineered landscapes. Water Resour Res2012, 48:W03528.

171. Fisher GB, Bookhagen B, Amos CB. Channel planformgeometry and slopes from freely available high-spatialresolution imagery and DEM fusion: implications for

channel width scalings, erosion proxies, and fluvial sig-natures in tectonically active landscapes. Geomorphol-ogy 2013, 194:46–56.

172. Hilldale RC, Raff D. Assessing the ability of airborneLiDAR to map river bathymetry. Earth Surf ProcessLandf 2008, 33:773–783.

173. Gao J. Bathymetric mapping by means of remotesensing: methods, accuracy and limitations. Prog PhysGeogr 2009, 33:103–116.

174. Carling, Gölz, Orr, Radecki-Pawlik. The morphody-namics of fluvial sand dunes in the River Rhine, nearMainz, Germany. I. Sedimentology and morphology.Sedimentology 2000, 47:227–252.

175. Schmitt T, Mitchell NC, Ramsay ATS. Characteriz-ing uncertainties for quantifying bathymetry changebetween time-separated multibeam echo-sounder sur-veys. Cont Shelf Res 2008, 28:1166–1176.

176. Cserkesz-Nagy A, Toth T, Vajk O, Sztano O. Erosionalscours and meander development in response to riverengineering: middle Tisza region, Hungary. Proc GeolAssoc 2010, 121:238–247.

177. Winterbottom SJ, Gilvear DJ. Quantification of chan-nel bed morphology in gravel-bed rivers using airbornemultispectral imagery and aerial photography. RegulRivers Res Manage 1997, 13:489–499.

178. Carbonneau PE, Lane SN, Bergeron N. Feature basedimage processing methods applied to bathymetricmeasurements from airborne remote sensing in flu-vial environments. Earth Surf Process Landf 2006,31:1413–1423.

179. Legleiter CJ. Mapping river depth from publicly avail-able aerial images. River Res Appl 2013, 29:760–780.

180. Roberts ACB. Shallow water bathymetry using inte-grated airborne multi-spectral remote sensing. Int JRemote Sens 1999, 20:497–510.

181. Whited D, Stanford JA, Kimball JS. Application ofairborne multispectral digital imagery to quantifyriverine habitats at different base flows. River Res Appl2002, 18:583–594.

182. Legleiter CJ, Overstreet BT. Mapping gravel bedriver bathymetry from space. J Geophys Res 2012,117:F04024.

183. Legleiter CJ, Roberts DA, Lawrence RL. Spectrallybased remote sensing of river bathymetry. Earth SurfProcess Landf 2009, 34:1039–1059.

184. Rickenmann D, Turowski JM, Fritschi B, KlaiberA, Ludwig A. Bedload transport measurements atthe Erlenbach stream with geophones and automatedbasket samplers. Earth Surf Process Landf 2012,37:1000–1011.

185. Chen Z, Li J, Shen H, Zhanghua W. Yangtze River ofChina: historical analysis of discharge variability andsediment flux. Geomorphology 2001, 41:77–91.

186. Boix-Fayos C, Barbera GG, Lopez-Bermudez F,Castillo VM. Effects of check dams, reforestation and



land-use changes on river channel morphology: casestudy of the Rogativa catchment (Murcia, Spain).Geomorphology 2007, 91:103–123.

187. Martin-Vide JP, Ferrer-Boix C, Ollero A. Incisiondue to gravel mining: modeling a case study fromthe Gallego River, Spain. Geomorphology 2010,117:261–271.

188. Wishart D, Warburton J, Bracken L. Gravel extractionand planform change in a wandering gravel-bed river:the River Wear, Northern England. Geomorphology2008, 94:131–152.

189. Ham DG, Church M. Bed-material transport esti-mated from channel morphodynamics: ChilliwackRiver, British Columbia. Earth Surf Process Landf2000, 25:1123–1142.

190. Ritchie JC, Zimba PV, Everitt JH. Remote sensingtechniques to assess water quality. Photogramm EngRemote Sens 2003, 69:695–704.

191. Kilham NE, Roberts D, Singer MB. Remote sensing ofsuspended sediment concentration during turbid floodconditions on the Feather River, California-A model-ing approach. Water Resour Res 2012, 48:W01521.

192. Ashmore PE, Church M. Sediment transport and rivermorphology: a paradigm for study. In: Klingeman PC,Beschta RL, Komar PD, Bradley JB, eds. Gravel-BedRivers in the Environment. Highlands Ranch, CO:Water Resources Publications; 1998, 115–148.

193. Martin Y, Church M. Bed-material transportestimated from channel surveys—Vedder River,British-Columbia. Earth Surf Process Landf 1995,20:347–361.

194. McLean DG, Church M. Sediment transport alonglower Fraser River—2. Estimates based on thelong-term gravel budget. Water Resour Res 1999,35:2549–2559.

195. Hicks DM, Gomez B. Sediment transport. In: KondolfGM, Piegay H, eds. Tools in fluvial geomorphology.Chichester, UK: John Wiley & Sons; 2003, 425–461.

196. Brasington J, Langham J, Rumsby B. Methodolog-ical sensitivity of morphometric estimates of coarsefluvial sediment transport. Geomorphology 2003,53:299–316.

197. Ibbeken H, Schleyer R. Photo-sieving: a method forgrain-size analysis of coarse-grained, unconsolidatedbedding surfaces. Earth Surf Process Landf 1986,11:59–77.

198. Butler JB, Lane SN, Chandler JH. Automated extrac-tion of grain-size data from gravel surfaces usingdigital image processing. J Hydraul Res 2001,39:519–529.

199. Carbonneau PE, Lane SN, Bergeron NE.Catchment-scale mapping of surface grain size ingravel bed rivers using airborne digital imagery. WaterResour Res 2004, 40:W07202.

200. Carbonneau PE, Bergeron N, Lane SN. Automatedgrain size measurements from airborne remote sensing

for long profile measurements of fluvial grain sizes.Water Resour Res 2005, 41:W11426.

201. Dugdale SJ, Carbonneau PE, Campbell D. Aerialphotosieving of exposed gravel bars for the rapidcalibration of airborne grain size maps. Earth SurfProcess Landf 2010, 35:627–639.

202. Buscombe D, Rubin DM, Warrick JA. A universalapproximation of grain size from images of noncohe-sive sediment. J Geophys Res 2010, 115:F02015.

203. Tucci M, Giordano A. Positional accuracy, positionaluncertainty, and feature change detection in historicalmaps: results of an experiment. Comput EnvironUrban Syst 2011, 35:452–463.

204. Bolstad PV, Smith JL. Errors in GIS. J For 1992,90:21–29.

205. Perez-Hoyos A, Garcia-Haro FJ, San-Miguel-AyanzJ. Conventional and fuzzy comparisons of largescale land cover products: application to CORINE,GLC2000, MODIS and GlobCover in Europe. ISPRSJ Photogramm Remote Sens 2012, 74:185–201.

206. Hooke JM, Kain RJP. Historical Change in the Physi-cal Environment: A Guide to Sources and Techniques.London, UK: Butterworth Scientific; 1982, 236 p.

207. Jensen JR. Remote Sensing of the Environment: AnEarth Resource Perspective. Upper Saddle River,NJ/London: Prentice Hall; 2000, 544 p.

208. Hu B. Application of geographic information systems(GIS) in the history of cartography. World Acad SciEng Technol 2010, 42:185–189.

209. Manzano-Agugliaro F, San-Antonio-Gómez C, LópezS, Montoya FG, Gil C. Pareto-based evolutionary algo-rithms for the calculation of transformation parame-ters and accuracy assessment of historical maps. Com-put Geosci 2013, 57:124–132.

210. James LA, Hodgson ME, Ghoshal S, Latiolais MM.Geomorphic change detection using historic maps andDEM differencing: the temporal dimension of geospa-tial analysis. Geomorphology 2012, 137:181–198.

211. Mount N, Louis J. Estimation and propagation oferror in measurements of river channel movementfrom aerial imagery. Earth Surf Process Landf 2005,30:635–643.

212. Lu D, Mausel P, Brondizio E, Moran E. Changedetection techniques. Int J Remote Sens 2004,25:2365–2407.

213. Congalton RG, Green K. Assessing the Accuracy ofRemotely Sensed Data: Principles and Practices. 2nded. Boca Raton, FL: CRC Press/Taylor & Francis;2009, 200 p.

214. Congalton RG. A review of assessing the accuracy ofclassifications of remotely sensed data. Remote SensEnviron 1991, 37:35–46.

215. Fichera CR, Modica G, Pollino M. Land coverclassification and change-detection analysis usingmulti-temporal remote sensed imagery and landscapemetrics. Eur J Remote Sens 2012, 45:1–18.



216. Metternicht G. Change detection assessment usingfuzzy sets and remotely sensed data: an applicationof topographic map revision. ISPRS J PhotogrammRemote Sens 1999, 54:221–233.

217. Milan DJ, Heritage GL, Large ARG, Fuller IC. Fil-tering spatial error from DEMs: implications for mor-phological change estimation. Geomorphology 2011,125:160–171.

218. Jordan BA, Annable WK, Watson CC, Sen D. Con-trasting stream stability characteristics in adjacenturban watersheds: Santa Clara Valley, California.River Res Appl 2010, 26:1281–1297.

219. Fryirs KA, Brierley GJ. Geomorphic Analysis of RiverSystems: An Approach to Reading the Landscape.Chichester, UK: John Wiley & Sons; 2013.

220. Cluer B, Thorne C. A stream model evolution inte-grating habitat and ecosystem benefits. River Res Appl2014, 30:135–154.

221. Schumm AS, Harvey MD, Watson CC. IncisedChannel: Morphology, Dynamics and Control. Lit-tleton, CO: Water Resources Publications; 1984,200 p.

222. Simon A, Hupp CR. Geomorphic and vegetative recov-ery processes along modified Tennessee streams: aninterdisciplinary approach to disturbed systems. ForestHydrology and Watershed Management, IAHS-AISHPublications, 1986, 167 p.

223. Thorne CR. Channel types and morphological classi-fications. In: Hey RD, Thorne CR, Newson MD, eds.Applied Fluvial Geomorphology for River Engineeringand Management. Chichester, UK: John Wiley & Sons;1997.

224. Nicholas A. Morphodynamic diversity of the world’slargest rivers. Geology 2013, 41:475–478.

225. Ziliani L, Surian N, Coulthard TJ, Tarantola S.Reduced-complexity modeling of braided rivers:assessing model performance by sensitivity analysis,0calibration, and validation. J Geophys Res 2013,118:2243–2262.

FURTHER READING/RESOURCESREFORM (Restoring Rivers for Effective River Management) website and wiki, http://www.reformrivers.eu,http://www.wiki.reformrivers.eu.


http://www.reformrivers.eu

http://www.wiki.reformrivers.eu

Characterizing geomorphological change to support sustainable …€¦ · Advanced Review...

Documents

Transcript of Characterizing geomorphological change to support sustainable …€¦ · Advanced Review...