PUBLICATIONS · 2016-07-14 · 1981–2012). The MPB site also receives an average of 400 mm...

RESEARCH ARTICLE1010022013WR014994

Increased evaporation following widespread tree mortalitylimits streamflow responseJ A Biederman12 A A Harpold3 D J Gochis4 B E Ewers5 D E Reed56 S A Papuga7and P D Brooks18

1Department of Hydrology and Water Resources University of Arizona Tucson Arizona USA 2Now at USDA AgriculturalResearch Service Tucson Arizona USA 3Institute of Arctic and Alpine Research and University of Colorado BoulderColorado USA 4National Center for Atmospheric Research Boulder Colorado USA 5Department of Botany and Programin Ecology University of Wyoming Laramie Wyoming USA 6Department of Atmospheric Science and Program inEcology University of Wyoming Laramie Wyoming USA 7School of Natural Resources and the Environment Universityof Arizona Tucson Arizona USA 8Department of Geology and Geophysics University of Utah Salt Lake City Utah USA

Abstract A North American epidemic of mountain pine beetle (MPB) has disturbed over 5 million ha offorest containing headwater catchments crucial to water resources However there are limited observationsof MPB effects on partitioning of precipitation between vapor loss and streamflow and to our knowledgethese fluxes have not been observed simultaneously following disturbance We combined eddy covariancevapor loss (V) catchment streamflow (Q) and stable isotope indicators of evaporation (E) to quantify hydro-logic partitioning over 3 years in MPB-impacted and control sites Annual control V was conservative vary-ing only from 573 to 623 mm while MPB site V varied more widely from 570 to 700 mm During wetperiods MPB site V was greater than control V in spite of similar above-canopy potential evapotranspiration(PET) During a wet year annual MPB V was greater and annual Q was lower as compared to an averageyear while in a dry year essentially all water was partitioned to V Ratios of 2H and 18O in stream and soilwater showed no kinetic evaporation at the control site while MPB isotope ratios fell below the local mete-oric water line indicating greater E and snowpack sublimation (Ss) counteracted reductions in transpiration(T) and sublimation of canopy-intercepted snow (Sc) Increased E was possibly driven by reduced canopyshading of shortwave radiation which averaged 21 W m22 during summer under control forest as com-pared to 66 W m22 under MPB forest These results show that abiotic vapor losses may limit widelyexpected streamflow increases

1 Introduction

Forested watersheds are among the most reliable sources of clean water [Brown et al 2005] but montaneforests are undergoing unprecedented die-off due to increased drought fire and pathogen infestation [Wil-liams et al 2010 2013 Breshears et al 2009 Adams et al 2012 Westerling et al 2006 Huber 2005 Tokuchiet al 2004] Western North America is experiencing an unprecedented epidemic of mountain pine beetle(MPB Dendroctonus ponderoseae) [Raffa et al 2008 Hicke et al 2012] which has affected more than 5 mil-lion hectares in the western US and British Columbia [Meddens et al 2012] MPB introduce blue-stain fungithat inhibit sap flow and usually kill host trees within weeks to months [Hubbard et al 2013] Dead treesmay retain their needles for 1ndash3 years termed a red phase [Wulder et al 2006] Once needles have fallentrees are said to be in a gray phase which may last years to decades This progressive canopy loss withoutimmediate soil disturbance is different from harvest or fire challenging our ability to predict MPB effects onwater resources [Clow et al 2011 Mikkelson et al 2013a]

Extensive literature provides initial context for hydrologic response to MPB Forest disturbance studies haveestablished that streamflow (Q) usually increases in proportion to forest removed with lesser response indry regions [Stednick 1996 Brown et al 2005] However few disturbance studies have been conductedwithin interior conifer forests with annual precipitation of 600ndash1200 mm [Bosch and Hewlett 1982] repre-senting much of the North American forest impacted by MPB Effects of disturbance on Q are the weakestand most variable in such regions (r2 5 001) [Stednick 1996] Variable Q response reflects variability in vaporloss (V) due to snow sublimation (S) evaporation (E) and transpiration (T) [Troendle 1983 Brown et al 2013Hubbart et al 2007] which remain poorly understood

Key Points Water vapor loss remained high

following forest mortality Abiotic evaporation counteracted

reduced transpiration Streamflow did not increase in

contrast to expectations

Supporting Information Increased Evaporation Following Tree

Mortality - Supplemental

Correspondence toJ A Biedermanjoelbiedermanarsusdagov

CitationBiederman J A A A Harpold D JGochis B E Ewers D E Reed S APapuga and P D Brooks (2014)Increased evaporation followingwidespread tree mortality limitsstreamflow response Water ResourRes 50 5395ndash5409 doi1010022013WR014994

Received 6 NOV 2013

Accepted 7 JUN 2014

Accepted article online 13 JUN 2014

Published online 3 JUL 2014

BIEDERMAN ET AL VC 2014 American Geophysical Union All Rights Reserved 5395

Water Resources Research

PUBLICATIONS

Recent work has examined how progressive hydrologic partitioning response to gradual canopy loss mayaffect water resources and forest succession [Edburg et al 2012] Forests regulate partitioning through inter-ception transpiration and sheltering the land surface [Troendle and King 1987 Molotch et al 2009 Varholaet al 2010] Disturbance-driven partitioning changes may affect water resources including the amount andtiming of streamflow [Pugh and Gordon 2013 Brown et al 2005 Ellison et al 2012] and water availabilityfor successional vegetation [Edburg et al 2012] Succession is likely to depend on both partitioningbetween Q and V as well as how V comprises S E and T [Edburg et al 2012 Brown et al 2013 Romme et al1986] The expectations for response to MPB include an initial large decline in T followed by more gradualincreases in E from newly exposed soils and increased T by surviving and new vegetation [Edburg et al2012 Romme et al 1986] Winter precipitation is a dominant hydrologic input to many infested forests andthe reduction in sublimation of intercepted snow from the canopy (Sc) is expected to increase snowmeltvolumes [Pugh and Small 2012] The timing and relative magnitudes of these changes are unknown butthe prevailing expectation is for reduced V resulting in greater Q [Pugh and Gordon 2013]

There is limited and conflicting empirical evidence of Q response to MPB in support of these expectationsIncreased Q was reported by Potts [1984] and Bethlahmy [1974] although the increase reported by Beth-lahmy [1974] peaked 15 years post-MPB contrasting with the view that Q increases are immediate anddecline over years to decades [Bosch and Hewlett 1982 Stednick 1996 Brown et al 2005] A recent study ofMPB-impacted Colorado catchments found mixed Q responses attributed in part to variable interactionsamong topography and canopy loss in controlling V [Brooks et al 2012 Somor 2010] Snowpack observa-tions have shown either that total winter S declined or that increased snowpack sublimation (Ss) compen-sated for reduced Sc resulting in no net snowpack change [Boon 2012 Pugh and Small 2012 Biedermanet al 2014] Summer V has been relatively constant during several years following mortality [Brown et al2013] although land surface models predict reduced V with declining leaf area index under simulated dis-turbance [Bewley et al 2010 Pomeroy et al 2012 Mikkelson et al 2013b] Similarly remote-sensing prod-ucts such as MOD16 (httpwwwntsgumteduprojectmod16) that use plant-mediated algorithms [Muet al 2011] predict reduced V [Bright et al 2013 Maness et al 2013]

The variable Q response to MPB and lack of consistency between observations and models demonstratesour limited ability to predict water resources following insect-driven disturbance There is therefore a needto empirically quantify hydrologic partitioning in insect-disturbed headwater catchments Snow surveysquantify winter V over a range of spatial scales [Anderton et al 2004 Musselman et al 2008 Veatch et al2009] Eddy covariance (EC) quantifies V from a footprint of similar size to a headwater catchment (101 ha)[Wilson et al 2001 Scott 2010] Alternatively the difference between precipitation (P) and observed Q canbe used to estimate catchment-scale vapor loss (V) assuming negligible storage changes DS over annualor longer periods [Brutsaert 2005] Finally stable water isotope fractionation can identify kinetic E and dis-tinguish it from T which does not leave a fractionation signal in remaining water [Jasechko et al 2013 Clarkand Fritz 1997]

In this study we quantify hydrologic fluxes during 3 years following MPB infestation at a site with severeMPB-driven forest mortality and at a paired control site to answer the question How does MPB tree mortalityaffect partitioning of precipitation between vapor loss and water available for streamflow To our knowledgethis is the first effort combining onsite precipitation snow surveys eddy covariance vapor flux catchmentstreamflow and stable isotope indicators of evaporation to quantify hydrologic response to forestdisturbance

2 Site Description

An MPB-impacted site lsquolsquoMPBrsquorsquo and control site lsquolsquoUnimpactedrsquorsquo were identified in the Central Rocky Mountainsa region severely impacted by MPB [Meddens and Hicke 2014] Due to extensive tree mortality adjacentMPB and control sites having similar characteristics were not found However biophysical and meteorologi-cal observations at these sites showed them to be well paired (Table 1) The MPB site at Chimney Park WYis 100 km west of Cheyenne in the Medicine Bow National Forest and the headwaters of the Laramie River(Figure 1) The MPB site experienced severe infestation beginning in 2007 with peak mortality during2007ndash2009 The Unimpacted site (Figure 1) is 125 km SSE of the MPB site and 50 km NW of Denver CO

Water Resources Research 1010022013WR014994


within the Niwot Ridge Long-Term Ecological Research (LTER) observatory in the headwaters of BoulderCreek We did not observe MPB-related tree mortality at the Unimpacted site

To isolate MPB effects we selectedsites with gentle slopes (5ndash8)and similar elevation of 2750ndash3000 m (Table 1) The sites havecold winters with continuous snowcover typically lasting from Octoberuntil May or June and mean annualair temperatures of 1ndash3C TheUnimpacted site receives 800 mmmean annual precipitation (MAP)with 400 mm as snow (Table 1)[Monson et al 2002] (Niwot SNOTEL1981ndash2012) The MPB site alsoreceives an average of 400 mmsnowfall but drier summers result inlower MAP of 600ndash650 mm [Faheyet al 1985] Soils at both sites aresandy loam Overstory trees at bothsites were dominated by lodgepolepine (Pinus contorta) aged 110ndash140years since the prior stand replace-ment with mean tree heights ca11 m [Knight et al 1985] (httpamerifluxlblgov) Mean diametersat breast height were 121 and140 cm at the Unimpacted andMPB sites respectively while stemdensities were 3900 and 2500 stemsha21 [Biederman et al 2014]

3 Methods

31 Experimental DesignThe study period comprised hydro-logic years 2010ndash2012 divided intowinter and summer by the onset ofsnowmelt runoff The Unimpactedsite contained an existing above-canopy EC tower forming part of the

Table 1 Characteristics of the Unimpacted and MPB Study Sitesa

Site Elev (m) Lat LongMean Annual and Mean

Winter Precipitation (mm)Stand

Age (Year)Stem Density

(per ha)Mean StemHeight (m) DBH (cm) Soils

Unimpactedb 3000 40 020 800 100 3900 114 121 Cryochrepts2105 330 400 Sandy Loam

MPBc 2750 41 040 600ndash650 90ndash110 2500 111 140 Cryochrepts2106 070 400 Sandy Loam

aThe sites had similar elevation soil texture stand age and average tree sizes in lodgepole pine-dominated overstory Location andelevation are for the eddy covariance (EC) towers The Unimpacted site had greater stem density in the tower footprint

bScott-Denton et al [2003]cKnight et al [1985] and Fahey et al [1985]

Figure 1 Maps of the (a) MPB and (b) Unimpacted sites showing observation locationswith inset showing regional location of the sites From 2007 to 2011 MPB killed themajority of overstory trees across the area shown in Figure 1a Contour lines show eleva-tion (masl) While both sites had eddy covariance (EC) towers and precipitation gaugesthe Unimpacted site also contained the Niwot SNOTEL Only the MPB site had a gaugedcatchment and groundwater piezometer though isotope samples were collected fromstreams at both sites The MPB EC tower footprint overlapped the 15 ha MPB catchmentunder prevailing winds from the WNW Hillslope plots included subcanopy snow depthand shortwave radiation sensors soil moisture profiles and soil water lysimeters



AmeriFlux Core network (httpamerifluxlblgov) and the Niwot SNOTEL station (httpwwwwccnrcsusdagov) The MPB site was instrumented with a new tower and precipitation gauges in 2009 enabling compara-ble observations of V and meteorology We instrumented study plots at both sites to observe snow depthsubcanopy solar radiation soil moisture and soil water isotopes

Water isotopes were used for two purposes First d18O values in soil and stream water were compared tothose in rain snowfall and evolved snowpack to identify likely water sources Second differential fractiona-tion of d18O and d2H during vaporization into an unsaturated transition layer causes remaining water toplot along evaporation lines of lower slope than the local meteoric water line (LMWL) [Clark and Fritz 1997Gonfiantini 1986] Evaporation lines are imparted only by S or E not by T and we used evaporation lines toidentify abiotic V [Gustafson et al 2010 Biederman et al 2014 Jasechko et al 2013] In a previous paper wedescribed use of precipitation samples to establish LMWL and snow surveys to quantify winter S at thesesites [Biederman et al 2014]

Streamflow was gauged in a 15 ha MPB catchment overlapping the prevailing MPB tower footprint (Figure1) Although a stream of comparable size was used for isotope sampling at the Unimpacted site the Unim-pacted site did not include a gauged catchment To quantify annual changes in water storage we evaluatedsoil moisture at both sites and groundwater at the MPB site at the end of each year

32 Stand Characterization and MBP Canopy MortalityForest characteristics were obtained for the Unimpacted tower footprint using the AmeriFlux database(httpamerifluxlblgov) and for the MPB site by three methods (1) Annual ground-based surveysassessed structural metrics and quantified infected trees (2) A QuickBird image (Satellite Imaging Corp)acquired in 2011 quantified green red and gray canopy using true-color imagery to train and evaluate amaximum likelihood classifier with green band Normalized Difference Vegetation Index and Red-GreenIndex variables [Coops et al 2006] and (3) We used the annual Landsat (wwwlandsatusgsgov) mortalityclassification of Meddens and Hicke [2014] to quantify the percentage of dead canopy (red 1 gray) Basedon observations that infected trees died within months but retained their needles for ca 2 years we esti-mated the percentage of gray trees as a 2 year lag of the Landsat percentage dead which agreed wellwith the 2011 QuickBird image

33 Local ClimateAbove-canopy climate observations included temperature (T) vapor pressure deficit (VPD) wind speed(WS) and incident shortwave radiation (Rsw) For the Unimpacted site we obtained AmeriFlux Level 1 data(quality checked calibrations applied and gaps filled httpamerifluxlblgov) while we filled gaps at theMPB site (5) using linear regressions against hourly NLDAS forcing data (httpldasgsfcnasagovnldas)Wind speeds were corrected to the same height above canopy at each site using a logarithmic profile [Shut-tleworth 2012] Above-canopy potential evapotranspiration (PET) was calculated using the open water for-mulation of Shuttleworth [2012] based on daily mean VPD T WS and net radiation assuming negligibleadvection or changes in energy storage over the time scales of interest (days to months) While there aredifferences between open water and the top of a forest canopy our objective was reasonable comparisonof evaporative demand at each site

MPB site precipitation was observed using two Alter-shielded weighing gauges (ETI Systems) and a shieldedreference weighing gauge (T-200B Geonor Inc) with a wind correction for snowfall [Rasmussen et al 2012]Precipitation for the Unimpacted site was from an onsite NOAA Climate Reference Network station (wwwncdcnoaagovcrn) gap filled (lt5) with SNOTEL data in winter and a tipping bucket in summer

In 20 subcanopy plots (12 MPB and 8 Unimpacted) we recorded continuous snow depth and shortwaveradiation [Biederman et al 2014] Peak snow water equivalent (SWE) for 2010 and 2011 was determined bysnow surveys and used to estimate winter S (n 1500 site21 yr21) [Biederman et al 2014] In 2012 snow-melt commenced before surveys could be performed but winter V was quantified by EC observation

34 Soil Moisture and GroundwaterVolumetric soil moisture was observed beginning after snowmelt 2010 in six profiles per site at depths of10 30 and 60 cm (EC-5 and 5-TE Decagon Corp) Groundwater level was observed from snowmelt 2010 toJuly 2012 at the MPB site only in a piezometer 114 cm deep located 55 m east of and 3 m elevation abovethe main channel ca 500 m downstream from the MPB catchment outlet (Figure 1)



35 Eddy Covariance Vapor LossUnimpacted site water vapor loss (V) was provided as a gap-filled Level 1 product (httpamerifluxlblgov)by P Blanken and SP Burns MPB site V was measured with an open-path infrared gas analyzer (LI-7500 Li-Cor Inc) and sonic anemometer (CSAT3 Campbell Scientific Inc) Ten-Hz data were processed to 30 min Vfollowing Lee et al [2004] and periods of insufficient turbulence were removed [Gu et al 2005] Seasonal Vwere computed by three alternate methods (1) The mean of all 30 min V observations was scaled over theseason (2) Stability-screened subdaily periods were filled using lookup tables of V and weather observations(VPD and net radiation produced similar results) in 15 day periods with separate tables for morning andafternoon [Falge et al 2001] which increased V estimates lt2 and (3) Multiday gaps remaining aftermethod 2 were filled by linear interpolation of the adjacent 3 day means which increased seasonal V esti-mates lt4 While time series shown use method 2 for visual clarity reported MPB site seasonal V usemethod 1 which was most conservative with respect to our conclusion of higher-than-expected V Wetested for the uncertainty introduced by MPB gaps by imposing them on the Unimpacted data set and fill-ing by method 1 which did not affect seasonal V further suggesting that sampling uncertainty related togaps was small [Goulden et al 1996 2012 Stoy et al 2006]

Eddy covariance systems in mountain terrain are thought to suffer from uncertainty due to larger-scaleatmospheric motions such as cold-air drainage [Goulden et al 2012 Turnipseed et al 2003] However suchproblems are most severe at night and of greater concern for carbon flux (eg respiration) than for V whichis predominantly a daytime flux Research has shown that for daytime fluxes mountain EC sites may becomparably accurate to flatter sites [Turnipseed et al 2002 2003] Observations and data processing at bothsites were performed according to AmeriFlux guidelines reducing potential site bias Energy balance clo-sure at each site averaged about 80 which is typical for tall-canopy EC systems including those at moreideal sites [Stoy et al 2013 Wilson et al 2002] and energy closure was not forced

At both EC towers we differenced annual P and V as an estimate of the residual water available for stream-flow Q assuming annual changes in catchment water storage DS were negligible in comparison to P and V[Wilson et al 2001 Scott 2010] Precipitation and peak SWE were assumed spatially uniform across thetower footprints based on their small dimensions (lt15 ha) gentle terrain and low spatial variability quanti-fied by snow surveys performed in 2010 and 2011 [Biederman et al 2014 Figures 5 and 6 Table 3]

36 Stable Isotope Indicators of EvaporationLocal meteoric water lines (LMWL) on plots of d2H and d18O including both winter and summer P were pre-viously established for these sites [Biederman et al 2014] Soil water was collected from the study plots atthe same depths as soil moisture observations in four profiles at the MPB site and three profiles at theUnimpacted site Water was extracted from porous tension lysimeters (Prenart Mini Prenart Corp) duringand after snowmelt until the soils became too dry to produce samples under 70 kPa tension Stream sam-ples were collected daily by auto sampler (ISCO 3700 Teledyne ISCO Corp) during snowmelt and manuallyevery 1ndash2 weeks during low flows Isotopic signatures were not altered during up to 7 days storage insidethe auto sampler Stable isotope ratios (d2H and d18O) were determined at the University of Arizona using aliquid water isotope analyzer (DLT-100 Los Gatos Research Inc) with standard uncertainties of 098 and036 amp for d2H and d18O respectively Stream and soil water samples were placed on plots of d2H and d18Oand one-way analysis of covariance was used to test whether samples plotted along evaporation lines withslopes significantly lower than the LMWL (Matlab 2012a Mathworks Corp)

37 MPB Catchment StreamflowStreamflow was gaged in a first-order 15 ha MPB catchment overlapping the EC tower footprint under pre-vailing Westerly winds (Figure 1) The main channel passed through two culverts the lower of which wasgauged with a water level logger (Hobo U20 Onset Corp) and verified with manual stage observations sev-eral times during runoff This outlet culvert flowed partially full (tranquil flow throughout) and a ratingcurve was developed using outlet stage (Type 4 culvert flow) [Chow 1959] We assessed sensitivity ofannual Q to uncertainty in stage culvert roughness and slope using a Monte Carlo approach with 10000iterations randomly sampling the range of reasonable values determined for each input to the dischargecalculation Standard deviations of resulting annual streamflow varied from 5 to 49 mm but since mostuncertainty varied similarly among years it affected interannual Q comparisons lt10 Catchment delinea-tion was performed in ARC-GIS 101 with the TauDEM toolkit (httphydrologyuwrlusuedutaudem



taudem50) using a 1 m airbornelaser swath map and verified withfield observations We differencedannual P and Q to estimatecatchment-scale vapor loss V [Brut-saert 2005] again assuming negligi-ble change in storage DS at theannual time scale As a regional ref-erence streamflow for the 1100 km2

Laramie River above Woods Landingto which the MPB site drains wasobtained from USGS gauge 6659500(httpwaterdatausgsgov)

4 Results

41 MPB InfestationThe MPB site infestation began in2007 and by August 2011 overstoryinfection reached 77 of trees in theEC tower footprint (Figure 2) Treesreached gray phase approximately 2

years after infection and mortality and gray-phase estimates for 2010ndash2012 were 16 37 and 50 oftrees in the footprint Understory appeared to respond rapidly to the release of resources with cover reach-ing 28ndash37 by summer 2011 in gray-phase stands as compared to 3 in uninfected stands at the MPBsite (U Norton et al 2014 Nitrous oxide and methane fluxes following beetle infestation in lodgepole pineforests submitted to Soil Science Society of America Journal)

42 Local ClimateIn the 3 years of this study (2010ndash2012) precipitation at the Unimpacted site was 100 128 and 90 ofaverage (Niwot SNOTEL 1981ndash2012) while the Cinnabar Park SNOTEL 20 km distant from the MPB site indi-cated precipitation for 2010ndash2012 was 121 127 and 74 of the 2004ndash2012 average Winter precipita-tion was similar at the two sites differing by only 5ndash9 (Table 2) The Unimpacted site experienced moresummer precipitation with the largest difference in summer 2012 (125 mm at MPB and 428 mm atUnimpacted)

Patterns of snow accumulation and ablation were similar at the two sites (Figure 3) Peak SWE rangedfrom 212 to 287 mm and differed between the two sites in any year by 3ndash24 (Table 2) Springsnowmelt commenced at the same time at each site and ended about 1 week earlier at the MPB site(Figure 3)

Above-canopy T VPD WS and Rsw showed similar seasonal means (Table 2) and temporal patterns (Figure3) Mean T ranged from 240 to 253C in winter and 75ndash105C in summer Mean VPD averaged 013ndash027kPa in winter and 058ndash090 kPa in summer Mean daily WS ranged from 29 to 61 m s21 and was greater atthe Unimpacted site by 03ndash16 m s21 While above-canopy Rsw was similar at the two sites subcanopy Rsw

under MPB forest was greater by three times or more averaging 61ndash70 W m22 in summers 2010 and 2011as compared to 20ndash21 W m22 under Unimpacted forest (Table 2)

43 Soil Moisture and GroundwaterMean volumetric soil moisture reached similar maximum values of 36ndash39 during snowmelt each year atboth sites (Table S1) Trench observations showed saturation so peak moistures likely indicate porositySnowmelt was the dominant input to soil moisture though a few small increases were observed followingsummer rainfall Mean soil moisture usually declined to annually consistent minima at the end of eachsummer ca 25 for MPB and 14 for Unimpacted indicating no net change in soil moisture storage overtime for each site However at the end of the dry summer 2012 MPB soil moisture reached a lower mean of18 indicating a decline in annual storage

Figure 2 Canopy mortality status in the MPB tower footprint and MPB catchmentInfected canopy was quantified by annual ground surveys and included both livingand dead trees Dead canopy was quantified by annual Landsat imagery Gray-phasecanopy in 2011 (open circle) was quantified with a single QuickBird image Gray-phasecanopy in other years (dashed line) was estimated using a 2 year time lag of the deadcanopy



The onset of snowmelt runoff at the MPB site occurred 9 April 25 March and 9 March in the years 2010ndash2012 respectively (Figure 4) Groundwater responded to snowmelt with the water table reaching theground surface (data not shown) Summer groundwater recession did not show observable responses torain consistent with minimal response in soil moisture End-of-year groundwater elevation increased 9 cmin 2011 as compared to 2010 By the time the sensor malfunctioned in July 2012 the elevation had alreadydeclined to the ending level from 2011 Given the very dry summer it is possible that groundwater leveldeclined further by the end of 2012

44 Eddy Covariance Vapor LossVapor loss time series showed similar patterns at each site both seasonally and daily corresponding tosimilar PET dynamics (Figure 5) In the wet year 2011 MPB tower V averaged 166 of Unimpacted Vduring snowmelt and V remained greater at the MPB site throughout most of the relatively wetsummer (Figures 5 and 6) During snowmelt 2012 MPB tower V averaged 204 of Unimpacted V but amuch drier summer at the MPB site appeared to curtail V (Figures 5 and 6) During winter 2012 thesole winter of continuous operation MPB tower V was similar to Unimpacted V consistent with snowsurveys in 2010 and 2011 showing similar winter S at each site [Biederman et al 2014] PET remainedsimilar at the two sites so higher MPB site V during wet seasons meant that vapor loss rates moreclosely approached PET (Figure 5 inset)

Annual Unimpacted V was 573 612 and 623 mm in 2010ndash2012 an annual variability of only 8 so themain influence of annual P variability on partitioning was reflected in the estimated residual available forstreamflow Q (Table 3) Annual V was more variable at the MPB site In 2010 when only an estimated 16of MPB tower footprint trees had lost their needles and progressed to gray phase (Figure 2) annual V was570 mm very similar to the Unimpacted site In the wet year 2011 when MPB canopy reached 37 grayphase annual V was 700 mm or 23 greater than 2010 reducing Q by 53 In the dry year 2012 essen-tially all available water was partitioned to V with annual Vgt P leaving no water for estimated residualstreamflow Q

45 Stable Isotope Indicators of EvaporationAt both sites d18O ratios in stream and soil water were closer to evolved snowpack than summer Psuggesting snowpack as the primary water source (Table S2) and summer precipitation did not producea response in stream water isotopes Beginning with the signature of evolved snowpack MPB soil waterand stream samples plotted along d2H versus d18O evaporation lines with slopes below the LMWL(plt 005) signifying kinetic abiotic evaporation while the Unimpacted site did not indicate kineticevaporation (Figure 7) MPB evaporation lines were not different (pgt 005) between soil and streamwater so these were combined into annual lines Some equilibrium evaporation (saturated transitionlayer) likely occurred at the Unimpacted site but the net evaporation rate would be reduced by simulta-neous condensation [Clark and Fritz 1997]

Table 2 Local Climate for the Unimpacted and MPB Sitesa

Year

SWE (mm) P (mm) T (C) VPD (kPa) Wind Speed (m s21)Rsw Above Canopy

(W m22)Rsw Below Canopy

(W m22)

YR W S YR W S W S W S W S W S

Unimpacted2010 230 (2) 420 381 801 248 (59) 90 (62) 024 (019) 064 (040) 46 (26) 31 (13) 137 (63) 251 (79) 14 (9) 21 (12)2011 270 (2) 459 566 1025 240 (71) 75 (73) 024 (020) 058 (040) 54 (28) 36 (19) 129 (59) 250 (81) 20 (12)2012 222 (6) 295 428 722 240 (59) 90 (60) 027 (019) 069 (041) 53 (31) 32 (1 2) 131 (50) 240 (79)MPB2010 286 (1) 440 311 751 251 (55) 102 (62) 013 (015) 082 (045) 32 (10) 29 (08) 131 (62) 275 (74) 25 (15) 61 (39)2011 270 (3) 430 356 786 240 (70) 91 (76) 019 (022) 075 (051) 38 (12) 31 (10) 123 (55) 261 (85) 70 (40)2012 266 (10) 323 125 448 253 (59) 105 (62) 020 (016) 090 (047) 41 (13) 29 (09) 124 (45) 270 (74)

aHydrometeorological variables for the water year (YR) winter (W) and summer (S) including peak snow water equivalent (SWE) precipitation (P) temperature (T) vapor pressuredeficit (VPD) wind speed (WS) and above and below-canopy incident shortwave radiation (Rsw) The 2010 was a year of average P while 2011 was wet and 2012 was warmer anddrier than average especially at the MPB site Above-canopy weather was similar at the two sites although summer P was lower at the MPB site particularly in 2012 Below-canopyRsw averaged 70 greater in winter and 300ndash350 greater in summer at the MPB site as compared to Unimpacted For SWE the mean and one standard error of spatially distrib-uted observations are shown Precipitation values are period totals All other values shown are mean (std dev) of daily values



46 MPB CatchmentStreamflowWe do not have mortality datafor the Laramie River (1100 km2)but given the heterogeneity ofinsect infestation across largecatchments we expect that therelative rate of disturbance waslow compared to our MPB siteConsistent with variability inregional precipitation the Lara-mie River produced 200 280 and50 mm of Q in 2010ndash2012 corre-sponding to greater P in 2011and a very dry 2012 across theregion (httpwwwwccnrcsusdagov) In marked contrastMPB catchment Q declined from232 mm in 2010 to only 75 mmin 2011 mirroring greaterobserved V (Table 3) Only 5 mmof MPB catchment Q was meas-ured in the dry year 2012 indicat-ing that nearly all P waspartitioned to V

5 Discussion

Following severe insect-driven forest disturbance in the central Rocky Mountains tower vapor loss (V) washigher and streamflow (Q) was lower than expected Kinetic fractionation of stream and soil water isotopessuggested that higher evaporation (E) and snowpack sublimation (Ss) counteracted likely reduction in tran-spiration (T) and canopy sublimation (Sc) Results from these multiple lines of observation contrast with

prior studies based only onstreamflow [Potts 1984 Beth-lahmy 1974] model simulations[Bewley et al 2010 Mikkelsonet al 2013b Pomeroy et al2012] or remote sensing [Brightet al 2013 Maness et al 2013]which suggest reduced V andincreased Q Given that theobserved response contradictsexpectations we first discusseach line of evidence andassociated uncertainties Nextwe draw inferences regardingmechanisms and drivers of VFinally we discuss how observedV exceeding expectationsimpacts our understanding ofhydrologic response todisturbance

While absolute accuracy of Vcould affect water budget

Figure 3 Precipitation (P) at the (a) Unimpacted and (b) MPB sites (c) snow depth (d)temperature (T) (e) vapor pressure deficit (VPD) and (f) incoming shortwave radiation(Rsw) divided by winter (W) and summer (S) In Figure 3b dark blue shading indicatessnowmelt (SM) Patterns of winter P were similar at the two sites except during summer2012 when the MPB site was drier Snow depth time series show similar patterns of accu-mulation and peak amounts although the MPB site snowmelt ended ca 1 week earlierthan at the Unimpacted site Temporal patterns of T VPD and Rsw were very similarbetween the two sites for most of the year

Figure 4 Streamflow for the MPB catchment (15 ha) in the headwaters of the LaramieRiver (1100 km2) which is shown for regional reference and otherwise as in Figure 3Over 90 of annual streamflow (Q) in the headwater MPB catchment occurred duringsnowmelt The reduction in MPB streamflow from 2010 to 2011 was not explained byprecipitation which increased Instead lower streamflow was consistent with greatereddy covariance vapor loss (V) Shown are 5 day means



comparisons the inferences of thisstudy are based on how the tem-poral progression at the MPB sitediffers from the expected behaviorobserved at the Unimpacted siteErrors in V are known to depend onthe magnitudes of V and radiativeforcing with secondary depend-ence on vegetation type and windspeed [Richardson et al 2006]These factors are alike among site-years (Figure 2) likely increasingerror similarity and improving ourconfidence in site-year compari-sons Since energy closure was sim-ilar (ca 80) for all site-years notforcing closure would not impactsuch comparisons

Sources of uncertainty in MPB catch-ment Q include stage measurementand the stage-discharge relation-ship Monte Carlo simulationshowed comparison among yearswas robust to these uncertainties

Most importantly the amount that Q was lower in 2011 than 2010 was relatively insensitive varying between62 and 74

While the study conclusions rely primarily on direct observations of V and Q the comparison of catchmentand tower water budgets provides somewhat rare and useful cross validation [Wilson et al 2001 Goulden

et al 2012 Scott 2010] Thoughnegligible annual DS is com-monly assumed for water budgetanalyses [eg Wilson et al 2001Hubbart et al 2007 Scott 2010]this study offers relevant obser-vations First MPB streamflowreached similar low summer val-ues (lt02 mm d21) each year(Figure 4) and isotopes reflectedonly the signature of snowmeltprogressively enriched by kineticfractionation providing no evi-dence of a shift in sources (Figure7) Second soil moisture reachedsimilar low values at the end ofmost years (Table S1) indicatingno DS in soil The exception wasthe MPB site in the dry year 2012when an end-of-year reductionof 7 would suggest DS ca270 mm in the top meter of soilconsistent with observations ofYaseef at al [2010] during the dri-est years at an arid site Third the

Figure 5 (a) Vapor loss (V) and (b) potential evapotranspiration (PET) and otherwise asin Figure 3 Inset plots show the summer 15 day mean ratio VPET In the wet year 2011when 36 of MPB trees had lost their needles V was greater at the MPB site throughoutsnowmelt and the rest of summer In 2012 when 50 of trees had lost their needlesMPB site V was greater during snowmelt but less during the rest of summer which wasmuch drier at the MPB site The 2012 winter vapor losses were similar for the two sitesconsistent with snow surveys showing equal amounts of winter V in 2010 and 2011Above-canopy PET was similar at the two sites and periods of higher V at the MPB sitecorresponded to VPET closer to one than at the Unimpacted site

Figure 6 Cumulative annual precipitation (P) and vapor loss (V) and otherwise as in Fig-ure 3 When precipitation was similar at each site in 2010 annual V was similar In thewettest year 2011 MPB tower V (slope of cumulative V) which was larger than Unim-pacted V during snowmelt and throughout the summer led to 18 greater annual V In2012 larger V at the MPB site during snowmelt was counteracted by lower V during latesummer which was much drier for the MPB site resulting in a similar annual V to 2010For visual clarity gaps in the V record were filled with the mean flux of the given seasondescried as method 1 in the Methods text



small changes in end-of-year groundwater level indicated minimal DS Taking the difference of the maxi-mum and minimum spatially averaged soil moistures during the study as an estimate of drainable porositythe 9 cm water table increase in 2011 corresponds to DS of only 15 mm Finally the residual between MPBcatchment Q and tower-estimated Q ranged from 5 to 50 mm (1ndash7 of P) comparable to Wilson et al[2001] We interpret these collective observations to support the assumption of negligible DS in all site-years except MPB in the dry year 2012 when observations suggest ca 70 mm release from soil waterstorage

MPB V exceeded P by 143 mm in the dry year 2012 (Table 3) as reported during dry years at other arid andsemiarid sites [eg Dore et al 2010 Yaseef et al 2010] Vgt P is consistent with estimated DS of 270 mmbut could additionally indicate water advection into the footprint or observational uncertainty Releasefrom soil water storage could contribute to V by T of deeply rooted surviving trees and understory and by Eand T of water reaching the surface in convergent areas of the tower footprint (ie near the stream) Sincesubsurface flow paths do not necessarily conform to terrain-based catchment delineation we cannotexclude subsurface flow into the tower footprint [MacKay et al 2002 Thompson et al 2011] although weobserved no corresponding increase in streamflow While there is uncertainty in eddy covariance V asdescribed above errors such as lack of energy closure and failure to capture fluxes from large-scale atmos-pheric motions would tend to result in underestimates [Stoy et al 2006 2013] Although snow surveys indi-cated relatively uniform SWE over large scales in 2010 and 2011 [Biederman et al 2014] effective P couldhave been underestimated in 2012 when a snow survey was not performed impacting the water budget[Flerchinger and Cooley 2000] Uncertainties in observations and assumptions highlight the value of concur-rent streamflow and eddy covariance observations but they do not alter the surprising result that whateverwater was available in 2012 nearly all of it was vaporized

Stand-scale evaluations of forest disturbance often focus on reduced canopy sublimation (Sc) of interceptedsnow [Troendle 1983 Pugh and Small 2012] However process-based research has shown that intermediateforest density can minimize total winter S and enhance soil moisture through an optimal balance of inter-ception and shading [Veatch et al 2009 Gustafson et al 2010 Gray et al 2002] Accordingly severe canopyloss can drive increases in snowpack sublimation (Ss) canceling or outweighing reduced Sc [Biedermanet al 2014 Harpold et al 2014] In a study at the MPB site although Sc diminished to zero in gray-phasestands snow surveys showed total S was 36ndash38 of winter snowfall illustrating counteracting changes inSc and Ss [Biederman et al 2014 Figures 3 and 6] Kinetic fractionation of MPB snowpack confirmed at least50ndash66 of the total S was Ss [Biederman et al 2014 Figure 4] In the present study greater MPB site V ascompared to Unimpacted during snowmelt 2011 and 2012 (Figure 5) suggests that E or Ss was large asdemonstrated by Schelker et al [2013]

Vapor loss was not partitioned into T and E in this study but available evidence suggests T declined and Eincreased maintaining or possibly increasing total V (Table 3) While understory growth and increased Trates in surviving vegetation may be stimulated by increased availability of water light and nutrients [Koz-lowski 2002 McMillin 2003 Van Pelt 1999 Hubbard et al 2013 Romme et al 1986] it is very likely that in

Table 3 Water Budget Fluxes at the Two Eddy Covariance Towers and the MPB Catchmenta

Year P Year

Tower Water Budget MPB Catchment Water Budget

Vapor Loss Residual Streamflow Q Streamflow Q Residual Vapor Loss V

W S Year Year Year Year

Unimpacted2010 801 190 383 573 228 No catchment observations2011 1025 189 423 612 4132012 722 194 429 623 99MPB2010 751 154 416 570 182 232 5192011 786 160 540 700 86 75 7112012 448 207 384 591 0b 5 443

aAnnual eddy covariance vapor loss V varied by only 8 at the Unimpacted tower as compared to 23 at the MPB tower MPB vapor loss was 130 mm greater in 2011 than 2010while MPB catchment streamflow (Q) was 157 mm less despite more precipitation (P) In the dry year 2012 both tower and catchment water balances showed essentially all availablewater was partitioned to V MPB tower and catchment water balances showed good agreement in temporal patterns although V-V varied from 151 to 2148 mm

bVgt P in 2012 consistent with very low catchment Q and an annual decline in soil moisture



this study stand-scale T declined Observations following severe MPB mortality of lodgepole pine in AlbertaCanada showed increased T of 21ndash40 per unit of surviving leaf area but stand-scale T fell by 47 [Pi~na2013] With 77 mortality in the present study even a 50 T rate increase by surviving vegetation wouldresult in 66 reduction of stand-scale T Accordingly stable isotope fractionation of soil water which doesnot occur during T indicated greater kinetic E at the MPB site (Figure 7)

In light of the existing expectations for more streamflow our observations of higher V (Figure 5) duein part to greater E (Figure 7) are initially surprising After 2010 MPB site V was a greater fraction of

PET during wet periods specifi-cally snowmelt 2011 and 2012and summer 2011 Since above-canopy PET was quite similarbetween sites (Figure 5)increased Ss and E from thewet surface were likelyenhanced by canopy openingDespite similar above-canopyRsw at the two sites the MPBsite received more than 300greater subcanopy Rsw (Table2) Increases in remotely sensedwinter albedo following MPBinfestation suggest reducedshortwave shading may driveincreased Ss and E from thesurface for years to a decade ormore postdisturbance [Brightet al 2013 Vanderhoof et al2013 OrsquoHalloran et al 2012]While trees are efficient

Figure 7 Stable water isotopes in soil water and streams and the local meteoric water line (LMWL) for the (a) Unimpacted and (b) MPBsites Stream and soil water were not distinguishable from the LMWL at the Unimpacted site (pgt 005) At the MPB site evaporation lineslopes less than the LMWL (plt 005) demonstrated kinetic fractionation as a result of abiotic evaporation

Figure 8 Annual observations of hydrologic partitioning to vapor loss V and streamflowQ expressed as a percentage of precipitation P (ie VP 5 vapor loss coefficient andQP 5 runoff coefficient) Partitioning was less variable at the Unimpacted site and fol-lowed the pattern of interannual climate with lower vapor loss coefficient in wetter years(Table 3) The MPB site showed greater partitioning to V and less to Q in 2011 as com-pared to 2010 in spite of larger P In 2012 which was very dry at the MPB site tower Vexceeded P in part due to a release of stored water and very little Q was observed



evaporators and are known to transpire during snowmelt [Moore et al 2008] cold soil temperaturesmay suppress T [Smirnova et al 2000 Mellander et al 2004] possibly contributing to lower VPET atthe Unimpacted site during snowmelt (Figure 5 inset) We hypothesize that during warm dry periodssoil E would be limited by transport of water to the soil surface at the MPB site while Unimpactedsite T would more effectively vaporize deeper water sources Unfortunately large site differences in Pduring summer 2012 obscured evaluation of this hypothesis and further research is required to clarifycontrols on the seasonal dynamics of E and T following disturbance

At the Unimpacted site partitioning followed annual climate (Figure 8) with VP greatest in the driestwarmest year (2012) and lowest in the coolest wettest year (2011) Such expected response to annual cli-mate was not observed at the MPB site where VP was greater and QP was lower in the wet year 2011 ascompared to 2010 (Figure 8) These results contrast with prevailing expectations and modeled hydrologicresponses to MPB [Pugh and Gordon 2013 Bewley et al 2010 Bright et al 2013 Maness et al 2013] butprior observations have shown Q response may indeed be small following forest disturbance in semiaridmountains [Bosch and Hewlett 1982 Brooks et al 2012 Somor 2010] including even zero Q response with100 forest removal [Stednick 1996]

This study clarifies that following severe forest disturbance (1) vapor loss may be conserved or possiblyincrease particularly during periods when water is abundant at the land surface and (2) increased abiotic Ecounteracts some portion of reduced vapor loss by Sc and T These results agree with the hydrologicresponse suggested by Edburg et al [2012] but demonstrate that increases in vapor loss terms (Ss E and Tby survivors) may collectively be as large as or larger than reductions in Sc and overstory T due to canopyloss constraining expected streamflow response

Our results highlight several priorities for future research First it is necessary to determine the applic-ability of results across the range of latitude and climate within the MPB epidemic from Canada [egBrown et al 2013] to Arizona [eg Morehouse et al 2008] and to other disturbance types [Frank et al2014 Hicke et al 2012] MPB site V appeared to depend on availability of water at the surface so Vcould be less following tree removal at sites with less snow cover or more well-drained soils Secondcomplex topography has been shown to influence hydrologic partitioning [Scott 2010 Flerchinger andCooley 2000 Flerchinger et al 2010] yet it is unclear how forest disturbance impacts vary with topog-raphy We hypothesize that forest hillslopes with greater exposure to incident solar radiation may pro-duce greater abiotic V following disturbance as suggested by Rinehart et al [2008] Third ourobservations of counteracting V increases contrast with land surface models [Bewley et al 2010 Mikkel-son et al 2013b Pomeroy et al 2012] and remote-sensing products [Bright et al 2013 Maness et al2013] demonstrating a need to improve representation of V Fourth resource managers remain uncer-tain about how interactions among climate change forest disturbance and forest management willaffect water for forests [Grant et al 2013] and downstream water resources [Clow 2010 Harpold et al2012] Our observations of increased subcanopy solar radiation and greater V during periods of wetland surface suggest forest disturbance may increase the sensitivity of hydrologic partitioning to cli-mate but further work is needed to quantify the controls on S E and T their contributions to total Vand how seasonal timing of these fluxes may change [Bearup et al 2014] Finally it is possible thatincreased abiotic evaporation may limit water availability in disturbed forests with feedbacks to pro-gression of mortality and forest succession [Kaiser et al 2012 Grant et al 2013]

6 Conclusions

Numerous forest disturbance studies [see Brown et al 2005] suggest that streamflow increase will be largeinitially and return to baseline with regrowth over 10ndash15 years Surprisingly we are unaware of any observa-tions demonstrating increased streamflow during the present MPB epidemic (post-1994) While ongoingmortality may continue to alter hydrologic partitioning conterminous eddy covariance and streamflowobservations provided evidence that total vapor loss could be conserved or even increase constrainingexpected streamflow increases while stable isotope fractionation confirmed increased abiotic evaporationThe present results help clarify the lack of streamflow response observed at the larger scales relevant towater resources



ReferencesAdams H D C H Luce D D Breshears C D Allen M Weiler V C Hale and T E Huxman (2012) Ecohydrological consequences of

drought- and infestation-triggered tree die-off Insights and hypotheses Ecohydrology 5(2) 145ndash159 doi101002eco233Anderton S P S M White and B Alvera (2004) Evaluation of spatial variability in snow water equivalent for a high mountain catchment

Hydrol Processes 18(3) 435ndash453 doi101002hyp1319Bearup L A R M Maxwell D W Clow and J E McCray (2014) Hydrological effects of forest transpiration loss in bark beetle-impacted

watersheds Nat Clim Change 4(6) 481ndash486 doi101038nclimate2198Bethlahmy N (1974) More streamflow after a bark-beetle epidemic J Hydrol 23(3ndash4) 185ndash189 doi1010160022-1694(74)90001-8Bewley D Y Alila and A Varhola (2010) Variability of snow water equivalent and snow energetics across a large catchment subject to

Mountain Pine Beetle infestation and rapid salvage logging J Hydrol 388(3ndash4) 464ndash479 doi101016jjhydrol201005031Biederman J A P D Brooks A A Harpold E Gutmann D J Gochis D E Reed and E Pendall (2014) Multi-scale observations of snow

accumulation and peak snowpack following widespread insect-induced Lodgepole Pine Mortality Ecohydrology 7(1) 150ndash162 doi101002eco1342

Boon S (2012) Snow accumulation following forest disturbance Ecohydrology 5(3) 279ndash285 doi101002eco212Bosch J M and J D Hewlett (1982) A review of catchment experiments to determine the effect of vegetation changes on water yield

and evapo-transpiration J Hydrol 55(1ndash4) 3ndash23 doi1010160022-1694(82)90117-2Breshears D D O B Myers C W Meyer F J Barnes C B Zou C D Allen and W T Pockman (2009) Tree die-off in response to global

change-type drought Mortality insights from a decade of plant water potential measurements Frontiers Ecol Environ 7(4) 185ndash189doi101890080016

Bright B C J A Hicke and A J H Meddens (2013) Effects of bark beetle-caused tree mortality on biogeochemical and biogeophysicalMODIS products J Geophys Res Lett Biogeosci 118 974ndash982 doi101002jgrg20078

Brooks P D A A Harpold J A Biederman M E Litvak P D Broxton D J Gochis and B E Ewers (2012) Insect fires and climate changeImplications for snow cover water resources and ecosystem recovery in western North America Abstract B21I-02 presented at 2012 FallMeeting AGU San Francisco Calif 3ndash7 Dec

Brown A E L Zhang T A McMahon A W Western and R A Vertessy (2005) A review of paired catchment studies for determiningchanges in water yield resulting from alterations in vegetation J Hydrol 310(1ndash4) 28ndash61 doi101016jjhydrol200412010

Brown M G T A Black Z Nesic V N Foord D L Spittlehouse A L Fredeen and G Meyer (2013) Evapotranspiration and canopy charac-teristics of two lodgepole pine stands following mountain pine beetle attack Hydrological Processes 28(8) 3326ndash3340 doi101002hyp9870

Brutsaert W (2005) Hydrology An Introduction Cambridge Univ Press N YChow V T (1959) Open-Channel Hydraulics McGraw-Hill N YClark I and P Fritz (1997) Environmental Isotopes in Hydrogeology Lewis Publ Boca Raton FlaClow D W (2010) Changes in the timing of snowmelt and streamflow in Colorado A response to recent warming J Clim 23(9) 2293ndash

2306 doi1011752009jcli29511Clow D W C Rhoades J Briggs M Caldwell and W M Lewis Jr (2011) Responses of soil and water chemistry to mountain pine beetle

induced tree mortality in Grand County Colorado USA Appl Geochem Suppl 26 S174ndashS178 doi101016japgeochem201103096Coops N C M Johnson M A Wulder and J C White (2006) Assessment of QuickBird high spatial resolution imagery to detect red attack

damage due to mountain pine beetle infestation Remote Sens Environ 103(1) 67ndash80 doi101016jrse200603012Dore S T Kolb M Montes-Helu S Eckert B Sullivan B Hungate and A Finkral (2010) Carbon and water fluxes from ponderosa pine for-

ests disturbed by wildfire and thinning Ecol Appl 20(3) 663ndash683Edburg S L J A Hicke P D Brooks E G Pendall B E Ewers U Norton and A J H Meddens (2012) Cascading impacts of bark beetle-

caused tree mortality on coupled biogeophysical and biogeochemical processes Frontiers Ecol Environ 10(8) 416ndash424 doi101890110173

Ellison D M N Futter and K Bishop (2012) On the forest cover-water yield debate From demand- to supply-side thinking Global ChangeBiol 18(3) 806ndash820 doi101111j1365ndash2486201102589x

Fahey T J J B Yavitt J A Pearson and D H Knight (1985) The nitrogen cycle in lodgepole pine forests southeastern Wyoming Biogeo-chemistry 1(3) 257ndash275

Falge E D Baldocchi R Olson P Anthoni M Aubinet C Bernhofer and H Dolman (2001) Gap filling strategies for defensible annualsums of net ecosystem exchange Agric For Meteorol 107(1) 43ndash69

Flerchinger G N and K R Cooley (2000) A ten-year water balance of a mountainous semi-arid watershed J Hydrol 237(1ndash2) 86ndash99 doi101016s0022-1694(00)00299-7

Flerchinger G N D Marks M L Reba Q Yu and M S Seyfried (2010) Surface fluxes and water balance of spatially varying vegetationwithin a small mountainous headwater catchment Hydrol Earth Syst Sci 14(6) 965ndash978 doi105194hess-14-965-2010

Frank J M W J Massman B E Ewers L S Huckaby and J F Negron (2014) Ecosystem CO2H2O fluxes are explained by hydraulically lim-ited gas exchange during tree mortality from spruce bark beetles J Geophys Res Biogeosci 119 doi1010022013JG002597

Gonfiantini R (1986) Environmental isotopes in lake studies in Handbook of Environmental Isotope Geochemistry The Terrestrial Environ-ment edited by P Fritz and J-C Fontes pp 113ndash168 Elsevier Amsterdam

Goulden M R Anderson R Bales A Kelly M Meadows and G Winston (2012) Evapotranspiration along an elevation gradient in Califor-niarsquos Sierra Nevada J Geophys Res 117 G03028 doi1010292012JG002027

Goulden M L J W Munger S M Fan B C Daube and S C Wofsy (1996) Measurements of carbon sequestration by long-term eddycovariance Methods and a critical evaluation of accuracy Global Change Biol 2(3) 169ndash182 doi101111j1365-24861996tb00070x

Grant G E C L Tague and C D Allen (2013) Watering the forest for the trees An emerging priority for managing water in forest land-scapes Frontiers Ecol Environ 11(6) 314ndash321 doi101890120209

Gray A N T A Spies and M J Easter (2002) Microclimatic and soil moisture responses to gap formation in coastal Douglas-fir forestsCan J For Res 32(2) 332ndash343 doi101139x01-200

Gu L H E M Falge T Boden D D Baldocchi T A Black S R Saleska and L K Xu (2005) Objective threshold determination for nighttimeeddy flux filtering Agric For Meteorol 128(3ndash4) 179ndash197 doi101016jagrformet200411006

Gustafson J R P D Brooks N P Molotch and W C Veatch (2010) Estimating snow sublimation using natural chemical and isotopic trac-ers across a gradient of solar radiation Water Resour Res 46 W12511 doi1010292009WR009060

AcknowledgmentsThis work was supported by fundingfrom the National Science FoundationEAR-0910831 United States GeologicalService Wyoming AgriculturalResearch Station throughMcIntire-Stennis and the WyomingWater Development CommissionScience Foundation Arizona theArizona Water Sustainability Programand the NCALM Seed grant programprovided support to JB The NationalScience Foundation Niwot Ridge Long-Term Ecological Research project andthe University of Colorado MountainResearch Station provided logisticalsupport and data The Niwot RidgeAmeriFlux tower has been supportedby the US Department of Energy(DOE) the National Institute forClimate Change Research (NICCR) andTerrestrial Carbon Processes Program(TCP) and the National ScienceFoundation (NSF) Long-Term Researchin Environmental Biology (LTREB) Wethank B Bright A Meddens and JHicke of the University of Idaho forsharing QuickBird and Landsat landcover classification of the MPB siteand S Peckham of the University ofWyoming for mapping of mortality inthe MPB tower footprint Field data(supporting information) wereprovided for Niwot by J Knowles RCowie and M Williams and forChimney Park by J Frank and theGlacier Lakes Ecosystem ExperimentsSite The National Center forAtmospheric Research is supportedthrough a cooperative agreementfrom the National Science FoundationThe authors thank T Meixner B BrightZ Guido and three anonymousreviewers for insightful comments thatclarified and improved the manuscript



Harpold A P Brooks S Rajagopal I Heidbuchel A Jardine and C Stielstra (2012) Changes in snowpack accumulation and ablation inthe intermountain west Water Resour Res 48 W11501 doi1010292012WR011949

Harpold A A J A Biederman K Condon M Merino Y Korgaonkar T Nan and P D Brooks (2014) Changes in snow accumulation andablation following the Las Conchas Forest Fire New Mexico USA Ecohydrology 7(2) 440ndash452 doi101002eco1363

Hicke J A C D Allen A R Desai M C Dietze R J Hall E H Hogg and J Vogelmann (2012) Effects of biotic disturbances on forest car-bon cycling in the United States and Canada Global Change Biol 18(1) 7ndash34 doi101111j1365ndash2486201102543x

Hubbard R M C C Rhoades K Elder and J Negron (2013) Changes in transpiration and foliage growth in lodgepole pine trees followingmountain pine beetle attack and mechanical girdling For Ecol Manage 289 312ndash317 doi101016jforeco201209028

Hubbart J A T E Link J A Gravelle and W J Elliot (2007) Timber harvest impacts on water yield in the continentalmaritime hydrocli-matic region of the United States For Sci 53(2) 169ndash180 [Available at httpwwwingentaconnectcom]

Huber C (2005) Long lasting nitrate leaching after bark beetle attack in the highlands of the Bavarian Forest National Park J EnvironQual 34(5) 1772ndash1779 doi102134jeq20040210

Jasechko S Z D Sharp J J Gibson S J Birks Y Yi and P J Fawcett (2013) Terrestrial water fluxes dominated by transpiration Nature496(7445) 347ndash350 doi101038nature11983

Kaiser K E B L McGlynn and R E Emmanual (2012) Ecohydrology of an outbreak Mountain pine beetle impacts trees in drier landscapepositions first Ecohydrology 6(3) 444ndash454 doi101002eco1286

Knight D H T J Fahey and S W Running (1985) Water and nutrient outflow from contrasting lodgepole pine forests in Wyoming EcolMonogr 55(1) 29ndash48

Kozlowski T T (2002) Physiological ecology of natural regeneration of harvested and disturbed forest stands Implications for forest man-agement For Ecol Manage 158(1ndash3) 195ndash221 doi101016s0378ndash1127(00)00712-x

Lee X W J Massman and B Law (2004) Handbook of Micrometeorology A Guide for Surface Flux Measurement and Analysis Springer Dor-drecht Netherlands

MacKay D S D E Ahl B E Ewers S T Gower S N Burrows S Samanta and K J Davis (2002) Effects of aggregated classifications of for-est composition on estimates of evapotranspiration in a northern Wisconsin forest Global Change Biol 8(12) 1253ndash1265 doi101046j1365-2486200200554x

Maness H P J Kushner and I Fung (2013) Summertime climate response to mountain pine beetle disturbance in British Columbia NatGeosci 6(1) 65ndash70 doi101038ngeo1642

McMillin J D and K K Allen (2003) Effects of Douglas-fir beetle (Coleoptera Scolytidae) infestations on forest overstory and understoryconditions in western Wyoming West North Am Nat 63(4) 498ndash506

Meddens A J and J A Hicke (2014) Spatial and temporal patterns of Landsat-based detection of tree mortality caused by a mountainpine beetle outbreak in Colorado USA Forest Ecology and Management 322 78ndash88 doi101016jforeco201402037

Meddens A J H J A Hicke and C A Ferguson (2012) Spatiotemporal patterns of observed bark beetle-caused tree mortality in BritishColumbia and the western United States Ecol Appl 22(7) 1876ndash1891

Mellander P-E K Bishop and T Lundmark (2004) The influence of soil temperature on transpiration A plot scale manipulation in a youngScots pine stand For Ecol Manage 195(1) 15ndash28

Mikkelson K L A Bearup R M Maxwell J D Stednick J E McCray and J O Sharp (2013) Bark beetle infestation impacts on nutrientcycling water quality and interdependent hydrological effects Biogeochemistry 115 1ndash21 doi101007s10533-013-9875-8

Mikkelson K M R M Maxwell I Ferguson J D Stednick J E McCray and J O Sharp (2013) Mountain pine beetle infestation impactsModeling water and energy budgets at the hill-slope scale Ecohydrology 6(1) 64ndash72 doi101002eco278

Molotch N P P D Brooks S P Burns M Litvak R K Monson J R McConnell and K Musselman (2009) Ecohydrological controls onsnowmelt partitioning in mixed-conifer sub-alpine forests Ecohydrology 2(2) 129ndash142 doi101002eco48

Monson R K A A Turnipseed J P Sparks P C Harley L E Scott-Denton K Sparks and T E Huxman (2002) Carbon sequestration in ahigh-elevation subalpine forest Global Change Biol 8(5) 459ndash478 doi101046j1365-2486200200480x

Moore D J J Hu W J Sacks D S Schimel and R K Monson (2008) Estimating transpiration and the sensitivity of carbon uptake to wateravailability in a subalpine forest using a simple ecosystem process model informed by measured net COlt subgt 2 and Hlt subgt 2 Ofluxes Agric For Meteorol 148(10) 1467ndash1477

Morehouse K T Johns J Kaye and A Kaye (2008) Carbon and nitrogen cycling immediately following bark beetle outbreaks in south-western ponderosa pine forests For Ecol Manage 255(7) 2698ndash2708 doi101016jforeco200801050

Mu Q Z M S Zhao and S W Running (2011) Improvements to a MODIS global terrestrial evapotranspiration algorithm Remote SensEnviron 115(8) 1781ndash1800 doi101016jrse201102019

Musselman K N N P Molotch and P D Brooks (2008) Effects of vegetation on snow accumulation and ablation in a mid-latitude sub-alpine forest Hydrol Processes 22(15) 2767ndash2776 doi101002hyp7050

OrsquoHalloran T L B E Law M L Goulden Z S Wang J G Barr C Schaaf and V Engel (2012) Radiative forcing of natural forest disturban-ces Global Change Biol 18(2) 555ndash565 doi101111j1365-2486201102577x

Pi~na P (2013) Impact of mountain pine beetle attack on water balance of lodgepole pine forests in Alberta PhD thesis Univ of AlbertaEdmonton Alberta [Available at httperalibraryualbertacapublicviewitemuuidea5892c7-d472ndash4609-b785-5ba515df1265]

Pomeroy J X Fang and C Ellis (2012) Sensitivity of snowmelt hydrology in Marmot Creek Alberta to forest cover disturbance HydrolProcesses 26(12) 1892ndash1905 doi101002hyp9248

Potts D F (1984) Hydrologic impacts of a large-scale mountain pine beetle (Dendroctonus Ponderosae Hopkins) epidemic Water ResourBull 20(3) 373ndash377 doi101111j1752-16881984tb04719x

Pugh E and E Gordon (2013) A conceptual model of water yield effects from beetle-induced tree death in snow-dominated lodgepolepine forests Hydrol Processes 27 2048ndash2060 doi101002hyp9312

Pugh E and E Small (2012) The impact of pine beetle infestation on snow accumulation and melt in the headwaters of the ColoradoRiver Ecohydrology 5(4) 467ndash477 doi101002eco239

Raffa K F B H Aukema B J Bentz A L Carroll J A Hicke M G Turner and W H Romme (2008) Cross-scale drivers of natural disturban-ces prone to anthropogenic amplification The dynamics of bark beetle eruptions Bioscience 58(6) 501ndash517 doi101641b580607

Rasmussen R B Baker J Kochendorfer T Meyers S Landolt A P Fischer and E Gutmann (2012) How well are we measuring snow TheNOAAFAANCAR winter precipitation test bed Bull Am Meteorol Soc 93(6) 811ndash829 doi101175bams-d-11-00052

Richardson A D D Y Hollinger G G Burba K J Davis L B Flanagan G G Katul and A E Suyker (2006) A multi-site analysis of randomerror in tower-based measurements of carbon and energy fluxes Agric For Meteorol 136(1) 1ndash18

Rinehart A J E R Vivoni and P D Brooks (2008) Effects of vegetation albedo and solar radiation sheltering on the distribution of snowin the Valles Caldera New Mexico Ecohydrology 1(3) 253ndash270 doi101002eco26



Romme W H D H Knight and J B Yavitt (1986) Mountain pine beetle outbreaks in the Rocky Mountains Regulators of primary produc-tivity Am Nat 127(4) 484ndash494 [Available at httpwwwjstororgstable2461578]

Schelker J L Kuglerova K Eklof K Bishop and H Laudon (2013) Hydrological effects of clear-cutting in a boreal forestmdashSnowpackdynamics snowmelt and streamflow responses J Hydrol 484 105ndash114 doi101016jjhydrol201301015

Scott R L (2010) Using watershed water balance to evaluate the accuracy of eddy covariance evaporation measurements for three semi-arid ecosystems Agric For Meteorol 150(2) 219ndash225 doi101016jagrformet200911002

Scott-Denton L E K L Sparks and R K Monson (2003) Spatial and temporal controls of soil respiration rate in a high-elevation subalpineforest Soil Biol Biochem 35(4) 525ndash534 doi101016s0038-0717(03)00007-5

Shuttleworth W J (2012) Terrestrial Hydrometeorology John Wiley Oxford U KSmirnova T G J M Brown S G Benjamin and D Kim (2000) Parameterization of cold-season processes in the MAPS land-surface

scheme J Geophys Res 105(D3) 4077ndash4086 doi1010291999JD901047Somor A J (2010) Quantifying streamflow change following bark beetle outbreak in multiple central Colorado catchments MS thesis

Univ of Ariz TucsonStednick J D (1996) Monitoring the effects of timber harvest on annual water yield J Hydrol 176(1ndash4) 79ndash95 doi1010160022-

1694(95)02780-7Stoy P C G G Katul M Siqueira J Y Juang K A Novick H R McCarthy and R Oren (2006) Separating the effects of climate and vegeta-

tion on evapotranspiration along a successional chronosequence in the southeastern US Global Change Biol 12(11) 2115ndash2135 doi101111j1365ndash2486200601244x

Stoy P C M Mauder T Foken B Marcolla E Boegh A Ibrom and C Bernhofer (2013) A data-driven analysis of energy balance closureacross FLUXNET research sites The role of landscape scale heterogeneity Agric For Meteorol 171 137ndash152 doi101016jagrformet201211004

Thompson S E C J Harman P A Troch P D Brooks and M Sivapalan (2011) Spatial scale dependence of ecohydrologically mediatedwater balance partitioning A synthesis framework for catchment ecohydrology Water Resour Res 47 W00J03 doi1010292010WR009998

Tokuchi N N Ohte S Hobara S J Kim and K Masanori (2004) Changes in biogeochemical cycling following forest defoliation by pinewilt disease in Kiryu experimental catchment in Japan Hydrol Processes 18(14) 2727ndash2736 doi101002hyp5578

Troendle C A (1983) The potential for water yield augmentation from forest management in the Rocky Mountain Region Water ResourBull 19(3) 359ndash373 doi101111j1752-16881983tb04593x

Troendle C A and R M King (1987) The effect of partial and clearcutting on streamflow at Deadhorse Creek Colorado J Hydrol 90(1ndash2)145ndash157 doi1010160022-1694(87)90177-6

Turnipseed A A P Blanken D Anderson and R Monson (2002) Energy budget above a high-elevation subalpine forest in complextopography Agric For Meteorol 110(3) 177ndash201 doi101016S0168ndash1923(01)00290-8

Turnipseed A A D E Anderson P D Blanken W M Baugh and R K Monson (2003) Airflows and turbulent flux measurements in moun-tainous terrain Part 1 Canopy and local effects Agric For Meteorol 119(1) 1ndash21 doi101016S0168-1923(03)00136-9

Van Pelt R and J F Franklin (1999) Response of understory trees to experimental gaps in old-growth Douglas-fir forests Ecol Appl 9(2)504ndash512 doi1018901051-0761(1999)009[0504routte]20co2

Vanderhoof M C A Williams B Ghimire and J Rogan (2013) Impact of mountain pine beetle outbreaks on forest albedo and radiativeforcing as derived from moderate resolution imaging spectroradiometer Rocky Mountains USA J Geophys Res Biogeosci 118 1461ndash1471 doi101002jgrg20120

Varhola A N C Coops C W Bater P Teti S Boon and M Weiler (2010) The influence of ground- and lidar-derived forest structure met-rics on snow accumulation and ablation in disturbed forests Can J For Res 40(4) 812ndash821 doi101139x10-008

Veatch W P D Brooks J R Gustafson and N P Molotch (2009) Quantifying the effects of forest canopy cover on net snow accumulationat a continental mid-latitude site Ecohydrology 2(2) 115ndash128 doi101002eco45

Westerling A L H G Hidalgo D R Cayan and T W Swetnam (2006) Warming and earlier spring increase western US forest wildfire activ-ity Science 313(5789) 940ndash943 doi101126science1128834

Williams A P C D Allen C I Millar T W Swetnam J Michaelsen C J Still and S W Leavitt (2010) Forest responses to increasing aridityand warmth in the southwestern United States Proc Natl Acad Sci USA 107(50) 21289ndash21294 doi101073pnas0914211107

Williams A P C D Allen A K Macalady D Griffin C A Woodhouse D M Meko and N G McDowell (2013) Temperature as a potentdriver of regional forest drought stress and tree mortality Nat Clim Change 3(3) 292ndash297 doi101038nclimate1693

Wilson K A Goldstein E Falge M Aubinet D Baldocchi P Berbigier and S Verma (2002) Energy balance closure at FLUXNET sites AgricFor Meteorol 113(1ndash4) 223ndash243 doi101016s0168-1923(02)00109-0

Wilson K B P J Hanson P J Mulholland D D Baldocchi and S D Wullschleger (2001) A comparison of methods for determining forestevapotranspiration and its components Sap-flow soil water budget eddy covariance and catchment water balance Agric For Mete-orol 106(2) 153ndash168 doi101016s0168-1923(00)00199-4

Wulder M A C C Dymond J C White D G Leckie and A L Carroll (2006) Surveying mountain pine beetle damage of forests A reviewof remote sensing opportunities For Ecol Manage 221(1ndash3) 27ndash41 doi101016jforeco200509021

Yaseef N R D Yakir E Rotenberg G Schiller and S Cohen (2010) Ecohydrology of a semi-arid forest Partitioning among water balancecomponents and its implications for predicted precipitation changes Ecohydrology 3(2) 143ndash154 doi101002eco65



l

l

l

l

l

Recent work has examined how progressive hydrologic partitioning response to gradual canopy loss mayaffect water resources and forest succession [Edburg et al 2012] Forests regulate partitioning through inter-ception transpiration and sheltering the land surface [Troendle and King 1987 Molotch et al 2009 Varholaet al 2010] Disturbance-driven partitioning changes may affect water resources including the amount andtiming of streamflow [Pugh and Gordon 2013 Brown et al 2005 Ellison et al 2012] and water availabilityfor successional vegetation [Edburg et al 2012] Succession is likely to depend on both partitioningbetween Q and V as well as how V comprises S E and T [Edburg et al 2012 Brown et al 2013 Romme et al1986] The expectations for response to MPB include an initial large decline in T followed by more gradualincreases in E from newly exposed soils and increased T by surviving and new vegetation [Edburg et al2012 Romme et al 1986] Winter precipitation is a dominant hydrologic input to many infested forests andthe reduction in sublimation of intercepted snow from the canopy (Sc) is expected to increase snowmeltvolumes [Pugh and Small 2012] The timing and relative magnitudes of these changes are unknown butthe prevailing expectation is for reduced V resulting in greater Q [Pugh and Gordon 2013]

There is limited and conflicting empirical evidence of Q response to MPB in support of these expectationsIncreased Q was reported by Potts [1984] and Bethlahmy [1974] although the increase reported by Beth-lahmy [1974] peaked 15 years post-MPB contrasting with the view that Q increases are immediate anddecline over years to decades [Bosch and Hewlett 1982 Stednick 1996 Brown et al 2005] A recent study ofMPB-impacted Colorado catchments found mixed Q responses attributed in part to variable interactionsamong topography and canopy loss in controlling V [Brooks et al 2012 Somor 2010] Snowpack observa-tions have shown either that total winter S declined or that increased snowpack sublimation (Ss) compen-sated for reduced Sc resulting in no net snowpack change [Boon 2012 Pugh and Small 2012 Biedermanet al 2014] Summer V has been relatively constant during several years following mortality [Brown et al2013] although land surface models predict reduced V with declining leaf area index under simulated dis-turbance [Bewley et al 2010 Pomeroy et al 2012 Mikkelson et al 2013b] Similarly remote-sensing prod-ucts such as MOD16 (httpwwwntsgumteduprojectmod16) that use plant-mediated algorithms [Muet al 2011] predict reduced V [Bright et al 2013 Maness et al 2013]

The variable Q response to MPB and lack of consistency between observations and models demonstratesour limited ability to predict water resources following insect-driven disturbance There is therefore a needto empirically quantify hydrologic partitioning in insect-disturbed headwater catchments Snow surveysquantify winter V over a range of spatial scales [Anderton et al 2004 Musselman et al 2008 Veatch et al2009] Eddy covariance (EC) quantifies V from a footprint of similar size to a headwater catchment (101 ha)[Wilson et al 2001 Scott 2010] Alternatively the difference between precipitation (P) and observed Q canbe used to estimate catchment-scale vapor loss (V) assuming negligible storage changes DS over annualor longer periods [Brutsaert 2005] Finally stable water isotope fractionation can identify kinetic E and dis-tinguish it from T which does not leave a fractionation signal in remaining water [Jasechko et al 2013 Clarkand Fritz 1997]

In this study we quantify hydrologic fluxes during 3 years following MPB infestation at a site with severeMPB-driven forest mortality and at a paired control site to answer the question How does MPB tree mortalityaffect partitioning of precipitation between vapor loss and water available for streamflow To our knowledgethis is the first effort combining onsite precipitation snow surveys eddy covariance vapor flux catchmentstreamflow and stable isotope indicators of evaporation to quantify hydrologic response to forestdisturbance

2 Site Description

An MPB-impacted site lsquolsquoMPBrsquorsquo and control site lsquolsquoUnimpactedrsquorsquo were identified in the Central Rocky Mountainsa region severely impacted by MPB [Meddens and Hicke 2014] Due to extensive tree mortality adjacentMPB and control sites having similar characteristics were not found However biophysical and meteorologi-cal observations at these sites showed them to be well paired (Table 1) The MPB site at Chimney Park WYis 100 km west of Cheyenne in the Medicine Bow National Forest and the headwaters of the Laramie River(Figure 1) The MPB site experienced severe infestation beginning in 2007 with peak mortality during2007ndash2009 The Unimpacted site (Figure 1) is 125 km SSE of the MPB site and 50 km NW of Denver CO





3 Methods

































4 Results
















Year



(W m22)


Unimpacted2010 230 (2) 420 381 801 248 (59) 90 (62) 024 (019) 064 (040) 46 (26) 31 (13) 137 (63) 251 (79) 14 (9) 21 (12)2011 270 (2) 459 566 1025 240 (71) 75 (73) 024 (020) 058 (040) 54 (28) 36 (19) 129 (59) 250 (81) 20 (12)2012 222 (6) 295 428 722 240 (59) 90 (60) 027 (019) 069 (041) 53 (31) 32 (1 2) 131 (50) 240 (79)MPB2010 286 (1) 440 311 751 251 (55) 102 (62) 013 (015) 082 (045) 32 (10) 29 (08) 131 (62) 275 (74) 25 (15) 61 (39)2011 270 (3) 430 356 786 240 (70) 91 (76) 019 (022) 075 (051) 38 (12) 31 (10) 123 (55) 261 (85) 70 (40)2012 266 (10) 323 125 448 253 (59) 105 (62) 020 (016) 090 (047) 41 (13) 29 (09) 124 (45) 270 (74)





5 Discussion






















Year P Year




















6 Conclusions









































































































l

l

l

l

l



3 Methods

































4 Results
















Year



(W m22)


Unimpacted2010 230 (2) 420 381 801 248 (59) 90 (62) 024 (019) 064 (040) 46 (26) 31 (13) 137 (63) 251 (79) 14 (9) 21 (12)2011 270 (2) 459 566 1025 240 (71) 75 (73) 024 (020) 058 (040) 54 (28) 36 (19) 129 (59) 250 (81) 20 (12)2012 222 (6) 295 428 722 240 (59) 90 (60) 027 (019) 069 (041) 53 (31) 32 (1 2) 131 (50) 240 (79)MPB2010 286 (1) 440 311 751 251 (55) 102 (62) 013 (015) 082 (045) 32 (10) 29 (08) 131 (62) 275 (74) 25 (15) 61 (39)2011 270 (3) 430 356 786 240 (70) 91 (76) 019 (022) 075 (051) 38 (12) 31 (10) 123 (55) 261 (85) 70 (40)2012 266 (10) 323 125 448 253 (59) 105 (62) 020 (016) 090 (047) 41 (13) 29 (09) 124 (45) 270 (74)





5 Discussion






















Year P Year




















6 Conclusions









































































































l

l

l

l

l




















4 Results
















Year



(W m22)


Unimpacted2010 230 (2) 420 381 801 248 (59) 90 (62) 024 (019) 064 (040) 46 (26) 31 (13) 137 (63) 251 (79) 14 (9) 21 (12)2011 270 (2) 459 566 1025 240 (71) 75 (73) 024 (020) 058 (040) 54 (28) 36 (19) 129 (59) 250 (81) 20 (12)2012 222 (6) 295 428 722 240 (59) 90 (60) 027 (019) 069 (041) 53 (31) 32 (1 2) 131 (50) 240 (79)MPB2010 286 (1) 440 311 751 251 (55) 102 (62) 013 (015) 082 (045) 32 (10) 29 (08) 131 (62) 275 (74) 25 (15) 61 (39)2011 270 (3) 430 356 786 240 (70) 91 (76) 019 (022) 075 (051) 38 (12) 31 (10) 123 (55) 261 (85) 70 (40)2012 266 (10) 323 125 448 253 (59) 105 (62) 020 (016) 090 (047) 41 (13) 29 (09) 124 (45) 270 (74)





5 Discussion






















Year P Year




















6 Conclusions









































































































l

l

l

l

l






Year



(W m22)


Unimpacted2010 230 (2) 420 381 801 248 (59) 90 (62) 024 (019) 064 (040) 46 (26) 31 (13) 137 (63) 251 (79) 14 (9) 21 (12)2011 270 (2) 459 566 1025 240 (71) 75 (73) 024 (020) 058 (040) 54 (28) 36 (19) 129 (59) 250 (81) 20 (12)2012 222 (6) 295 428 722 240 (59) 90 (60) 027 (019) 069 (041) 53 (31) 32 (1 2) 131 (50) 240 (79)MPB2010 286 (1) 440 311 751 251 (55) 102 (62) 013 (015) 082 (045) 32 (10) 29 (08) 131 (62) 275 (74) 25 (15) 61 (39)2011 270 (3) 430 356 786 240 (70) 91 (76) 019 (022) 075 (051) 38 (12) 31 (10) 123 (55) 261 (85) 70 (40)2012 266 (10) 323 125 448 253 (59) 105 (62) 020 (016) 090 (047) 41 (13) 29 (09) 124 (45) 270 (74)





5 Discussion






















Year P Year




















6 Conclusions









































































































l

l

l

l

l


5 Discussion






















Year P Year




















6 Conclusions









































































































l

l

l

l

l















Year P Year




















6 Conclusions









































































































l

l

l

l

l












6 Conclusions









































































































l

l

l

l

l






































































































l

l

l

l

l

PUBLICATIONS · 2016-07-14 · 1981–2012). The MPB site also receives an average of 400 mm...

Documents

Transcript of PUBLICATIONS · 2016-07-14 · 1981–2012). The MPB site also receives an average of 400 mm...