Research Article Downscaling Maximum …downloads.hindawi.com/journals/amete/2014/594965.pdfResearch...

Research ArticleDownscaling Maximum Temperatures to SubkilometerResolutions in the Shenandoah National Park of Virginia USA

Temple R Lee1 Stephan F J De Wekker1 and John E B Wofford2

1 Department of Environmental Sciences Clark Hall University of Virginia PO Box 400123 Charlottesville VA 22904-0123 USA2 Shenandoah National Park Luray VA 22835 USA

Correspondence should be addressed to Temple R Lee templevirginiaedu

Received 25 April 2014 Accepted 6 August 2014 Published 31 August 2014

Academic Editor Richard Anyah

Copyright copy 2014 Temple R Lee et al This is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

Downscaling future temperature projections to mountainous regions is vital for many applications including ecological and waterresource management In this study we demonstrate a method to downscale maximum temperatures to subkilometer resolutionsusing the Parameter-elevation Regression on Independent Slopes Model (PRISM) We evaluate the downscaling method withobservations from a network of temperature sensors deployed along western and eastern slopes of Virginiarsquos Shenandoah NationalPark in the southern Appalachian Mountains We find that the method overestimates mean July maximum temperatures by about2∘C (4∘C) along the western (eastern) slopes Based on this knowledge we introduce corrections to generate maps of current andfuture maximum temperatures in the Shenandoah National Park

1 Introduction

General circulation models (GCMs) predict changing tem-perature and moisture patterns in many regions of theworld due to increases in atmospheric carbon dioxide (CO

2)

concentration [1] The most significant changes are expectedat high elevations [2 3] which include the habitats of manyendangered species [4 5] One example is the Shenandoahsalamander Plethodon shenandoah (P shenandoah) whosehabitats are located in a 6 km2 area along ridgetops inthe Shenandoah National Park (SNP) [6] There is largeuncertainty in how climate change will affect the local climatein these habitats partly because the grid spacing of GCMson the order of a few hundred kilometers cannot resolvethe impacts of climate change at local to regional scales [2]Even regional climate models (RCMs) with resolutions ofa few tens of kilometers are still too coarse to assess theeffects of climate change on the scale of local habitats Forexample a 50 km grid spacing typical of RCMs [7 8] averagesthe elevation changes across the grid thereby smoothingthe topography in mountainous regions [9] including theBlue Ridge Mountains in and around SNP This topographicsmoothing makes it very difficult to understand projectedclimate changes in these mountainous regions

To reduce the uncertainty of climate change projectionsin mountainous regions RCMs must be further downscaledwhich is done using either dynamical andor statistical tech-niques [10 11] Dynamical downscaling techniques use high-resolution meteorological models which are computationallyintensive and can have difficulties resolving atmosphericprocesses in complex terrain at subkilometer resolutions[12] Statistical downscaling techniques include regressiontechniques and stochastic modeling [13] Many statisticaldownscaling techniques assume a constant lapse rate todownscale temperatures to spatial scales on the order of20 km [14] Assuming a constant lapse rate is not justifiedfor many regions [15] because lapse rates exhibit significantspatial variability that are a function of season [16 17]slope position (ie leeward versus windward) [15] and slopeazimuth [18 19]

On the other hand gridded climate data sets at highresolution (on the scale of 1ndash4 km) have been developedfrom which spatially and temporally varying lapse rates canbe derived Two widely known examples of these griddedclimate data sets that are based on networks of surfaceobservations are the Parameter-elevation Regression onIndependent Slopes Model (PRISM) [20 21] and the DailySurface Weather and Climatological Summary (DAYMET)

Hindawi Publishing CorporationAdvances in MeteorologyVolume 2014 Article ID 594965 9 pageshttpdxdoiorg1011552014594965

2 Advances in Meteorology

[22] which have been used in for example ecologicalapplications [23] and climate studies [24]

To obtain accurate temperature projections at subkilo-meter scales statistical downscaling methods could usethe spatially and temporally varying lapse rates that canbe derived from high-resolution gridded data sets In thisstudy we develop such a downscaling method and use thismethod to downscale near-surface temperatures maps atthe high elevation habitats of P shenandoah in SNP undercurrent and future climate conditions We limit ourselvesto a demonstration of our downscaling method for Julymaximum temperatures motivated by the importance ofchanges in summer maximum temperatures to the survivalof P shenandoah

2 Methods

21 Site Description SNP is located along the crest of theVirginia Blue Ridge Mountains on the eastern flank ofthe southern Appalachians in the eastern US The park isoriented southwest-northeast and extends 115 km from nearWaynesboro Virginia to Front Royal Virginia Elevationsin SNP range from 200m above mean sea level (msl) in thevalleys to 1000ndash1200mmsl along the ridgeline East of SNP isthe Virginia Piedmont and the Shenandoah Valley is west ofSNPThe climate of SNP ranges fromhumid subtropical at thelower elevations to humid continental along the ridgetops

22 Downscaling Method Our downscaling method usesthe gridded temperature data sets from PRISM PRISMassimilates observations of temperature (precipitation) fromapproximately 10000 (13000) surface stations in the conter-minous US and linearly interpolates these observations tohigh spatial resolutions using elevation ocean proximity andtopographic facet [20 25] PRISM outputs two data sets (1)mean monthly maximum and minimum temperature andtotal precipitation from 1895 through 2013 at a 4 km resolu-tion and (2) 30-year monthly climate means for maximumand minimum temperature and total precipitation averagedover 1971ndash2000 (and as of 2014 1981ndash2010) at 800m spatialresolution [20 21] PRISM gridded data sets are availableonline from httpprismoregonstateedu

The 4 km and 800m spatial resolutions of the PRISMdatasets are too coarse to resolve temperature variability acrossthe habitats of P shenandoah and thus further downscalingis required In our downscaling method (Figure 1) we down-scale the 4 km PRISMmeanmonthly maximum temperatureto a 15m resolution digital elevation model (DEM) whichis the highest-resolution DEM available for the region usingspatially and temporally varying lapse rates These lapse ratesrepresent the line of best fit between elevation and the 30-year 800m mean monthly maximum temperature for the25 elevation-temperature pairs within each 4 km grid cellThroughout SNP this relationship is high (119903 gt 098 119875 lt0001) Thus the downscaled maximum temperature can bewritten as follows

119879Downscaled = 1198794 km + (1198854 km minus 119885DEM) sdot LR08 km (1)

PRISM mean monthly maximum

resolution)

Downscaled PRISM temperatures

monthly climate

temperature

Digital elevation PRISM 30-year

means for maximumresolution)

Correction factor from slope

measurements

Corrected PRISM temperatures

temperature (4km

(800m resolution)

model (15m

(15m resolution)

(15m resolution)

T4 km Z4 km ZDEMLR08 km

Tdownscaled = T4 km + (Z4 km minus ZDEM ) middot LR08 km

Figure 1 Summary of our downscaling method in which wedownscale the monthly 4 km mean monthly maximum PRISMtemperature (119879

4 km) to a 15m resolution DEM based on the lapserates derived from the 800mPRISM30-yearmonthly climatemeansfor maximum temperature over the period 1971ndash2000 (LR

08 km)and the difference between the mean elevation of the 4 km gridbox (119885

4 km) and elevation of each 15m grid cell (119885DEM) to obtaindownscaled PRISM (119879Downscaled) at a 15m resolution We thenapply a correction to 119879Downscaled based on the slope temperaturemeasurements to obtain more realistic high-resolution estimates oftemperature in Shenandoah National Park

where 119879Downscaled is the PRISM mean monthly maximumtemperature downscaled to a 15m resolution DEM 119879

4 kmis the 4 km PRISM mean monthly maximum temperature1198854 km is the mean elevation of the 4 km grid box 119885DEM is

the elevation of each 15m grid cell and LR08 km is the lapse

rate based on the 800m PRISM 30-year monthly climatemeans for maximum temperature where a positive lapserate indicates a decrease in temperature with height Thisdownscaling method is evaluated using observations froma network of temperature sensors discussed in the nextsection

23 Slope Measurements The observations which we use toevaluate our downscaling method come from a network oftemperature sensors deployed in SNP We use Onset HOBOPro V2 temperature-relative humidity sensors which havea manufacturer-state accuracy of plusmn021∘C over the range 0∘to 50∘C for temperature and plusmn25 from 10 to 90 forrelative humidity These sensors have been tested and usedsuccessfully in many applications to investigate near-surfacetemperature variations [26 27]

In April 2011 we deployed 37 sensors along 3 transects inSNP The transect east of the ridgeline included 18 sensorsand the transects west of the ridgeline included 19 sensors(Figure 2) For all transects sensors were deployed along aline extending from near the ridgeline (about 1000mmsl) toabout 600mmsl Of the 19 sensors along the west side of the

Advances in Meteorology 3

SensorsSNP boundary

Elevation (m)

0 1 2 3 405(km)

300

-400

400

-500

500

-600

600

-700

700

-800

800

-900

900

-1000

1000

-1100

78∘20

9984000998400998400W78

∘22

9984000998400998400W78

∘26

9984000998400998400W 78

∘24

9984000998400998400W

38∘36

9984000998400998400N

38∘32

9984000998400998400N

38∘34

9984000998400998400N

lt300

gt1100

N

N

Figure 2 Location of experimental site within Virginia (black boxin the inset map) and map of the study area The red triangles andyellow line indicate the locations of the temperature sensors and theShenandoahNational Park boundary respectivelyThe two ridgetopclimatemonitoring stations BigMeadows andPinnacles are labeledand shown with a black triangle Shading indicates elevation

ridgeline 14 were deployed along one transect and 5 weredeployed along another transect 6 km to the north Along thetransect east of the ridgeline we deployed sensors in pairs atnine different elevations to assess the impact of different slopeazimuths on the temperature measurements Sensors weredeployed approximately 50m apart at each elevation so thatthe effect of slope azimuth on temperature could be assessed

We deployed all sensors adjacent to hiking trails for easeof access and to minimize disturbances to the area Theywere attached to a fence post installed 15m above groundlevel (agl) and enclosed within a radiation shield (type RS-3 from the Onset Corporation) to reduce radiation errorsAll sensors were visited approximately every 2-3 months todownload data 60min means from 10min samples werecomputed and the daily maximum temperature was deter-mined from the 60minmeansThemeanmonthly119879max is themean of all daily maximum temperatures during one monthOur analyses in this study are based on data collected betweenJuly 2011 and July 2013 The data set for this period is mostlycomplete but there are occasional gaps in the record dueto various factors such as water short-circuiting the sensorelectronics damage by wildlife and difficulties downloadingdata from the sensor

Table 1 Mean bias error between downscaled PRISM and theobservations along the west and east slope of SNP in July 2011 2012and 2013

Month West slope MBE (∘C) East slope MBE (∘C)July 2011 221 466July 2012 191 350July 2013 134 385

24 Long-Term Climate Stations In addition to evaluatingour downscaling method using slope temperature measure-ments we use 2m air temperature data from two long-term monitoring sites along the ridgetop of SNP Big Mead-ows (3853N 7844W 1079mmsl) and Pinnacles (3862N7835W 1017mmsl) The monitoring site at Big Meadows islocated in an enclosed grassy field approximately 15m from amixed deciduous forest canopy Observations of daily max-imum and minimum temperature as well as precipitationbegan in 1935 and are obtained from the National ClimateData Center (NCDC) Hourly meteorological observationsbegan in 1988 as part of the Environmental ProtectionAgency(EPA) Clean Air Status and Trends (CASTNET) networkBeginning in 2008 half hour meteorological measurementsbegan at Pinnacles along a 17m walkup tower located ina forest with a mean height of approximately 14m [28]Temperature data from aHOBOProV2 sensor installed 40mnorth of the tower in a small grassy field at 15m agl are alsoused

25 Regional Climate Models We apply our downscalingmethod to RCM output from the North American RegionalClimate Change Assessment Program (NARCCAP) NARC-CAP uses 4 GCMs that provide boundary conditions for 6different RCMs [7 8] Because not all GCM-RCM combina-tions are available we use output from 7 GCM-RCM com-binations in this study All GCM-RCM combinations thatcomprise NARCCAP have a 50 times 50 km2 spatial resolutionover North America and are run for the past (1971ndash2000) andfuture (2041ndash2070) using the Special Report on EmissionsScenarios (SRES) A2 emissions scenario [7]The A2 scenarioassumes rapid population growth and slow economic growthwhich result in greater anthropogenic CO

2emissions than in

the other scenarios used by the Intergovernmental Panel onClimate Change (IPCC) [29]

3 Results and Discussion

31 Evaluation of the Downscaling Method When we applythe downscaling method discussed in Section 22 we findthat downscaled PRISM consistently overestimates meanmonthly119879max in July 2011 2012 and 2013 (Table 1)The largestoverestimates occur along the east slope of SNP where themean bias error (MBE) between downscaled PRISM andthe slope observations of mean 119879max is 35ndash47

∘C (Figure 3)Along the west slope of SNP the warm bias is smaller butdownscaled PRISM still overestimates mean July 119879max by 13ndash22∘C


22 24 26 28 30 3222

24

26

28

30

32

Observed temperature (∘C)

Dow

nsca

led

PRIS

M te

mpe

ratu

re (∘

C)

(a)

22 24 26 28 30 3222

24

26

28

30

32


Dow

nsca

led

PRIS

M te

mpe

ratu

re (∘

C)(b)

West slopeEast slope

Big MeadowsPinnacles

20 22 24 26 28 3020

22

24

26

28

30


Dow

nsca

led

PRIS

M te

mpe

ratu

re (∘

C)

(c)

Figure 3 Comparison between observed mean July 119879max from the sensor network and downscaled PRISM for July 2011 (a) July 2012 (b)and July 2013 (c) using PRISM-derived lapse rates Open circles indicate comparisons along the west slope filled circles indicate comparisonsalong the east slope ldquo+rdquo and ldquoXrdquo denote Big Meadows and Pinnacles respectively Black dotted line is the 1 1 line Note that Big Meadowsmeasurements are unavailable in 2013 Also note the same range but different scales on the 119909-axis and 119910-axis in panel (c)

There are also differences in mean July 119879max betweenthe sensor pairs deployed at different azimuths along theeast slope with mean July 119879max along the southeast-facingslopes on average 05∘C higher than the northeast-facingslopes These differences are lower than those reported in

previous studies by for example Bolstad et al [19] in theSmoky Mountains where mean daily 119879max was 14

∘C higheralong south-facing slopes than along northwest-facing slopesregardless of season The larger differences in the Bolstad etal study occur because neither of the slopes on which our


Table 2 4 km PRISM mean elevation 4 km PRISM lapse rate based on 800m PRISM climate means and July 2011 2012 and 2013 4 kmPRISM mean July 119879max for the locations at which temperature sensors were deployed along the west slope and east slope of SNP

PRISM variable West slope East slope4 km mean elevation from PRISM (m) 64669 888894 km PRISM lapse rate based on 800m PRISM climate means (∘Ckmminus1) 784 7802011 mean July 119879max from PRISM (∘C) 3087 28722012 mean July 119879max from PRISM (∘C) 3008 28192013 mean July 119879max from PRISM (∘C) 2746 2561

sensors were deployed was oriented either directly north ordirectly south and because the sensors in the Bolstad et alstudy were installed above the forest canopy

32 Causes for Differences between Maximum Tempera-tures from the Downscaling Method and Observations Theobserved differences between mean July 119879max from the mea-sured slope temperatures and downscaled PRISM highlightsome of the challenges that arise when using PRISM inmountainous regions Differences can occur because of errorsin PRISM andor observational errors The warm bias indownscaled PRISM can arise because of the interpolationof climatological data in data-sparse regions PRISM inter-polates data from climate stations by calculating a linearrelationship (lapse rate in the case of temperature) betweenelevation and the variable of interest which is linearlyextrapolated in locations where there are no high-elevationstations [21 25] Thus in regions with nearby observationstations at similar elevations PRISM performs well At SNPHeadquarters in Luray VA (3867N 7837W 359mmsl)which is the nearest low elevation station to SNP and islocated about 5 km northwest of the transect of sensorsalong the west slope downscaled PRISM and observed meanJuly 119879max agree to within about 05∘C The warm bias indownscaled PRISM may occur because the mean monthly4 km PRISM July 119879max grids are based on temperatures madeat relatively low elevation stationsWe note that the lapse ratesof 7-8∘Ckmminus1 based on the 30-year climate means (Table 2)are comparable to the mean July 119879max lapse rates observedalong the slope The overestimates in the 4 km July 119879maxPRISM grids are most evident in July 2011 and July 2012 whenthe largest warm bias exists In July 2013 when the 4 km July119879max is about 3

∘C lower there is better agreement along thewest slope but slightly worse agreement along the east slopeWe suspect that the warm bias in downscaled PRISM is partlycaused by the lack of high elevation stations in the region thatcan help constrain temperature estimates along the slopes ofSNP

A warm bias in downscaled PRISM may also occurbecause of canopy cover differences between the low eleva-tion sites (used in PRISM) and the mountain slopes of SNPAll of the locations along the slopes at which we deployedtemperature sensors have canopy cover gt60 based onArcGIS analyses and on our own observations Deploymentin locations with dense canopies helped to minimize radia-tion errors that can occur in the naturally ventilated radiationshields used in this study [30] However many of the climate

stations used in PRISM are not located in forests but insteadlocated in grassy fields which typically have higher daytimemaximum temperatures [31 32] At Pinnacles mean July119879max measured 2m agl within the forest canopy is on average2∘C cooler than the temperature in the open grassy field40m from the tower Because of this difference the July119879max MBE is much smaller between downscaled PRISM andthe measurements from the grassy field at Pinnacles thanbetween downscaled PRISM and the measurements fromwithin the forest canopy at Pinnacles Similarly downscaledPRISM agrees well with temperature observations from BigMeadows (cf Figure 3) which is a monitoring station moretypical of the stations used to generate the PRISM data sets

Differences in slope azimuth angle may also explain someof the bias between downscaled PRISM and the observa-tions The west slope has greater afternoon exposure andthus higher average near-surface incoming solar radiation[33] than the east slope resulting in higher mean monthly119879max than the east slope and thus better agreement withdownscaled PRISM

Thus we conclude that the warm bias in downscaledPRISM is due to a combination of factors including a lack ofhigh elevation stations in the region to constrain temperatureestimates along the slopes of SNP and differences in canopycover and slope azimuth To further investigate the potentialeffect of canopy cover and slope azimuth we apply our down-scaling method to January when these factors exert a smallerinfluence on temperature In January the overestimates indownscaled PRISM temperature are smaller than in July onthe order of 10ndash15∘C and the biases along the eastern andwestern slope are comparable to each other Furthermorethe differences in mean January 119879max between the sensor15m agl along the tower at Pinnacles and the sensor in theadjacent field are negligible and are within themanufacturerrsquosstated accuracy of the sensors Because downscaled PRISMagrees better with the observations in the winter when theeffects of canopy cover and azimuth are reduced we suspectthat PRISMrsquos warm bias in mean July 119879max along the slopesin SNP occurs because (1) the stations used to generate themonthly 4 km PRISM grids in and around SNP are locatedat low elevation nonforested sites and (2) there is a lack ofsurface monitoring stations at high elevations to constraintemperature estimates

33 Corrections to Downscaled PRISM for SNP To summa-rize the monthly 4 km PRISM mean July 119879max (Figure 4(a))does not account for the fine-scale temperature variations at


Big Meadows

Pinnacles

78∘26

9984000998400998400W 78

∘24

9984000998400998400W 78

∘22

9984000998400998400W 78

∘20

9984000998400998400W

38∘38

9984000998400998400N

38∘36

9984000998400998400N

38∘34

9984000998400998400N

N

38∘32

9984000998400998400N

200

300

300

400

500

500

600

600

600

700

700700

800

800

900

900

9001000

1000

1100

(a)

38∘38

9984000998400998400N

38∘36

9984000998400998400N

38∘34

9984000998400998400N

38∘32

9984000998400998400N

Big Meadows

Pinnacles

78∘26

9984000998400998400W 78

∘24

9984000998400998400W 78

∘22

9984000998400998400W 78

∘20

9984000998400998400W

N

200

300400

500

500

600

600

600

700700

700

800

300

800

900

900

900

1000

1000

1100

(b)

SNP boundarySensors 0 2 4 6 81

(km)

lt22

22

-23

23

-24

24

-25

25

-26

26

-27

27

-28

28

-29

29

-30

30

-31

31

-32

gt32

Temperature (∘C)

38∘38

9984000998400998400N

38∘36

9984000998400998400N

38∘34

9984000998400998400N

38∘32

9984000998400998400N

Big Meadows

Pinnacles

78∘26

9984000998400998400W 78

∘24

9984000998400998400W 78

∘22

9984000998400998400W 78

∘20

9984000998400998400W

N

200

300

300

400

500

500

600

600

600700

700

700

800

800

900

900

900

1000

1000

1100

(c)


(km)

lt22

22

-23

23

-24

24

-25

25

-26

26

-27

27

-28

28

-29

29

-30

30

-31

31

-32

gt32

Temperature (∘C)

38∘38

9984000998400998400N

38∘36

9984000998400998400N

38∘34

9984000998400998400N

38∘32

9984000998400998400N

Big Meadows

Pinnacles

78∘26

9984000998400998400W 78

∘24

9984000998400998400W 78

∘22

9984000998400998400W 78

∘20

9984000998400998400W

N

200

300

300

400

500

500

600

600

600

700

700700

800

800

900

900

900

1000

1000

1100

(d)

Figure 4Maps of mean119879max for July 2012 resulting from the different steps in our downscalingmethod Panel (a) shows 4 km PRISM outputpanel (b) shows PRISM downscaled to 15m in SNP using the 30-year PRISM lapse rates and without any corrections applied panel (c) showsPRISM downscaled to 15m in SNP with our corrections from Table 1 applied to each 15m grid box panel (d) shows estimated mean 119879maxin SNP assuming a temperature increase of 33∘C (average projected temperature increase from NARCCAP RCMs for SNP) applied to July2012 mean 119879max Thin black lines denote elevation contours at 100m spacing with contour lines labeled solid black line indicates the SNPboundary black triangles mark temperature sensors and the two ridgetop climate monitoring stations Big Meadows and Pinnacles arelabeled Note that the downscaling method is only applied to a 15m DEM inside the SNP boundary (panels (b)ndash(d))

high elevation habitats in the regionDownscaling using lapserates based on the 30-year climate means (Figure 4(b)) yieldstemperature estimates that have a warm bias and thus donot agree well with the observationsTherefore we introducecorrections to the downscaled PRISM to provide more

accurate estimates of present-daymeanmonthly119879max in SNPThe corrections that we apply are a function of (1) locationrelative to the ridgeline which we define using catchmentdelineation in ArcGIS and (2) slope azimuth obtained froma 15m DEM The corrections are thus the MBE between the


Table 3 Summary of empirical corrections applied to downscaled PRISM mean July 119879max

Correction applied to downscaled PRISMmean July 119879max along south-facing slopes

(∘C)

Correction applied to downscaled PRISMmean July 119879max along non-south-facing

slopes (∘C)West slope in SNP 132 182East slope in SNP 350 400

observations and downscaled PRISMaveraged over July 2011July 2012 and July 2013 (Table 3) Therefore locations west(east) of the ridgeline and locations that are south-facing (notsouth-facing) have a smaller (larger) correction The use ofthese corrections enables us to produce high-resolutionmapsof present-day mean monthly 119879max in SNP (Figure 4(c)) thatare more accurate than the maps shown in Figures 4(a) and4(b)

34 Climate Change Projections for SNP Using RCMs Thecorrected fine-scale present-day maps of mean monthly 119879maxfrom Figure 4(c) can be used to estimate future temperaturesin the habitat of P shenandoah and in SNP using a tem-perature change obtained from the NARCCAP RCMs Toprovide confidence in the NARCCAP RCMs we compareobserved monthly 119879max from the Big Meadows NCDC dataset withNARCCAPoutput over the period 1971ndash2000Due tothe coarse resolution of the NARCCAP RCMs the elevationof the NARCCAP grid box containing Big Meadows isabout 800m lower than Big Meadowsrsquo actual elevationThusall NARCCAP RCMs overestimate the temperature at BigMeadows The GFDL-RCM3 model has the smallest MBEwhereas the CRCM-CCSMmodel has the largest MBE

All models project an increase in mean July 119879max of 2ndash4∘C (03ndash06∘C decademinus1) through 2055 for the grid boxcontaining the P shenandoah habitat and Big Meadows TheCGCM-WRFG (GFDL-HRM3) model suggests the smallest(largest) increase (029∘C decademinus1) (+060∘C decademinus1)whereas the GFDL-RCM3 model projects an increase of044∘C decademinus1 (Table 4)

To estimate the projected changes in mean July 119879maxbetween the present-day and future NARCCAP simulationswe apply the mean temperature increase in NARCCAPRCMs (33∘C) to present-day mean July 119879max (Figure 4(d))These types of maps can be used by biologists and resourcemanagers to assess and mitigate the effects of climate changeon habitats such as those of P shenandoah

35 Application of the Downscaling Method to other Grid-ded Climate Data Sets While we have focused on usingPRISM in our downscaling method other gridded climatedata sets could be used as well One example is DAYMETDAYMET uses a weighted least-squares regression to inter-polate temperature precipitation moisture and radiationmeasurements from about 8000 surface monitoring stationsto a 1 km resolution over the conterminousUS [22]Whenweuse DAYMET in our downscaling method for mean monthly119879max we find a bias similar to PRISM Whereas downscaled

Table 4 Projected change in mean July 119879max in SNP Shownin parentheses is the projected change in mean July 119879max in∘C decademinus1 The full names of the GCM-RCM combinationsare CCSM-CRCM community climate system model-Canadianregional climate model CGCM-CRCM coupled general circu-lation model-Canadian regional climate model CCSM-WRFGcommunity climate systemmodel-weather research and forecastingmodel CGCM-WRFG coupled general circulation model-weatherresearch and forecasting model GFDL-HRM3 geophysical fluiddynamics model-Hadley regional model version 3 CGCM-RCM3Coupled General Circulationmodel-regional climatemodel version3 GFDL-HRM3 geophysical fluid dynamicsmodel-Hadley regionalmodel version 3

GCM-RCM Change in mean July 119879max (∘C) (∘C

decademinus1)CCSM-CRCM +337 (+048)CGCM-CRCM +391 (+056)CCSM-WRFG +344 (+049)CGCM-WRFG +204 (+029)GFDL-HRM3 +423 (+060)CGCM-RCM3 +274 (+039)GFDL-RCM3 +311 (+044)

mean monthly 119879max using DAYMET and PRISM correlatewell (119903 = 085 119875 lt 001) downscaled mean monthly 119879maxusing DAYMET also overestimates temperatures along boththe west slope and east slope with a bias in mean July 119879max ofabout 4∘C (6∘C) along the west (east) slope and a 1-2∘C bias inwinter Thus the use of DAYMET for mean July 119879max is alsounreliable in SNP without using corrections

36 Limitations of theDownscalingMethod Thedownscalingmethod presented in this paper is an important step towardmaking gridded climate data sets such as PRISMorDAYMETmore reliable in regions of complex terrain However thereare some limitations and caveats that are not explicitlyconsidered with our downscaling method

(1) Our downscaling method does not consider year-to-year lapse rate variations but instead downscalesassuming a lapse rate based on 30-year climatemeansAlthough mean maximum temperatures were about2-3∘C higher in July 2011 than in July 2013 at the siteswhere temperature was measured in SNP the lapserates showed relatively little variability in the threeJuly months of interest and compared well with the30-year mean PRISM lapse rates


(2) Our downscalingmethod does not explicitly incorpo-rate other factors that affect near-surface temperatureon subkilometer spatial scales such as vegetationcover soil moisture soil type vegetation type [2634] and shading by adjacent mountain ridges [33]Further understanding and explicitly quantifyingthese effects and their relative importance over thespatial scale of SNP is very difficult and requiresextensive networks of high-density observations Inaddition high-resolution atmospheric model simu-lations could be used to help isolate some of theother important drivers of near-surface spatiotempo-ral temperature variability in this region

In addition to limitations with the downscaling methoditself there are several limitations with the corrections thatwe introduce in this study to make PRISM more reliable inSNP

(1) The corrections do not vary with elevation We foundthat the difference between observed and downscaledmean July119879max for the sensors deployed along the eastslope did not vary as a function of elevation Along thewest slope of SNP there was more variability in thistemperature difference because of greater variabilityin local slope characteristics that affect temperaturefor example inclination and azimuth [18 19 26] thanalong the east slope

(2) There is a discontinuity in downscaled temperaturepresent along the ridgeline To remove this dis-continuity the application of our correction shoulddecrease with elevation in the immediate vicinity ofthe ridgeline Due to logistical constraints such asthe accessibility to some areas permission issuesand manpower requirements each transect does notbegin right at the ridgeline Therefore we cannotfully assess the necessary corrections that should beapplied within forested locations near the ridgeline

(3) Because of the logistical constraints mentioned pre-viously our corrections for slope azimuth are notbased on measurements made along slopes that wereeither directly north-facing or directly south-facingInstead our correction for azimuth is based onmeasurements from northeast-facing and southeast-facing slopes Based on previous studies [19] weexpect a larger temperature difference between slopesthat are truly north-facing and slopes that are trulysouth-facing

(4) The correction factors are based on data from threeJuly months Although data from more July monthswould help to refine the empirical correction factorsthat we apply to PRISM we expect that even withadditional years of data these correction factorswill change by no more than about 05∘C which ismuch smaller than the range of projected temperaturechanges from the suite of NARCCAP RCMs (cfTable 4)

4 Conclusions and Implications

In this study we have presented a new method to downscalemaximum temperatures in a mountainous region usingPRISM and have evaluated our method with temperatureobservations frommountain slopes in SNPWe found that thedownscaled maximum temperatures using PRISM as wellas using other gridded climate data sets have a warm biasin SNP Several reasons for this warm bias were suggestedincluding the lack of high-elevation stations in the regionto constrain temperature estimates and the effects of canopycover and slope azimuth To reduce the warm bias we intro-duced corrections that depend on the location relative to theridgeline and the locationrsquos azimuth Including corrections inthe downscaling method provides more accurate estimates ofnear-surface temperature than could be obtained using eithera fixed environmental lapse rate or a spatially varying lapserate based on 30-year PRISM climate means Though thereare several limitations the application of our corrections todownscaled PRISM represents an important step to improvetemperature projections at subkilometer spatial scales in SNPand potentially in othermountainous areas as wellThereforefuture work should involve the testing and application of ourdownscaling method to other mountain slopes Ongoing andfuture work also involves using high-resolution atmosphericmesoscale models to dynamically downscale RCM output toSNP and to evaluate how dynamical downscaling methodscompare with the method presented here

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

Acknowledgments

This research and maintenance of the temperature sensornetwork was funded by the National Park Service Cli-mate Change Response Program Cooperative Agreement484010004 Support for student assistantships that helpedmaintain and analyze data from this network was alsoprovided by NSF-CAREER Grant ATM-1151445 The authorsthank Evan Grant and Adrianne Brand from the US Geo-logical Survey for insightful discussions and feedback onthe design of this study They also thank students at theDepartment of Environmental Sciences at the Universityof Virginia for helping to maintain data collection fromthe sensor network especially Stephanie Phelps and MarkSghiatti Thanks also to Zeljko Vecenaj from the Universityof Zagreb Croatia for on-site support in 2012

References

[1] IPCC ldquoClimate change 2007 The physical science basisrdquo inContribution of Working Group I to the Fourth AssessmentReport of the Intergovernmental Panel on Climate Change SSolomonDQinMManning et al Eds pp 1ndash996 CambridgeUniversity Press Cambridge UK 2007


[2] F Giorgi J W Hurrell M R Marinucci and M BenistonldquoElevation dependency of the surface climate change signal amodel studyrdquo Journal of Climate vol 10 no 2 pp 288ndash296 1997

[3] R S Bradley M Vuille H F Diaz and W Vergara ldquoThreats towater supplies in the tropical andesrdquo Science vol 312 no 5781pp 1755ndash1756 2006

[4] E A Beever P F Brussard and J Berger ldquoPatterns of apparentextirpation among isolated populations of pikas (Ochotonaprinceps) in the Great Basinrdquo Journal of Mammalogy vol 84no 1 pp 37ndash54 2003

[5] N L Rodenhouse S N Matthews K P McFarland et alldquoPotential effects of climate change on birds of the NortheastrdquoMitigation and Adaptation Strategies for Global Change vol 13no 5-6 pp 517ndash540 2008

[6] R Highton and R D Worthington ldquoA new salamander of thegenus Plethodon fromVirginiardquoCopeia vol 1967 no 3 pp 617ndash626 1967

[7] L O Mearns W Gutowski R Jones et al ldquoA regional climatechange assessment program for North Americardquo Eos Transac-tions American Geophysical Union vol 90 no 36 pp 311ndash3132009

[8] L O Mearns R Arritt S Biner et al ldquoThe north americanregional climate change assessment program overview of phasei resultsrdquoBulletin of the AmericanMeteorological Society vol 93no 9 pp 1337ndash1362 2012

[9] M Beniston H F Diaz and R S Bradley ldquoClimatic change athigh elevation sites an overviewrdquo Climatic Change vol 36 no3-4 pp 233ndash251 1997

[10] S Antic R Laprise B Denis and R de Elıa ldquoTesting thedownscaling ability of a one-way nested regional climate modelin regions of complex topographyrdquo Climate Dynamics vol 26no 2-3 pp 305ndash325 2006

[11] N Salzmann J Notzli C Hauck S Gruber M Hoelzle andWHaeberli ldquoGround surface temperature scenarios in complexhigh-mountain topography based on regional climate modelresultsrdquo Journal of Geophysical Research F Earth Surface vol112 no 2 Article ID F02S12 2007

[12] S Zong and F K Chow ldquoMeso- and fine-scale modelingover complex terrain parameterizations and applicationsrdquo inMountainWeather Research and Forecasting Recent Progress andCurrent Challenges F K Chow S F J de Wekker and B JSynder Eds pp 591ndash653 Springer New York NY USA 2013

[13] R L Wilby and T M L Wigley ldquoDownscaling general cir-culation model output a review of methods and limitationsrdquoProgress in Physical Geography vol 21 no 4 pp 530ndash548 1997

[14] R Bordoy and P Burlando ldquoBias correction of regional climatemodel simulations in a region of complex orographyrdquo Journal ofApplied Meteorology and Climatology vol 52 no 1 pp 82ndash1012013

[15] J R Minder P WMote and J D Lundquist ldquoSurface tempera-ture lapse rates over complex terrain lessons from the CascadeMountainsrdquo Journal of Geophysical Research vol 115 Article IDD14122 2010

[16] C Rolland ldquoSpatial and seasonal variations of air temperaturelapse rates in alpine regionsrdquo Journal of Climate vol 16 pp1032ndash1046 2003

[17] T R Blandford K S Humes B J Harshburger B C Moore VPWalden andH Ye ldquoSeasonal and synoptic variations in near-surface air temperature lapse rates in a mountainous basinrdquoJournal of Applied Meteorology and Climatology vol 47 no 1pp 249ndash261 2008

[18] Z Tang and J Fang ldquoTemperature variation along the northernand southern slopes of Mt Taibai Chinardquo Agricultural andForest Meteorology vol 139 no 3-4 pp 200ndash207 2006

[19] P V Bolstad L Swift F Collins and J Regniere ldquoMeasuredand predicted air temperatures at basin to regional scales inthe southern Appalachian mountainsrdquo Agricultural and ForestMeteorology vol 91 no 3-4 pp 161ndash176 1998

[20] C Daly G H Taylor W P Gibson T W Parzybok G LJohnson and P A Pasteris ldquoHigh-quality spatial climate datasets for the United States and beyondrdquo Transactions of theAmerican Society of Agricultural Engineers vol 43 no 6 pp1957ndash1962 2000

[21] C Daly W P Gibson G H Taylor G L Johnson andP Pasteris ldquoA knowledge-based approach to the statisticalmapping of climaterdquo Climate Research vol 22 no 2 pp 99ndash1132002

[22] P E Thornton S W Running and M A White ldquoGeneratingsurfaces of daily meteorological variables over large regions ofcomplex terrainrdquo Journal of Hydrology vol 190 no 3-4 pp 214ndash251 1997

[23] J T Abatzoglou ldquoDevelopment of gridded surface meteorolog-ical data for ecological applications and modellingrdquo Interna-tional Journal of Climatology vol 33 no 1 pp 121ndash131 2013

[24] J LWeiss C L Castro and J TOverpeck ldquoDistinguishing pro-nounced droughts in the southwestern united states seasonalityand effects of warmer temperaturesrdquo Journal of Climate vol 22no 22 pp 5918ndash5932 2009

[25] C Daly M Halbleib J I Smith et al ldquoPhysiographically sen-sitive mapping of climatological temperature and precipitationacross the conterminous United Statesrdquo International Journal ofClimatology vol 28 no 15 pp 2031ndash2064 2008

[26] J D Fridley ldquoDownscaling climate over complex terrainHigh finescale (lt1000 m) spatial variation of near-groundtemperatures in a montane forested landscape (Great SmokyMountains)rdquo Journal of Applied Meteorology and Climatologyvol 48 no 5 pp 1033ndash1049 2009

[27] C D Whiteman J M Hubbe and W J Shaw ldquoEvaluationof an inexpensive temperature datalogger for meteorologicalapplicationsrdquo Journal of Atmospheric and Oceanic Technologyvol 17 no 1 pp 77ndash81 2000

[28] T R Lee S F J de Wekker A E Andrews J Kofler andJ Williams ldquoCarbon dioxide variability during cold frontpassages and fair weather days at a forested mountaintop siterdquoAtmospheric Environment vol 46 pp 405ndash416 2012

[29] N Nakicenovic and R Swart Special Report on EmissionsScenarios A Special Report of Working Group III of the Inter-governmental Panel on Climate Change Cambridge UniversityPress Cambridge UK 2000

[30] R Nakamura and L Mahrt ldquoAir temperature measurementerrors in naturally ventilated radiation shieldsrdquo Journal ofAtmospheric and Oceanic Technology vol 22 no 7 pp 1046ndash1058 2005

[31] R Geiger The Climate Near the Ground Harvard UniversityPress Cambridge Mass USA 1965

[32] M D Morecroft M E Taylor and H R Oliver ldquoAir and soilmicroclimates of deciduous woodland compared to an opensiterdquo Agricultural and Forest Meteorology vol 90 no 1-2 pp141ndash156 1998

[33] C D Whiteman Mountain Meteorology Fundamentals andApplications Oxford University Press 2000

[34] T R Oke Boundary Layer Climates Halsted Press New YorkNY USA 2nd edition 1988

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

ClimatologyJournal of

EcologyInternational Journal of


EarthquakesJournal of


Hindawi Publishing Corporationhttpwwwhindawicom

Applied ampEnvironmentalSoil Science

Volume 2014

Mining


Journal of

Hindawi Publishing Corporation httpwwwhindawicom Volume 2014

International Journal of

Geophysics

OceanographyInternational Journal of


Journal of Computational Environmental SciencesHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal ofPetroleum Engineering


GeochemistryHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

Atmospheric SciencesInternational Journal of


OceanographyHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Advances in


MineralogyInternational Journal of


MeteorologyAdvances in

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Paleontology JournalHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

ScientificaHindawi Publishing Corporationhttpwwwhindawicom Volume 2014


Geological ResearchJournal of


Geology Advances in


[22] which have been used in for example ecologicalapplications [23] and climate studies [24]

To obtain accurate temperature projections at subkilo-meter scales statistical downscaling methods could usethe spatially and temporally varying lapse rates that canbe derived from high-resolution gridded data sets In thisstudy we develop such a downscaling method and use thismethod to downscale near-surface temperatures maps atthe high elevation habitats of P shenandoah in SNP undercurrent and future climate conditions We limit ourselvesto a demonstration of our downscaling method for Julymaximum temperatures motivated by the importance ofchanges in summer maximum temperatures to the survivalof P shenandoah

2 Methods

21 Site Description SNP is located along the crest of theVirginia Blue Ridge Mountains on the eastern flank ofthe southern Appalachians in the eastern US The park isoriented southwest-northeast and extends 115 km from nearWaynesboro Virginia to Front Royal Virginia Elevationsin SNP range from 200m above mean sea level (msl) in thevalleys to 1000ndash1200mmsl along the ridgeline East of SNP isthe Virginia Piedmont and the Shenandoah Valley is west ofSNPThe climate of SNP ranges fromhumid subtropical at thelower elevations to humid continental along the ridgetops

22 Downscaling Method Our downscaling method usesthe gridded temperature data sets from PRISM PRISMassimilates observations of temperature (precipitation) fromapproximately 10000 (13000) surface stations in the conter-minous US and linearly interpolates these observations tohigh spatial resolutions using elevation ocean proximity andtopographic facet [20 25] PRISM outputs two data sets (1)mean monthly maximum and minimum temperature andtotal precipitation from 1895 through 2013 at a 4 km resolu-tion and (2) 30-year monthly climate means for maximumand minimum temperature and total precipitation averagedover 1971ndash2000 (and as of 2014 1981ndash2010) at 800m spatialresolution [20 21] PRISM gridded data sets are availableonline from httpprismoregonstateedu

The 4 km and 800m spatial resolutions of the PRISMdatasets are too coarse to resolve temperature variability acrossthe habitats of P shenandoah and thus further downscalingis required In our downscaling method (Figure 1) we down-scale the 4 km PRISMmeanmonthly maximum temperatureto a 15m resolution digital elevation model (DEM) whichis the highest-resolution DEM available for the region usingspatially and temporally varying lapse rates These lapse ratesrepresent the line of best fit between elevation and the 30-year 800m mean monthly maximum temperature for the25 elevation-temperature pairs within each 4 km grid cellThroughout SNP this relationship is high (119903 gt 098 119875 lt0001) Thus the downscaled maximum temperature can bewritten as follows

119879Downscaled = 1198794 km + (1198854 km minus 119885DEM) sdot LR08 km (1)

PRISM mean monthly maximum

resolution)

Downscaled PRISM temperatures

monthly climate

temperature

Digital elevation PRISM 30-year

means for maximumresolution)

Correction factor from slope

measurements

Corrected PRISM temperatures

temperature (4km

(800m resolution)

model (15m

(15m resolution)

(15m resolution)

T4 km Z4 km ZDEMLR08 km

Tdownscaled = T4 km + (Z4 km minus ZDEM ) middot LR08 km

Figure 1 Summary of our downscaling method in which wedownscale the monthly 4 km mean monthly maximum PRISMtemperature (119879

4 km) to a 15m resolution DEM based on the lapserates derived from the 800mPRISM30-yearmonthly climatemeansfor maximum temperature over the period 1971ndash2000 (LR

08 km)and the difference between the mean elevation of the 4 km gridbox (119885

4 km) and elevation of each 15m grid cell (119885DEM) to obtaindownscaled PRISM (119879Downscaled) at a 15m resolution We thenapply a correction to 119879Downscaled based on the slope temperaturemeasurements to obtain more realistic high-resolution estimates oftemperature in Shenandoah National Park

where 119879Downscaled is the PRISM mean monthly maximumtemperature downscaled to a 15m resolution DEM 119879

4 kmis the 4 km PRISM mean monthly maximum temperature1198854 km is the mean elevation of the 4 km grid box 119885DEM is

the elevation of each 15m grid cell and LR08 km is the lapse

rate based on the 800m PRISM 30-year monthly climatemeans for maximum temperature where a positive lapserate indicates a decrease in temperature with height Thisdownscaling method is evaluated using observations froma network of temperature sensors discussed in the nextsection

23 Slope Measurements The observations which we use toevaluate our downscaling method come from a network oftemperature sensors deployed in SNP We use Onset HOBOPro V2 temperature-relative humidity sensors which havea manufacturer-state accuracy of plusmn021∘C over the range 0∘to 50∘C for temperature and plusmn25 from 10 to 90 forrelative humidity These sensors have been tested and usedsuccessfully in many applications to investigate near-surfacetemperature variations [26 27]

In April 2011 we deployed 37 sensors along 3 transects inSNP The transect east of the ridgeline included 18 sensorsand the transects west of the ridgeline included 19 sensors(Figure 2) For all transects sensors were deployed along aline extending from near the ridgeline (about 1000mmsl) toabout 600mmsl Of the 19 sensors along the west side of the


SensorsSNP boundary

Elevation (m)

0 1 2 3 405(km)

300

-400

400

-500

500

-600

600

-700

700

-800

800

-900

900

-1000

1000

-1100

78∘20

9984000998400998400W78

∘22

9984000998400998400W78

∘26

9984000998400998400W 78

∘24

9984000998400998400W

38∘36

9984000998400998400N

38∘32

9984000998400998400N

38∘34

9984000998400998400N

lt300

gt1100

N

N








2emissions than in






22 24 26 28 30 3222

24

26

28

30

32


Dow

nsca

led

PRIS

M te

mpe

ratu

re (∘

C)

(a)

22 24 26 28 30 3222

24

26

28

30

32


Dow

nsca

led

PRIS

M te

mpe

ratu

re (∘

C)(b)



20 22 24 26 28 3020

22

24

26

28

30


Dow

nsca

led

PRIS

M te

mpe

ratu

re (∘

C)

(c)

















Big Meadows

Pinnacles

78∘26

9984000998400998400W 78

∘24

9984000998400998400W 78

∘22

9984000998400998400W 78

∘20

9984000998400998400W

38∘38

9984000998400998400N

38∘36

9984000998400998400N

38∘34

9984000998400998400N

N

38∘32

9984000998400998400N

200

300

300

400

500

500

600

600

600

700

700700

800

800

900

900

9001000

1000

1100

(a)

38∘38

9984000998400998400N

38∘36

9984000998400998400N

38∘34

9984000998400998400N

38∘32

9984000998400998400N

Big Meadows

Pinnacles

78∘26

9984000998400998400W 78

∘24

9984000998400998400W 78

∘22

9984000998400998400W 78

∘20

9984000998400998400W

N

200

300400

500

500

600

600

600

700700

700

800

300

800

900

900

900

1000

1000

1100

(b)


(km)

lt22

22

-23

23

-24

24

-25

25

-26

26

-27

27

-28

28

-29

29

-30

30

-31

31

-32

gt32

Temperature (∘C)

38∘38

9984000998400998400N

38∘36

9984000998400998400N

38∘34

9984000998400998400N

38∘32

9984000998400998400N

Big Meadows

Pinnacles

78∘26

9984000998400998400W 78

∘24

9984000998400998400W 78

∘22

9984000998400998400W 78

∘20

9984000998400998400W

N

200

300

300

400

500

500

600

600

600700

700

700

800

800

900

900

900

1000

1000

1100

(c)


(km)

lt22

22

-23

23

-24

24

-25

25

-26

26

-27

27

-28

28

-29

29

-30

30

-31

31

-32

gt32

Temperature (∘C)

38∘38

9984000998400998400N

38∘36

9984000998400998400N

38∘34

9984000998400998400N

38∘32

9984000998400998400N

Big Meadows

Pinnacles

78∘26

9984000998400998400W 78

∘24

9984000998400998400W 78

∘22

9984000998400998400W 78

∘20

9984000998400998400W

N

200

300

300

400

500

500

600

600

600

700

700700

800

800

900

900

900

1000

1000

1100

(d)







(∘C)

























Acknowledgments


References













































Volume 2014

Mining


Journal of



Geophysics







Journal of




Advances in











Geology Advances in


SensorsSNP boundary

Elevation (m)

0 1 2 3 405(km)

300

-400

400

-500

500

-600

600

-700

700

-800

800

-900

900

-1000

1000

-1100

78∘20

9984000998400998400W78

∘22

9984000998400998400W78

∘26

9984000998400998400W 78

∘24

9984000998400998400W

38∘36

9984000998400998400N

38∘32

9984000998400998400N

38∘34

9984000998400998400N

lt300

gt1100

N

N








2emissions than in






22 24 26 28 30 3222

24

26

28

30

32


Dow

nsca

led

PRIS

M te

mpe

ratu

re (∘

C)

(a)

22 24 26 28 30 3222

24

26

28

30

32


Dow

nsca

led

PRIS

M te

mpe

ratu

re (∘

C)(b)



20 22 24 26 28 3020

22

24

26

28

30


Dow

nsca

led

PRIS

M te

mpe

ratu

re (∘

C)

(c)

















Big Meadows

Pinnacles

78∘26

9984000998400998400W 78

∘24

9984000998400998400W 78

∘22

9984000998400998400W 78

∘20

9984000998400998400W

38∘38

9984000998400998400N

38∘36

9984000998400998400N

38∘34

9984000998400998400N

N

38∘32

9984000998400998400N

200

300

300

400

500

500

600

600

600

700

700700

800

800

900

900

9001000

1000

1100

(a)

38∘38

9984000998400998400N

38∘36

9984000998400998400N

38∘34

9984000998400998400N

38∘32

9984000998400998400N

Big Meadows

Pinnacles

78∘26

9984000998400998400W 78

∘24

9984000998400998400W 78

∘22

9984000998400998400W 78

∘20

9984000998400998400W

N

200

300400

500

500

600

600

600

700700

700

800

300

800

900

900

900

1000

1000

1100

(b)


(km)

lt22

22

-23

23

-24

24

-25

25

-26

26

-27

27

-28

28

-29

29

-30

30

-31

31

-32

gt32

Temperature (∘C)

38∘38

9984000998400998400N

38∘36

9984000998400998400N

38∘34

9984000998400998400N

38∘32

9984000998400998400N

Big Meadows

Pinnacles

78∘26

9984000998400998400W 78

∘24

9984000998400998400W 78

∘22

9984000998400998400W 78

∘20

9984000998400998400W

N

200

300

300

400

500

500

600

600

600700

700

700

800

800

900

900

900

1000

1000

1100

(c)


(km)

lt22

22

-23

23

-24

24

-25

25

-26

26

-27

27

-28

28

-29

29

-30

30

-31

31

-32

gt32

Temperature (∘C)

38∘38

9984000998400998400N

38∘36

9984000998400998400N

38∘34

9984000998400998400N

38∘32

9984000998400998400N

Big Meadows

Pinnacles

78∘26

9984000998400998400W 78

∘24

9984000998400998400W 78

∘22

9984000998400998400W 78

∘20

9984000998400998400W

N

200

300

300

400

500

500

600

600

600

700

700700

800

800

900

900

900

1000

1000

1100

(d)







(∘C)

























Acknowledgments


References













































Volume 2014

Mining


Journal of



Geophysics







Journal of




Advances in











Geology Advances in


22 24 26 28 30 3222

24

26

28

30

32


Dow

nsca

led

PRIS

M te

mpe

ratu

re (∘

C)

(a)

22 24 26 28 30 3222

24

26

28

30

32


Dow

nsca

led

PRIS

M te

mpe

ratu

re (∘

C)(b)



20 22 24 26 28 3020

22

24

26

28

30


Dow

nsca

led

PRIS

M te

mpe

ratu

re (∘

C)

(c)

















Big Meadows

Pinnacles

78∘26

9984000998400998400W 78

∘24

9984000998400998400W 78

∘22

9984000998400998400W 78

∘20

9984000998400998400W

38∘38

9984000998400998400N

38∘36

9984000998400998400N

38∘34

9984000998400998400N

N

38∘32

9984000998400998400N

200

300

300

400

500

500

600

600

600

700

700700

800

800

900

900

9001000

1000

1100

(a)

38∘38

9984000998400998400N

38∘36

9984000998400998400N

38∘34

9984000998400998400N

38∘32

9984000998400998400N

Big Meadows

Pinnacles

78∘26

9984000998400998400W 78

∘24

9984000998400998400W 78

∘22

9984000998400998400W 78

∘20

9984000998400998400W

N

200

300400

500

500

600

600

600

700700

700

800

300

800

900

900

900

1000

1000

1100

(b)


(km)

lt22

22

-23

23

-24

24

-25

25

-26

26

-27

27

-28

28

-29

29

-30

30

-31

31

-32

gt32

Temperature (∘C)

38∘38

9984000998400998400N

38∘36

9984000998400998400N

38∘34

9984000998400998400N

38∘32

9984000998400998400N

Big Meadows

Pinnacles

78∘26

9984000998400998400W 78

∘24

9984000998400998400W 78

∘22

9984000998400998400W 78

∘20

9984000998400998400W

N

200

300

300

400

500

500

600

600

600700

700

700

800

800

900

900

900

1000

1000

1100

(c)


(km)

lt22

22

-23

23

-24

24

-25

25

-26

26

-27

27

-28

28

-29

29

-30

30

-31

31

-32

gt32

Temperature (∘C)

38∘38

9984000998400998400N

38∘36

9984000998400998400N

38∘34

9984000998400998400N

38∘32

9984000998400998400N

Big Meadows

Pinnacles

78∘26

9984000998400998400W 78

∘24

9984000998400998400W 78

∘22

9984000998400998400W 78

∘20

9984000998400998400W

N

200

300

300

400

500

500

600

600

600

700

700700

800

800

900

900

900

1000

1000

1100

(d)







(∘C)

























Acknowledgments


References













































Volume 2014

Mining


Journal of



Geophysics







Journal of




Advances in











Geology Advances in













Big Meadows

Pinnacles

78∘26

9984000998400998400W 78

∘24

9984000998400998400W 78

∘22

9984000998400998400W 78

∘20

9984000998400998400W

38∘38

9984000998400998400N

38∘36

9984000998400998400N

38∘34

9984000998400998400N

N

38∘32

9984000998400998400N

200

300

300

400

500

500

600

600

600

700

700700

800

800

900

900

9001000

1000

1100

(a)

38∘38

9984000998400998400N

38∘36

9984000998400998400N

38∘34

9984000998400998400N

38∘32

9984000998400998400N

Big Meadows

Pinnacles

78∘26

9984000998400998400W 78

∘24

9984000998400998400W 78

∘22

9984000998400998400W 78

∘20

9984000998400998400W

N

200

300400

500

500

600

600

600

700700

700

800

300

800

900

900

900

1000

1000

1100

(b)


(km)

lt22

22

-23

23

-24

24

-25

25

-26

26

-27

27

-28

28

-29

29

-30

30

-31

31

-32

gt32

Temperature (∘C)

38∘38

9984000998400998400N

38∘36

9984000998400998400N

38∘34

9984000998400998400N

38∘32

9984000998400998400N

Big Meadows

Pinnacles

78∘26

9984000998400998400W 78

∘24

9984000998400998400W 78

∘22

9984000998400998400W 78

∘20

9984000998400998400W

N

200

300

300

400

500

500

600

600

600700

700

700

800

800

900

900

900

1000

1000

1100

(c)


(km)

lt22

22

-23

23

-24

24

-25

25

-26

26

-27

27

-28

28

-29

29

-30

30

-31

31

-32

gt32

Temperature (∘C)

38∘38

9984000998400998400N

38∘36

9984000998400998400N

38∘34

9984000998400998400N

38∘32

9984000998400998400N

Big Meadows

Pinnacles

78∘26

9984000998400998400W 78

∘24

9984000998400998400W 78

∘22

9984000998400998400W 78

∘20

9984000998400998400W

N

200

300

300

400

500

500

600

600

600

700

700700

800

800

900

900

900

1000

1000

1100

(d)







(∘C)

























Acknowledgments


References













































Volume 2014

Mining


Journal of



Geophysics







Journal of




Advances in











Geology Advances in




(∘C)

























Acknowledgments


References













































Volume 2014

Mining


Journal of



Geophysics







Journal of




Advances in











Geology Advances in












Acknowledgments


References













































Volume 2014

Mining


Journal of



Geophysics







Journal of




Advances in











Geology Advances in












































Volume 2014

Mining


Journal of



Geophysics







Journal of




Advances in











Geology Advances in

Research Article Downscaling Maximum …downloads.hindawi.com/journals/amete/2014/594965.pdfResearch...

Documents

Transcript of Research Article Downscaling Maximum …downloads.hindawi.com/journals/amete/2014/594965.pdfResearch...