Table Of Content15DECEMBER2002 ATKINSON AND GAJEWSKI 3601
High-Resolution Estimation of Summer Surface Air Temperature in the Canadian
Arctic Archipelago
DAVID E. ATKINSON AND K. GAJEWSKI
LaboratoryofPaleoclimatologyandClimatology,DepartmentofGeography,UniversityofOttawa,Ottawa,Ontario,Canada
(Manuscriptreceived27August2001,infinalform4June2002)
ABSTRACT
IntheCanadianhighArcticpatternsoftemperaturearepoorlyresolvedatthemesoscale.Thisissueisaddressed
using a model to estimate mean summer surface air temperature at high spatial resolution. The effects on
temperatureofsiteelevationandcoastalproximitywereselectedforparameterization.Thespatialbasisisa1-
km resolution digital elevation model of the region. Lapse rates and resultant wind estimates were obtained
from upper-air ascents. These were used to estimate the change in temperature with elevation based on the
digitalelevationmodel.Advectioneffectsarehandledusingresultantwinds,airtemperatureabovetheocean,
anddistancetocoast.Modelresultsfor14-dayrunswerecomparedtoobserveddata.Thetwoeffectscaptured
muchofthemesoscalevariabilityoftheArcticclimate,asshownbyverificationwithpointobservationaldata.
Sensitivityanalyseswereperformedonthemodeltodetermineresponsetoalterationsinlapseratecalculation,
seasurfacetemperature,andwindfieldgeneration.Themodelwasmostsensitivetothelapseratecalculation.
The best results were obtained using a moderate lapse rate calculation, moderate wind field, and variable sea
surfacetemperature.
1. Introduction free,whichvariessignificantlyonanannualbasis.Land
surfaces near the coast experience a typical pattern of
Interactionsoftheearth’ssurfacewiththeatmosphere
maritime attenuation, whereastheinteriorsoflargeris-
areparticularlyevidentintheArctic.Plantsurvivaland
lands exhibit continental conditions. Exceptions arear-
growth is closely tied to the climate (Arft et al. 1999)
eas near snowpacks or extended ice fields, which are
and the presence of permafrost is a major influence on
cooledinthesummer.Topographiccomplexityalsocon-
landscape dynamics (Williams and Smith 1989). Un-
tributes to mesoscalevariabilityintemperature,precip-
derstanding such climate–surfaceinteractionsisimpor-
itation, and cloudiness. These factors serve to render
tant,asfutureclimatechangesarepredictedtobegreater
questionable results taken from surface air temperature
here than in most areas of the world with a potentially
plotsthatarebasedoninterpolationfromthefewavail-
large impact on the landscape (Watson et al. 1995).
able meteorological stations.
However, at the present time, environmental and pa-
Improving the spatial resolution of surface air tem-
leoenvironmentalresearchintheArcticishamperedby
perature estimates is thus an important contribution to
a lackofmesoscaleclimate,andmostimportantlytem-
an understanding of surface climate in this region, and
perature, data.
The Canadian Arctic Archipelago (CAA; Fig. 1) is one that is not forthcoming from the existing obser-
served by few meteorological stations. The mean sep- vational network. Two mechanisms exist to better un-
aration between stations of the Meteorological Service derstand mesoscale temperature: integrating alternate
of Canada (MSC) is 500 km and therepresentativeness data or information into an analysis (Atkinson et al.
ofallstationssuffersduetolocalcoastalbiasandhighly 2000;Atkinson2000;Kahletal.1992;AltandMaxwell
variedtopographyandsurfacetypes.Physiographycon- 1990)orusingempirical(WillmottandMatsuura1995;
tributes to temperature pattern variability at the meso- Daly et al. 1994) or physical models (Trenberth 1992)
scale for various reasons. The archipelago is heavily of the atmosphere to augment traditional analysesand/
fiorded, exposing landareastoanoceanthatcanbeice or data sources.
covered, contain isolated floating ice floes, or be ice Maxwell (1980, 1982) used information fromhistor-
ical short-term stations and his own experience to sub-
jectively modify isotherms to depict cooler ice field/
Correspondingauthoraddress:Dr.DavidE.Atkinson,Geological uplandregions.AltandMaxwell(1990)employednon-
Survey of Canada (Atlantic), Bedford Institute of Oceanography, 1
standard,short-termweatherobservationdatafromsev-
ChallengerDr.,P.O.Box1006,DartmouthNSB2Y4A2,Canada.
E-mail:[email protected] eral, more recent sources(e.g.,Atkinsonetal.2000)to
(cid:1)2002AmericanMeteorologicalSociety
3602 JOURNAL OF CLIMATE VOLUME15
uationwiththeunderlyingsurface.Themodeldescribed
below takes as input the synoptic-scale features of the
temperature field, asestimatedfromtheMSCupper-air
stations,andmodifiestheirsignalusingelevationaland
coastal proximity data derived from the DEM.
2. Data and model description
The model was implemented at a spatial resolution
of 1 km (cid:1) 1 km. Physical processes accommodatedin
this model areas follows:
• Themeanenvironmentallapseratespecifictothetime
periodbeingmodeled,derivedusingtemperaturedata
fromrawinsondeascentsatMSCupper-airstationsin
the study region, is used to define the rate of tem-
perature change with elevation.
FIG.1.TheCAA.Upper-airstationsoperatedbytheMSCare • The mean, low-level wind direction and velocity,de-
indicated. rived from rawinsonde ascents, is used to determine
the extent to which coastal zones are modified by
onshore advective flow.
increasespatialdetailofaJulytemperaturenormalplot
• Surface temperatures for locations possessing major
for the Queen Elizabeth Islands. Jacobs (1990) linked
ice fields are stipulated using a linearmodificationof
an automatic weather station to MSC weather stations
the base temperature estimate.
using transferfunctionsallowingthegenerationofdata
at a ‘‘virtual’’ station. Other studies have used the ap- The spatial basis of the model is a DEM of the Ca-
proach of guidedtemperatureestimationusingadigital nadianArcticArchipelago,organizedasamatrixof1996
elevationmodel(DEM)inconjunctionwithalapserate columns by 1833 rows, subsetfromtheU.S.Geological
for detailed climate work (Daly et al. 1994; Willmott Survey GTOPO30 DEM of the world (available online
andMatsuura1995;Dalyetal.1997;DalyandJohston at http://edcdaac.usgs.gov/gtopo30/gtopo30.html). Each
1998; Johnson et al. 2000) or to support other types of point represents approximately 1 km2.
research (Santibanez et al. 1997; Goodale et al. 1998; The first step in estimating surface air temperature
Dodson and Marks 1997). valuesforeachpointwastoobtainmeanenvironmental
In this paper, we describe a semiempirical model of lapseratesforeachstation.Theseweregeneratedusing
the mesoscale summer temperature climate of the Arc- vertical profiles of dry-bulb temperature obtainedfrom
tic. The conceptual basis for the model is that much of twice-daily rawinsonde ascents at stations throughout
the spatial variability of the Arctic surfacetemperature the region (Table 1). The mean ascent curve was de-
regime can be accounted forbyseveralprocesses.Spe- scribed using a fifth-order polynomial. A high-order
cifically, we hypothesized that the two most important polynomial was used because it was felt important to
contributors to the spatial variability of surface tem- model a shallow, surface inversion that was found to
perature patterns at the mesoscale (horizontal scale of be present in many of the ascent profiles (Figs. 2a,b),
tens to hundreds of kilometers) are 1)variationoftem- which are discussed below.
perature with elevation, and 2) location with respectto The inversions (Table 1) were smaller in magnitude
advective sources of air temperaturemodification,such thanthoseobservedinwinter(Bradleyetal.1992;Max-
as large bodies of water or ice fields. well 1980). Their likely cause is advective, ratherthan
Elevationaleffectsweretargetedbecausemanyofthe radiative, given that the summer net surface radiation
islands consist of large central plateaus with a small balance is positive. It was thus assumed that a summer
coastal zone. In the northern and eastern parts of the surface inversion at a coastal location is a local-scale
archipelago,significantmountainousregionsarefound. effect that must be removed before using the environ-
Concurrent lapse rates applied to site elevations were mental lapse rate to represent interior sites.
felt to be the best way to improve estimates of tem- Removal of the inversion involved firstdetectingthe
perature in these areas. Advective effects were also inflection point on the curve above the inversion using
modeled because many of the islands in the CAA are aglobal-maximumdetectionalgorithm(McCrackenand
large enough to possess a coast-to-interior heatinggra- Dorn 1964). Next, data were extrapolated from this
dientthatrangesfromunimpededsurfaceheatinginthe pointtothesurfaceusingtherateofchangethatexisted
interior to coastal locations completely dominated by inthecurveabovetheinversion.Thenewascentseries
maritime air. than had the polynomial equation refit to it (Fig. 2c).
Ingeneral,thesurfacetemperatureclimateattheme- This procedure was verified by comparing estimatesof
soscale is formed by the interaction of thesynopticsit- surface temperature made by the refit polynomial to
15DECEMBER2002 ATKINSON AND GAJEWSKI 3603
TABLE 1. Upper-air stations used to generate regional estimates of environmental lapse rate. Frequency of inversionsobserved in mean
ascentcurvesduringmodelruns(1974–88,1990)arelisted.Valueclassisheightoftheinversionmaximuminmabovetheground.Here
GTrefersto‘‘greaterthan700m’’(observedonlyattheAlaskastations).
No
inver- (cid:2)100 (cid:2)200 (cid:2)300 (cid:2)400 (cid:2)500 (cid:2)600 (cid:2)700 GT
Upper-airstation Lat(N) Lon(W) sion (m) (m) (m) (m) (m) (m) (m) 700m
Alert 82(cid:3)20(cid:4) 62(cid:3)30(cid:4) 10 2 — 2 2 — — 1 —
BarrowPoint(Alaska) 80(cid:3)00(cid:4) 85(cid:3)56(cid:4) 4 — — — 1 5 1 2 4
BarterIsland(Alaska) 76(cid:3)14(cid:4) 119(cid:3)20(cid:4) 2 1 — — 2 2 2 3 5
CambridgeBay 74(cid:3)43(cid:4) 94(cid:3)59(cid:4) 14 — 3 — — — — — —
Eureka 63(cid:3)45(cid:4) 68(cid:3)33(cid:4) 4 3 5 5 — — — — —
Iqaluit(FrobisherBay) 68(cid:3)47(cid:4) 81(cid:3)15(cid:4) 9 1 4 2 — 1 — — —
HallBeach 69(cid:3)07(cid:4) 105(cid:3)01(cid:4) 6 3 2 2 2 2 — — —
MouldBay 71(cid:3)18(cid:4) 156(cid:3)47(cid:4) 14 1 1 — — — — 1 —
ResoluteBay 70(cid:3)05(cid:4) 143(cid:3)36(cid:4) 8 2 1 3 1 1 — 1 —
Totals 71 13 16 14 8 11 3 8 9 153
observations from summer research camps at inland niques (e.g., McCullagh 1981; Shepard 1968).Temper-
sites (Atkinson 2000). aturevalueswerethenobtainedbysolvingtheequation
Next, the polynomial coefficients representing the at each grid point using elevation data as the indepen-
lapse rate at each station were interpolated throughout dent value. This gave a regionwide estimate of surface
the DEM grid. Each coefficient was interpolated indi- temperature that reflected the environmental lapse rate
vidually onto the grid using an inverse distanceweight without a coastal signal. A concern when using upper-
procedure with decay set to a factor of 2; this was se- air temperature data to estimate near-surface air tem-
lectedtoprovideabalancebetweenlocalweightingand peratures is the potential for underestimation of near-
range of influence. The paucity of observing sites and surface temperatures; however, consistent bias or large
a lack of spatial structure(e.g.,nopointclustering)did departures from verification data were not observed
not warrant use of more specialized interpolationtech- (Figs. 3a–d).
FIG.2.Typicalverticaltemperatureprofilesforindividualascentsshowingtheinversion(thin
line with black dots) and the fitted polynomial curve (heavier black line): (a) 2 Jul 1987 0000
UTC; (b) 23 Jul 1987 1200 UTC. (c) Mean rawinsonde ascent profiles for Jul generated by
averagingallpolynomialestimatesforeachascentoverthemonthofJulforagivenyear(solid
line). Inversion is removed by extrapolating the straight portion of the original curve (dashed
line) to the surface (‘‘high-slope’’inversion removal algorithm).All profilesobtainedfromthe
Eurekaupper-airstation.
3604 JOURNAL OF CLIMATE VOLUME15
TABLE2.Winddirectionandvelocityclassificationcategories.
Direction Velocity
((cid:3)trueN) Class (kmh(cid:5)1) Class
316–45 North 0 Calm
46–135 East 1–13 Low
136–225 South 14–26 Medium
226–315 West 27(cid:6) High
Temperatures for ice fields were then estimated.Ha-
vens et al. (1965) demonstrated an average ‘‘ice field
cooling’’ factor of about 3(cid:3)C using data from two sta-
tions, one on top of an ice field and the other nearby
on a nonice surface. In the model, ice field locations
were assigned a new value that consisted of the initial
temperature estimate minus this cooling factor.
Next thecoastal effectwasparameterized.Theinflu-
ence of wind for this model is expressed as a mixing
ofthebaselandestimates,obtainedasdescribedabove,
withthetemperatureovertheocean.Windvelocityfrom
the 90-kPa level was extracted from upper-air ascents.
The90-kPalevelwasselectedbecauseitishighenough
((cid:7)900 m) to be above most topography and to possess
the steady characteristics of winds at higher levels, yet
low enough to reasonably represent the direction and
speedofwindsfeltatthesurface.Basedonthesevalues
the image was classified into four direction and speed
classes, giving a total of 13 categories (12 categories
when speeds were (cid:8)0, and 1 category for 0 wind
speeds) (Table 2). Velocity classification was based on
a breakdown of observed wind speeds such that the
majority of wind events fell into the ‘‘low’’ category
andprogressivelyfewerintothe‘‘medium’’and‘‘high’’
categories. These wind categories formed the basis for
theselectionofa‘‘matrixfilter’’(Bonham-Carter1994)
that was applied to abinary representationoftheDEM
inwhichlandpixelsareassignedavalueof1andocean
pixels are assigned 0. A matrix filter is a small, square
matrix composed of values that are symmetric and op-
posite. This filter is placed over a given pixel on the
binary DEM. The neighborhood around the pixel that
matches the filter in size is extracted from the binary
DEM and multiplied, pixel by pixel, with the filter.All
the values in the resulting matrix are then summed to
arrive at asinglevalue;thisvaluerepresentsthepoten-
tial wind influence on a pixel, which is used in Eq. (1)
below.Thefilterisarrangedsuchthatthelargestvalues
arenearthemiddle,representingcloseproximitytothe
ocean,withasteadydecaytotheedgeofthefilter.Thus,
pixels near the ocean will feel the greatest potential
influence of an onshore flow, decreasing with distance
from the coast. The effects of a stronger flow, which
are greater potential impact on the near-shore environ-
FIG.3.Dailytemperaturedataobservedfromanautomaticweather ment and farther potential inland penetration, is repre-
station (dashed lines) and estimated by the model for the same lo-
sented by a filter that has both largervalues,tocapture
cation (solid lines) for the years and periods indicated. All plotted
data series have been filtered using a five-point Gaussian kernel. greater impact, and larger physical size, to represent a
Automatic weather station was located at the PCSP Hot Weather greater inland penetration. Greater physical size of the
Creek research camp, 30-km inland from the upper-air station at filter is used to represent the increased range of effect
EurekaonEllesmereIsland.
15DECEMBER2002 ATKINSON AND GAJEWSKI 3605
TABLE3.Modelrunperiodsselectedforthe‘‘original’’ TABLE4.Sensitivityanalyses.
parameterizationset.
Total
Year Dates Parameter Natureofalteration runs
1974 9–22Jul Inversionremoval High-sloperemoval(original) 17
1975 8–21Jul Low-sloperemoval 5
1976 4–17Jul Peak-pointremoval 5
1977 19Jul–1Aug Nomodificationtotemperature 5
1978 21Jul–3Aug profile
1979 9–22Jul Seasurfacetemperature Constantoverentireregion 17
1980 23Jul–5Aug (original)
1981 29Jun–12Jul Variableoverregionwithlow- 5
1982 8–21Jul sloperemoval
1983 28Jun–11Jul Variableoverregionwithpeak- 5
1984 16–29Jul pointremoval
1985 6–19Jul Coastalwindeffect Moderate effect application (orig- 17
1986 9–22Jul inal)
1987 5–18Jul Maximumeffectapplicationwith 5
1988 16–29Jul low-sloperemovaland
1989 13–26Jul constantSST
1990 9–22Jul Maximumeffectapplicationwith 5
low-sloperemovaland
variableSST
of a stronger flow because in the DEM the wind filter
cannot be applied to an area that it does not physically Three parameterizations were targeted for sensitivity
reach. analysis: the inversion removal algorithm, the sea sur-
The result of application of the wind filter was a face temperature (SST), and the wind effect (Table 4).
‘‘resultant wind effect’’ parameter that represented a Each sensitivity combination was tested on 5 separate
potential modification of the base temperatureestimate years;thesame5yearswereusedineachcasetopermit
atasite.Themaximumvalueforaresultantwindeffect comparison (Table 5). Three additional approaches to
is 100, indicating that 100% of the temperature at that dealing with the inversion were considered: a ‘‘peak-
pixel is a result of ocean influence, and the minimum point’’ removal, in which the slope removal line was
is 0, for no modification due to wind. The wind effect drawn vertically down from the point of maximum
parameterandthevaluesfromthetemperatureestimates warming to the surface; a ‘‘low-slope’’ removal, in
image were combined using which the slope removal line was drawn from a point
roughly halfway between the original model and the
T (cid:9) W (cid:1) T (cid:6) (1 (cid:5) W) (cid:1) T , (1) peak-pointremoval;and‘‘none,’’inwhichnoalteration
r s L
to the observed lapse rate was performed. These rep-
whereT (cid:9)resultanttemperaturevalueatagivenpoint,
r resent a gradation in the magnitude of inversion re-
W(cid:9)windmodificationvalue(%),T (cid:9)airtemperature
s moval, from a maximum in the original model (‘‘high-
over the ocean surface, and T (cid:9) air temperature over
L slope’’ removal) to no alteration (none). For SST, the
the land surface obtained from the polynomial-based
constant value was replaced by values derived from a
estimate.ValuesforT weresetat2(cid:3)Corrangedbetween
s map of mean observed SST (Maxwell 1982). The ex-
0(cid:3) and 4(cid:3)C depending on the type of model run being
isting wind effect was increased in strength, such that
conducted. Topographic modification of wind was not
its influence could be felt twice as far inland as in the
explicitly parameterized.
original model.
The final output of a model run was a 1 km (cid:1) 1 km
For verification, mean temperature values were cal-
grid of estimates of the mean surface air temperature
culated using available surface stations present during
for the period of the model run. Values were estimated
the period of the run. This included both MSC and
foralllandsurfacesoveraregionencompassingallthe
islands in the Canadian Arctic Archipelago, Boothia
Peninsula,andsomeofthenorthcoastofthemainland.
TABLE5.Periodsforwhichsensitivityanalyseswererun.
Temperatures were estimated for 2-week averaging
periods using an initial set of parameterizations (iden- Period Reasonforselecting
tifiedas‘‘original’’),onerunforeachof17years(Table 9–22Jul1974 Largezoneofnegativeresidualinorig-
3), although the model can be run for any averaging inalmodel
21Jul–3Aug1978 Lackofinversionsforthetimeperiod
period. Dates of application varied from year to year
28Jun–11Jul1983 Largezoneofpositiveresidualinorig-
and were chosen to maximize the availability ofobser- inalmode
vationaldata(Atkinson2000)forverificationpurposes. 16–29Jul1984 Largenumberofstationsavailablefor
Based on these results various permutations of model verification
16–29Jul1988 Climatologicallywarmsummer
parameterizationchangeswererunonsubsetsof5years.
3606 JOURNAL OF CLIMATE VOLUME15
FIG.4.Modelresultsfor9–22Jul1974.Resultsareestimatesofmeansurfaceairtemperature((cid:3)C),integratingthe
specifiedtimeperiod,arrangedona1km(cid:1)1kmgrid.Locationsofallverificationsiteshavebeenplottedasblack
dots.
nonstandard data from thePolarContinentalShelfPro- persistentpenetrationofcoolairsoutheastfromtheArc-
ject archives(Atkinsonetal.2000).Modelestimatesat tic Ocean into the central CAA.
grid points coinciding with the station locations were The primary diagnostictool forassessingmodelper-
extractedandresidualswerecalculated.Residualswith- formancewasasetofresidualsobtainedbysubtracting
intherange(cid:10)1.4(cid:3)Cwereconsideredacceptable,asthis modelestimatesfromobservedstationtemperaturedata.
is the minimum standard deviation of two-weekmeans Model estimates were obtained from the grid points
obtained from theobservationaldata.Valuesoutsideof closest to a given station, andtheobservedstationdata
this range were mapped to gauge the performance of were averaged for the time period coincident with the
the model by revealing regions of systematic over- or model run. These residuals were processed as a com-
underestimation. pleteset(i.e.,andnotbyyear)toobtainameanabsolute
errorandwerebothplottedagainststationdistancefrom
coast and elevation to look forsituationalbiasesandin
3. Results
a mapped form to identify spatial zones of model in-
Output from selected years is presented in Figs. 4–6 consistency. An overall mean absolute error of 1.5(cid:3)C
and Table 6. Temperature values have been rounded to was obtained on the 386 residual values available for
the nearest whole degree Celsius. As expected, cooler the 17 two-week model runs. Consideringtheresiduals
temperatureswerefoundathigherelevationsintheeast- separatelyforeachrun,theresidualsrangedfromhaving
ern Arctic. A general north–south temperaturegradient a mean that was close to zero with low variation (e.g.,
was also captured, as was a ‘‘northwest cool bulge,’’ 1976) to a mean that deviated significantly from zero
which is a typical temperature pattern caused by the with large variation (e.g., 1974). Overall, negative re-
15DECEMBER2002 ATKINSON AND GAJEWSKI 3607
FIG.5.Modelresultsfor4–17Jul1976.Resultsareestimatesofmeansurfaceairtemperature((cid:3)C),integratingthe
specifiedtimeperiod,arrangedona1km(cid:1)1kmgrid.Locationsofallverificationsiteshavebeenplottedasblack
dots.
siduals(modeloverestimation)weremorecommon.All Axel Heiberg Islands (1977, 1978, 1982, 1983, 1984,
availableresidualswereplottedagainststationelevation 1988,1990),andoftenintheeasternpartsofEllesmere
and distance from coast (Fig. 7). Biases in the model (1977, 1988, 1990).
estimates were not apparent. In general, none of the sensitivity combinations in-
A more detailed assessment of model performance vestigated yielded a clearly superior result (Table 8);
wasdeterminedbyconsideringthesizeofzonesformed however, they all yielded results that were superior to
by residuals of a given sign. A large residual zone of the original inversion removal algorithm (high slope).
a given sign is more likely a result of a systematic Applying different inversion removal algorithms while
shortcominginthemodel,whereassmall,discontinuous maintaining the constant sea surface temperature re-
residualzonesofbothsignsindicatelocalforcingagents sulted in skewed residual groupings: skewed negative
or random fluctuations. The 129 residuals that fell out- using the high-slope and low-slope removals, and
side the acceptable range formed a total of 57 zones; skewed positive using the peak-point and none resid-
those zones, which possessed three or more stations, uals,withthetotalnumberofresidualsintheacceptable
represented only 24% of all observed zones (Table 7). category changing little each time. Similar skewed re-
Seven residual zones possessed five or more stations; sults were observed when using thestrongerwindfield
of these six were negative and in all cases the residual withthelow-sloperemoval.Themostevendistribution
zones were situated largely in the northern part of the ofresidualswasobtainedusingthelow-slopeinversion
archipelago. Several persistent features were noted in removal with a variable sea surface temperature; how-
the residuals plots. Large and small zones of positive ever, itmustbenoted thatthismethodalsoyieldedone
residual were frequent in the north, on Ellesmere and of the lowest numbers of residuals in the acceptable
3608 JOURNAL OF CLIMATE VOLUME15
FIG.6.Modelresultsfor19Jul–1Aug1977.Resultsareestimatesofmeansurfaceairtemperature((cid:3)C),integrating
the specified time period, arranged on a 1 km (cid:1) 1 km grid. Locations of all verification sites havebeen plotted as
blackdots.
TABLE6.Surfaceairtemperature((cid:3)C)meanandstddev,andsamplesizeofobservedvalues,modelestimates,andresidualsforeach
yearof‘‘original’’modelrunandaveragedforallyears.
Observed Estimate Residual
Mean Stddev Mean Stddev Mean Stddev N
1974 5.5 2.6 7.8 1.8 (cid:5)2.3 2.5 26
1975 3.4 1.9 4.1 1.3 (cid:5)0.6 1.4 27
1976 2.9 1.8 2.9 1.3 0 1.1 23
1977 5.9 2 6 1.6 0 2 27
1978 4.9 1.7 5.4 1.8 (cid:5)0.5 2.3 23
1979 4.2 1.8 5 1.5 (cid:5)0.8 1.5 26
1980 4.7 1.3 5.3 1.2 (cid:5)0.6 1 24
1981 6.4 2.2 7.4 1.6 (cid:5)1 2.3 24
1982 6.8 1.6 7.3 1.4 (cid:5)0.4 1.8 21
1983 3.5 1.9 2.9 1.8 0.6 1.3 23
1984 5 2.4 4.5 2.1 0.5 1.5 33
1985 5.9 2 7.4 2.2 (cid:5)1.5 1.8 24
1986 4.3 2.4 4.8 2.2 (cid:5)0.4 1.2 26
1987 6.2 2.6 5.8 2.4 0.4 1.5 14
1988 8.3 2.2 9.6 2.4 (cid:5)1.3 2.8 23
1990 6.3 2.1 5.7 2.6 0.6 1.7 22
Overall 5.21 2.4 5.69 2.5 (cid:5)0.5 2 386
15DECEMBER2002 ATKINSON AND GAJEWSKI 3609
FIG. 7. (a) Residual values for all model runs (n (cid:9) 386) plotted against distance from coast of the verification
station.Distanceaxisisinkmplottedonalogscale.(b)Residualvaluesforallmodelruns(n(cid:9)386)plottedagainst
elevationoftheverificationstation.Elevationaxisisinmetersplottedonalogscale.
range.Thelargestnumberofresidualsintheacceptable Altering the inversion removal to reduce the mag-
range was obtained using no inversion removal; how- nitudeoflapseratecorrectionresultedinthelargeneg-
ever,italsogeneratedthemosthighlyskewedresiduals ativeresidualzonesbeingreducedinsizeand/orbroken
set. An examination of the residuals of the sensitivity up into smaller zones (e.g., Figs. 8a–h). Changesintro-
analyses shows how the different alterations affected duced by alteration of inversion removal were more
specific years (Table 9). significant thatthoseresultingfromalteringtheSSTor
3610 JOURNAL OF CLIMATE VOLUME15
TABLE7.Frequencyofoccurrenceofresidualzonespossessingcertainnumberofstations,byresidualtype.
Overall Negativeresidual Positiveresidual
No.ofstations No.ofzonesof No.ofzonesof No.ofzonesof
inzone thissizeinplots % thissizeinplots % thissizeinplots %
1 34 60 19 56 15 65
2 9 16 3 9 6 27
3 4 7 4 11 — —
4 3 5 2 6 1 4
5 3 5 2 6 1 4
6 — — — — — —
7 3 5 3 9 — —
14 1 2 1 3 — —
Total 57 34 23
wind fields, which tended to result more in sporadic, signed for the interiors of large islands, thus overesti-
low-magnitude changes. mated at these locations. This can be remedied by al-
lowing the model to take into account the size of the
land area a given location is situated in, and adjusting
4. Discussion
the lapse rate accordingly. Finally, too few wind direc-
Areaswellrepresentedintheoriginalmodelincluded tionoptionsinthewindfilterwasanothercauseofwind-
the central, south-central, and west-southwest regions. relatedmodeloverestimationatcoastallocations.Using
When the model was in error it tended to overestimate eight, instead of four, wind directions would improve
temperatures. In several years large zones of model this. Another problem that may contribute to a model
overestimation were observed in the northwest and overestimation in the ‘‘original’’ model runs in the
along the eastern edge of the archipelago. Modelover- northwest isthevalueusedfortheairtemperatureover
estimation occurred in the presence of unusually deep the ocean. That is, 2(cid:3)C is too high for an ocean that is
and widespread inversions or when a shallow, surface usually ice covered. This was altered for some of the
inversion has undergone a slight surface warming.The sensitivityrunsinwhichavariableairtemperatureover
largezonesofsystematicoverestimation(zonesofseven the ocean was used and found to be an improvement.
ormorestations)accountedfor27%ofallresidualval- Residuals along the eastern edge of the archipelago
ues and were confined to four specific years in the 17- mostlikelyoccurredbecausenoneoftheupper-airpro-
yr run period. When these four years were excluded, filesarecharacteristicofthisregion.Alert,whileonthe
theresidualsshowednoparticulartendencytowardpos- coast near the eastern coastal region, is located at the
itive or negative skewing. extreme northern limit of this area, which limits the
In several cases the model overestimated when the representativenessofitsprofile.Furthermore,thenature
resultantwindwaszero.Thisoccurredbecause,without of the interpolation procedure is such that, for muchof
an onshore wind component, the model did not apply the central-east coast of EllesmereIsland,theinfluence
any cooling to coastal areas. This can be remedied by exerted by Eureka’s vertical profile, a station poorly
allowing somecoolingforareasclosetothecoasteven suited to guide estimates in this area, will exceed that
during conditions of zero wind. A related problem is of Alert. Incorporation of vertical profiles from Thule
that the model overestimated temperatures on small is- AirForceBaseinwesternGreenlandcouldimprovethis
lands,suchasPrinceLeopoldIslandorSeymourIsland. situation.
Inthesecasestheradiativeheatingcapacityofthesmall Overestimation occurred more frequently than un-
landareaoftheislandisinsufficienttomodifythecool derestimation,andthemagnitudeofmostofthepositive
lowest levels of the atmosphere. Application of a cor- residuals was small. That fact, coupled with the spatial
rected vertical temperature profile, which has been de- distribution of residuals, did not indicate systematic
TABLE8.Residualtotalsbymodelfactorparameterizationset.
Factorparameterizationset
Maxwind, Maxwind,
VariableSST,VariableSST, constSST, variableSST,
Residual Original Lowslope Peakpoint None lowslope peakslope lowslope lowslope
Lessthan(cid:5)1.4 33 31 18 15 31 16 33 33
(cid:5)1.4to(cid:6)1.4 76 87 89 92 80 86 83 85
Greaterthan(cid:6)1.4 19 19 30 30 26 35 21 19
Totalobs 128 137 137 137 137 137 137 137
Description:Introduction. Interactions of the tant, as future climate changes are predicted to be greater data or information into an analysis (Atkinson et al. 2000