Dp. Turner et al., COMPARISON OF ALTERNATIVE SPATIAL RESOLUTIONS IN THE APPLICATION OF ASPATIALLY DISTRIBUTED BIOGEOCHEMICAL MODEL OVER COMPLEX TERRAIN, Ecological modelling, 90(1), 1996, pp. 53-67
Spatially distributed biogeochemical models may be applied over grids
at a range of spatial resolutions, however, evaluation of potential er
rors and loss of information at relatively coarse resolutions is rare,
In this study, a georeferenced database at the l-km spatial resolutio
n was developed to initialize and drive a process-based model (Forest-
BGC) of water and carbon balance over a gridded 54976 km(2) area cover
ing two river basins in mountainous western Oregon. Corresponding data
sets were also prepared at 10-km and 50-km spatial resolutions using
commonly employed aggregation schemes. Estimates were made at each gri
d cell for climate variables including daily solar radiation, air temp
erature, humidity, and precipitation, The topographic structure, water
holding capacity, vegetation type and leaf area index were likewise e
stimated for initial conditions. The daily time series for the climati
c drivers was developed from interpolations of meteorological station
data for the water year 1990 (1 October 1989-30 September 1990), Model
outputs at the l-km resolution showed good agreement with observed pa
tterns in runoff and productivity. The ranges for model inputs at the
10-km and 50-km resolutions tended to contract because of the smoothed
topography. Estimates for mean evapotranspiration and runoff were rel
atively insensitive to changing the spatial resolution of the grid whe
reas estimates of mean annual net primary production varied by 11%, Th
e designation of a vegetation type and leaf area at the 50-km resoluti
on often subsumed significant heterogeneity in vegetation, and this fa
ctor accounted for much of the difference in the mean values for the c
arbon flux variables. Although area-wide means for model outputs were
generally similar across resolutions, difference maps often revealed l
arge areas of disagreement. Relatively high spatial resolution analyse
s of biogeochemical cycling are desirable from several perspectives an
d may be particularly important in the study of the potential impacts
of climate change.