The Different Spatial-scaling Effect of DEM Resolution on the Simulation of Temperature Spatial Distribution in Mountainous Area: A Case Study in Xianju County, Zhejiang Province
In mountainous areas, the terrain factors (e.g., elevation, slope, aspect, hillshade) are the main factors that affect the spatial distribution of temperature. Calculating the temperature over rugged terrain is very diffi- cult. In the mountainous regions of complicated landform and great height difference, the observation data of the existing weather stations are far from enough to reflect the spatial distribution of the air temperature. How- ever, in recent twenty years, digital elevation model (DEM) and digital topography analysis of GIS technology provide a better way to research the spatial distribution of temperature in mountainous areas. Previous studies laave shown that the accuracy of topographic factors derived from DEM with different resolutions varies great- ly due to the scaling effect. Accordingly, the simulation of temperature spatial distribution also is influenced by DEM at different spatial scales. In this paper, Xianju County in Zhejiang Province is selected as the study area. The topography-adjusted statistical model is applied to simulate annual temperature with different resolutions based on temperature data from 10 meteorological stations and different resolutions (5 m, 30 m, 90 m, and 900 m) DEM data derived from 1 : 10 000 digital topographic map, Aster GDEM, SRTM 90 m and GTOPO30', re- spectively. It is hypothesized that the temperature map with 5 m resolution is relatively accurate and the errors between the other temperature map with different resolutions and 5 m temperature map are compared and ana- lyzed. Moreover, the temperatures with different grid cell sizes differ significantly with topography in macro and micro scale. The topographic and spatial-scaling effects are analyzed on the temperature simulation in Xianju County. The results show that the spatial distribution of the simulated temperature based on different resolution DEM has great different characteristics. The simulated temperature error increases gradually with the decrease of DEM resolution and the spatial heterogeneity decreases. The terrain factors such as slope gradi- ent and aspect in micro topography scale have significant change with DEM resolution and then have great ef- fect on the spatial distribution of temperature. With the decrease of DEM resolution, terrain is smoothing, the mean value of slope decreases, and lapse rate of temperature decreases. The minimum effect of DEM resolu- tion on temperature simulation is in low elevation (elevation 〈 100 m) area. The error of the simulated tempera- ture is only less than 0.3℃. In the area, the elevation range is from 100 m to 200 m, and the error is about 1.1℃. The temperature error will reach 2.1-2.4℃ in higher elevation (elevation〉200 m). It is found that the coarser the resolution is, the smoother the terrain is. Moreover, with the DEM grid size increasing, variation of aspect and slope cause the decrease in the spatial heterogeneity of the simulated temperature. The maximum temperature difference in different slope gradient and aspect may reach 10-12.5℃, while the minimum differ- ence is only about 1.9-2.6℃.
NETL
NSTL
万方数据
维普
