我国地理可照时数空间分布模拟中尺度效应研究

不同空间分辨率的DEM会对地理可照时数模拟结果产生显著影响。为了分析DEM尺度效应对我国不同区域地理可照时数的影响，选取小兴安岭、太行山区、秦岭、青藏高原弧形山区、横断山区和东南丘陵6个实验区，基于3种空间分辨率的DEM和地理可照时数分布式模型，利用GIS的多层面复合分析功能，模拟了典型月份不同空间尺度的地理可照时数，定量分析了典型地形因子与模拟结果的关系，并探讨了DEM尺度效应的影响。研究结果表明：（1）我国地理可照时数差异明显，1月、7月及其差异主要受剖面曲率、海拔、纬度、地形阴影和地形起伏度等因子的影响；（2）我国地理可照时数与典型地理和地形因子之间极显著相关，复相关系数在0.610 0～0.804 3，正相关因子为海拔高度和水平曲率，负相关因子为坡度、地形起伏度、剖面曲率和地表粗糙度；（3）不同地形因子的空间尺度效应存在差异，平面曲率和剖面曲率均非常敏感，坡度和地形起伏度次之，海拔、地表粗糙度和地形阴影相对较弱；（4）地理可照时数模拟结果随空间分辨率的减小而增大，1月增幅在8.2%～38.9%，7月增幅在5.9%～17.6%，标准差和冬夏季节差异值均随空间分辨率的减小而减小，空间异质性明显降低。

Study on spatial scale effect of DEM resolution on spatial simulation of duration of possible sunshine in China.

The effect of digital elevation model (DEM) with different spatial resolution ratios causes great uncertainty to the numerical simulation of duration of possible sunshine (PSD). To analyze the spatial-scaling effect on PSD in China, 6 typical terrain areas were selected and their spatial distribution of PSD with three spatial resolutions (30 m, 90 m, and 900 m) in January and July were calculated using a distributed statistical model and DEM data derived from Aster GDEM, SRTM and GTOPO30′, respectively. The influence of spatial scale effect of DEM resolution on simulation results were compared and analyzed. The results show that: (1) Due to the effect of terrain factors, the difference of spatial distribution of PSD is significant in China. The average value of PSD in January, July and their difference were mainly affected by some terrain factors, such as profile curvature, elevation, latitude, hill shade, and topographic relief. (2) The correlation between PSD and terrain factors was extremely significant in China. The multiple correlation coefficient varied from 0.610 0 to 0.804 3. There was a positive correlation between PSD and elevation and plan curvature. But PSD had significant negative correlation with slope, topographic relief, profile curvature, and surface roughness， respectively. (3) Different terrain factors derived from DEM with three spatial resolution had a different degree of scale effect. For example, plan curvature and profile curvature were very sensitive to spatial resolution, while slope and topographic relief were secondary. The spatial scale had relatively weak effect on elevation, surface roughness, and hill shade. (4) The mean value of PSD in study areas increased with the decrease of spatial resolution. The increase range of PSD was from 8.2% to 38.9% in January, while from 5.9% to 17.6% in July. However, the standard deviation and difference of PSD in January and July decreased with the decrease of spatial resolution. Moreover, the spatial heterogeneity of PSD decreased obviously.