基于数字地形信息的拉萨河流域河宽预测模型研究

河道宽度是重要的水文特征，对水文计算具有重要意义。以拉萨河流域为研究区，依据Google Earth影像，测量了1 000个河段的河宽，计算了相应测段的集水面积、局部坡度、坡度原点矩、河道曲率4项地形要素。使用其中800个河段的数据判别河宽与各地形要素的相关关系，选取与河宽相关性最好的集水面积和坡度原点矩两要素，构建了集水面积模型、坡度原点矩模型及4种河宽因子模型，使用另外200个河段的数据进行模型验证。结果表明：在集水面积小于2 100 km²的拉萨河上游河段，基于集水面积和坡度原点矩的河宽因子模型效果最优，在集水面积大于2 100 km²的拉萨河下游河段，基于集水面积的模型效果最优。因此，预测拉萨河流域河宽时应以2 100 km²集水面积为阈值，在上游使用河宽因子模型，在下游使用集水面积模型，采用该方法预测的上、下游河宽误差均小于20%。

Channel width prediction model based on digital terrain in Lhasa River Basin

Channel width is an important hydrological feature, which is of great significance to the hydrological calculation of the rivers in Tibet Plateau. The channel widths of 1000 rivers in Lhasa River Basin were measured using Google Earth images, and the corresponding catchment area, local slope, origin moment of slope and channel curvature of each channel section was calculated accordingly. The correlation between channel widths and these topographic factors were evaluated based on the data of 800 channel sections. Then channel width models based on catchment area, origin moment of slope and four channel width factors were constructed subsequently. The other 200 channel width data were used to verify the precision of the models. The results show that the channel width factor model based on catchment area and origin moment of slope with the catchment area less than 2 100 km² in the upstream section has the best performance and the model based on catchment area with the catchment area more than 2 100 km² in the downstream has the best performance. Thus, the catchment area threshold of 2100 km² should be applied in the prediction of the river width in Lhasa River Basin, with the application of channel width factor model to the upstream channels, and catchment area model to the downstream channels. The average deviation of the prediction result is less than 20%.