Zhang Lu, Li Bing Zhang, Guo Ke Ji, Liu Feng, Zong Ga, Li Xin Yu, Lyu Yong Lei, Ouyang Zhi Yun
State Key Laboratory of Urban and Regional Ecology, Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, Beijing 100085, China.
Forest Inventory and Planning Institute of Tibet Autonomous Region, Lhasa 850000, China.
Ying Yong Sheng Tai Xue Bao. 2019 Aug;30(8):2793-2802. doi: 10.13287/j.1001-9332.201908.035.
This study assessed the dynamics of lakes in the north Tanggula Mountains in Tibet (NT area) and predicted the spatial changes. We used object-oriented classification and a spectral-angle vector change detection method to generate ecosystem distribution data for the NT area at five-year interval between 2000 and 2015. Based on this dataset, we measured the spatial pattern of lake dynamics and related geophysical and meteorological factors. A boost regression tree model was used to examine the contribution of these variables to lake area change, which were further used in a GEOMOD model to predict the distribution of lakes in 2030. The results showed that the area of lakes increased by 14.2% between 2000 and 2015, which was one of the main forms of ecosystem change in the NT area. In the study region, the area of 10 lakes out of the total 15 lakes with an area greater than 10 km increased and that of another five lakes decreased with a relatively low shrinkage ratio. Both the mean site area and site number of lakes increased, and the percentage of large lake sites increased slightly. Lakes located in areas with high altitude, high slope, low tempe-rature, low precipitation, or near glaciers had a higher probability of size increase. Locales near the current lake sites with low temperature, precipitation, and slope tended to become new parts of extant lakes. Following the trend of the last 15 years, lakes in the NT area would increase by 119 km by 2030, with the main form of lake expansion changing from inundation of area around large lakes to area around smaller sites.
本研究评估了西藏唐古拉山北部地区（NT地区）湖泊的动态变化，并预测了其空间变化。我们采用面向对象分类法和光谱角向量变化检测方法，生成了2000年至2015年期间NT地区每隔五年的生态系统分布数据。基于该数据集，我们测量了湖泊动态变化的空间格局以及相关的地球物理和气象因素。使用提升回归树模型来检验这些变量对湖泊面积变化的贡献，并进一步将其用于GEOMOD模型中，以预测2030年湖泊的分布情况。结果表明，2000年至2015年期间，湖泊面积增加了14.2%，这是NT地区生态系统变化的主要形式之一。在研究区域内，面积大于10平方千米的15个湖泊中，有10个湖泊面积增加，另外5个湖泊面积减少，且萎缩比例相对较低。湖泊的平均面积和数量均有所增加，大型湖泊的比例略有上升。位于高海拔、高坡度、低温、低降水或靠近冰川地区的湖泊面积增加的可能性更高。当前湖泊周边温度低、降水少且坡度小的区域，往往会成为现有湖泊的新组成部分。按照过去15年的趋势，到2030年NT地区的湖泊面积将增加119平方千米，湖泊扩张的主要形式将从大湖周边区域被淹没转变为小湖泊周边区域被淹没。
