1961-2010年内蒙古草原植被分布和生产力变化——基于MaxEnt模型和综合模型的模拟分析

定量评估气候变化对内蒙古草原植被分布及其净第一性生产力的影响有助于理解干旱区域生态系统结构和功能对气候变化的响应。基于最大熵模型（MaxEnt）评价了气候因子的重要性，进而模拟了1961-2010年内蒙古草原植被的地理分布，同时应用综合模型模拟了净第一性生产力变化。研究表明，湿润指数（MI）、年降水量（P）、最暖月平均温度（T_w）和最冷月平均温度（T_c）是决定草原植被分布的主导气候因子。1961-2010年内蒙古草甸草原、典型草原和荒漠草原分布面积分别减少了5%、1%和62%，草原面积整体减少了11%，预示着草原向着荒漠化的方向发展。降水是决定内蒙古草原净第一性生产力变化的最重要因素。

Changes in distribution and productivity of steppe vegetation in Inner Mongolia during 1961 to 2010: Analysis based on MaxEnt model and synthetic model

Global temperature and precipitation are undergoing unprecedented change, which has seriously affected the structure and function of steppe ecosystems in Inner Mongolia. The vegetation of this region has been suffering from severe salinization, desertification, and decline in productivity, resulting in severely ecological problems and economic losses. Although numerous studies indicate that desertification is intensifying currently, but the history of vegetation cover and geographical distribution patterns of vegetation are poorly understood. This research attempts to simulate the temporal and spatial changes in the geographical distribution of steppe vegetation based on MaxEnt and changes of net primary productivity (NPP) based on synthetic model in Inner Mongolia during 1961 to 2010. The aim of this research is to provide a valuable method for predicting the potential impact of climate change on vegetation in the future. The results show that moisture index (MI), annual precipitation (P), mean temperature of the warmest month (T_w), and mean temperature of the coldest month (T_c) are the dominant climatic factors determining the geographical distribution of steppes. The MaxEnt model performs well in simulating the geographical distribution of steppe vegetation in Inner Mongolia. The AUC of meadow steppe, typical steppe, and desert steppe are about 0.90 (excellent), 0.85 (very good), and 0.95 (excellent), respectively. The kappa consistencies of meadow steppe, typical steppe, and desert steppe are about 0.95 (excellent), 0.70 (very good), and 0.57 (very good), respectively. The overall kappa value of the entire steppe vegetation is about 0.76 (very good). The areas characterized by steppe vegetation in Inner Mongolia have shrunk during 1961 to 2010. The distribution area decreased to about 5% for meadow steppe, 1% for typical steppe, 62% for desert steppe, and 11% for the entire steppe. The gravity centers of both meadow steppe and typical steppe have moved towards southwest, while the desert steppe moved towards southwest on the initial stage and to the northeast later. Due to these shifts the entire range of steppe are confined within narrow distribution boundary. The area converted from meadow steppe to typical steppe (2.38 × 10⁴ km²) is less than the area of transformation of the latter to the former (3.57 × 10⁴ km²). The area transformed from desert steppe to typical steppe accounts for 15% of the total area of desert steppe. The change in areas shows that the Inner Mongolia steppe is undergoing desertification. The NPP of the entire steppe, meadow steppe, typical steppe, and desert steppe were 312 gDWm^-2a^-1, 341 gDWm^-2a^-1, 309 gDWm^-2a^-1, and 260 gDWm^-2a^-1 in Inner Mongolia during 1961 to 2010. The change in NPP of the entire steppe followed a decline-increase-decline trend during 1961 to 2010. Regional precipitation is the primary determining factor for the NPP of steppes in Inner Mongolia. The MaxEnt model is one of the bioclimatic envelope models (BEMs) based on the niche theory. BEMs are often used to solve problems in conservation biology, biogeography, ecology, and taxonomy at species level. We have successfully used the MaxEnt model to simulate the geographical distribution of the different steppe vegetation types. This study extends the range of application of BEMs, which has facilitated the understanding of the functional diversities of ecological systems and advanced classification units (above the individual and community level).