提出了用于区域土地利用变化预测改进型的空间分析预测模型,以半干旱环境下典型生态脆弱区松嫩平原西部通榆县为例,用改进型空间分析测算模型分析了全县1992~2001年间土地盐碱化演化过程,并在此基础上预测全县未来29 a内土地盐碱化发展趋势。结果表明:(1)改进型空间分析测算模型可较准确预测区域未来某一时间点盐碱地数量与结构;(2)通榆县2030年以前将以非盐碱地快速盐碱化为主,其中低覆盖度草地和沼泽地是新增盐碱地主要来源;(3)随着气候变化和人类干预,全县土地盐碱化过程有一定可逆性,以轻、中度盐碱地向耕地、高覆盖草地、农村居民用地和城镇用地转化为主。

The application of traditional spatial analysis model of land use is greatly restricted due to the complex logical structure itself and the lack of rules for restricting the maximum amount of each type of land use in future land use scenario. In this study, we brought forward an improved spatial analysis model for predicting the local land use change. Since land saline-alkailization is becoming one of the most serious geological-environmental disasters in the world and threatening the sustainable development in ecology, economy and society, thus, we selected Tongyu County, a typical zone of the western Songnen Plain in semi-arid areas as study area to test the reliability of the impraved spatial analysis model of land use. We firstly evaluated the performance of the improved spatial analysis model of land use in predicting the change of saline-alkaili land through comparing the statistic area of saline-alkaili land in 2001 from remote sensing image with those from improved spatial analysis model of land use based on the characteristic of land use change during 1979 to 1992. Then, the saline-alkaili land in Tongyu County in 2030 was predicted using the improved spatial analysis model of land use based on the characteristic of land use change during 1992 to 2001. The results revealed that the improved spatial analysis model of land use is capable of predicting the composition of future land use through combining the changed and unchanged parts of land use. The results also showed that the degree of land saline-alkailization was becoming more serious and saline-alkailization was emerging in the major types of land uses (e.g. grassland, swampland). Until the year of 2030, land saline-alkailization will continuously speed up through the transition of the lowly covered grassland and swampland to saline-alkaili land. However, it was also observed that, with the change of climate and human impact, the saline-alkaili land could be transferred to non-salin-alkaili land in Tongyu County in the next coming 29 years. In this process, the low- and median- level saline-alkaili lands would be transferred to highly covered grassland, swampland and residence land, while only a few of, high- level saline-alkaili land would be transferred to residence land. In conclusion, the improved spatial analysis model of land use proposed in this study expands the applications of traditional spatial analysis model of land use and makes it possible for predicting the local land use change based on characteristic of previous land use change. Human would have to fight for survival and development with saline-alkaili land clue to the increasing confliction between human and natural environment in the arid and semi-arid areas in the near future.