首页 | 本学科首页   官方微博 | 高级检索  
     检索      

基于ANN-CA的银川平原湿地景观演化驱动力情景模拟分析
引用本文:张美美,张荣群,郝晋珉,艾东.基于ANN-CA的银川平原湿地景观演化驱动力情景模拟分析[J].地球信息科学,2014(3):418-425.
作者姓名:张美美  张荣群  郝晋珉  艾东
作者单位:[1]中国科学院遥感与数字地球研究所,北京100094 [2]中国农业大学信息与电气工程学院,北京100083 [3]中国农业大学资源与环境学院,北京100193
基金项目:国家自然科学基金项目(41271419).
摘    要:本文通过对湿地景观的时空动态发展过程其形成空间格局的分析,构建了基于ANN-CA的银川平原湿地景观时空模拟模型,并对湿地景观格局过程与主要驱动力因子间的响应关系进行了情景模拟。研究结果表明:年降水量对天然湿地中的河流湿地和湖泊湿地的驱动作用呈正相关关系,对水稻田和坑塘湿地的影响不显著;人口密度对人工湿地的驱动作用呈正相关,随着人口密度的增加,水稻田和坑塘向各个方向大面积蔓延,河流和湖泊等天然湿地的面积则逐渐减少;随着农业生产活动的加强、农业总产值的增加,河流和湖泊缓慢减少,水稻田和坑塘等人工湿地分布迅速扩张。

关 键 词:元胞自动机  神经网络  湿地  驱动力  情景模拟  银川平原

The Scenarios Simulation Analysis of Driving Forces of Wetland Landscape Evolution Using ANN-CA in Yinchuan Plain
ZHANG Meimei,ZHANG Rongqun,HAO Jinmin,and AI Dong.The Scenarios Simulation Analysis of Driving Forces of Wetland Landscape Evolution Using ANN-CA in Yinchuan Plain[J].Geo-information Science,2014(3):418-425.
Authors:ZHANG Meimei  ZHANG Rongqun  HAO Jinmin  and AI Dong
Institution:3 ( 1. Institute of Remote Sensing and Digital Earth, Beijing 100094, China; 2. College oflnformation and Electrical Engineering China Agricultural University, Beijing 100083, China; 3. College of Resources and Environmental Sciences, China Agricultural University, Beijing 100193, China )
Abstract:Wetland landscape spatio-temporal dynamic development process is more important than the ultimate form of its spatial pattern. Only clearly understand wetland dynamic development process, the theory and deci- sion support of wetland resources protection and sustainable utilization can be provided. In this paper, Yinchuan Plain wetland landscape evolution driving force analysis model was established, full considering the causal rela- tionship between the geographical phenomena in space and time. The transform rules of cellular automata (CA) were built with the model of artificial neural network (ANN), which reduced the man-made subjective factors, and improved the accuracy. Comparing the prediction results with actual wetland types, it concludes that the pre- diction accuracy reaches about 84.24%. Three driving force factors as annual precipitation, population density and agriculture gross output value were selected for the scenarios simulation of wetland landscape pattern. The scenarios simulation results show that, average annual rainfall has more significant driving force to natural wet- land, in the process of reduced by 10% to increased by 10%, the area of river and lake wetlands continues to in- crease, with river wetland increased 26.3844 km2 and lake wetland 22.4100km2. Rice paddies and ponds main- tain a steady growth. Population density has more significant driving force to artificial wetland. With the growth rate of population density changing from 8 %o to 18.7 %0, rice paddies and ponds expanded greatly, i.e. 19.4364 km2 and 18.2088 km2, respectively. But the area of natural wetlands (river and lake wetlands) decreased gradual- ly, and the construction land increased markedly. Total agricultural output also has more significant driving force to artificial wetlands, but slow reverse inhibition force to natural wetlands. When the growth rate of total agricul- tural output changes from 4.5% to 6.5%, artificial wetlands such as rice paddies and ponds expand rapidly, in- creasing 21.5604 km2 and 19.1880 km2, respectively; river and lake wetlands decrease slowly; and the construc- tion land and the Yellow River washland remain basically unchanged.
Keywords:cellular automata  neural network  wetland  driving force  scenarios simulation  Yinchuan Plain
本文献已被 CNKI 维普 等数据库收录!
点击此处可从《地球信息科学》浏览原始摘要信息
点击此处可从《地球信息科学》下载免费的PDF全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号