基于自组织特征映射神经网络算法的生态服务功能分区(英文)

A Self-organization Mapping Neural Network Algorithm and Its Application to Identify Ecosystem Service Zones

探索了一种基于自组织特征映射神经网络算法识别区域尺度生态系统服务功能分区的新方法。在此基础上,依据新千年生态系统评估框架构建了生态服务功能评价指标体系,并运用自组织特征映射神经网络算法开展了生态服务功能空间聚类分析,在1km栅格上识别并排定了各类生态服务功能的重要性。在案例区锡林郭勒盟的研究表明,利用基于自组织特征映射神经网络算法划分出的该区6个生态服务功能分区比较科学、合理,所形成的分区结论为案例区生态系统的可持续管理提供有时空针对性的决策参考信息。

The self-organization mapping （SOM） neural network algorithm is a new method used to identify the ecosystem service zones at regional extent. According to the ecosystem assessment framework of Millennium Ecosystem Assessment （ MA）, this paper develops an indicator system and conducts a spatial cluster analysis at the 1km by I km grid pixel scale with the SOM neural network algorithm to sort the core ecosystem services over the vertical and horizontal dimensions. A case study was carried out in Xilingol League. The ecosystem services in Xilingol League could be divided to six different ecological zones. The SOM neural network algorithm was capable of identifying the similarities among the input data automatically. The research provides both spatially and temporally valuable information targeted sustainable ecosystem management for decision-makers.