首页|ANN-FLUS模型在历史耕地分布变化重建上的应用——以1885-2000年的越南为例

ANN-FLUS模型在历史耕地分布变化重建上的应用——以1885-2000年的越南为例

扫码查看
历史土地利用及其覆被的变化对于理解不同时期的人地关系具有重要指导意义。传统的耕地开垦优先度模型未考虑土地利用转换规则及邻域效应,重建精度不高。基于多源史料提取越南1885-2000年5个关键年份耕地总量及其9个驱动因素,并应用有大数据处理和土地转换规律挖掘优势的ANN-FLUS模型探讨耕地分布变化。结果表明:(1)耕地总量由1532千hm2增至6200千hm2,其变化以1940年代为转折阶段而表现出前快、后慢的差异;(2)耕地历来集中在红河、湄公河三角洲以及中部沿海平原,其在全国尺度上开发模式由"聚点式"缓慢延伸逐渐转变为"散点"式迅速扩张;(3)ANN-FLUS模型在越南1950年耕地分布格局上的重建精度为Kappa系数0。65,显著优于开垦优先度模型(0。54)及HYDE 3。2数据集模型(0。50)。
The application of ANN-FLUS model in reconstructing historical cropland distribution changes:A case study of Vietnam from 1885 to 2000
The historical changes in land use and land cover hold significant value for under-standing the dynamics of human-environment interactions across various eras.Traditional land reclamation priority models have often overlooked the principles of land use conversion and the neighborhood effects,resulting in inaccuracies in the reconstruction process.This study utilizes a diverse range of historical documents to quantify the total cropland area and nine influential factors over five pivotal years,spanning from 1885 to 2000 in Vietnam.Employing the ANN-FLUS model,known for its prowess in huge-volume data processing and extraction of land con-version rules,the analysis delves into cropland distribution changes.The findings revealed that:(1)The cropland area experienced a growth from 1,5 3 2x103 hectares to 6,200x 103 hectares,characterized by an initial rapid expansion and a subsequent deceleration after the 1940s.(2)Historically,the cropland has predominantly been distributed in the Red River Delta,Me-kong River Delta and the Central Coastal Plain.Nationally,the pattern of cropland reclamation has transitioned from a concentrated,gradual spread to a more dispersed and swifter enlarge-ment.(3)The ANN-FLUS model demonstrates its efficacy in the reconstruction of cropland dis-tribution in Vietnam for the year 1950,achieving a Kappa coefficient of 0.65,a notable im-provement over the reclamation priority model's 0.54 and the HYDE 3.2 dataset model's 0.50.

historical croplanddistribution changemachine learningANN-FLUS modelre-constructionVietnam

刘娴、刘浩龙

展开 >

中国科学院地理科学与资源研究所,陆地表层格局与模拟重点实验室,北京 100101

中国科学院大学,北京 100049

历史耕地 分布变化 机器学习 ANN-FLUS模型 重建 越南

中国科学院战略性先导科技专项国家重点研发计划全球变化重点专项

XDA23100403122017YFA0603304

2024

自然资源学报
中国自然资源学会 中国科学院地理科学与资源研究所

自然资源学报

CSTPCDCSSCICHSSCD北大核心
影响因子:2.073
ISSN:1000-3037
年,卷(期):2024.39(6)
  • 71