地理科学进展 ›› 2022, Vol. 41 ›› Issue (12): 2342-2355.doi: 10.18306/dlkxjz.2022.12.012

• 研究论文 • 上一篇    下一篇

基于因果网络的黄河流域植被变化归因分析

兰志洋1(), 梁伟1,*(), 傅伯杰2, 吕一河2, 严建武1, 纪秋磊2   

  1. 1.陕西师范大学地理科学与旅游学院,西安 710119
    2.中国科学院生态环境研究中心,北京 100085
  • 收稿日期:2022-03-07 修回日期:2022-08-18 出版日期:2022-12-28 发布日期:2022-12-31
  • 通讯作者: *梁伟(1981— ),男,陕西周至人,博士,教授,研究方向为多尺度生态水文过程遥感观测与模拟、复杂地理系统。E-mail: liangwei@snnu.edu.cn
  • 作者简介:兰志洋(1998— ),男,内蒙古赤峰人,硕士生,研究方向为人地耦合。E-mail: zhiyangl1998@snnu.edu.cn
  • 基金资助:
    国家自然科学基金项目(42071144)

Attribution analysis of vegetation change in the Yellow River Basin based on causal network

LAN Zhiyang1(), LIANG Wei1,*(), FU Bojie2, LV Yihe2, YAN Jianwu1, JI Qiulei2   

  1. 1. School of Geographical Sciences and Tourism, Shaanxi Normal University, Xi'an 710119, China
    2. Research Center for Eco-Environmental Sciences, CAS, Beijing 100085, China
  • Received:2022-03-07 Revised:2022-08-18 Online:2022-12-28 Published:2022-12-31
  • Supported by:
    National Natural Science Foundation of China(42071144)

摘要:

人类与自然耦合系统是一个充满大量非线性互馈关系的复杂巨系统。当前对系统内单要素和静态特征的研究,虽有助于理解某个时刻的系统状态,却不能完整地表达其内部不同要素间错综复杂的关联。论文着眼于黄河流域区县尺度植被变化归因(以叶面积指数LAI作为表征植被的指标),综合利用人类—自然耦合系统多源数据和因果诊断方法,构建黄河流域LAI复杂因果网络。通过复杂网络统计方法对网络节点、结构特征进行分解,揭示影响流域LAI变化的关键要素及其空间分布特征。研究结果表明:① 1990—2018年流域LAI以1.3%/a的速率增长,增长速率在空间上由东南向西北逐渐递减。② 在网络视角下,降水量、气温、饱和水汽压差、农业用地、城镇化率和粮食产量为影响流域LAI变化的关键要素。③ 自然要素(如降水量和气温等)主导了流域内259个区县的LAI变化;社会经济(如城镇化率和粮食产量等)和土地利用类型(如林地和草地等)要素主导了流域内76个区县的LAI变化,主要集中于黄土高原且影响强度远大于自然要素。研究在人类—自然耦合系统框架下构建了黄河流域植被变化多层互馈网络,综合认知流域植被演变的自然和社会驱动力,为理解自然—社会系统复杂的互馈关系提供了一种新的思路和方法。

关键词: 叶面积指数, 因果诊断, 复杂网络, 黄河流域

Abstract:

Coupled human and natural systems are complex and open giant system with plenty of nonlinear mutual feedback relationships. Although current studies regarding single element and static characteristics in a system are helpful for understanding its state at a certain moment, such study cannot fully express the intricate relationship between different elements inside the system. This study focused on attributing vegetation dynamics (leaf area index, LAI, is applied as an indicator to characterize vegetation dynamics) at the district and county scale in the Yellow River Basin based on multi-source datasets and causal diagnostic methods. A basin-wide complex causal network of LAI is constructed. The spatial distribution of dominant factors influencing LAI variations is finally identified through decomposing the nodes and structural characteristics of the constructed complex network. The results show that: 1) Basin-wide LAI increased at an annual rate of 1.3% during 1990-2018. The growth rate decreased gradually from the southeast to the northwest of the basin. 2) From the perspective of the network, precipitation, temperature, saturated water vapor pressure deficit, agricultural land use, urbanization rate, and grain yield are key factors affecting LAI changes in the basin. 3) Natural elements (for example, precipitation and temperature) dominated LAI changes of 259 districts and counties inside the basin; socioeconomic (for example, urbanization rate and grain yield) and land use (for example, forest and grassland use) elements dominated LAI changes of 76 districts and counties, which are mostly concentrated on the Loess Plateau. The influence intensity of socioeconomic and land use factors on LAI variations is much greater than that of natural factors. In this study, we constructed a multilayer mutual feedback network under the coupled human and natural system framework to comprehensively examine vegetation dynamics and their natural and social drivers in the Yellow River Basin. It provides a new idea for understanding complex mutual feedback relationships in coupled natural-social systems.

Key words: leaf area index, causal diagnostics, complex network, Yellow River Basin