干旱/半干旱草原灌丛化遥感动态监测与模拟研究

Under the background of global climate change and increasing human activities, shrub encroachment in arid and semi-arid grassland has become a worldwide ecological problem and has significant impacts on grassland ecosystem and local economies. Factors affected shrub encroachment have been addressed, such as climate change, fire frequency, graze and feedbacks related to shrub-grass competition. High spatial resolution aerial photos and remote sensing data have been used to extract shrub and monitor its expansion; however, the existing researches still lack effective methods for dynamic monitoring of shrub encroachment at regional scale, which limits the quantitative analysis of the shrub extension mechanisms and the modeling and simulation of the shrub encroachment. This project aims to study the shrub (Caragana) encroachment process and dynamics in the arid and semi-arid grassland located in Xilin Gol, Inner Mongolia, China. The objectives of the project include: (1) exploring the methods to identify shrub within a ‘shrub-grass-bare soil’ mixture and estimate its coverage using multi-scale and multi-temporal remotely sensed data, based on the characteristics of phonological difference and precipitation sensitivity difference between shrub and grass; (2) developing a ‘shrub-soil’ coverage-based indicator to evaluate the states and differentiate the stages of shrub encroachment; (3) quantitatively analyzing the mechanism of the factors such as climate, fire and graze and their effects on shrub encroachment dynamics at various spatiotemporal scales and in the stage transformation; (4) modeling, simulating and forecasting the shrub encroachment in grassland based on vegetation growth model and spatial expansion model. This research can help to reveal the history, status and future of shrub encroachment, understand its process and mechanism, and provide the strategies for control and manage the shrub encroachment in the arid and semi-arid grassland.

在全球变化及人类活动综合影响下，干旱和半干旱草原的灌丛化过程已成为全世界广泛关注的生态问题，对草原生态系统及其经济发展造成重大影响。目前的研究还缺乏利用遥感对大范围草原灌丛化进行有效动态监测的方法和数据，大大限制了对灌丛化机制及其过程建模和模拟的研究。本研究拟选取内蒙古锡林郭勒盟的干旱/半干旱草原为研究区域，以锦鸡儿属灌丛化为研究对象，利用多尺度、多时相遥感数据，根据灌草物候差异以及对降水敏感性差异等特征，探索灌丛化多尺度遥感动态监测的方法，提供灌丛化动态数据；发展基于遥感监测结果的灌丛化指标，以揭示灌丛化状态及用于灌丛化阶段划分；考虑气候、火灾、放牧等影响因素，定量分析草原灌丛化在不同时空尺度下的影响机制，进而对灌丛化过程进行空间建模、模拟和预测，为掌握该区域灌丛化历史和现状、理解灌丛化的过程和机制、并管理和控制草原灌丛化提供理论和方法支撑。

针对干旱/半干旱草原灌丛化问题，目前的研究还缺乏利用遥感对大范围草原灌丛化进行有效监测的方法和数据，大大限制了对灌丛化机制及其过程建模和模拟的研究。本项目以高时空分辨率遥感数据重建、景观及区域尺度的灌木识别与盖度反演算法研究、灌丛化阶段定性划分与制图、灌丛化过程建模与分析为研究主线，在内蒙古锡林郭勒草原为主要研究区开展研究。主要研究内容包括：通过研发面向对象的多尺度分割组合算法，利用高空间分辨率遥感影像进行灌木提取；研究基于时空加权的高时空分辨率NDVI重建方法，利用灌草物候差异，在中等分辨率尺度进行大范围灌木盖度遥感制图；通过构建基于“灌-草-土”覆盖度的灌丛化指标，对灌丛化阶段进行划分和空间制图；针对高分影像历史资料缺乏、中等分辨率遥感反演精度受限的问题，研究通过灌木年龄估计模型进而拟合区域种群结构模型的方法，从而分析灌丛化的影响因素、历史过程和未来趋势。项目成果为掌握该区域灌丛化历史和现状、理解灌丛化的过程和影响因素、并管理和控制草原灌丛化提供理论和方法支撑。