集成遥感数据的陆地净初级生产力估算模型研究综述

Overview on estimation models of land net primary productivity integrating remote sensing data

净初级生产力(NPP)是衡量碳循环、指导土地利用、评估生态安全、指示环境变化、反映粮食安全等的重要参量,其估算受模型构建机理和生态系统关键地表参数输入的影响。近年来,随着遥感数据的不断丰富和遥感处理技术的快速发展,集成遥感数据的NPP估算模型相较于仅采用气候、土壤等传统观测数据的非遥感模型,在分析时空异质性等方面的优势日益凸显。本文基于Web of Science和CNKI两大数据库,采用文献统计分析方法,系统回顾NPP研究概况及国内外集成遥感数据的NPP估算模型的近期进展;并将集成遥感数据进行NPP估算的模型分为统计模型、光能利用率模型、过程模型及耦合模型四类;重点阐述了各类遥感估算模型的机理、差异性、适宜性及局限性;最后,在分析NPP遥感估算面临困境和科学挑战的基础上,从机理与影响因素、数据基础、参数反演、时空尺度拓展、软硬件支撑等方面对未来研究进行了展望。

As a general term that describes the accumulation of organic matters within specific temporal and spatial scopes on crop land, woodland, grassland, or other types of lands, land net primary productivity（NPP） is considered an important parameter to measure carbon cycle, guide land use, assess ecological security, reflect environmental changes, and indicate the level of food security. The estimation precision of NPP is significantly influenced by the type of models and input of key surface parameters of ecosystems. In recent years, with the continuous growth of remote sensing data and the rapid development of remote sensing data processing technologies, NPP estimation models based on remote sensing data, as compared to NPP estimation using traditional observation data such as climate and soil data with coarse spatiotemporal resolutions, have become very prominent in analyzing temporal and spatial heterogeneity. Based on the Web of Science and CNKI databases and statistical analysis methods, this study systematically reviewed research on NPP and its estimation models integrating remote sensing data in China and internationally. The commonly used models can be divided into four categories： statistical models, light use efficiency models, process models, and coupling models. We examined the mechanisms, differences, suitability, and limitation of the various kinds of models, Based on an analysis of the difficulties and scientific challenges that face integrating remote sensing data into NPP estimation models, research prospects are put forward with regard to model mechanism, influencing factors, data provision,parameter derivation, expansion of spatiotemporal scales, and hardware and software supports.