1982-2020年中国森林凋落物碳密度数据集

基于大量站点观测的森林凋落物碳密度数据，利用长时期遥感观测 NDVI 数据驱动构建的随机森林模型，模拟评估1982-2020 年中国森林凋落物碳密度的时空变化特征，得到该数据集。凋落物碳密度模拟结果显著高正相关于观测值（r=0.65，p<0.001，n=4882），其误差百分率较小（0.96%），表明随机森林模型模拟精度较高。1982-2020 年凋落物碳密度的时空变化特征分析表明，中国森林凋落物碳密度整体呈显著的增加趋势（r=0.81，p<0.001），其中常绿针叶林、落叶阔叶林和落叶针叶林的凋落物碳密度增加显著。该数据集是论文《1982-2020 NDVI 数据驱动随机森林模型模拟中国森林凋落物碳密度的研究》（2021年11月发表于《遥感技术与应用》）的支撑数据。

This dataset was generated by developing a random forest model driven by long-term remotely sensed Normalized Difference Vegetation Index (NDVI) observations, based on a large volume of field-measured forest litter carbon density data, to simulate and evaluate the spatio-temporal variation characteristics of forest litter carbon density in China between 1982 and 2020. The simulated litter carbon density results exhibited a significantly strong positive correlation with in-situ observed values (r=0.65, p<0.001, n=4882), with a low relative error percentage of 0.96%, demonstrating that the random forest model had high simulation accuracy. Analysis of the spatio-temporal variation of litter carbon density from 1982 to 2020 showed that the overall forest litter carbon density in China displayed a significant increasing trend (r=0.81, p<0.001), with notably increased litter carbon densities in evergreen coniferous forests, deciduous broad-leaved forests, and deciduous coniferous forests. This dataset is the supporting data for the paper titled *Study on Simulating Forest Litter Carbon Density in China Using Random Forest Model Driven by NDVI Data from 1982 to 2020*, which was published in *Remote Sensing Technology and Application* in November 2021.

搜集汇总

背景与挑战

背景概述

该数据集基于大量站点观测和遥感NDVI数据，通过随机森林模型模拟了1982-2020年中国森林凋落物碳密度的时空变化，模拟精度较高（r=0.65，p<0.001）。结果显示，在此期间中国森林凋落物碳密度整体呈显著增加趋势，尤其在常绿针叶林、落叶阔叶林和落叶针叶林中增加明显，数据格式为栅格图像，存储量7.24MB，适用于地球科学、生态学和林学等领域的研究。