高时空分辨率林地叶面积指数遥感快速估测方法及在碳循环模型中的应用

Time-series LAI is an important driving data in forest ecosystem carbon cycle simulation. According to the low spatial resolution of current time-series LAI products based on remote sensing, leading to large error of carbon cycle simulation at small and medium regional scale, a fast estimation method for forest leaf area index with high spatial and temporal resolution based on remote sensing was proposed. Under similar ecological environment, forest LAI in the same forest type and the similar age shows similar growth trends. Based on the principles mentioned above, the relationship was constructed between the pixel LAI value in time-series LAI images and pixel LAI values in the high spatial resolution LAI image. In this study, the normalized LAI growth curves of main forest types in North China were extracted from MODIS LAI products with higher temporal resolution, which eliminated the effect of stand density. Meanwhile high spatial resolution LAI band was estimated using images of GF series satellites and the traditional remote sensing statistical models. Then the normalized LAI growth curves were used to extend the high spatial resolution LAI in the whole growth cycle, to establish high spatial and temporal resolution LAI data sets of the main forest types in North China, which provided basic data with high precision for the further study of forest ecosystem carbon cycle simulation in North China.

时间序列LAI是森林生态系统碳循环模拟中重要的驱动数据，针对当前时间序列LAI 遥感产品空间分辨率较低导致中小区域尺度碳循环模拟误差较大的问题，该研究提出了一种高时空分辨率林地叶面积指数遥感快速估测方法。在相似的生态环境下，相同林分类型且林龄相近的植被LAI具有相似的生长趋势，基于此原理建立时间序列LAI像元值与高空间分辨率对应像元LAI值的相互关系。本研究首先利用高时间分辨率的MODIS LAI产品提取华北地区主要林分类型的归一化LAI生长曲线，消除了林分密度效应，同时基于高空间分辨率的国产GF系列卫星数据和传统遥感统计模型，建立高空间分辨率LAI，然后利用归一化LAI生长曲线将高空间分辨率的LAI数据扩展到全生长周期，建立华北地区主要林分类型高时空分辨率LAI数据集，为华北地区森林生态系统碳循环模拟的深入研究提供高精度基础数据支持。

叶面积指数（leaf area index，LAI）是植被冠层最重要的结构性参数之一，时间序列LAI 表征了植被的生长动态，是碳循环、水循环等生态过程模型最重要的驱动数据之一，快速精确的获取高时空分辨率的LAI 数据对降低碳循环等模型模拟的不确定性至关重要。主要研究进展与研究结果如下：（1）提取了不同龄组的华北落叶松LAI生长曲线，归一化后的生长曲线消除了林分密度效应及枝干的遮光效应，更能代表植被叶子的生长；（2）基于高空间分辨率的哨兵及国产GF系列卫星数据，提出一种新的高时空分辨率LAI 快速估测算法，建立了塞罕坝机械林场主要林分类型中小区域尺度的逐日LAI 数据集；（3）构建了华北落叶松与白桦针阔混交林树高曲线及树高与胸径关系的预测模型，基于主要影响因子和包含哑变量的非线性混合效应树高与胸径关系预测模型，能够有效解决混交林中树种及主要因子对树高与胸径关系的影响，提高了预测模型的适用性及预测精度，修正了碳循环模型混交林生物量估测方程；（4）在生长季不同时间对华北落叶松进行了光合观测，模拟落叶松最大羧化速率随着时间变化方程，完善了原来光合作用模型中最大羧化速率不随季节变化的缺陷。本研究为探究华北地区森林生态系统碳收支、碳源/汇分布奠定基础。