基于Landsat和MODIS数据融合的农牧区NPP模拟

NPP Simulation of Agricultural and Pastoral Areas Based on Landsat and MODIS Data Fusion

天山北坡是中国重要的农牧业发展基地,利用遥感数据准确获取植被净初级生产力(Net primary productivity,NPP)的时空信息,对于合理分配农牧业草地资源具有重要意义。由于受到天气影响及卫星传感器受到时间分辨率和空间分辨率的限制,获取既具有中空间分辨率、又具有高时间分辨率的遥感数据比较困难。本文基于中空间分辨率Landsat 8 OLI数据与高时间分辨率MODIS数据,采用遥感数据时空融合STARFM算法,获取中空间分辨率和高时间分辨率序列的遥感数据,以天山北坡中段区域为实验区,结合CASA模型,对区域内植被NPP进行模拟。结果表明,2016年内8个时期,融合后的NDVI数据与对应时刻的Landsat 8 OLI NDVI数据的相关系数不小于0.759,偏差在0.0062~0.0094之间,均方根误差在0.074~0.135之间;利用融合数据与CASA模型协同模拟的NPP具有良好的空间细节信息,NPP模拟值与野外实测值决定系数R^2为0.8601,表明两者具有较好的相关性。本研究为多源遥感影像融合技术与光能利用率模型协同模拟NPP提供了新的思路。

Northern Tianshan of China is an important development base of agricultural and animal husbandry.The space-time information of net primary productivity(NPP)was accurately obtained based on remote sensing data,the grassland resources of agricultural and pastoral areas can be rationally allocated.It has important and actual significance to the development of northern Tianshan.Due to the influence of weather and the limitation of time and space resolution of satellite sensors,it is difficult to obtain remote sensing data with medium space resolution and high time resolution series.Based on the middle spatial resolution Landsat 8 OLI data and the high time resolution series MODIS data,the middle spatial resolution and high time resolution series remote sensing data were obtained by using STARFM algorithm for space-time fusion of remote sensing data.Then,the vegetation NPP in the middle section of the northern Tianshan was simulated through the CASA model.The results showed that the relationship coefficient r of the fused NDVI data and Landsat 8 OLI NDVI data was not more than 0.759,the Bias was between 0.0062 and 0.0094 and RMSE was between 0.074 and 0.135 in eight periods in 2016.There were good space detail information of the NPP simulated of the fusion data and CASA model.The R^2 of NPP simulation value and field measurement value was 0.8601.The research results can provide method for the collaborative simulation of NPP by multi-source remote sensing image fusion technology and light utilization model.