Evaluation and Validation of the Net Primary Productivity of the Zoige Wetland Based on Grazing Coupled Remote Sensing Process Model

Zoige wetland is located in the southeastern edge of the Qinghai-Tibet Plateau, in recent years, the wetland local area grassland productivity decline, land sanding, and other serious degradation phenomenon. Using remote sensing and hydrology data combined with the basic principles of vegetation ecology, we applied the Carnegie-Ames-Stanford Approach (CASA) model and Boreal Ecosystem Productivity Simulator (BEPS) model to invert the net primary productivity (NPP) of the Zoige wetland ecosystem and then identified the deficiencies of the two models. In this article, we tested the factors that influence the light use efficiency of vegetation and considered the effects of livestock grazing on grassland ecosystems and wetland moisture content on vegetation growth. A new grazing coupled remote sensing process (GCRSP) model suitable for the productivity of the Zoige wetland was proposed. Field survey data were combined with the CASA model, BEPS model, and Moderate Resolution Imaging Spectroradiometer (MODIS) NPP product, and the accuracy of the proposed GCRSP model was analyzed at regional level. The results show that after fitting the field measurements to the GCRSP model, the fitted variance between the model and the measured data is 0.84, which is larger than the variance between the other two methods and the measured data. In addition, the variance of the linear fit between the GCRSP model and the MODIS NPP product was 0.11, which was relatively small, indicating that the regional-scale GCRSP model was relatively accurate for the Zoige wetland NPP. Therefore, this model was suitable for the inversion of regional-scale NPP. The results indicated that the average grassland NPP of the Zoige wetland was 487.57 gC/m(2), and larger NPP values were distributed near the rivers and wetland and smaller NPP values were distributed in the mountain glacier and desertified land areas. These results were consistent with the regional characteristics. This article provided technical and theoretical support for the desertification management of the Zoige grassland and a new method for studying vegetation NPP, and the findings could help improve the accuracy of regional-scale vegetation NPP estimates.

Automated summary

This study used remote sensing and hydrology data to propose a new grazing coupled remote sensing process model for the Zoige wetland, which could accurately estimate the regional-scale vegetation net primary productivity. The model provided technical and theoretical support for wetland ecosystem management and offered a new method for studying vegetation NPP, contributing to improving the accuracy of regional-scale vegetation NPP estimates.