Publications

Li, L; Zhao, YL; Fu, YC; Xin, QC (2018). Satellite-Based Models Need Improvements on Simulating Annual Gross Primary Productivity: A Comparison of Six Models for Regional Modeling of Deciduous Broadleaf Forests. REMOTE SENSING, 10(7), 1008.

Abstract
Modeling vegetation gross primary productivity (GPP) is crucial to understanding the land-atmosphere interactions and, hence, the global carbon cycle. While studies have demonstrated that satellite-based models could well simulate intra-annual variation of vegetation GPP, there is a need to understand our ability to capture interannual GPP variability. This study compares the spatiotemporal performance of six satellite-based models in regional modeling of annual GPP for deciduous broadleaf forests across the eastern United States. The 2001-2012 average annual gross primary productivities (AAGPPs) derived from different models have mismatched spatial patterns with divergent changing trends along both latitude and longitude. Evaluation using flux tower data indicates that some models could have considerable biases on a yearly basis. All tested models, despite performing well on the 8-day basis because of the underlying strong seasonality in vegetation productivity, fail to capture interannual variation of GPP across sites and years. Our study identifies considerable modeling uncertainties on a yearly basis even for an extensively studied biome of deciduous broadleaf forest at both site and large scales. Improvements to the current satellite-based models have to be made to capture interannual GPP variation in addition to intra-annual variation.

DOI:
10.3390/rs10071008

ISSN:
2072-4292