Research advances in model simulation and data assimilation of water and carbon fluxes in land surface vegetation

Abstract

Accurately estimating water and carbon fluxes is of great significance for the research in land surface water and carbon cycles. However, it is very challenging. The estimation accuracy needs further improvement. Both traditional model simulation and site observation methods have advantages and disadvantages, which need to be examined in combination. Data assimilation integrates observations into models based on physics laws to obtain the optimal estimates of model state variables and parameters as much as possible, and provides an effective way for their combination. In this review, we traced the research progress for process models assimilated with multi-source observational data of land surface water carbon fluxes and analyzed the domestic and foreign research status of land surface process models focused on water carbon fluxes, data assimilation algorithms, and assimilation of land surface carbon flux data. We summaried problems in this research area, including insufficient coordination of multi-source observation data, relatively simple assimilation strategy, lacking fusion of assimilation models, and limited assimilation scale. The future development directions and trends were analyzed and prospected from five aspects, including assimilation strategy, model selection, data expansion, scale effect, and scientific calculation. This work would provide more comprehensive background information for scholars in this field, and arouse common concerns.