Estimating net radiation flux in the Tibetan Plateau by assimilating MODIS LST products with an ensemble Kalman filter and particle filter

Abstract

Net radiation flux is a key physical variable of energy and water exchange processes in land–atmosphere interactions. Among the methods to estimate net radiation flux, measurements and estimates from modeling have their limitations. Data assimilation methods have potential to estimate net radiation flux more accurately by merging model dynamics and observations. In this paper, two data assimilation schemes are developed based on the Common Land Model (CoLM) using the ensemble Kalman filter (EnKF) and particle filter (PF) algorithms, respectively, to improve predictions of surface net radiation flux in the Tibetan Plateau. First, Moderate Resolution Imaging Spectroradiometer (MODIS) leaf area index (LAI) products were used to update the LAI in CoLM. It was found that sparse vegetation in the Tibetan Plateau has an obvious influence on the surface temperature. Then, the experiments of assimilating MODIS land surface temperature (LST) products were carried out with the EnKF and PF algorithms, respectively. The two schemes were tested and validated with observations from two automatic weather stations (AWSs) in Muztaga in the Tibetan Plateau during the period of October 1, 2009–September 30, 2010. The results show that the two assimilation schemes are both feasible and an improvement over the predictions of cases with no assimilation for the estimations of surface temperature and net radiation flux. Furthermore, the different surface types and the accuracy of MODIS LST products have an important influence on the assimilation results. In addition, the performances of EnKF and PF are almost same in the study.