Assimilation of MODIS Chlorophyll-a Data Into a Coupled Hydrodynamic-Biological Model of Taihu Lake
Assimilation of MODIS Chlorophyll-a Data Into a Coupled Hydrodynamic-Biological Model of Taihu Lake
复制标题
DOI：
10.1109/jstars.2013.2280815
复制
发表时间：
2014-05
影响因子：
5.5
通讯作者：
Lin Qi;R. Ma;Weiping Hu;S. Loiselle
中科院分区：
文献类型：
--
作者：
Lin Qi;R. Ma;Weiping Hu;S. Loiselle
利用最优插值方法将MODIS叶绿素a浓度数据同化到水动力-生物耦合模式中。利用覆盖太湖的MODIS数据进行了模拟，2009年5月，当藻类水华通常开始发生。同化方法的结果表明，在空间一致性和时间连续性的估计叶绿素a分布的改进。同化偏差（同化后运行的模型）为5.1%，RMSE为49.7%。相比之下，自由运行（没有同化的模型运行）的偏倚为-34.9%，RMSE为176.5%。用于比较的原位数据显示，减少RMSE和同化的偏差。两个敏感性实验被用来确定合适的相关长度尺度相对于观测数据的精度。结果表明，500 m是构造背景误差协方差矩阵的最佳尺度。观测数据精度的敏感性试验也表明，更准确的观测数据允许更好的同化结果。
MODIS chlorophyll-a concentration (Chla) data were assimilated into a coupled hydrodynamic-biological model using an Optimal Interpolation method. Simulations were conducted using MODIS data covering Taihu Lake in May 2009, when algal blooms typically begin to occur. The results of the assimilation approach showed improvements in the estimation of Chla distributions in spatial coherency and temporal continuity. Bias of assimilation (model run after assimilation) was 5.1%, with a RMSE of 49.7%. In comparison, the free run (model run without assimilation) had a bias of -34.9% and RMSE of 176.5%. In situ data used for comparison showed reduced RMSE and the Bias for assimilation. Two sensitivity experiments were used to determine the suitable correlation length scale with respect to observation data accuracy. The result showed that 500m is the optimum scale to construct the background error covariance matrix. The sensitivity experiment of observational data accuracy also showed that more accurate observation data allowed for better assimilation results.