论文部分内容阅读
Impacts of the four-dimensional variation (4DVAR) assimilation of Tropical Rainfall Measuring Mission (TRMM)/TMI surface rainfall rate (SRR) on the initialization and simulation of TC Danas in 2001 are studied. A number of sensitivity experiments are designed to evaluate the performance of assimilation during the particular stage when Danas developed from a tropical depression towards a typhoon, as well as an abrupt turning. Under the dynamic and thermodynamic constraints of meso-scale model and the 4DVAR, the TRMM SRR data are directly assimilated in high horizontal resolution (18 km). On the basis of the adjustment of initial model field, 4DVAR of TRMM SRR considerably improved the depiction of TC’s structure and rainfall. The same explanation can also be given to the enhancement of simulation on TC’s intensity and track. In comparison with the popular idea of twice steps of SRR assimilation, namely, 4DVAR plus 1DVAR, the retrieval error is expected to be avoided in the procedure of direct 4DVAR. In addition, the TRMM SRR 4DVAR within a high-resolution model can take into account as much information as observed and is expected to be performed in operational use if the observed rainfall data are routinely available.
Impacts of the four-dimensional variation (4DVAR) assimilation of Tropical Rainfall Measuring Mission (TRMM) / TMI surface rainfall rate (SRR) on the initialization and simulation of TC Danas in 2001 are studied. A number of sensitivity experiments are designed to evaluate the performance of assimilation during the particular stage when Danas developed from a tropical depression towards a typhoon, as well as an abrupt turning. Under the dynamic and thermodynamic constraints of meso-scale model and the 4DVAR, the TRMM SRR data are directly assimilated in high horizontal resolution (18 km). On the basis of the adjustment of initial model field, 4DVAR of TRMM SRR considerably improved the depiction of TC’s structure and rainfall. The same explanation can also be given to the enhancement of simulation on TC’s intensity and track. In comparison with the popular idea of twice steps of SRR assimilation, namely, 4DVAR plus 1DVAR, the retrieval error is expected to be avoided in the procedure of direct 4DVAR. In addition, the TRMM SRR 4DVAR within a high-resolution model can take into account as much information as observed and expected expected to be performed in operational use if the observed rainfall data are routinely available.