The impact of spatial and temporal distribution of satellite observations on tropical cyclone data assimilation: Preliminary results

Preliminary work gauging the impact of varying the spatial and temporal resolution of Cloud Drift Wind (CDW) data using different assimilation techniques is presented, particularly within the framework of a generalised inverse data assimilation scheme. Results are presented for the NW Pacific and Australian regions. There were three main findings for the cases examined. Firstly, hourly and 12-hourly highdensity CDWs produced lower mean forecast errors (relative to intermittent assimilation) than those produced using operational CDWs from the Global Telecommunication System (GTS). This is consistent with Le Marshall et al. (1994). Secondly, intermittent assimilation was significantly worse than both the nudging and variational procedures. Finally, there was little improvement using the variational as opposed to the nudging scheme when using 12-hourly data insertion and operational winds from the GTS. There, the variational procedure proved to be about 4 per cent superior to the nudging. This result is of note, given the variational procedure takes about an order of magnitude longer to produce the initial field than does nudging, and should be contrasted with an earlier finding by one of the authors (LML) where the variational procedure was found to be clearly superior to the nudging approach (Bennett et al., 1993) where an enhanced CDW field was used.