Conclusions

Application of SSM/I satellite data to a hurricane simulation

“…With a similar retrieval algorithm (Wentz 1993), different numbers for the standard deviation have been used, such as 1.6 m s −1 for SSW by Boutin and Etcheto (1996), 1 kg m −2 for V by Hou et al (2000), and about 4-5 kg m −2 for V by Deblonde (1999). In the present work, the standard deviations for SSW and V are assigned as 2.5 m s −1 and 2 kg m −2 respectively.…”

“…Advanced data-assimilation techniques based on one-dimensional (1D-Var) (Phalippou 1996;Gérard and Saunders 1999;Joiner and Rokke 2000), three-dimensional (3D-Var) (Deblonde 1999;Hou et al 2000;Fillion 2002;MacDonald et al 2002), or four-dimensional variational algorithms (4D-Var) (Gerard and Saunders 1999;Xiao et al 2000;Pu and Braun 2001;Zupanski et al 2002) have been used to assimilate different types of data, such as those from the TIROS Operational Vertical Sounder (TOVS) (Joiner and Rokke 2000), from the Global Positioning System (MacDonald et al 2002) as well as precipitation data (Fillion 2002) and SSM/I data. Variational assimilation is able to deal with indirect measurements and therefore avoids the need to estimate the retrieval's error and its correlation with the background error of the dataassimilation system.…”

“…On the other hand, directly assimilating observational information using an observation operator may be slightly more expensive, but is physically based (e.g., through use of a physical observational operator) and reduces the dependence of the result on statistical relationships that may or may not be representative of the region in question. It is, in particular, able to reduce systematic biases sometimes present in retrieved data (Deblonde 1999). For example, Phalippou (1996) pointed out that the total-column water-vapour (V ) retrieval-algorithm of Alishouse et al (1990) is biased toward underestimating high V values and, to a lesser extent, overestimating low V values.…”

“…In the past decade, scientists have been using retrieved SSM/I products, such as V (Filiberti et al 1994;Deblonde 1999;Gérard and Saunders 1999;Hou et al 2000;Xiao et al 2000), sea-surface wind (SSW) (Boutin and Etcheto 1996;Yu et al 1997;Wick et al 2000), liquid water, and precipitation rate (Hou et al 2000;Xiao et al 2000), to improve weather analyses and model forecasts/simulations. SSM/I retrievals have also been treated as verification data for validating model performance (V ) (Vesperini 1998).…”

“…SSM/I retrievals have also been treated as verification data for validating model performance (V ) (Vesperini 1998). The SSM/I V has been applied to numerous data-assimilation studies (Filiberti et al 1994;Wu and Derber 1994;Ledvina and Pfaendtner 1995;Aonashi and Shibata 1996;Deblonde 1999;Gérard and Saunders 1999;Hou et al 2000;Xiao et al 2000), and the results show some improvement of analyses and model simulations/forecasts.…”