(Invited talk) Assimilation of satellite data in improving high-impact weather forecasting: Progress, Challenge and Development

As witnessed in recent years, the social, ecological and economic impacts of high impact weather systems (e.g., tropical cyclones, floods) can be devastating. Accurate forecasts of these high-impact weather systems much rely on the improved understanding of these systems and better representation of them in numerical models and initial conditions. Satellites provide a very useful source of data for improving our ability in these aspects. Over the last decade, significant progresses have been made in satellite data assimilation. Satellite data become a major data sources in the numerical weather prediction. However, compared with the general success in the use of satellite data in improving weather forecasts (e.g., forecast of large-scale synoptic environment, hurricane track forecast), accurate forecasts of the intensity and structure of these high-impact weather systems (e.g., hurricane intensity and structure, quantitative precipitation forecasting) remain a challenge problem. In addition to inadequate understanding of the physical mechanism and processes that control the intensity changes and evolutions of these systems, our limited ability to fully assimilate these satellite data contributes to the limited forecast skill. For instance, many data assimilation systems can only assimilate the clear-sky radiances owing to the difficulties in assimilating satellite data over the cloud and precipitation areas. As a consequence, few observations in the storm core areas can be assimilated into the model, thus the model may not specify the initial structure of the storm accurately and then cause the failure in the subsequent forecast. In the mean time, biases corrections and the background error statistics have also been posted as additional challenges in the satellite data assimilation respect to high-impact weather systems. Recent development in advanced radiative transfer models, bias correction and data assimilation methods bring hope to improving our ability in fully use of available satellite data in numerical weather prediction. It is anticipated that the future satellite data assimilation will make significant contribution to improve the skill of high-impact weather forecasting. The progress, challenge and development will be overview and discussed with most recent research results.