Atmospheric motion vector observations in the ECMWF system : 1-year report

Atmospheric motion vector observations in the ECMWF system : 1-year report
Date Published
EUMETSAT/ECMWF Fellowship Programme Research Reports
Kirsti Salonen
Abstract The research work towards a general enhancement of the assimilation of AMV observations in the ECMWF system is ongoing. This includes detailed studies on the observation error characteristics, and how to better take them into account in the assigned observation errors. Two main sources of AMV observation errors are errors in the wind vector derivation, and errors in the height assignment of the tracers. Situation dependent observation errors have been derived and their impact on model analysis and forecasts is carefully considered. The new approach allows to down-weight observations in regions with high vertical wind shear where errors in height assignment are problematic. Preliminary results are encouraging. In order to estimate the AMV observation error due to error in height assignment, a height error estimate is required. Model best-fit pressure is one option. In the context of estimating the new observation errors for AMVs, the model best-fit pressure statistics have been compared between the Met Office and ECMWF global assimilation systems. The obtained results show that the statistics are similar for both systems in terms of bias and standard deviation. The results are also in good agreement with earlier findings about the characteristics of height assignment methods, and quality of the AMVs. In 2011, a significant change in the operationally used AMVs has been that JMA/MSC started to disseminate hourly-derived AMVs from MTSAT-2. Earlier the AMVs were disseminated every 3-hours on the northern hemisphere and every 6-hours on the southern hemisphere. Also the image intervals used in the AMV derivation were changed. Results on monitoring the quality of the new 1-hourly AMVs, and on impact studies are presented in the report. The quality of the 1-hourly MTSAT-2 AMVs is comparable to the quality of the old MTSAT-2 AMVs. Impact studies indicate improvements in the forecasts when MTSAT-2 AMVs are used in general, and when 1-hourly AMVs are used in particular.