TY - GEN
AU - Philippe Lopez
AB - Four-dimensional variational data assimilation (4D-Var) experiments with 6-hourly rain gauge accumulations observed at synoptic stations around the globe have been run over several months, both at high resolution in an ECMWF operations-like framework and at lower resolution in an early or mid-20th century reanalysis style with the reference observational coverage reduced to surface pressure data only. The key aspects of the technical implementation of rain gauge data assimilation in 4D-Var are described, which includes the specification of observation errors, bias correction procedures, screening and quality control. Results from experiments indicate that the positive impact of rain gauges on forecast scores remains limited in the operations-like context because of their competition with all other observations already available. In contrast, when only synoptic station surface pressure observations are assimilated in the reanalysis-like control experiment, the additional assimilation of rain gauge measurements substantially improves not only surface precipitation scores, but also analysis and forecast scores of temperature, geopotential, wind and humidity at most atmospheric levels and for forecast ranges up to 10 days. The verification against Meteosat infrared imagery also shows a clear improvement in the spatial distribution of clouds. This suggests that assimilating rain gauge data available during data sparse periods of the past might help to improve the quality of future reanalyses and subsequent forecasts.
BT - ECMWF Technical Memoranda
DA - 01/2012
DO - 10.21957/0dnmprv72
LA - eng
M1 - 661
N2 - Four-dimensional variational data assimilation (4D-Var) experiments with 6-hourly rain gauge accumulations observed at synoptic stations around the globe have been run over several months, both at high resolution in an ECMWF operations-like framework and at lower resolution in an early or mid-20th century reanalysis style with the reference observational coverage reduced to surface pressure data only. The key aspects of the technical implementation of rain gauge data assimilation in 4D-Var are described, which includes the specification of observation errors, bias correction procedures, screening and quality control. Results from experiments indicate that the positive impact of rain gauges on forecast scores remains limited in the operations-like context because of their competition with all other observations already available. In contrast, when only synoptic station surface pressure observations are assimilated in the reanalysis-like control experiment, the additional assimilation of rain gauge measurements substantially improves not only surface precipitation scores, but also analysis and forecast scores of temperature, geopotential, wind and humidity at most atmospheric levels and for forecast ranges up to 10 days. The verification against Meteosat infrared imagery also shows a clear improvement in the spatial distribution of clouds. This suggests that assimilating rain gauge data available during data sparse periods of the past might help to improve the quality of future reanalyses and subsequent forecasts.
PB - ECMWF
PY - 2012
EP - 25
T2 - ECMWF Technical Memoranda
TI - Experimental 4D-Var assimilation of SYNOP rain gauge data at ECMWF
UR - https://www.ecmwf.int/node/10790
ER -