The Performance of the NCEP and ECMWF Operational Models for Quantitative Precipitation Forecasts during 2015-2016

We assess the quantitative precipitation forecast (QPF) skill of the NCEP Global Forecast System (GFS) and ECMWF Integrated Forecast System (IFS) operational forecast models in twice daily forecasts for the period January 2015 through March 2016. Forecast skill was quantified using a variety of metrics including space-time correlations for various latitude bands, space-time spectra of forecast precipitation, mean-square errors of precipitation forecasts over the global tropics and extratropics, and fractional skill scores. Results reveal that, in general, initial conditions are reasonably well estimated in both forecast systems, as indicated by relatively good scores for the 6-12 hour forecasts. Since precipitation is not directly assimilated into either forecast system, this skill comes almost entirely from the reasonably good initialization of the mass circulation. Model skill is substantially better in the extratropics, however, tropical QPF in both systems is not considered useful by typical metrics much beyond a few days, although this is situational dependent. To quantify the relationship between QPF and dynamical skill, we calculate space-time spectra and coherence between rainfall and divergence fields. It is shown that while tropical variability is too weak overall in both models, the IFS performs better in allowing tropical waves to propagate for longer lead times. This study indicates that differences in physical parameterizations used in each system, and in particular, how moist convective processes are coupled to the large scale flow, appear to be major sources of forecast errors.