Equatorially trapped waves account for a large portion of the perturbations within the tropical atmosphere and ocean. In the atmosphere, these disturbances are coupled to convection and determine a significant amount of rainfall variability on synoptic to intraseasonal time scales. Numerical models used for both weather and climate forecasting universally still have great difficulty simulating these convectively coupled disturbances. We assess the quantitative precipitation forecasts (QPF) skill of NOAA's Global Forecast System (GFS) and the European Centre for Medium Range Weather Forecasting Integrated Forecast System (IFS) operational models used for short term forecasts out to 10 days. Forecast skill was assessed by comparison with virtually independent GPM and CMORPH satellite precipitation estimates. Skill was quantified using a variety of metrics including pattern correlations for various latitude bands, temporal correlation at individual grid points, and space-time spectra of forecast precipitation over the global tropics and extratropics. 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 estimates are not directly assimilated into these systems, this indicates that the initialization of dynamical and thermodynamical fields is able to produce a reasonable QPF field, at least for the larger scales. We present evidence that the specification of the mass circulation rather than the moisture field is the primary source of this initial skill. Model skill is substantially better overall in the extratropics, however, tropical QPF in both systems is not considered useful by typical metrics much beyond a few days. A portion of this lack of tropical skill in can be traced back to inadequate treatment of equatorial wave activity coupled to convection. It is also demonstrated that extratropical forecast skill is positively correlated to preceding tropical skill, strongly suggesting that improvements in the treatment of tropics will lead to improved extratropical forecasts on the weekly and longer timescale.
