Are cloud forecasts improving?

 

Authors

Ewan James O'Connor — University of Reading
Robin J. Hogan — University of Reading

Category

Modeling

Description

The ARM Climate Research Facility now has nearly a decade of weather forecasts from NCEP, ECMWF, and the UK Met Office archived over its observational sites, a period over which we might expect to see some improvement in the forecasts with respect to the observations. In this poster we analyze the performance of these models over SGP and Darwin in a number of different ways, looking at skill and bias versus height, time of day, season, forecast lead time, and mid-troposphere vertical wind. We also investigate whether there has been a detectable improvement in cloud forecasts over the decade in terms of reduced bias, increased skill, or increased forecast half-life (the lead-time into a forecast it takes until some measure of skill falls to half its initial value). This work is part of the FASTER project.