The NWP Testbed for the FASTER Project
Model intercomparison for cloud fraction
The links on this page summarise our initial results for the model intercomparison of the following properties: cloud fraction, ice water content, and liquid water content. Our primary focus is on the skill various NWP models have in forecasting cloud, and to investigate this we use the symmetric extreme dependency skill score, SEDS.
The first set of figures display a timeseries intercomparison of the vertically-averaged value of SEDS for cloud fraction. For models with a range of forecast lead times we can also investigate how the model forecast degrades with time.
Some seasons are often much easier to forecast than others. This has been investigated by constructing seasonal composites over all available years. Composites have been created to illustrate the performance of various models at forecasting cloud fraction using SEDS, and have also been created for both observations and models for cloud fraction, ice water content and liquid water content.
For similar plots describing individual years, or for more detailed comparisons with the observations for individual models (e.g. mean, frequency of occurrence, amount when present and PDFs), click here.
Timeseries: cloud fraction skill (SEDS) intercomparison
Intercomparisons are also displayed for individual models where the different forecast lead times are available.
SGP (taken from 2004-2009)
Darwin (taken from 2005-2009)