Prediction of Global Cloud Cover with a Very High Resolution Global Spectral Model
Abstract
Currently in numerical weather prediction, two avenues for cloud forecasting are being pursued by the research community. The conventional one defines cloud coverage as a function of prevailing relative humidity. The new method explicitly predicts clouds as a variable of the model. Our research effort covers both avenues. The major results of our research are: (1) The threshold relative humidity approach exhibits a decay of cloud fractions during the medium range weather forecasts. The major errors in the prediction appear to occur in the first 24 hours, an initialization problem. Observed clouds appear to exhibit more of a resilience than is demonstrated by the models. Long lasting cloud debris (i.e., non precipitating elements) are not reasonably handled by the model. This deficiency is related to the strong selection rules imposed by the model for the existence of clouds; (2) The explicit treatment of clouds where the cloud water mixing ratio is used as a dependent variable of the model, appears to handle long lasting clouds in a more realistic manner. It does not show the rapid spin-down feature present in the threshold relative humidity approach; and (3) A large effort on physical initialization is currently underway in our global modeling effort. This provides a consistent analysis of the humidity variable with respect to the rain rates (as seen from satellite based measurements).
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 29, 1992
- Accession Number
- ADA261047
Entities
People
- T. N. Krishnamurti
Organizations
- Florida State University