Search Results
You are looking at 1 - 2 of 2 items for
- Author or Editor: Albert Thomasell x
- Refine by Access: All Content x
Abstract
The impact of satellite sounding data on the systematic errors of the numerical weather prediction model of the Israel Meteorological Service has been investigated. In general, satellite data have been shown to reduce systematic error, and in particular, the greatest impact is near where the data have been introduced in the vicinity of low pressure systems.
Abstract
The impact of satellite sounding data on the systematic errors of the numerical weather prediction model of the Israel Meteorological Service has been investigated. In general, satellite data have been shown to reduce systematic error, and in particular, the greatest impact is near where the data have been introduced in the vicinity of low pressure systems.
Abstract
Predictions of pressure, temperature, precipitation, cloudiness, and visibility were prepared using linear regression methods. Factor-analysis techniques were employed for the purpose of isolating from a complex of variables a small number of factors which accounted for most of the variation of the predictand. A screening procedure was used to select from the same complex of variables those few variables which, when combined linearly, provided, the ‘best’ forecasts. Forecast equations were derived by both of the above methods using all of the available data. The data were then stratified on the basis of a precipitation criterion, and new forecast equations were derived. Four sets of forecasts were prepared and compared. No significant differences were noted. It was concluded that screening would be preferred because of its simplicity, that stratification was not helpful in this case, that the equations performed almost as well on new data as they did on the dependent data, and that the addition of cloudiness and precipitation to the list of predictors was rewarding.
Abstract
Predictions of pressure, temperature, precipitation, cloudiness, and visibility were prepared using linear regression methods. Factor-analysis techniques were employed for the purpose of isolating from a complex of variables a small number of factors which accounted for most of the variation of the predictand. A screening procedure was used to select from the same complex of variables those few variables which, when combined linearly, provided, the ‘best’ forecasts. Forecast equations were derived by both of the above methods using all of the available data. The data were then stratified on the basis of a precipitation criterion, and new forecast equations were derived. Four sets of forecasts were prepared and compared. No significant differences were noted. It was concluded that screening would be preferred because of its simplicity, that stratification was not helpful in this case, that the equations performed almost as well on new data as they did on the dependent data, and that the addition of cloudiness and precipitation to the list of predictors was rewarding.
