An Evaluation of Discretized Conditional Probability and Linear Regression Threshold Techniques in Model Output Statistics Forecasting of Cloud Amount and Ceiling Over the North Atlantic Ocean.

• Main Author: Wooster,M H
• Other Authors: NAVAL POSTGRADUATE SCHOOL MONTEREY CA
• Format: Text
• Language: English
• Published: 1984
• Subjects: Meteorology; *WEATHER FORECASTING; *CEILING; COMPUTER PROGRAMS; MATHEMATICAL MODELS; OUTPUT; CLOUDS; METHODOLOGY; COASTAL REGIONS; PREDICTIONS; MODELS; PARAMETERS; THRESHOLD EFFECTS; PROBABILITY; STATISTICS; FORECASTING; THESES; REGRESSION ANALYSIS; CLUSTERING; MATHEMATICAL PREDICTION; VISIBILITY; TABLES(DATA); HORIZONTAL ORIENTATION; STATISTICAL ANALYSIS; OCEANS; LINEAR REGRESSION ANALYSIS; OCEAN ENVIRONMENTS; NORTH ATLANTIC OCEAN; SELECTION; OPEN WATER; North Atlantic
• Online Access: http://www.dtic.mil/docs/citations/ADA155100; http://oai.dtic.mil/oai/oai?&verb=getRecord&metadataPrefix=html&identifier=ADA155100

This thesis describes the application and evaluation of several statistical models in the forecasting of cloud amount and ceiling over coastal and open ocean areas of the North Atlantic Ocean. The focus of this study is to evaluate the applicability of previous Naval Postgraduate School model output statistics research in the area of horizontal marine visibility to the forecasting of cloud amount and ceiling over ocean areas. The models, including minimum probable error linear regression threshold techniques, maximum conditional probability and natural regression, utilize observed visibility data and model output parameters from the Navy Operational Global Atmospheric Prediction System (NOGAPS). Results show statistically similar results for the linear regression and maximum conditional probability models. Also included is the result of additional experimentation on the application of several measures of separability and cluster analysis to predictor selection. Additional keywords: mathematical prediction; computer programs; tables (data). (Author).