Off-campus South Dakota State University users: To download campus access theses, please use the following link to log into our proxy server with your South Dakota State University ID and password.

Non-South Dakota State University users: Please talk to your librarian about requesting this thesis through interlibrary loan.

Document Type

Thesis - University Access Only

Award Date


Degree Name

Master of Science (MS)

Department / School

Mathematics and Statistics

First Advisor

Gary Hatfield


The Climate Prediction Center (CPC) issues seasonal 30-60-90 day forecasts in the United States and are used for planning purposes on a nation-wide scale. Categorical forecasting using the concept of equal chance (EC) determines whether a forecast is above normal (A) or below normal (B) for future monthly outlooks. CPC forecasting is too broad in scope for local forecasting systems. Thus, the need for more effective small scale local forecasting is the subject of this analysis. The purpose of this study is to predict and compare a neural and analog network model relative to the CPC seasonal forecasts. The tools generated from the model are referenced against historical sixty-year half standard deviation values to determine categorical forecasts. The results are paired with actual observed categorical forecasts to populate a 3x3 contingency table for each climate division in the study. Measures of the skill of forecasting are initially measured using the Heidke skill score. Furthermore, due to the ordinal nature of the forecast variables, the marginal probability scores are weighted giving penalties to forecasts that are farther away from the actual observed values to test the validity of the prediction measures.

Library of Congress Subject Headings

Long-range weather forecasting
Long-range weather forecasting -- South Dakota
Neural networks (Computer science)


Includes bibliographic references (pages 44-45)



Number of Pages



South Dakota State University


In Copyright - Non-Commercial Use Permitted