Earth and Atmospheric Sciences, Department of


Date of this Version



A THESIS Presented to the Faculty of The Graduate College at the University of Nebraska In Partial Fulfillment of Requirements For the Degree of Master of Science, Major: Earth and Atmospheric Sciences, Under the Supervision of Professor Merlin Lawson. Lincoln, Nebraska: August 2011

Copyright 2011 Paul John Wayne Fajman


Verification is the process of determining the quality of forecast information. Office and personal forecast verifications are significantly lacking throughout the National Weather Service for many reasons. The primary reasons are that verification is time consuming, tedious, and monotonous. This research attempted to ease that process by creating new computer procedures to automate the verification process. The new procedures were tested using two years of forecasting data from November 2007 to November 2009 from the Omaha/Valley Weather Forecasting Office to serve as a framework for future verifications. Point Forecast Matrices (PFM) produced by the National Weather Service twice daily and the GFS (Global Forecasting System) served as the forecasting data for this research. The analysis of the forecast data can provide valuable feedback to the Omaha/Valley Weather Forecasting Office. The GFS was very competitive against the PFM in terms of accuracy, but the PFM were an improvement for most forecasting situations.

More difficult forecasting situations, such as snow cover and temperatures near climatic temperature records, received additional scrutiny. Snow cover forecasts were divided into non-freezing and freezing day forecasts. The division revealed the PFM to be more accurate for freezing days and the GFS to be more accurate for non-freezing days. Analysis of near climatic temperature records showed that the GFS handled warmer than normal temperatures well and the PFM were better at handling cooler than normal temperatures. In addition to analyzing accuracy, forecast consistency was also studied. The Forecast Convergence Score, a statistic which measures how often and how much a forecast changes from forecast to forecast, was used to measure forecast consistency. PFM Forecast Convergence Scores are a vast improvement over GFS Forecast Convergence Scores for all forecasting situations. When consistency is combined with accuracy, the use of PFM proves to be the most trusted forecasting system over the entire forecasting database.

Adviser: Merlin Lawson