A Semi-Markov Weather Model.
Abstract
The study traces the development of a weather model based on a real transition Semi-Markov process. The data consisted of eighteen years of taped hourly weather observations from Bitburg Air Base, Germany. Maximum likelihood estimators were used to construct the basic parameters of the model from the taped data. The results consist of weather models for each season. The model for the winter season was used to develop both short and long-term weather statistics. The results confirmed that the Semi-Markov process may be used to develop a weather model that requires few assumptions, is small in size, independent of geographic location, and is distribution free. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 01, 1974
- Accession Number
- AD0778873
Entities
People
- Kenneth O. Merrill
Organizations
- Air Force Institute of Technology