Daily rainfall simulation : a comparison of four approaches /
| Main Author: | |
|---|---|
| Other Authors: | , |
| Format: | Thesis Book |
| Language: | English |
| Published: |
1989.
|
| Subjects: | |
| Online Access: | Link to OAKTrust copy |
| Abstract: | A simple Markov chain process (SMC), a transition probability matrix methodology (TMP), a lattice matrix methodology (LMM), and a Multi-Poisson methodology (MPM) were evaluated for their ability to accurately reproduce daily precipitation for a wide range of climatic conditions. Excluding the TMP, a simple exponential function was used to determine rainfall amounts for the sites used in this study. Data from 3 Texas sites, 3 other U.S. sites, and 3 Australian sites were used to compare simulation methodologies. The MPM provided excellent reproduction of the rainfall events per month, mean dry spell length, longest dry spell, mean wet spell length, and longest wet spell per month for the sites. The TMP preserved the mean rainfall amount per month in all sites possessing a non-uniform annual rainfall distribution. Overall, the SMC provided good reproduction of a few parameters for all sites; it emerged as the best model for only the Seattle study site. The performance of the LMM was poor for all sites and parameters. All simulation methodologies suppressed extreme rainfall amounts. Selection of a simulation methodology was shown to be dependent upon climatic type and associated rainfall distribution; the MPM should be used for sites receiving more than 30 inches (760 mm) of annual rainfall. For sites possessing 18 to 30 inches (450 to 760 mm) of annual rainfall, the TMP provided the best overall preservation of historical characteristics. Below 18 inches (450 mm) of annual rainfall, either the MPM or TMP may be used for simulation. |
|---|---|
| Item Description: | Typescript (photocopy). Vita. "Major subject: Agronomy." |
| Physical Description: | xiv, 291 leaves : illustrations ; 29 cm |
| Bibliography: | Includes bibliographical references. |