A probabilistic evaluation of the harvesting behavior of private timberland owners in eastern Texas /
| Main Author: | |
|---|---|
| Other Authors: | , , |
| Format: | Thesis Book |
| Language: | English |
| Published: |
1993.
|
| Subjects: | |
| Online Access: | ProQuest, Abstract Link to OAKTrust copy |
| Abstract: | This study was conducted to provide a mechanism for improving local timber supply modeling capability by estimating the probabilities of softwood harvest for the two private timberland ownership classes under varying biological conditions. The FIA data for eastern Texas were examined to determine their viability for use in harvest behavior modeling on small geographic levels and filtering procedures were developed to use the data. Models of softwood harvesting behavior for private timberland owners for two timbershed areas in eastern Texas during the period between 1975 and 1986 were specified and calibrated using the filtered FIA data. The usefulness of logistic regression diagnostic procedures for harvest behavior modeling was examined. Close examination of the FIA data revealed suspect data values caused by measurement or recording errors, plots with negative periodic gross growth, and other problems that are not usually considered in analyses over small geographic areas using FIA data. Lost plots and substitute points were problematic when estimating the models. After filtering, the probability of harvest for both industrial and nonindustrial owners was found to increase as initial standing volume increased and as periodic growth and periodic mortality decreased. Data from the two timbersheds was pooled, and the dummy variable created to test for differences between the two areas while controlling the stand variables was significant at the p=0.10 level, supporting the notion that differences between the timbersheds could be seen in stand age structure and types of processing facilities. The diagnostic procedures were useful in identifying influential observations that affected the model fit. The magnitude of the coefficients on the continuous variables in both timbersheds increased after reestimation of the model. Unless there is evidence that measurement or recording errors occurred, however, it is not recommended that such plots be deleted from calibrating the model. The logistic model also offers opportunities to explore other factors affecting timber management behavior that may be useful for forest policy analysis purposes. |
|---|---|
| Item Description: | Vita. "Major subject: Forestry." |
| Physical Description: | x, 126 leaves : illustrations ; 28 cm |
| Bibliography: | Includes bibliographical references. |