The use of remotely sensed imagery (e.g. Landsat TM) for developing forest inventory strata has become increasingly more common in recent years as data have become more readily available. Errors are inherent with the use of this technology, either from user mis-classification of conditions represented in the imagery or due to flaws in the technology. Knowledge of these errors is important, as they can inflate the variance of inventory estimates. Forest inventory estimates from the Mississippi Institute for Forest Inventory (MIFI) were applied to determine the extent that classification errors affect volume and area estimates. Forest strata (e.g. hardwood, mixed, and pine) determined by the classification of imagery and used for inventory design were compared with field verification data obtained during the inventory. Mis-classified plots were reallocated to their correct strata and both area and volume estimates were obtained for both scenarios (i.e. mis-classified and correctly classified plots). The standard error estimates for mean and total volume decreased when plots were re-allocated to their correct strata. Mis-classification scenarios were then performed, introducing various levels of mis-classification in each stratum. When the scenarios were performed for the Doyle volume unit the statistical efficiencies were larger than for cubic foot volume. Care should be taken when utilizing moderate resolution satellite imagery such as Landsat TM as image mis-classification could lead to large losses in the precision of volume estimates. The increased efficiency obtained from a correct classification/forest stratification scheme, as demonstrated here, could lead to the exploration of additional image classification methods or the use of higher resolution satellite data. Knowledge of these errors in advance could be useful to investors seeking a minimum-risk area for a forest products mill location.
| Identifer | oai:union.ndltd.org:MSSTATE/oai:scholarsjunction.msstate.edu:td-2155 |
| Date | 30 April 2011 |
| Creators | Crosby, Michael Keith |
| Publisher | Scholars Junction |
| Source Sets | Mississippi State University |
| Detected Language | English |
| Type | text |
| Format | application/pdf |
| Source | Theses and Dissertations |
Page generated in 0.002 seconds