A remote sensing methodology for testing the applicability and implementation of landscape partitioning systems with Advanced Very High Resolution Radiometer satellite data.

摘要

The Southern Research Station, Forest Inventory and Analysis Program (SRS-FIA) has the mandate to conduct forest inventories for all southern states from Virginia to Texas. Except for sparsely forested regions in west Texas and west Oklahoma, forest land in the South has been field inventoried over several cycles in recent history. An assessment of these western woodland resources is needed to meet the SRS-FIA's inventory mandate. Advanced Very High Resolution Radiometer (AVHRR) weather satellite data was examined for its usefullness in fulfilling this mandate.; Satellite data analysis for natural resource assessment involves extracting useful information from the data. This study was developed to enhance existing methodologies designed to extract information from AVHRR data. Understanding the landscape processes which dictate plant distributions is essential to this process.; A review of the literature indicated that landscape partitioning systems have been used with AVHRR data in pre-classification and post-classification scenarios. Decisions regarding the need for partitioning and the choice of a particular partitioning system have been subjectively made.; This study yielded quantitative evidence that confirms the usefulness of partitioning in general, and the superiority of Bailey's landscape partitioning system for the west Texas and west Oklahoma study area. Statistical tests revealed that variability of Jeffries-Matusita (J-M) distances were within acceptable limits, and that a sample based approach could be used for generating spectral signatures for assessing partitioning effectiveness.