Moderate-Scale Mapping Methods of Aspen Stand Types: A Case Study for Cedar Mountain in Southern Utah.

摘要

Quaking aspen (Populus tremuloides Michx.) are the most widely distributed tree species across North America, but its dominance is declining in many areas of the western United States, with certain areas experiencing rapid mortality events over the past decade. The loss of aspen from western landscapes will continue to profoundly impact biological, commercial, and aesthetic resources associated with aspen. However, many options are available for its restoration. Advances in remote sensing technologies offer cost-effective means to produce spatial and quantitative information on the distribution and severity of declining aspen at many scales. This report describes the development and application of transferable remote sensing and geographic information system methodologies to accurately classify aspen condition within areas of delineated aspen woodland cover. These methodologies were applied on Cedar Mountain in southern Utah within the Colorado Plateau to map three aspen stand conditions (healthy, damaged, and seral) successfully. Using moderate-scale imagery (2008 Landsat TM data), digital elevation model derivatives, high-resolution National Agriculture Imagery Program imagery, and a decision tree modeling approach, a spatially explicit 2008 landscape assessment of Cedar Mountain aspen was produced with an overall accuracy of 81.3% (Kappa (e) or KHAT accuracy measure = 0.69, n = 445). Of the total area mapped as aspen within the 12,139-ha study area, healthy aspen was the most abundant with 49% (5960 ha), followed by damaged with 35% (4210 ha), and seral with an estimated 16% (1968 ha) coverage. Aspen classification maps, derived from remotely sensed digital imagery and ancillary datasets, can offer objective management information to land managers to utilize when planning, implementing, and evaluating aspen restoration activities.