Vegetation classification and mapping project report, Little Bighorn Battlefield National Monument.

Rice, P.; Schweiger, E.; Gustafson, W.; Lea, C.; Manier, D.; Shorrock, D.; Frakes, B.; O'Gan, L.

The vegetation mapping effort at Little Bighorn Battlefield National Monument is part of the National Park Service Inventory and Monitoring program and provides core or ‘baseline’ information that park managers need to effectively manage and protect park resources. This vegetation inventory was conducted in accordance with specified protocols and quality assurance standards. Data obtained through this inventory are compatible with other efforts, allowing for synthesis and analysis at broader levels. These data comply with standards and protocols of the USGS-NPS Vegetation Inventory Program. To effectively classify and map the wide range of vegetation within the mapping boundary required a multi-year approach and consisted of several linked phases: (1) vegetation classification using the U.S. National Vegetation Classification applied to field data collected by this project, (2) digital vegetation map production, (3) map accuracy assessment, and (4) revision of vegetation map and production of final products. The project was a cooperative effort between researchers from the University of Montana and National Park Service staff. Constant communication and coordinated efforts between these parties allowed for consistency in both accuracy and classification. The Little Bighorn Battlefield National Monument lies on the northern Great Plains in southeast Montana. The terrain consists of sloping uplands dissected by narrow eroded draws opening on the floodplain of the Little Bighorn River. The project mapping area is 780 acres (316 ha) and encompasses both the larger Custer Battlefield unit on the northwest end and the smaller Reno-Benteen Battlefield unit to the southeast. A preliminary list of potential vegetation types was developed from previous descriptive vegetation inventory work in the area. Spectral analysis of aerial imagery from the National Agricultural Imagery Program was used in combination with soils data to delineate the boundaries of polygons with apparently similar vegetation. Because of the small total area of the Monument and limited number of polygons it was decided to census every polygon as to community type rather than sample a subset of polygons and conduct a supervised classification. Polygons were censused by field crews in 2006; crew members assigned each polygon to a vegetation or landcover type from the preliminary list. Accuracy assessment field work in 2007 collected data at 154 sample points which were used to assess the validity of the initial map assignments. The initial vegetation mapping recognized 27 community types and an additional 3 landcover types. The initial fine-scale classification and mapping scheme had a pooled accuracy rate of only 56.2%. Some of the initial types were similar or overlapping in composition and their dominant or indicator species. The initial classification also used indicator species for some types based on patch dominance rather than stand formation – leading to more detail than could be reliably documented in the accuracy assessment. The original 27 types were aggregated into 22 final map classes representing 15 USNVC plant associations, 4 park-specific vegetation types, and 3 landcover classes. This consolidation resulted in a 78.1% accuracy rate (90% confidence interval 72.1% to 84.1%). Map classes were also assigned to four Ecological Systems; this level of grouping represents the Monument vegetation with an overall map accuracy of 97.9% (90% confidence interval 95.7% to 100%).

Vegetation classification and mapping project report, Little Bighorn Battlefield National Monument.

Series