The Land Condition-Trend Analysis (LCTA) program was developed by the U.S. Army to ass ist in the sustainable management of natural resources on U.S. Army lands. The LCTA program applies a standardized procedure in order to select long-term monitoring sites. The LCTA monitoring program was applied to Camp Williams, a National Army Guard training site located in central Utah. Due to the criteria set by the LCTA monitoring program, 6 I percent of Camp Williams was explicitly excluded from the LCTA monitoring protocol because it appeared to be more heterogeneous, which would make it difficult to locate monitoring sites in the field. This study compared the monitored landscape with the unmonitored landscape to determine how the two landscapes differed. The expectation was that the monitored landscape would contain larger, less numerous patches compared with the unmonitored landscape, which was expected to contain smaller, more numerous patches. Accordingly, the landscape structures of the included and excluded lands were compared. The landscape metrics utilized to quantify landscape structure were largest patch index (percent), number of patches, patch density (#/1 00 ha), mean patch size (ha), double log fractal dimension, Simpson 's diversity index, Simpson 's evenness index, interspersion (percent), and contagion. Small differences did occur between the two landscapes, though the population variance showed that the two landscapes were more alike than different for all metrics, except interspersion and contagion which did show small differences. Due to the criteria set by the LCTA monitoring program, these results were not expected. Since it was shown for the majority of landscape metrics that the two landscapes were more alike than different, the 61 percent of Camp Williams excluded from monitoring consideration was not greatly different. However, important features such as riparian areas and recent small burns were largely contained within the areas excluded by the LCTA program. Further investi gation of landscape metrics is encouraged because previously unmonitored features of wildlands can only be assessed by examination of these coarse-scale characteristics.

How plants from a common species pool form community has been considered from a variety of approaches. A promising approach involves the search for assembly rules based on plant functional traits. This approach has potential to provide insight into community and ecosystem processes In this research. a general and simple conceptual model based on life forms and independent of species is used as a framework for assessing the internal structure of plant communities. Plant functional traits are used to identify patterns within and between plant communities in the contrasting environments of Camp Williams, Utah, and Camp Grayling, Michigan. The conceptual model has three different functional types formed by one to three functional groups. A functional group, made up of species with similar life form, is analogous to a vegetation stratum. A functional type, consisting of one or more functional groups. is analogous to a community or vegetation type. Correspondence analysis (CA) and analysis of variance (ANOVA) indicate that richness, species diversity, and trait diversity are essentially independent of functional type and are, for example, fairly consistent regardless of climatic regime or structural complexity. Cover. on the other hand. increases with the number of functional groups in a functional type. Consistent patterns and trends for sets of functional traits support the view that assembly rules may account for internal structure in plant communities. The consistent association of sets of traits with functional groups even in taxonomically dissimilar communities suggests that the functional traits are related to fundamental ecological processes that shape these communities. Ambiguity in some of the results might be explained by extending the analysis to additional installations that replicate the climatic conditions found at Camp Williams and Camp Grayling.

Current vegetation classes were generated from remotely sensed data to provide coarse-level information for an ecosystem management plan developed at Camp Williams, Utah. Vegetation trend from 1973 - 1993 was also examined via satellite imagery. The data set consisted of Landsat Multispectral Scanner (MSS) and Thematic Mapper (TM) images from July or August of 1973, 1975, 1980, 1988, and 1993. Two approaches were used to detect vegetation change. The first approach determined overall and cover type trend from standard digital image differencing of soil-adjusted vegetation index (SAVI) images. The second approach used an unsupervised classification of a composite SAVI image of all dates. The first approach defined areas of increase, decrease, and no significant change in SAVI and differences in trend for tree versus shrub cover types. The second approach resulted in an ecological classification that defined new environmental patterns based on vegetation trend.

Soil erosion was assessed at Camp Williams National Guard Base by creating an erosion risk classification map and comparing the erosion impact of disturbance regimes on different hillslopes. Soil erosion does not appear to be a problem for most of Camp Williams. The Revised Universal Soil Loss Equation was applied using GIS to create a soil erosion risk map for the entire Camp Williams facility. The map indicated where problem areas occurred and showed relative erosion risk, but its lack of quantitative accuracy should be noted. Areas of concern included landscapes with little or no protective vegetation such as roads, abandoned agricultural fields, and sensitive riparian areas where gullies tend to form and expand. The Water Erosion Prediction Project model was used to evaluate the erosion impacts of various disturbances on five study hillslopes. The model did not appear to function well on the Camp Williams study hillslopes because the distribution of infiltration rates could not be satisfactorily represented. However, hydraulic conductivity measurements collected for this task were useful in providing insight into some of the physical processes of erosion. The hydraulic conductivity measurements showed some of the impacts of military activities, grazing, and wildfire on soil properties. Erosion bridges were also used on the five study hillslopes in an attempt to measure soil Joss and deposition. However, the bridges Jacked the capability of measuring the low rates of erosion during the time period set for this experiment. The bridges showed potential for measuring erosion in rills, gullies, highly disturbed areas, or in longer duration experiments.

Fire occurrence and behavior were determined by collecting and analyzing fuel , weather , and fire history ii data. Fuel plots were used to measure average fuel loading by vegetation type and integrated with weather to make worst - case fire behavior prediction s . A fire history was developed using oakbrush (Quercus qaffibelli Nutt . ) sprouts to determine age and the Global Positioning System (GPS) for mapping the burned areas . Average fuel loading was highest in the oakbrush fuel type with 16.8 t/ha , then juniper (Juniperus ost eosoerma Torrey ) with 6 . 72 t/ha , and the lowest was in sagebrush (Artemisia triden~at a Nutt. ) with 4 . 93 t/ha . Fire behavior predictions were similar for all fuel types. The fire rotation for the study area was calculated to be 30 years. The fire history showed the most hectares burned were in the oakbrush fuel type due to fuel loading and horizontal continuity. Prescribed burns and negative fuel breaks were suggested as management alternatives. The effect of fire on deer mouse (Peromyscus maniculatus) density in oakbrush was determined by using a trapping web design with distance sampling techniques . Webs were set in four pairs with one web of each pair being in 7- year- old burned oakbrush and the other web in unburned oakbrush. Variables such as shrub height and litter depth were recorded in order to reduce variance . Trapping occurred in June 1994 with each web set for two consecutive nights using 80 Museum Special snap traps spaced 6 m apart on eight lines . Density estimates were determined by using a computer program called DI STANCE and then analyzed using analysis of variance with a randomized block design . No significant differences between deer mouse densities were detected between burned and unburned oakbrush. Although litter depth and shrub height were both significantly less in burned sites, it did not affect deer mouse density . The conclusion from these result s was that after 7 years oakbrush had recovered to a point that the effect of fire on deer mouse density was negligible.