Approaches to Identify and Measure Landscape-Scale Attributes for Assessing Rangeland Health

Updated 02/27/2009

X. Ben Wu
Department of Rangeland Ecology and Management
Texas A&M University, College Station, TX

Landscape can be defined as a heterogeneous area composed of different ecosystems, or a mosaic of patches of different type (landscape elements). Landscape structure (compositional and spatial patterns of the landscape elements) affects landscape function (flows of nutrient, energy and organisms, and disturbances across the landscape) and in turn modified by landscape processes.

The emphasis of landscape-scale studies is on the structure of spatial heterogeneity and its interaction with landscape processes. This is significantly different from community/site-scale studies that often focus on the structure and function of spatially homogeneous systems. Landscape-scale approaches also place a special emphasis on the issue of scales. Landscape patterns and processes are dependent on spatial and temporal scales, i.e., the spatial/temporal resolution of the data ("grain") and the area/duration encompassed by the data ("extent"), and may have characteristic "domains of scale".

Rangeland ecosystems can be considered as landscape elements in a mosaic of different ecosystems such as cropland, aquatic systems, urban areas, transportation corridors, etc. This landscape context may affect the structure and function (biophysical processes) of the rangeland ecosystems such as fragmentation, disturbance regime, and hydrology. The significance of the rangeland ecosystems to the biophysical processes in encompassing landscape is determined to a degree by the landscape context. The landscape context, which reflects the socio-economic pattern and demographic structure, may also influence the societal values for the rangeland ecosystems. Many rangeland ecosystems can be considered as landscapes themselves, such as savanna or parkland systems with discrete woody, herbaceous, and bare ground patches. The composition and spatial configuration of these patches affect the processes and dynamics of the rangeland ecosystem.

Many of the societal goals for reangeland health, such as water yield, water quality, soil erosion, and habitat conservation, are landscape or regional-scale issues in nature. Approaches of landscape analysis and modeling developed in the recent decades, together with remote sensing and GIS technology, can provide effective and efficient tools for assessment and monitoring of rangeland health.

Quantification of landscape pattern

There are two major groups of methods for quantifying landscape pattern based on the types of data. Many landscape data come in the form of categorical data often derived from remote sensing, such as maps of soil types and vegetation types, or distribution of woody patches in a rangeland ecosystem. A landscape is represented by a mosaic of discrete patches of different categories (landscape elements). Many landscape metrics (or indices) have been developed for quantifying the compositional and spatial structure of a landscape represented by a discrete categorical map. The compositional structure of a landscape can be described using the number of categories, proportions of the area in categories, diversity indices measuring the richness and evenness of the categories, and dominance measuring the degree of dominance by a subset of categories. The spatial attributes of patches of a given category and their spatial configuration can be quantified using metrics such as patch density, patch size, patch shape, edge density, amount of interior area, and connectivity measured by proximity, nearest neighbor distance, and patch cohesion. The overall landscape spatial structure can be quantified using contagion that measures the degree of aggregation, lacunarity that is a multiscale assessment of the texture or "gappiness" of the landscape and its domains of scale, and fractal dimension that reflect the complexity of the landscape structure and reveal domains of scale.

Some landscape attributes, such as topography, soil organic content and habitat suitability, are continuous in nature or can be represented by continuous variables. In addition to many traditional methods, such as trend surface analysis and spectral analysis, geostatistical analysis and modeling methods, such as variography, kriging, and stochastic conditional simulation, provide a powerful tool for landscape analysis and modeling. Variography is used to analyze and model the spatial structure of a landscape attribute in terms of pattern of its spatial autocorrelation as a function distance between sampling points. Kriging, a set of spatial interpolation methods based on the spatial structure determined with variography, can be used to predict the values of the landscape attribute in areas without sample data. Stochastic conditional simulation can be used to assess the distribution and uncertainty of landscape attributes based on the spatial structure determined with variography. Many other spatially explicit methods, such as terrain and hydrologic analysis and modeling, can be used to analyze and model landscape processes such as hydrology and soil erosion.

Possible landscape-scale approaches and attributes for assessing rangeland health

Fragmentation of rangelands

Rangelands are increasingly fragmented, in both the reductions of the total area and sizes, by land conversion and subdivision. It has significantly impacted many rangeland values such as economic feasibility of ranching, vegetation management for water yield, and habitat quality and connectivity. The degree of fragmentation can be quantified with landscape metrics such as proportion, patch density, size, shape, and connectivity.

Soil erosion potential

Soil erosion potential and its spatial distribution can be evaluated using GIS-based soil and topography data and empirical models to delineate areas of high sensitivity to vegetation disturbances, and develop an efficient system for monitoring soil erosion risks on large-scales based on the soil erosion potential data and remote sensing-based land cover data.

Water yield and water quality

Watershed and regional scale of water yield and water quality can be assessed using hydrologic models coupled with GIS, based on the climatic and topoedaphic characteristics and vegetation composition and spatial configuration in different range/ecological sites.

Habitat suitability

Habitat suitability and spatial distribution can be assessed using spatially explicit habitat suitability models based on composition and spatial configuration of vegetation types, other life requisites, and disturbances. A set of indicator species with conflicting habitat requirements can be used to evaluate the effect of rangeland management on their habitat suitability. These species represent not only the present societal goals for the rangeland health but also management options for other potential goals that society may choose in the future.

Aesthetic and recreational values

The composition and especially spatial configuration and landscape elements (e.g., vegetation types) determine aesthetic values of rangeland landscape. Visibility analysis can be used to evaluate the aesthetic value of rangeland landscape and other recreational values such as huntability for game species that could be considerably different from the habitat suitability for the species.

Identification of thresholds

Landscape analysis and modeling approaches can help to determine the relationship between composition and spatial configuration of vegetation/cover types and thresholds for ecosystem state transition, thresholds for economic feasibility of management options, minimum habitat requirements for wildlife species, etc.

Organization of community/site-scale assessment and monitoring

Landscape analysis can also help to improve the effectiveness and efficiency of community/site-scale assessment and monitoring. It can be used to identify critical areas such as riparian areas and potential habitat for endangered species, and improve the distribution of intensive field sampling sites based on the spatial pattern of physical landscape attributes and stratification of land cover types. Existing geo-referenced point/site data, strategically expanded if necessary, can be used to detect domains of scale and determine the pattern of spatial and temporal autocorrelation for improving sampling design, using spatial statistical methods.

< Back to Agenda