No abstract

Changes in climate and land management practices in the San Pedro River basin have altered the vegetation patterns and dynamics. Therefore, there is a need to map the spatial and temporal distribution of the vegetation community in order to understand how climate and human activities affect the ecosystem in the arid and semi-arid region. Remote sensing provides a means to derive vegetation properties such as fractional green vegetation cover (fc) and green leaf area index (GLAI). However, to map such vegetation properties using multitemporal remote sensing imagery requires ancillary data for atmospheric corrections that are often not available. In this study, we developed a new approach to circumvent atmospheric effects in deriving spatial and temporal distributions of fc and GLAI. The proposed approach employed a concept, analogous to the pseudo invariant object method that uses objects void of vegetation as a baseline to adjust multitemporal images. Imagery acquired with Landsat TM, SPOT 4 VEGETATION, and aircraft based sensors was used in this study to map the spatial and temporal distribution of fractional green vegetation cover and green leaf area index of the San Pedro River riparian corridor and southwest United States. The results suggest that remote sensing imagery can provide a reasonable estimate of vegetation dynamics using multitemporal remote sensing imagery without atmospheric corrections.2

The Semi-Arid Land-Surface-Atmosphere Program (SALSA) is a multi-agency, multinational research effort that seeks to evaluate the consequences of natural and human-induced environmental change in semi-arid regions. The ultimate goal of SALSA is to advance scientific understanding of the semi-arid portion of the hydrosphere-biosphere interface in order to provide reliable information for environmental decision making. SALSA approaches this goal through a program of long-term, integrated observations, process research, modeling, assessment, and information management that is sustained by cooperation among scientists and information users. In this preface to the SALSA special issue, general program background information and the critical nature of semi-arid regions is presented. A brief description of the Upper San Pedro River Basin, the initial location for focused SALSA research follows. Several overarching research objectives under which much of the interdisciplinary research contained in the special issue was undertaken are discussed. Principal methods, primary research sites and data collection used by numerous investigators during 1997-1999 are then presented. Scientists from about 20 US, five European (four French and one Dutch), and three Mexican agencies and institutions have collaborated closely to make the research leading to this special issue a reality. The SALSA Program has served as a model of interagency cooperation by breaking new ground in the approach to large scale interdisciplinary science with relatively limited resources. Published by Elsevier Science B.V.

We surveyed a group of rangeland managers in the Southwest about vegetation monitoring needs on grassland. Based on their responses, the objective of the RANGES (Rangeland Analysis Utilizing Geospatial Information Science) project was defined to be the accurate conversion of remotely sensed data (satellite imagery) to quantitative estimates of total (green and senescent) standing cover and biomass on grasslands and semidesert grasslands. Although remote sensing has been used to estimate green vegetation cover, in arid grasslands herbaceous vegetation is senescent much of the year and is not detected by current remote sensing techniques. We developed a ground truth protocol compatible with both range management requirements and Landsat’s 30 m resolution imagery. The resulting ground-truth data were then used to develop image processing algorithms that quantified total herbaceous vegetation cover, height, and biomass. Cover was calculated based on a newly developed Soil Adjusted Total Vegetation Index (SATVI), and height and biomass were estimated based on reflectance in the near infrared (NIR) band. Comparison of the remotely sensed estimates with independent ground measurements produced r2 values of 0.80, 0.85, and 0.77 and Nash Sutcliffe values of 0.78, 0.70, and 0.77 for the cover, plant height, and biomass, respectively. The approach for estimating plant height and biomass did not work for sites where forbs comprised more than 30% of total vegetative cover. The ground reconnaissance protocol and image processing techniques together offer land managers accurate and timely methods for monitoring extensive grasslands. The time-consuming requirement to collect concurrent data in the field for each image implies a need to share the high fixed costs of processing an image across multiple users to reduce the costs for individual rangeland managers.