Remote sensing of riparian habitat in the Colorado River delta

   "Arid and semi-arid zone riparian corridors are among the most important yet threatened natural ecosystems on Earth. They are important because they provide water, food and migration routes for wildlife in otherwise dry habitats; they are threatened because diversion of water for human use and flow-regulation to control flooding have severely degraded the habitat value of virtually all the world's perennial dryland rivers (Glenn et al., 2001b, in press)." Riparian corridors are linear landscape features that connect ecosystems across regions (Dynesius and Nilsson, 1994). Hence, the deterioration of riparian habitat may affect populations of migratory species at continental or even hemispherical scales (Huerta et al., 1999; EDF, 1999). Remote sensing should be among methods of choice for monitoring riparian zones, since these corridors stretch over thousands of kilometers, cross national borders and are difficult to survey on the ground (Nagler et al., 2001, in press). However, remote sensing and monitoring for the management of riparian zones has primarily taken place for the San Pedro watershed (Goodrich et al., 2000a) and the Rio Grande (Coonrod, 2001).

   Measurement of riparian plant cover is an important tool in evaluating the effectiveness of water conservation practices in the Colorado River delta and serves to protect biologically diverse natural resources and habitats for endangered species. Current methods for measuring plant cover in the delta are difficult due to inaccessibility in the vast, desert areas, and tedious due to the variability and number of plant transect surveys required, and also subjective due to inherent human error. Remote sensing methods for estimating plant cover have largely replaced manual methods of determining vegetation percent cover (including other factors such as plant type, height, and density), but this conversion of methods has primarily been for non-riparian biomes, such as grasslands, coniferous and deciduous forests, and agricultural settings. Although remote sensing provides consistent, large-area coverage of riparian regions, which are difficult to access by land due to a lack of roads, Huete et al. (1992) state that the technique in such areas is hampered by large areas of bare soil, shadowing effects, and nonlinear relationships between the measured signal and the areal extent and leaf density of shrubs. They describe the difficulty of making accurate quantification estimates of percent vegetation cover in semi-arid environments is mostly due to the effect of bare soil on the measurement of green leaf area and density (these problems are cited in detail in section 3.3). Furthermore, riparian cooridors have been considered difficult targets for analysis by satellite imagery because they are narrow features with complex mixes of vegetation, water and soil.

   Although difficult to access, Glenn et al. (2001) were awarded a NASA Grant (CARBON-0000-0114) to study carbon cycle science and the biology and biogeochemistry of ecosystems and applications, specifically for studying aspects of the delta of the Colorado River in Mexico using remote sensing. As part of this proposal, Glenn et al. (2001, NASA application) intend to use ground-, aerial- remote sensing techniques, and satellite images, to develop vegetation and habitat maps of the Colorado River delta in Mexico and to model the surface hydrology of the flood plain and delta in different flood stages. They proposed using geographically comprehensive data acquisition strategies at the ground, aerial and satellite levels (e.g., integrating locational data (i.e, GPS), transect data, ancillary maps at different spatial and temporal scales in a GIS, laser altimetry, multi-band digital cameras and radiometers from aircraft, and satellite sensors). The GIS is in cooperation with the currently used Bureau of Reclamation (BoR) Lower Colorado Accounting System (LCRAS). The satellite sensors are the Enhanced Thematic Mapper (ETM+) and the Moderate Resolution Imaging Spectrometer (MODIS). They will produce the hydrological model using ground data, historic flow data, laser altimetry from aircraft (present) and LightSAR from satellite (future), and higher spatial resolution sensor data to include both ETM+ and MODIS for comparing flood magnitudes at different scales.

   Additionally, the remote sensing plan includes the analyses of the change in vegetation response (consequence) to flood flows (cause) (Glenn et al., 2001, NASA application). They proposed using ground vegetation surveys, historic flood flow event reports, and vegetation indices from satellite images to determine not only the extent and magnitude of vegetation change, but also to create a predictive model of these land-cover, land-use change dynamics and to make assessments of endangered species habitat.

   A remote sensing method to validate the estimate of water stress and evapo-transpiration of vegetation in the delta is included in this proposal to NASA. They intend to obtain a water balance map product using ground-based sap flow meters and sensors on aircraft and satellites which have visible, near-infrared, and thermal channels. The objective of this application is to produce a predictive surface hydrology - vegetation - habitat model that uses as input flow releases from the United States to Mexico and has output predictions of extent of vegetation cover and of specific vegetation units associated with wildlife habitat values (Glenn et al., 2001, NASA application).

As with many land cover and land use change applications, the remote sensing component includes an on-going monitoring protocol. In the delta, this includes monitoring to (i) improve the management of this semiarid ecosystem for both sustainability and resilience of the natural resources (i.e., water) and (ii) further the understanding of the consequences of land-cover and land-use change on habitat value using remote sensing methods (Glenn et al., 2001, NASA application).