Defining ecosystem processes of the Australian Great Artesian Basin springs from multi-sensor synergies

Abstract

The Australian Great Artesian Basin (GAB) supports a unique and diverse range of ecologically significant groundwater-fed wetland ecosystems termed GAB springs. The springs are of great national and international importance for their ecological, scientific and economic values, and are culturally significant to indigenous Australians. The ecological sustainability of the springs has become uncertain in recent times due to increased mining operations and associated groundwater extractions from the GAB. The impacts of existing water extractions from the time of European settlement, pastoral activities and more recently mining are largely unknown. This situation is compounded by the likelihood of future increasing demand of water extractions for mining operations. The GAB springs are spatially disparate ecosystems located within the arid interior of South Australia, akin to islands in their ecological setting. The springs exhibit a diverse range of geomorphology, hydrogeology, surface expressions and vegetation community composition over a wide range of spatial scales. A suite of remote sensing technologies were used to capture the range of scales of the spring wetlands and their surface expressions. This multi-sensor approach enabled definition of the spatial and temporal responses of dominant plant species, communities and entire wetlands. To validate the suite of satellite and airborne imagery several comprehensive field campaigns were conducted, capturing the variation in spring vegetation communities and surface expressions. This paper provides a review highlighting the sensor synergies that can be drawn from research conducted from the Australian National Water Commission flagship research program, Allocating Water and Maintaining Springs in the Great Artesian Basin, which has developed new spatial and temporal tools for monitoring indicators of GAB spring response to water allocations and land use (Lewis et al., 2013). The main objectives of this study included: mapping the location and elevation of western GAB springs using high-precision DGPS; development of protocols for ground-based image validation data; develop techniques for detection and monitoring of the surface characteristics of spring-fed wetlands and surrounding environments using fine spatial and spectral resolution imagery; define the short and long-term temporal dynamics of the springs of indicative vegetation types and entire wetlands; use these remote sensing techniques to provide objective and quantitative information about the spring environments and ecological processes. The sensors and image analyses employed to address these objectives included: MODIS NDVI time series (annual and seasonal traces of entire wetlands and dominant vegetation types); very high resolution multispectral satellite (detailed delineation of wetland extents using NDVI thresholds) and airborne hyperspectral imagery (detailed discrimination of spring plant communities and surrounding substrate using spectral matching algorithms); supported by concurrent colour digital aerial photography and collection of near-concurrent in-situ ecological and spectral data for image calibration and validation purposes. The main focus of this review paper is to draw synergies from the image analyses and research findings that can uniquely be provided using this suite of image data in combination, over differing temporal and spatial scales, to provide new understanding of the drivers and ecological processes underpinning the springs. The multi-sensor approach revealed for the first time the spatial and temporal responses of these unique ecosystems to changing climatic conditions, land use change and groundwater extractions. Our results reveal that long-term variation is an inherent signature of the wetl ands, with distinct phenological responses for differing vegetation species being driven by seasonal temperature and rainfall. In addition the springs were found to change over short time periods (2-...