Remote Sensing with physical models of soil moisture status to monitor land degradation and drought
Tim R Mc Vicar, David LB Jupp, Joe Walker
CSIRO Division of Water Resources, Canberra, Australia
Tian Guoliang
Institute of Remote Sensing Application, Chinese Academy of Sciences, Beijing, China
Intoduction
The Murray-Darling Basin (MDB), which covers over one million square km of Southeast Australia, has undergone extensive clearing for cropping and grazing, this and the use of ground water in irrigation alters the hydrologic balance. Reduced regional transpiration, due to land use change, increase recharge into local and regional ground waters. The subsequent rise in water tables causes water logging and has mobilizes soil salts, leading to dry land salinisation. Temporal change in soil moisture is a key parameter in locating sites of such land degradation, as well as in monitoring the effects of drought. Drought physically occurs when regional soil moisture stores fall below a certain threshold (Sivakumar 1991). Although this definition ignores the interactions with landuse and economics that characterize drought in social terms, accurate information on temporal and spatial changes in regional soil moisture stores and their relation to current landuse is critical for successful drought monitoring and declaration. Consequently, using Remote Sensing to monitor regional soil moisture has significant application for land degradation, drought and a variety of other environmental management issues such as crop forecasting, flood hazard and fire risk.
Regional water balance and soil moisture status may be studies using meteorological (met) data obtained from the sparse network of stations. The water balance model used may range from a simple index based on past rainfall, an Antecedent Precipitation Index (API), to complex models of water movement between soil layers and groundwater. Those used in drought studies in Australia have been summarized by White (1990). The difficulty with such modeling is that in simple models the parameters are aggregate and not easily associated with measurable field data, while complex models contain a large number of poorly or in-determined parameters. In each case, the problem is the lack of objective and goal-directed validation. Our aim is to use Remote Sensing to calibrate and validate regional water balance models and to disaggregate the regional soil moisture information spatially. This is done by capitalizing on the nature of remotely sensed observations of the energy balance as a means of validation, and its spatial image format as a tool for bringing the resulting information into coincidence with current land use.
The water and energy balances are coupled through evapotranspiration (ET) in which both water and energy are transported. This coupling is defined regionally by a landcover dependent “Operating Characteristic’ of the form:
E/Ep=f(W)
Where E is the (daily) ET, Ep is the potential ET and W is the available soil moisture. The ratio is called the ‘moisture availability’ (m
a). It is possible to generate daytime and diurnal surface temperature indices using suitable process model as a function of met data, assuming that W is known (Jupp et al. 1990a). Modified resistance energy balance models (REBM) can be used for the daytime data (Jupp, 1990) and the heat flow equation must be solved for the diurnal surface temperature range (van de Griend et al. 1985). The moisture availability derived from the water balance model is a key input to the energy balance models, which conversely may be used with a time series of observed remotely sensed surface temperatures to calibrate and/or validate the temporal series of moisture information. Such a time series is available for the MDB using remotely sensed AVHRR data.
Physically based Environmental Modeling
Spatial and temporal variation in both day and night AVHRR thermal data is due to many environmental factors, including vegetation, soils, geology, topographic influences, the presence of water and ET, (Jupp et al. 1990a). The regional water balance is governed by two general factors: capacitance (Storage of water in the soil) and resistance (loss of water to the atmosphere). Variation in daytime surface temperature as measured using AVHRR data is largely due to surface resistance to loss of water by ET. Nighttime surface temperature, when the air and surface temperatures have equilibrated, responds strongly to the amount of stored water (or capacity) present in the environment (Jupp et al. 1990a; van de Greind et al. 1985).
