AMSR-E Soil Moisture Validation Efforts in the Australian Arid Zone
Jeffrey P. Walker, Rodger B. Grayson, Rocco Panciera, Xiwu Zhan, and Paul R. Houser
Soil moisture content, a vital land surface parameter throughout a wide range of applications, is one of two land surface products to be produced from the AMSR-E (Advanced Microwave Scanning Radiometer for the Earth observing system) instrument on board NASA's Aqua satellite. Using the brightness temperature data at 6.6GHz, average soil moisture content for the surface 1cm layer will be produced on a 25km x 25km grid. In order for this information to be useful, it must first be "validated". In Australia, validation efforts are underway in the temperate and arid zones. This talk will concentrate on results from the arid zone, which comprises 70% of Australia. Four sub-areas were chosen for this exercise (Sturt Stony Desert, Simpson Desert, Western Lake Eyre Basin and Stuart's Shelf), rather than a single large area, so that a wider range of arid zone land cover and climatic conditions could be covered. Each site was approximately 60km x 60km, so that measurements were made across almost an entire sensor footprint (45km x 70km). Monitoring of each site was undertaken on two different occasions; June and September 2003. Measurements of soil moisture content (1cm thermogravimetric and 6cm Theta probes) and supporting data (fraction of bare soil, rock, grass, shrub and trees, vegetation water content and soil temperature) were made along station tracks and access roads with a spacing of approximately 5km. A base station continuously monitored air and soil temperature throughout the day. As many as 100 GPS located measurements were made across any one site. On several occasions rain fell overnight and a repeat sampling was undertaken on the following day. Validation of the AMSR-E soil moisture product includes direct comparison with the average ground-measured near-surface soil moisture content at each of the four sites on each of the field measurement days throughout the year. AMSR-E and ground measured near-surface soil moisture estimates are also compared with those from other passive microwave remote sensing soil moisture algorithms. Moreover, current algorithms do not make any explicit account for the presence of surface rock, a significant attribute in much of the arid zone landscape. This data set allows the effects of rock cover to be assessed.