A temporal stability analysis of the Australian SMAP mission validation site

The National Aeronautics and Space Administration’s (NASA) Soil Moisture Active Passive (SMAP) mission is a soil moisture dedicated mission scheduled for launch in October 2014. The payload consists of a combined L-band radar and radiometer system with the objective of mapping near surface soil moisture globally. The scientific rationale for SMAP is an improved accuracy and spatial resolution of the soil moisture estimates through a unique combination of high resolution (3 km) but noisy radar derived soil moisture information and more accurate yet lower resolution (36 km) radiometer derived soil moisture information, yielding a 9 km active/passive soil moisture product. Soil moisture estimates from remote sensing data have an inherent uncertainty due to the model required to convert the observed microwave signal into area-average soil moisture. Validation of the satellite products is thus required after launch to ensure their accuracy. This can be difficult due to the high spatio-temporal variability of soil moisture and the mismatch in scale of point measurements and satellite footprints. One method of overcoming this problem is to use hydrological monitoring networks that have been collecting data over a long period, by finding point measurement locations that are representative of the area-average values. This technique is known as temporal stability analysis. As part of the Australian contribution to the SMAP mission, a calibration/validation site has been developed at Yanco, New South Wales. The Yanco site is part of the OzNet soil moisture monitoring network and has been intensively monitored for remote sensing purposes since 2001 (www.oznet.org.au). This paper presents a temporal stability analysis using data from the Yanco site to assess the most suitable monitoring stations to be used as the test-bed for SMAP. The soil moisture stations are located in a specifically designed nested grid to assist with the validation of the 3 km, 9 km and 36 km SMAP soil moisture products. Results show that in the cropping area, YA4 and YA7 (3 km × 3 km) are best represented by YA4b and YA7d stations, respectively. A high standard deviation in soil moisture is found in the YA (9 km × 9 km) due irrigation and cropping practices. On the other hand, the grassland area YB7 (3 km × 3 km) is well represented by the YB7d station, with the YB3 station giving a good approximation of the average soil moisture for the entire YB area (9 km × 9 km) and the Y10 station a good approximation for the entire 36 km × 36 km area. A sensitivity analysis was then performed supporting the choice of the Y10 station and leading to the conclusion that approximately 1 year of data are required to determine the most representative station within a given area. Data from the Y10 station has been also compared to the average moisture from a number of OzNet monitoring stations and was found to provide an accurate estimate of their average value for soil moisture lower than 0.3 m 3 /m 3 .