|De Jeu, R|
Submitted to: American Geophysical Union
Publication Type: Abstract Only
Publication Acceptance Date: 1/6/2009
Publication Date: 1/20/2009
Citation: Lakshmi, V., Mladenova, I., Jackson, T.J., Walker, J., Merlin, O., De Jeu, R. 2009. The impact of standing water and irrigation on AMSR-E sensitivity to soil moisture over the NAFE'06 experiment area [abstract]. American Geophysical Union, Fall Meeting Supplements, 89(53):H41D-0908. Interpretive Summary:
Technical Abstract: AMSR-E sensitivity to soil moisture and its accuracy have been studied over a wide variety of surface conditions and weather regimes using both in situ measured data and aircraft derived soil moisture estimates. Several extensive soil moisture field campaigns involving ground and air-borne components have been undertaken in support of validation and development of soil moisture retrieval algorithms, addressing calibration issues, performing accuracy assessment studies etc. The lack of well established soil moisture networks worldwide to a certain extent limit the validation sites to areas located primarily throughout the continental US. In the present research we expand AMSR-E validation over an area with extensive irrigation and standing water (rice fields), located in the Murrumbidgee Catchment, SE Australia, using the National Airborne Field Experiment (NAFE'06) data set and an alternative AMSR-E derived SM product (product developed by the Vrije Universiteit, Amsterdam, the Netherlands) utilizing C-band measured brightness temperature. Using a surface reflectance threshold methodology it was determined that rice fields coverage was greater than 1% over one-third and greater than 4% over one-ninth of the domain. It has been previously shown that as little as 2.5% of standing water at 1 km pixel size can result in worse accuracy than the desired for soil moisture retrieval value of 0.04 v/v. Therefore it was expected that the presence of irrigation and standing water will affect the retrieval algorithm performance and result in lower accuracy over these fields as compared to the rest of the domain. The spatial correlations between AMSR-E and an aircraft product, derived from the Polarimetric L-band Microwave Radiometer flown during the campaign (average R = 0.94 and RMSE = 0.04 v/v), indicated significant AMSR-E sensitivity to changes in soil moisture caused by both precipitation and irrigation. Expectedly the irrigation area was depicted as being the wettest part of the domain. Consistent wetting and drying trends, evident in both products (AMSR-E and PLMR), indicated adequate response to temporal soil moisture variability and excellent agreement for the duration of the campaign and under the NAFE'06 site specific conditions. Most importantly, the AMSR-E SM - PLMR analysis over areas with standing water and irrigation revealed retrieval accuracy better than 0.04 v/v (4% rice field coverage at 25 km pixel size: R = 0.87 and RMSE = 0.02 v/v).