Processing of the AMSR-E/AMSR2 Unified data set includes both an updated Normalized Polarization Difference (NPD) algorithm and the first standard Single Channel Algorithm (SCA). NPD and SCA codes are integrated together to create consistent AMSR-E and AMSR2 soil moisture products. Further information regarding the algorithms will be available in the Algorithm Theoretical Basis Document when it is completed

NPD Algorithm

The NPD soil moisture algorithm uses Polarization Ratios (PR), which are sometimes called normalized polarization differences of the AMSR channel brightness temperatures. PR is the difference between the vertical and horizontal brightness temperatures at a given frequency divided by their sum. This effectively eliminates or reduces surface temperature effects, which is necessary since no dynamic ancillary surface temperature data are input to the algorithm. The algorithm first computes a vegetation/roughness parameter g using PR 10.7 GHz and PR 18.7 GHz, plus three empirical coefficients. Soil moisture is then computed using departures of PR 10.7 GHz from a baseline value, plus four additional coefficients. The baseline values for PR 10.7 GHz are based on monthly minima at each grid cell over an annual cycle.

The vegetation/roughness parameter g incorporates effects of vegetation and roughness together, because both have the same functional form (exponential) in their influence on the normalized polarization differences in the simplified model used in the retrieval algorithm. g may be interpreted as an equivalent vegetation water content with units of kg m^-2. In a desert with no vegetation, any value of g greater than zero is due to roughness only. The value of g reflects the influence of roughness on the polarization ratio as if equivalent vegetation of amount g (kg m^-2) were present. If the surface were smooth everywhere, then g would equal the vegetation water content in kg m^-2 since the roughness contribution would be zero. Vegetation water content and roughness cannot be determined independently from g, and it is computed primarily as a lumped correction factor for the soil moisture retrieval (Njoku and Chan 2006).

Refer to Njoku et al. (2004) for an assessment of calibration biases over land, and methods used to correct these biases.

SCA Algorithm

The SCA approach uses horizontally polarized (h-pol) brightness temperature observations from the lowest frequency channel due to its highest sensitivity to soil moisture observations. The SCA approach is based on the simplified radiative transfer model developed under the assumption of equal canopy and soil temperature (Jackson 1993). In the SCA approach, brightness temperatures are converted to emissivity using a surrogate for the effective physical temperature (T) of the emitting layer. The derived emissivity (e_obs) is corrected for vegetation and surface roughness to obtain the smooth soil emissivity (e_smooth). The Fresnel equation is then used to determine the dielectric constant of the soil-water mixture (k). Finally, a dielectric mixing model is used to obtain the soil moisture (SM).

Processing Steps

AMSR-U consists of the completed AMSR-E dataset and the ongoing AMSR2 dataset, processed using the JAXA L1R as input and joint algorithms (NPD and SCA) for processing. AMSR-U consists of the completed AMSR-E dataset and the ongoing AMSR2 dataset, processed using the JAXA L1R as input and joint algorithms (NPD and SCA) for processing. All L1R input data is ingested via internet transfer to the Science Investigator-led Processing System (SIPS) at the Global Hydrology Resource Center Distributed Active Archive Center (GHRC). The granule is defined as one-half of one orbit, the division being at the poles, so that a granule is descending (North Pole to South Pole) or ascending (South Pole to North Pole). There are approximately 29 Level 1R granules per day. The AMSR-U Level 2B land product is generated at the AMSR SIPS and includes metadata and browse imagery. These data are re-mapped to a 25 km Equal-Area Scalable Earth Grid (EASE-Grid).  Each granule is stored in HDF-EOS5. More detailed information is forthcoming.

Error Sources

Quality Assessment

Each HDF-EOS file contains core metadata with Quality Assessment (QA) metadata flags, RetrievalQualityFlagNPD and RetrievalQualityFlagSCA, that are set by the SIPS-GHRC prior to delivery to NSIDC. A separate metadata file (.xml file extension) is also delivered to NSIDC with the HDF-EOS file; it contains the same information as the core metadata. Three levels of QA are conducted with the AMSR-E Level 2: automatic, operational, and science QA. If a product does not fail QA, it is ready to be used for higher-level-processing, browse generation, active science QA, archive, and distribution. If a granule fails QA, SIPS does not send the granule to NSIDC until it is reprocessed (Conway 2002).