Left: 5-Day composite of the global soil moisture distribution derived from EUMETSAT MetOp-ASCAT data for August 18, 2007. Right: Retrieval noise. Pixels marked in red denote areas with unreliable data due to missing satellite data, mixed pixel effects, or weather and/or vegetation effects. Pixels marked in pink denote areas flagged as snow covered, frozen or saturated with water / flooded. Black areas over land denote lakes or permanent ice. Olive, light and dark blue mark low, medium and high soil moisture or soil moisture retrieval noise, respectively. Blue areas in the right image denote areas where the retrieved soil moisture values should be given less confidence because of a substantial retrieval noise (complex topography, dense vegetation like the rain forest in South America).

Period and temporal resolution:

• Jan. 1, 2007 - Aug. 17, 2011
• daily as 5-day composites

Coverage and spatial resolution:

• Global
• Spatial resolution: about 0.1° x 0.1°, cartesian grid
• Geographic longitude: 179.944°W to 179.944°E
• Geographic latitude: 89.9437°S to 89.9437°N
• Dimension: 1599 rows x 3207 columns
• Altitude: following terrain

Format:

• NetCDF

Original data are organized in 5° x 5° tiles. Each tile comprises a certain number of 12.5 km grid cells. The data are stored as time series per grid cell per tile in binary format.

This data set contains a number of quality flags and additional information. One is the retrieval noise. This is estimated using Gaussian error propagation of the input uncertainties like the variability of the used radar backscatter values due to measurement noise and the variability of the sensitivity of the radar backscatter values to soil moisture for different soil and vegetation types.

We note, that the soil moisture retrieval method used here, is of limited use particularly in polar regions and regions covered by dense rain forest like in South America. These areas are therefore often flagged as unreliable data and/or show a large retrieval noise.

Soil moisture retrieval is not possible for areas covered with snow and ice, for areas with frozen soil and for wetlands/lakes/rivers. In order to allow identification of dubious soil moisture values (which have perhaps not been flagged as being unreliable), the product contains the percentage fractions of snow, frozen soil, and wetlands for every grid cell. Those for wetland are static, assuming that the fraction does not change over time, while those of snow and frozen soil stem from a climatology and therefore vary with season but don't have interannual variation.

Regions of a strongly variable topography are also problematic because of the highly variable local incidence angle in these cases. This causes problems to normalize the measured radar backscatter values to the common incidence angle and thus to retrieve the soil moisture. Therefore, the data set contains the normalized standard deviation of the altitude in each grid cell (in relative units) as a measure of the topographic complexity.

We recommend to check the references (see below) for details.

Wolfgang Wagner

Institute of Photogrammetry and Remote Sensing (IPF)

Vienna University of  Technology (TU Wien)

E-Mail: ww@we dont want spamipf.tuwien.ac.at 

Stefan Kern 

ICDC, CliSAP, Universität Hamburg

E-Mail: stefan.kern@we dont want spamzmaw.de

• Data product sheet
• Definition of quality flags
• Bartalis et al. (2007)
• ASCAT Product Guide

Upon using this data please cite as follows:

ASCAT Soil Moisture 2007-2010, Institute of Photogrammetry and Remote Sensing (IPF), Vienna Institute of Technology (TU Wien), Vienna, Austria,

provided as 5-Day composites by: Integrated Climate Data Centre (ICDC, icdc.zmaw.de), University of Hamburg, Hamburg, Germany.