Emerging operational applications of spaceborne scatterometers for land surface monitoring

Emerging operational applications of spaceborne scatterometers for land surface monitoring


Volkmar Wismann
Institute for Applied Remote Sensing
Am Josefsbergle 2, D-79100 Freiburg, Germany

Spaceborne scatterometers are active microwave instruments for measuring the normalized radar cross section (NRCS) of the Earth surface. These measurements are independent of cloud coverage and illumination by the sun and provide a global coverage within 2 to 4 days. Thus they are well suited for a wide range of operational monitoring tasks. Presently, scatterometers are exclusively dedicated to the determination of the wind speed and direction over the oceans. However, it is increasingly acknowledged that, despite the coarse resolution, a variety of geophysical parameters can be measured and monitored over land surfaces and sea ice. As one consequence the European Organisation for the Exploitation of Meteorological Satellites (EUMETSAT) starts to develop operational land surface applications for the data of the Advanced Scatterometer (ASCAT) aboard the Metop satellite series. The launch of Metop-1 is scheduled for 2003. This paper gives a review of recent developments for obtaining information on snow properties, thawing of soils, soil moisture, and vegetation from spaceborne scatterometer data. On Greenland, summer melting of snow is common in large areas and its extent and intensity are expected to respond immediately to climate change. Volume scattering is the dominating backscattering mechanism for snow with volumetric moisture of less than 3%. The free liquid water in the snow causes high dielectric loss, which increases the absorption coefficient. Therefore, the normalized radar cross section of snow decreases with increasing snow wetness when the snow starts melting. NRCS measurements over Greenland were used to detect and monitor snowmelt for the period August 1991 to September 2000. The produced maps of snowmelt show considerable interannual variations in terms of melt extent and intensity. In regions without annual snowmelt the accumulation of dry snow can be monitored by measuring a temporal decrease of the NRCS, which can be attributed to increasing attenuation of the microwave signal due to a growing dry snow layer. Accumulation rates delineated from such NRCS data show good agreement with conventional precipitation data.

Information on the state (frozen/thawed) of soil and vegetation is of great interest for climate modeling and understanding the fluxes of heat, moisture and geo-chemical parameters between the land surface and the atmosphere. Radar measurements have a great potential to provide this information, because at microwave frequencies the dielectric properties vary considerably when going from the frozen to the thawed state. A technique was developed to retrieve the date of the transition from the frozen to thawed state based on measurements of the NRCS of the Earth’s surface by the ERS scatterometers. This technique was applied to the data obtained over Siberia for the years 1992 to 2000. The maps of the date of thawing show considerable temporal and spatial variations. This date varies locally by 1 month and also the extent of the thawed area varies considerably by up to 3 x 106 km2 from year to year for a given date.

The dielectric properties of soil strongly depend on its moisture content. However, due to the limited penetration depth of microwaves into the soil, scatterometer can only provide moisture estimates in the soil surface layer (0.5 – 5 cm). This layer is the interface between the highly dynamic atmosphere and the deeper soil layers, and hence its moisture content is subject to short-term fluctuations on temporal scales of less than one day, depending on the type of soil. Maps of the Iberian Peninsula were produced which show the spatial and temporal variability of soil moisture, which reveal the seasonal rain variability as well as interannual variations, e.g., droughts in 1992 and 1993 and flooding in winter 1995/1996. For vegetated areas the radar backscatter is governed by a combination of volume and surface scattering, whereby the amount of volume scattering depends on the canopy coverage and its density. A method, which combines the strength of the NRCS and its dependence on incidence angle, has been developed and was used to produce a map of canopy density for Siberia. This map reveals a lot of details, e.g., the dense forest along the Ural mountains, especially on the western slope, the variability of the vegetation along the river Ob in the West Siberian Lowland, and different forest types and densities in the Central Siberian Plateau. Also the small and isolated forests north of Semipalansk (51ON; 80OE) were detected.