ASCAT SCATTEROMETER OCEAN CALIBRATION

A new scatterometer, the so-called Advanced scatterometer (ASCAT), onboard MetOp-A satellite was successfully launched on October 19 2006. During the commissioning phase period one of the main goals is to accurately calibrate the instrument. Due to an unexpected delay in the set-up of the calibration transponders, calibration over the ocean played a central role in the calibration of ASCAT during commissioning. Using ocean calibration corrections, ASCAT winds are produced routinely at KNMI since March 2007 as the first MetOp-A geophysical product. Ocean calibration results and scatterometer wind speed statistics show that the KNMI wind product is of high quality.

The absolute calibration of the backscatter signal of a scatterometer is essential for the retrieval of optimum quality geophysical products. It is not simple to obtain an accurate absolute calibration for a scatterometer. For the calibration of the ASCAT scatterometer EUMETSAT performs the absolute calibration by a controlled radar return to the scatterometer from transponders when these are illuminated by one of the six scatterometer radar beams. In addition, an inter-beam comparison is planned over sea ice and rain forest where the radar cross section is known to be very stable and rather time independent. Furthermore, the incidence angle response is known to be smooth over sea ice and rain forest. However, calibration procedures over the ocean have the advantage that they may be applied over a large portion of the globe and consequently may provide accurate results over a relatively short period.

An important tool for ASCAT inter-beam calibration is the visualization of triplets of radar backscatter in the 3-dimensional measurement space. Every Wind Vector Cell (WVC) is illuminated by three antenna beams at different azimuth angles, For a given WVC number, i.e., position across the swath, it is shown that the ASCAT measured triplets are distributed around a well-defined “conical” surface and hence that the signal largely depends on just two geophysical parameters, i.e., wind speed and direction. Such cone visualization in the measurement space by the C-band Geophysical Model Function (GMF) CMOD5, i.e., a function that relates sea surface wind vector to backscatter, can in turn be used for ASCAT calibration. For ASCAT calibration, the same GMF can be used as for its predecessor, the European Remote-sensing Satellite (ERS), since both ERS scatterometer and ASCAT are C-band vertically-polarized fan-beam antenna systems. That is, for coincident ERS/ASCAT incidence angle ranges the ASCAT triplets are expected to be distributed around the cone in the same way as for the ERS scatterometer. Systematic displacements of the cloud of triplets in any direction of the 3D space are mainly due to absolute beam biases, which can be adequately removed, firstly by examining the distribution of the cloud around the wind cone, and secondly by adjusting the mean retrieved ASCAT wind speeds.

Within the framework of the Ocean & Sea Ice (OSI) Satellite Application Facility (SAF), KNMI has developed another ocean calibration tool, which can handle both real and simulated data. Simulations are useful to assess the accuracy of the method. As such, a simulation run with realistic “true” wind distribution and realistic measurement and NWP wind-component error values is performed. The simulations show that the impact of the wind component errors on the calibration results is large as compared to the impact of the measurement errors. Some absolute ocean calibration differences between real and simulated data still need to be further investigated, but relative beam calibration can be done with confidence. The calibration method is based on Fourier analysis of the data. The method consists of comparing the average measured backscatter from the antennae to the simulated backscatter from collocated Numerical Weather Prediction (NWP) vector winds, to assess the absolute values of the measurements and to show inter-beam biases. To simulate the backscatter, a forward model or GMF is needed. This ocean calibration needs only 6 hours of data to produce accurate results and is therefore also very suitable for monitoring purposes.

Currently a 3-transponder calibration campaign is being carried out in order to calibrate the ASCAT antenna backscatter. The KNMI winds have a pre-operational status, awaiting the final calibration results in February 2008. Both the ocean calibration and visualization tools are complementary and, as such, consistency between both of them is checked to improve the accuracy of the calibration.