SAR Sub-Band Splitting for Reducing Ionospheric Effects on L-Band InSAR Measurements
Li, Zhenhong1; Derauw, Dominique2; Feng, Wanpeng1
1University of Glasgow, UNITED KINGDOM; 2Centre Spatial de Liège, BELGIUM

The main advantage of L-band InSAR is that its coherence is generally better than C-band, particularly in heavily vegetated areas, due to its capability to penetrate more deeply in vegetation. However, since the ionosphere is dispersive, its impact on microwaves depends on the frequency of the signal and is ~17 times greater at L-band than at C-band. This indicates that caution needs to be exercised when interpreting L-band interferograms and ionospheric correction models should be explored.

Based on our intensive experience in InSAR atmospheric correction models and split-band InSAR, a split-band-based InSAR ionospheric correction model is investigated and validated using ALOS PALSAR data. Firstly, starting from an ALOS PALSAR interferometric pair, we split the full band into sub-bands and generate lower resolution images from each acquisition. Secondly, all the sub-images are interferometrically processed and multiple sub-band interferograms are generated, each being centred on slightly different frequencies; Thirdly, sub-band interferograms are combined into independent pairs; and finally, a relative total electron content (TEC) map is estimated using the differential dispersion in the sub-bands. The unique feature of this split-band-based technique is multiple sub-bands are generated from each acquisition and optimally combined to determine TEC maps.

Although the resultant TEC map has a reduced spatial (range) resolution because of the smaller bandwidth (and necessary spatial smoothing), ionospheric signals with a wavelength greater than tens of metres on L-band interferograms can be reduced. It is believed that the TEC maps generated using this split-band-based technique can also significantly benefit ionospheric studies.