| Abstract: |
Inundation and consequent anoxic condition induce methane release, which is one of the most potent greenhouse gases. Course-resolution radiometers are effective in mapping inundation over the continental scale of the methane emission. In comparison, the maps derived from high-resolution optical and radar data have restricted use due to clouds and limited coverage. Flood monitoring is another application of the radiometer-based inundation maps, when combined with high-resolution topography information. The goal of this work is to evaluate and improve the fidelity of the inundation extent derived from the brightness temperature observations of the L-band Soil Moisture Active Passive (SMAP) mission, to explore the L-band capabilities to penetrate clouds and vegetation at 3-day revisit.
Our previous evaluations using the high-resolution maps made with Radarsat and Landsat data reported the satisfactory difference of ~5 % (mean) and ~ 15 % (standard deviation) [Kim et al., 2017]. The evaluations were limited to Canadian Prairies and California Delta over five temporal instances because of the availability of the concurrent data sets. To overcome the limitation, this work compares the waterbody extents derived with SMAP synthetic aperture radar (SAR) and radiometer. This allows the examination over the global landcover, including dense vegetation regions, and over the 2.5 months of the SAR availability. The SMAP SAR provided sigma0 data with 3km resolution (3-day repeat). The preliminary comparison (Figure 1) shows the overall agreement in water signatures among 36km SMAP radiometer, 3km SMAP SAR, and 10m Radarsat. The SMAP SAR results are based on the multiple co-pol thresholding method [Kim et al., 2016]. Its waterbody map in Figure 1 was derived using the pre-launch threshold values, which may explain the underestimation of the extent by the SMAP radar data, with respect to the Radarsat map.
Future efforts are as follows. First, the SMAP SAR inundation maps will be calibrated by revising the threshold values. Second, sharpening sigma0 from 3km to 1km or 500m is feasible through temporal averaging. The higher resolution maps will help address the effects of antenna pattern on the radiometer-based water maps. Last, the SMAP’s L-band penetrates the vegetation layer better than X- or Ku-bands that were used in the existing inundation maps, and thus may identify the inundation under the vegetation more effectively. We will examine this capability by deriving water maps with 3km resolution or higher, and compare with the SMAP-radiometer maps.
Kim, S.B., B. Brisco, and H. Wu, Inundation extent monitored with SMAP data for carbon studies, in Proc. IGARSS, Fort Worth, 2017.
Kim, S.B., J. Oullette, J.J. van Zyl, and J.T. Johnson, Dual-copolarized approach to detect surface water extent using L-land radar for the Soil Moisture Active Passive Mission, IEEE Trans. Geosci. Remote Sens., 54, 3388-3399, 10.1109/TGRS.2016.2517010, 2016.
|
