Detection of inland water bodies under dense biomass by CYGNSS

Abstract

The reliable detection of inland water bodies under dense biomass is not possible with traditional spaceborne remote sensing systems. In this investigation, we demonstrate the capability of the National Aeronautics and Space Administration (NASA) Cyclone Global Navigation Satellite System (CYGNSS) mission to detect and quantify the extent of small water bodies under dense biomass up to 400 ton/ha in complex heterogenous scenes such as tropical rainforests. A novel inland water body product based on L1 CYGNSS data is developed using a combination of peak surface reflectivity Γp and its spread with respect to time delay and Doppler shift, as characterized by the power ratio Pratio. Only L1 CYGNSS mission data has the potential to study Earth's surface water dynamics because raw Intermediate Frequency (IF) data availability is limited. Several coherence indices are used to verify that the L-band Global Positioning System (GPS) signals are being coherently reflected by surface water bodies and attenuated by the intervening vegetation canopy. These indices include full and computationally fast versions of the entropy associated with Delay Doppler Maps (DDMs), Efull and Efast, as well as the time derivative of the phase φpeak of the reflected GPS signal. The capability to accurately resolve surface water extent under dense vegetation has been shown to improve the estimation by current state-of-art surface water extension datasets used by global methane CH4 emission models in wetlands.