Orbital river gauge from optical and passive microwave radiometry. A comparison of capabilities

Abstract. Climate change is impacting our everyday life and is altering the magnitude and frequency of weather-related extreme events such as precipitation, floods, and drought. For this reason, there is a great demand to study these phenomena in a changing environment on a large global spatial extent.
In the past, we developed a protocol for river gauge measurements using satellite passive microwave radiometer (PMR) data and applied over multiple river watersheds [2,1]. 
Exploiting the high sensitivity of microwave emission to water presence we use low frequency L-band (1–2 GHz) passive microwave radiometry (PMR) to monitor rivers and reservoirs and to compare over different microwave frequencies and polarization configurations. We successfully applied the methodology to ESA Soil Moisture and Sea Salinity (SMOS) sensor data reaching high correlation to in-situ discharge measurements over various river basins in different climate. 
In this paper, we compare the capabilities of PMR and optical EO to observe river flow over the Amazon and Niger-Volta watershed. We tested the Moderate Resolution Imaging Spectroradiometer (MODIS), a low resolution (250 m – 5 km) NASA satellite data to derive hydrological time series. To understand the performance of optical orbital river gauge from MODIS comparable to the PMR gauge measurements we analyzed 8-day composite of Terra MOD09A1 in Google Earth Engine environment. The product is surface spectral reflectance of Bands 1 through 7 corrected for atmospheric conditions and averaged over an 8-day. 
Results over tropical regions showed a significant obstacle of cloud cover for optical data (r2: 0.52, std: 0.33), where PMR has the potential to measure river streamflow (r2: 0.82, std: 0.12). Yet over regions with less clouds both optical and PMR can be good alternative to in-situ streamflow ground measurements.