River Flow Monitoring by Sentinel-3 OLCI and MODIS: Comparison and Combination

Angelica Tarpanelli,Luca Brocca,Jérôme Benveniste,Marco Restano,Filippo Iodice

doi:10.3390/rs12233867

Abstract

The monitoring of rivers by satellite is an up-to-date subject in hydrological studies as confirmed by the interest of space agencies to finance specific missions that respond to the quantification of surface water flows. We address the problem by using multi-spectral sensors, in the near-infrared (NIR) band, correlating the reflectance ratio between a dry and a wet pixel extracted from a time series of images, the C/M ratio, with five river flow-related variables: water level, river discharge, flow area, mean flow velocity and surface width. The innovative aspect of this study is the use of the Ocean and Land Colour Instrument (OLCI) on board Sentinel-3 satellites, compared to the Moderate Resolution Imaging Spectroradiometer (MODIS) used in previous studies. Our results show that the C/M ratio from OLCI and MODIS is more correlated with the mean flow velocity than with other variables. To improve the number of observations, OLCI and MODIS products are combined into multi-mission time series. The integration provides good quality data at around daily resolution, appropriate for the analysis of the Po River investigated in this study. Finally, the combination of only MODIS products outperforms the other configurations with a frequency slightly lower (~1.8 days).

Highlights

The estimation of river discharge is of primary importance for the hydrology community as it is fundamental: to close the water cycle, for the efficient management of water resources, and for flood forecasting, especially under severe climate change conditions [1,2,3]
We address the problem by using multi-spectral sensors, in the near-infrared (NIR) band, correlating the reflectance ratio between a dry and a wet pixel extracted from a time series of images, the C/M ratio, with five river flow-related variables: water level, river discharge, flow area, mean flow velocity and surface width
In the description of the results, different aspects were taken into account: (1) the analysis of the reflectance of M and C pixels; (2) the use of single and multiple pixels; (3) the sensitivity analysis relating the C/M ratio to the different river flow-related variables; (4) the possibility to combine several reflectance ratio time series into a single C/M time series to be used for monitoring the river flow at each site

Summary

Introduction

The estimation of river discharge is of primary importance for the hydrology community as it is fundamental: to close the water cycle, for the efficient management of water resources, and for flood forecasting, especially under severe climate change conditions [1,2,3]. In the near-infrared (NIR) region, the reflectance is much higher than that in the visible band due to the structure of the leaves. Based on these well-known principles, the NIR region is largely used by virtue of its high potential to discriminate between soil and water. Following studies of Brakenridge et al [15] and successively Tarpanelli et al [13], the ratio between the reflectance of a dry pixel (C for calibration) and a wet pixel (M for measurement), i.e., the C/M ratio, is expected to represent the river variation well both in terms of mean flow velocity and river discharge

Objectives

Methods

Results

Conclusion