Articles | Volume 21, issue 12
Hydrol. Earth Syst. Sci., 21, 6445–6459, 2017
Hydrol. Earth Syst. Sci., 21, 6445–6459, 2017

Research article 18 Dec 2017

Research article | 18 Dec 2017

Monitoring small reservoirs' storage with satellite remote sensing in inaccessible areas

Nicolas Avisse1, Amaury Tilmant1, Marc François Müller2, and Hua Zhang3 Nicolas Avisse et al.
  • 1Department of Civil Engineering and Water Engineering, Université Laval, Québec, QC G1V 0A6, Canada
  • 2Department of Civil & Environmental Engineering & Earth Science, University of Notre Dame, Notre Dame, IN 46556, USA
  • 3Department of Engineering, School of Engineering and Computing Sciences, Texas A & M University – Corpus Christi, Corpus Christi, TX 78412, USA

Abstract. In river basins with water storage facilities, the availability of regularly updated information on reservoir level and capacity is of paramount importance for the effective management of those systems. However, for the vast majority of reservoirs around the world, storage levels are either not measured or not readily available due to financial, political, or legal considerations. This paper proposes a novel approach using Landsat imagery and digital elevation models (DEMs) to retrieve information on storage variations in any inaccessible region. Unlike existing approaches, the method does not require any in situ measurement and is appropriate for monitoring small, and often undocumented, irrigation reservoirs. It consists of three recovery steps: (i) a 2-D dynamic classification of Landsat spectral band information to quantify the surface area of water, (ii) a statistical correction of DEM data to characterize the topography of each reservoir, and (iii) a 3-D reconstruction algorithm to correct for clouds and Landsat 7 Scan Line Corrector failure. The method is applied to quantify reservoir storage in the Yarmouk basin in southern Syria, where ground monitoring is impeded by the ongoing civil war. It is validated against available in situ measurements in neighbouring Jordanian reservoirs. Coefficients of determination range from 0.69 to 0.84, and the normalized root-mean-square error from 10 to 16 % for storage estimations on six Jordanian reservoirs with maximal water surface areas ranging from 0.59 to 3.79 km2.

Short summary
Information on small reservoir storage is crucial for water management in a river basin. However, it is most of the time not freely available in remote, ungauged, or conflict-torn areas. We propose a novel approach using satellite imagery information only to quantitatively estimate storage variations in such inaccessible areas. We apply the method to southern Syria, where ground monitoring is impeded by the ongoing civil war, and validate it against in situ measurements in neighbouring Jordan.