Determining water reservoir characteristics with global elevation data

Abstract

Quantification of human impact on water, sediment, and nutrient fluxes at the global scale demands characterization of reservoirs with an accuracy that is presently unavailable. This letter presents a new method, based on virtual dam placement, to make accurate estimations of area‐volume relationships of large reservoirs, using solely readily available elevation data. The new method is based on regional similarity of area‐volume relationships. The essence of the method is that virtual reservoirs are created in the vicinity of an existing reservoir to derive area‐volume relationships for the existing reservoir. The derived area‐volume relationships reproduced in situ bathymetric data well. An intercomparison for twelve reservoirs resulted in an average R2 = 0.93. This is a significant improvement on estimates using the best existing global regression model, which gives R2 = 0.54 for the same set of reservoirs.

Height–Area–Storage Functional Models for Evaporation-Loss Inclusion in Reservoir-Planning Analysis

“…where h and n are parameters and all the other symbols are as previously defined. The parameters of Equation (16) were also determined by least-squares regression fitting using the available height-storage data.…”

“…Given the popularity with which the GRanD models are used for regional water-resource assessment across the globe [16], it was felt that this study would not be complete without testing its efficacy at the case-study reservoirs. The generalised GRanD volume-area regression model was developed from the Global Reservoir and Dam database using 5824 reservoirs [11].…”

“…It should therefore not be surprising if the function performs poorly when used for the purpose of predicting the at-site area-storage relationship. For example, while Lehner et al [11] recorded an R 2 of 0.8 during the development of Equation 17, the use of the equation to predict the at-site area-storage relationship by van Bemmelen et al [16] only produced an R 2 of 0.54; however, most analysts still use the GRanD equation for the latter purpose. Indeed, the GRanD model is used for this same purpose in the current study to further demonstrate its poor performance in local at-site situations.…”

“…Just like the Pong reservoir in the Himalayas valley, the district experiences two seasonal regimes, monsoon (rainy) and post monsoon (dry). Mean annual rainfall for the catchment is approximately 1260 mm [16]; mean annual net evaporation is 571 mm.…”

Modelo área-volume para a Represa Guarapiranga empregando o índice NDWI

“…Em razão destas variações, o monitoramento dos volumes dos reservatórios é essencialmente importante para regular seu uso. O método áreavolume, utilizado para estimar o volume de água em reservatórios, se baseia na relação entre área e volume, em que a área da lâmina d'água permite estimar o volume de água no reservatório (LEHNER e DÖLL, 2004;RAN e LU, 2012;VAN BEMMELEN et al, 2016;CAI et al, 2016).…”