Evaporative losses in a dry tropical region: direct measurements, remote sensing assessment and impact of climate change on water availability in Ceará, Brazil

Abstract

The state of Ceará, in the Brazilian semiarid region, is supplied by a dense network of reservoirs (over 30,000) that are vulnerable to drought and water deficit. Precise knowledge of the evaporation process, which corresponds to roughly 30% to 50% of the water balance, is fundamental for the management of water resources and the assessment of probable impacts of climate change on water availability. This dissertation was developed in different study areas in Ceará. Its first part examines the Gavião reservoir and estimates lake evaporation by using the following four approaches: two equations (Penman and one adaptation of Dalton) and two direct measurement sensors (one based on pressure difference and the other one on acoustic signals). The sensors were installed in floating tanks and the equations drew on variables collected from a meteorological station on board a raft. The second part analyses the impact of climate change on evaporation in the Gavião, Pacoti and Riachão reservoirs and also estimates future water availability (2070-2099) in the Metropolitan Region of Fortaleza, supplied by these reservoirs. Historical evaporation (1961-2005) was estimated using the Penman equation with data from a meteorological station of the National Institute of Meteorology - INMET located in Fortaleza, and the remote sensing algorithm AquaSEBS was applied to Landsat 5 and 8 images to estimate lake evaporation. The impact on water availability was determined by using stochastic modelling. We used simulations of the regional climate models Eta-CanESM2 and Eta-MIROC5, developed by the National Institute for Space Research – INPE. These models generated four climate-change scenarios based on the simulated evaporation rates. This study’s third part is composed of four months of direct measurement of evaporation with a pressure sensor installed in two evaporation tanks. The aim was to gauge the effect of the presence of the floating macrophyte Eichhornia crassipes on free surface evapotranspiration. The main conclusions resulting from this dissertation are: (i) the approaches for evaporation estimation presented satisfactory performance, with emphasis on Dalton's adaptation for hourly steps, and on the acoustic sensor, which presented acceptable values for steps of 4 hours (r > 0.6) or more; (ii) one of the climate scenarios predicts an increase in the evaporative rate (12% by the end of the 21st century in relation to the historical period) and the others show stability (varying from -2% to +2%); this fact highlights the uncertainty in climate modelling even within the same region and data; (iii) the remote sensing algorithm presented acceptable results for estimating evaporation from water bodies, differing by 27% from the assessment made based on data from an on-land station; (iv) evaporation rates were found to be higher (≈ 16%) in the presence of aquatic plants. The findings of this work are important contributions for an accurate monitoring of water losses through evaporation and reservoir operations, particularly in dry regions.