Uncertainty assessment of open water evaporation measurements: case study in a tropical reservoir in Brazil

The Brazilian semiarid region covers about 1 million km² and is characterised by irregular rainfall and high potential evaporation (above 2000 mm per year of water loss). To overcome the lack of water, the population depends on man-made reservoirs, which are the main source of water during the dry season (8 months of the year). In the federal state of Ceará there is on average one dam every 5 km, totalling more than 20,000. It is, therefore, noteworthy how the accurate estimation of evaporation would improve the management of scarce water resources. Moreover, climate change scenarios predict impacts on water availability in this region and better accuracy in evaporation assessments are fundamental to foster such simulations, which itself works at an intra-day time resolution. We suggest the most precise measurements possible, close to the lake and preferably on top of the water body. We used between October-December 2019 two direct measurement sensors in floating pans on the reservoir. A newly developed one based on the echo sounder principle and a standard differential pressure meter. Both show that at the reservoir more than 6 mm evaporate per day on average in the dry period. Uncertainty analyses for a single measurement show good performance of the acoustic sensor (± 0.11 mm) compared to the standard sensor (± 0.6 mm), sensitive enough to record the daily course of evaporation. The field measurements show 3.7% of uncertainty comparing the acoustic sensor with the standard pressuremeter. The minimum number of measurements with the acoustic sensor to obtain such uncertainty is 307. In this case, for 1h step, the uncertainty is 1.1%. The wave movement caused by wind influences the quality and frequency of the measurement. These results serve as a basis for future approaches to assess spatial variation of evaporation and wind on the lake. This might highlight the minimal number of instruments needed to solve the issue of spatial variability of evaporation.