Unraveling evapotranspiration dynamics and processes in tropical Andean tussock grasslands

Abstract

The páramo biome provides water resources for many cities in the Andes. These resources are used for drinking water, irrigation, hydropower generation and for sustaining aquatic ecosystems. Notwithstanding mountainous terrains place difficulties for their study, due to its remoteness and data scarcity, knowledge about the functioning of this biome has improved lately. Precipitation (P) and runoff monitoring has increased dramatically, but this was not the case for evapotranspiration (ETa). In order to understand the components of the hydrological cycle, this study aimed at understanding the evapotranspiration process of this important biome by pursuing the following three objectives: (1) to quantify interception, transpiration and their contribution to evapotranspiration, (2) to find suitable methods for measuring and estimating evapotranspiration, and (3) to investigate the controls on evapotranspiration. Results show the high contribution of interception to the evapotranspiration process. The maximum capacity of tussock grasslands to intercept water was 2 mm. During small events (P < 2 mm), between 100 and 80 % of precipitation was intercepted and released back to the atmosphere as vapour; while during large events (P > 2 mm), interception loss decreased from 80 to 10 %. Interception was mainly driven by precipitation amount and secondary by relative humidity. During dry periods, transpiration rates were on average 1.7 mm/day (ranging between 0.7 and 2.7 mm/day) and on top, the fog and dew harvested by the vegetation contributed to the evapotranspiration in around 30 %. For measuring evapotranspiration, the eddy-covariance method is considered the most accurate and with the highest resolution. However, given the high cost of the method, complex installation, operation and maintenance, two hydrological models (HBV-light and PDM) and the calibrated Penman-Monteith equation were found robust alternative methods for the daily estimation of evapotranspiration. These alternative methods are accurate (Pearson’s correlation coefficient > 0.7 and bias percentage < 20 %), freely available and easy to implement. This study also showed that the commonly used water balance method was not suitable for estimating evapotranspiration at daily or monthly scale. Finally, it was found that the páramo biome has a relatively low evapotranspiration rate (annual ETa/P = 0.5) and is an energy-limited site, where net radiation is the primer control on evapotranspiration (annual ETa/Rn = 0.47). The secondary controls were wind speed, surface and aerodynamic conductance, especially important during dry periods.