Lysimeter assessment of the Simplified Two-Source Energy Balance model and eddy covariance system to estimate vineyard evapotranspiration

Abstract

Estimation of crop water needs plays a key role in the water resource management in arid and semi-arid regions. Actual evapotranspiration (ETa) becomes the key term in both water and energy balances at this point. In this work we focus on vineyard due to the significance of this crop for La Mancha region, Spain, with the greatest concentration of vineyards in the world. Eddy-covariance (EC) technique has been traditionally used for ground observations of ETa. One of the aims of this work is to assess the feasibility of an EC system under the challenging conditions of a small drip-irrigated vineyard in a semi-arid environment.Two-source energy balance modelling allows for ETa estimation, as well as soil evaporation (E) and canopy transpiration (T) partitioning, using radiometric temperatures as a basis. A Simplified version of this Two-Source Energy Balance approach (STSEB) has been previously tested in a variety of crops and environments. The second goal of this paper is to explore now the performance of the STSEB model in a vineyard structure.Two experiments were carried out during the growing season of 2014 and 2015 in a ˜4 ha row-crop drip-irrigated Tempranillo vineyard, in a semi-arid location in Castilla-La Mancha, Spain. As a novelty in this work, a 9-m2 monolithic large weighting lysimeter was available. An eddy-covariance flux tower was deployed together with net radiation and soil heat flux instrumentation. The residual technique was selected to force the closure after an analysis of the energy imbalance based on the comparison EC-Lysimeter. Good agreement between adjusted EC measurements and lysimeter data (RMSE of ±0.09 mm h−1 and ±0.5 mm d−1 at hourly and daily scales, respectively) supports the validity of eddy-covariance technique to monitor turbulent fluxes and accurate ETa in vineyards.A set of 5 thermal-infrared radiometers were assembled to guaranty an appropriate thermal characterization of the vineyard structure required in the STSEB approach. Surface energy fluxes were modeled every hour with average estimation errors lower than ±30 W m−2 for net radiation and soil heat flux, and around ±50 W m−2 for sensible and latent heat fluxes, with systematic deviations lower than 25 W m−2 in all fluxes. Comparison with lysimeter data showed an average underestimation of 0.04 mm d-1 with a RMSE value of ±0.6 mm d-1 in modeled daily ETa. In terms of accumulated ETa for the full experiments, an underestimation of 12% in 2014 and an overestimation of 7% in 2015 were observed. These results reinforce the feasibility of the STSEB approach to monitor hourly, daily and accumulated ETa in row-crops such as vineyards. Although no ground measurements were available to assess the partitioning, separated E–T values were obtained for the full campaigns, showing a significant ratio E/ETa of 0.35−0.40 for soil evaporation.