Disentangling the main sources of evapotranspiration in a vineyard

Evapotranspiration (ET) is a complex phenomenon that responds to soil water availability, plant development, weather variations and climate change in many magnitudes, from leaf to ecosystem scale. Disentangling the different sources contributing to the total ET at the ecosystem level could contribute to a better understanding of the single process and the overall ET dynamics.

To tackle this goal, we established an Eddy Covariance station in a vineyard in Caldaro, South Tyrol, Italy, where cv. Chardonnay and cv. Sauvignon blanc are cultivated. The vineyard soil is covered by grasses and drip irrigation is available. We attempted to partition the total evapotranspiration (ET_ec) data obtained into the vines transpiration (T_v), the vines’ canopy evaporation (E_v) and the understory evapotranspiration (ET_u), the latter comprising the soil evaporation and the transpiration of the ground-level vegetation. By this Ecophysiological Partitioning Approach (EPA) the ecosystem ET_EPA is the sum of T_v, E_v and ET_u.

T_v was estimated upscaling the sap flow rate measured via Sap Flow sensors (SFM1, ICT International,  Armidale, NSW, Australia; 3 sensors, 1 sensor per plant). ET_u was assessed with 3 transparent soil-ground-flux-chambers and a multiplexer (Li-8100 Licor Biosciences, Lincoln, NE, USA) in 6 campaigns of 72 hours each, with the chambers being moved to a new position every 24 hours to cover time and spatial variability. E_v was assessed by means of three leaf wetness sensors placed within grapevine canopy. All the measurements were set at 30-minutes intervals, to match the frequency of ET_ec.

Preliminary results of this ongoing project, which forsees two years of field measurements, showed that ET_ec amounted to 545 mm during the growing season 2021, with values ranging from 0.33 to 4.83 mm d^-1. ET_ec correlated well with net radiation and with ET_u. All sap flow sensors showed a similar trend across the season, consistent with ET_ec, but differed among each other in terms of flow quantities, likely due to wood specificities of each sampled grapevine which will require specific on-site calibrations.

E_v component, rarely considered in ET partitioning studies, was strongly dependent on precipitation pattern and we hypothesize it can offer a gain of more than 5% in explaining the ET dynamics in the experimental vineyard, wether compared to removing wet canopy moments from the dataset.  

Once the calibration of the soil-ground-flux-chambers system and the installed sap-flow sensor be improved, in-situ measurements of components of ET_EPA will contribute to a computational partitioning approach which improves the comprehension of the dynamics of the ecosystem ET sources under climate change.