Impacts on water availability for vitiviniculture worldwide using different potential evapotranspiration methods

Abstract

Context and purpose of the study ‐ Beyond the sole warming globally perceived and monitored, climate change impacts water availability. Increasing heatwaves frequency observed during the last decades and projected for the 21st century certainly result (or will result)in more water deficit stress for grapevine. Change in water availability throughout the season depends on the balance between precipitation and evapotranspiration. The latter is seldom assessed through potential evapotranspiration (ET0) calculated with empirical formulae relying on air temperature only. This study compares the changes in water availability estimates for viticulture using such formulae in comparison to the reference Penman‐ Monteith approach.

Material and methods – Monthly interpolated minimum and maximum temperature, precipitation and Penman‐Monteith (PM) ET0 data for land surfaces worldwide were collected from the CRU TS4.01 gridded dataset, from 1971 to 2017. Other ET0estimates were produced using the Thornthwaite (T) and the Hargreaves (H) temperature‐based as well as the Modified Hargreaves (M) temperature‐and‐ rainfall‐based methods. PM, T, H, M ET0 data were used to calculate the dryness index (DI), a monthly water balance‐based index for viticulture. Changes between the periods 1971‐2000 (HIST) and 2001‐ 2017 (PRES) in potential evapotranspiration and in DI were compared for each of the 4 ET0calculation methods. The changes were analyzed in wine producing regions using the vineyard geodatabase v1.2.3, a shapefile referencing 691 wine producing regions worldwide.

Results – All 4 methods compute an average increase (from HIST to PRES) in ET0 of about 20 mm during the grapevine growing season, i.e. April to September (October to March) for the northern (southern) hemisphere. The change (PRES ‐ HIST) differ substantially in space, according to the method used. For instance, a decrease in ET0 is shown in southwestern and central North America when using PM method, while T method indicates a weak to moderate raise in ET0 in these regions. Changes in dryness index th st from the late 20 to the early 21 century are large and highly variable in space: from ‐65 mm to +62 mm (0.05 and 0.95 percentiles), according to the location and to the ET0 calculation method. DI also strongly varies in space, but results are less sensitive to ET0 calculation method. PM shows a decrease in DI (PRES ‐ HIST) down to ‐75 mm in most regions but Australia, central Europe and Italy. While PM, H and M methods indicate a clear decrease of DI in France, Portugal and Spain, T method suggests an increase in DI in the northern part of France and in most of Spain. It is concluded that (1) ET0 has risen and contributed to DI decrease in many wine regions worldwide and (2) using T empirical method to derive ET0 from temperature can lead to different conclusions concerning changes in water availability for viticulture