Using atmospheric and statistical models to understand local climate and assess spatial temperature variability at a fine scale over the Stellenbosch wine district, South Africa

Atmospheric and statistical models were used to increase understanding of potential climatic impacts, resulting from mesoscale physical processes that cause significant temperature variability for viticulture within the Stellenbosch Wine of Origin district. Hourly temperature values from 16 automatic weather stations and 40 tinytag data loggers located in the vineyards were analysed. The 5th of March 2009 was selected as an example to study the cooling potential of the terroirs in radiative weather conditions during grape ripening time. Differences reached more than 10°C between vineyards and can be considered as significant for viticulture. Numerical simulations using the Regional Atmospheric Modeling System were performed. Results for a horizontal grid resolution of 200 m over the Stellenbosch wine region for the 5th of March 2009 showed that the temperature difference was due to cool air accumulation with land and downslope breezes. Surface temperature data recorded in the vineyards were used to produce, by means of multicriteria statistical modelling, which took environmental factors into account, a map of spatial distribution of the daily minimum temperature at a fine scale (90 m). The use of the two models represented an interesting tool to help in identifying the cooling potential of locations for viticulture and, at a later stage, studying the impacts of climate change at fine scales.