Aims: Climate change imposes increasingly warm and dry conditions in most winegrowing regions. Mediterranean vineyards are particularly vulnerable and have registered, in most situations, declining yields over the past years. Although a majority of Mediterranean vineyards are still dry-farmed, yields can be increased by the implementation of irrigation. However, irrigation has an impact on increasingly affected water resources. An alternative solution to irrigation can be the adaptation of training systems. As can be shown by water balance modeling, low density non-irrigated vineyards are much less vulnerable to climatic drought compared to medium or high density vineyards. And while yields tend to be lower in low density vineyards, so are production costs. The aim of this study is to investigate to what extent low density vineyards can be a sustainable and cost effective adaptation for grape growing in dry climates.Methods and results: A water balance model was applied to conceptual vineyards with different soil water holding capacities and different planting densities over recent past (1981-2010) and near future (2041-2070) climatic conditions for two winegrowing scenarios (Cabernet-Sauvignon in Bordeaux and Grenache in Avignon, Côtes du Rhône). Row spacings of 2.0, 3.0 and 4.0 m were investigated for vineyards with 100, 200 and 300 mm total transpirable soil water (TTSW), while inter-vine spacing, vine architecture, and canopy height were kept similar. Projected yields were estimated to vary according to vine density and water deficit based on a meta-analysis of data published in the literature. Production costs were calculated according to an operation-based costing methodology and compared among the different scenarios on a cost per hectare basis. Gross profit per hectare, defined as grape sales revenue minus production costs, was then computed for two grape sale revenue scenarios (1 €/kg and 3 €/kg). The modeled average fraction of transpirable soil water (FTSW) varied across the different winegrowing scenarios, climate periods (recent past or near future), and TTSW and row spacing assumptions. In soils with 200 or 300 mm TTSW, the 30-day average FTSW prior to modeled grape harvest roughly doubled when 4.0 m versus 2.0 m spacing was assumed in both the recent past and near future climate scenarios. In soils with 100 mm TTSW, water deficit was more severe overall and the effect of row spacing on average FTSW was less pronounced. Changes in projected yields were estimated as a function of vine density and FTSW based on relationships published in the literature. Yields decreased with decreasing vine density and increasing water deficits, while production costs decreased with decreasing vine density. When the assumed revenue from grape sales was lower (1 €/kg), the effect of reduced production cost savings outweighed the loss in revenue caused by reduced yields, leading to increased gross profit per hectare. On the other hand, when higher grape revenue was assumed (3 €/kg), the effect of reduced yield on revenue outweighed the associated reduction in production costs, leading to reduced gross profit per hectare.Conclusions: Lower density, dry-farmed vineyards will experience less water deficit under warmer and drier climate conditions, although this difference is less pronounced in soils with less water holding capacity. When considering differences in yields, revenues, and production costs, lower density vineyards producing lower value grapes (1 €/kg) may also experience an associated increase in gross profit, while such vineyards producing higher value grapes (3 €/kg) might experience a decrease in gross profit.Significance and impact of the study: The implementation of dry-farmed, low density vineyards provides a sustainable solution for grape growing by reducing the need for irrigation water. It allows maintaining vineyards in very dry areas where water is not readily available for irrigation and where other crops (except possibly olive trees) cannot be grown. Modeling of yield, revenue, and production costs shows that this solution is also economically viable, particularly for vineyards producing lower value (€/kg) grapes. Unlike goblet trained bush vine, low density trellised vineyards are perfectly adapted for mechanization.