Scots pine forests subjected to continental Mediterranean climates undergo cold winter temperatures and drought stress. Recent climatic trends towards warmer and drier conditions across the Mediterranean Basin might render some of these pine populations more vulnerable to drought-induced growth decline at the Southernmost limit of the species distribution. We investigated how cold winters and dry growing seasons drive the radial growth of Scots pine subject to continental Mediterranean climates by relating growth to climate variables at local (elevational gradient) and regional (latitudinal gradient) scales. Local climate-growth relationships were quantified on different time scales (5-, 10- and 15-days) to evaluate the relative role of elevation and specific site characteristics. A negative water balance driven by high maximum temperatures in June (low-elevation sites) and July (high-elevation sites) was the major constraint on growth, particularly on a 5- to 10-day time scale. Warm nocturnal conditions in January were associated with wider rings at the high-elevation sites. At the regional scale, Scots pine growth mainly responded positively to July precipitation, with a stronger association at lower elevations and higher latitudes. January minimum temperatures showed similar patterns but played a secondary role as a driver of tree growth. The balance between positive and negative effects of summer precipitation and winter temperature on radial growth depends on elevation and latitude, with low-elevation populations being more prone to suffer drought and heat stress; whereas, high-elevation populations may be favoured by warmer winter conditions. This negative impact of summer heat and drought has increased during the past decades. This interaction between climate and site conditions and local adaptations is therefore decisive for the future performance and persistence of Scots pine populations in continental Mediterranean climates. Forecasting changes in the Scots pine range due to climate change should include this site-related information to obtain more realistic predictions, particularly in Mediterranean rear-edge areas.