This paper provides new evidence on the impact of temperatures on tourism demand in the winter season. The analysis is based on time series data spanning from 1960 to 2015 for the South Tyrolean mountains in Italy. Since 1960, winter temperatures have increased by 0.4 degrees Celsius per decade, on average. A nonlinear autoregressive distributed lag (ARDL) model is employed for the estimations. This model allows two separate coefficients to impact tourism demand, following temperature changes (decreases or increases). Results reveal that an increase in winter temperatures by one degree Celsius leads to a decline in the number of accommodation guests (arrivals) by eight per cent, while temperature decreases have no effect on the number of arrivals. However, sensitivity to temperature increases has been declining since the early 1990s, probably due to the widespread usage of snowmaking facilities. The number of these facilities has increased by almost 10 per cent per year on average over the same period. In recent years (1986-2015), and as a consequence of these measures, temperature increases no longer have an effect on winter tourism demand. Conversely, decreases in temperatures lead to small increases in arrivals (by four per cent increase due to a one degree Celsius decrease).