Abstract Changes to seasonal precipitation can have dynamical and thermodynamic origins and disentangling these processes is challenging. To evaluate observed changes in UK winter precipitation from 1901 to 2023, we separate the signal into dynamical and nonâdynamical components by applying a dynamical adjustment using European weather patterns. We find longâterm changes in the nonâdynamical component are tied to temperature, with a 7% increase in UK winter precipitation per degree of local warming, in line with expected increases in atmospheric moisture content. We find this significant increase in nonâdynamical winter precipitation is attributable to anthropogenic forcing according to the CMIP6 global climate model simulations. Notably, most CMIP6 models produce a much weaker intensification of the nonâdynamic (thermodynamic) component relative to observations, with a model ensemble mean scaling of 4% °Câ1. The results show that projections based on CMIP6 models could underestimate the scaling of precipitation with temperature and increase in flood risk.