Aim Drought stress has focused on water availability during the growing season, thus primarily on summer. However, variation in rainfall continentality can produce striking vegetation differences. We aim to disentangle summer water balance from winter rainfall continentality, to better understand how climate regulates the distributions of woody plants in the western USA. Location Western USA. Time period Actual. Major taxa studied Angiosperms and conifers. Methods We used redundancy analysis (RDA) to investigate correlations between rainfall continentality, summer water balance, minimum winter temperature and length of growing season on the distributions of 130 tree and shrub species in 467 plots. Rainfall continentality was calculated using the Gams index, modified for winter precipitation, and summer water balance with the ratio of summer precipitation to temperature. We estimated actual evapotranspiration (AET), deficit (DEF), mean annual temperature and rainfall from global gridded data sets and correlated them with RDA axes. Results Rainfall continentality measured with the Gams index and minimum temperatures best explained the contrast between oceanic vegetation in the Pacific Coast Ranges and continental vegetation in the Intermountain Region and Rocky Mountains. Growing season length (GSL) was the second strongest factor correlated with vegetation distributions. Summer water balance, despite being the most widely used climatic factor to assess drought stress in biogeography, was the third strongest factor correlating with vegetation classes of the western US. AET was equally correlated with RDA axes 1 and 3, and, thus, could not discriminate between the contrasts in the RDA. Main conclusions Rainfall continentality measured with the winter Gams index provides a more precise metric than summer water balance for understanding the biogeography of woody plants in the western USA. Broadly integrating the Gams index of continentality into plant distributions may improve our understanding of biogeographical distributions and predictions of responses to climate change.< Réduire