摘要 : This study investigates how extreme precipitation scales with dewpoint temperature across the northeastern United States, both in the observational record (1948-2020) and in a set of downscaled climate projections in the state of ... 展开 This study investigates how extreme precipitation scales with dewpoint temperature across the northeastern United States, both in the observational record (1948-2020) and in a set of downscaled climate projections in the state of Massachusetts (2006-99). Spatiotemporal relationships between dewpoint temperature and extreme precipitation are assessed, and extreme precipitation-temperature scaling rates are evaluated on annual and seasonal scales using nonstationary extreme value analysis for annual maxima and partial duration series, respectively. A hierarchical Bayesian model is then developed to partially pool data across sites and estimate regional scaling rates, with uncertainty. Based on the observations, the estimated annual scaling rate is 5.5% degrees C-1, but this varies by season, with most nonzero scaling rates in summer and fall and the largest rates (similar to 7.3% degrees C-1) in the summer. Dewpoint temperatures and extreme precipitation also exhibit the most consistent regional relationships in the summer and fall. Downscaled climate projections exhibited different scaling rates compared to the observations, ranging between -2.5% and 6.2% degrees C-1 at an annual scale. These scaling rates are related to the consistency between trends in projected precipitation and dewpoint temperature over the twenty-first century. At the seasonal scale, climate models project larger scaling rates for the winter compared to the observations (1.6% degrees C-1). Overall, the observations suggest that extreme daily precipitation in the Northeast only thermodynamic scales with dewpoint temperature in the warm season, but climate projections indicate some degree of scaling is possible in the cold season under warming. Significance StatementA warmer climate will likely result in the intensification of extreme precipitation, with the potential to enhance flood and stormwater risk. However, the relationship between extreme precipitation and temperature (i.e., the precipitation-temperature scaling rate) remains uncertain, particularly at regional scales, inhibiting societal adaptation to extreme events. Using observations and climate projections of daily precipitation and dewpoint temperature across the northeastern United States, we demonstrate that extreme daily precipitation does indeed scale with dewpoint temperature, but the rate of scaling varies by season, with the strongest relationship in the warm season. 收起