Abstract: The human impacts of changes in heat events depend on changes in the joint behavior of temperature and humidity. Little is currently known about these complex joint changes, either in observations or projections from general circulation models (GCMs). Further, GCMs do not fully reproduce the observed joint distribution, implying a need for simulation methods that combine information from GCMs with observations for use in impact studies. We present an observation-based, conditional quantile mapping approach for the simulation of future temperature and humidity. A temperature simulation is first produced by transforming historical temperature observations to include projected changes in the mean and temporal covariance structure from a GCM. Next, a humidity simulation is produced by transforming humidity observations to account for projected changes in the conditional humidity distribution given temperature, using a quantile regression model. We use the Community Earth System Model Large Ensemble (CESM1-LE) to estimate future changes in summertime (June–August) temperature and humidity over the continental United States (CONUS), and then use the proposed method to create future simulations of temperature and humidity at stations in the Global Summary of the Day dataset. We find that CESM1-LE projects decreases in summertime humidity across CONUS for a given deviation in temperature from the forced trend, but increases in the risk of high dewpoint on historically hot days. In comparison with raw CESM1-LE output, our observation-based simulation largely projects smaller changes in the future risk of either high or low humidity on days with historically warm temperatures.

Abstract Estimates of the climate response to anthropogenic forcing contain irreducible uncertainty due to the presence of internal variability. Accurate quantification of this uncertainty is critical for both contextualizing historical trends and determining the spread of climate projections. The contribution of internal variability to uncertainty in trends can be estimated in models as the spread across an initial condition ensemble. However, internal variability simulated by a model may be inconsistent with observations due to model biases. Here, statistical resampling methods are applied to observations in order to quantify uncertainty in historical 50-yr (1966–2015) winter near-surface air temperature trends over North America related to incomplete sampling of internal variability. This estimate is compared with the simulated trend uncertainty in the NCAR CESM1 Large Ensemble (LENS). The comparison suggests that uncertainty in trends due to internal variability is largely overestimated in LENS, which has an average amplification of variability of 32% across North America. The amplification of variability is greatest in the western United States and Alaska. The observationally derived estimate of trend uncertainty is combined with the forced signal from LENS to produce an “Observational Large Ensemble” (OLENS). The members of OLENS indicate the range of observationally constrained, spatially consistent temperature trends that could have been observed over the past 50 years if a different sequence of internal variability had unfolded. The smaller trend uncertainty in OLENS suggests that is easier to detect the historical climate change signal in observations than in any given member of LENS.