Scatterplot of the low-level winds (ms−1) versus the low-level specific humidity (g kg−1) in the Historical (black) and RCP8.5 (red) runs of the NorESM1-M model
Figure 4. Scatterplot of the low-level winds (ms−1) versus the low-level specific humidity (g kg−1) in the Historical (black) and RCP8.5 (red) runs of the NorESM1-M model. Histograms of the marginal distributions are shown, highlighting the shift in the specific humidity distribution. The Pearson product moment correlation coefficient (left) and p-value (right) between the low-level winds and specific humidity is shown for the historical and RCP8.5 scenarios.
Within the warm conveyor belt of extra-tropical cyclones, atmospheric rivers (ARs) are the key synoptic features which deliver the majority of poleward water vapour transport, and are associated with episodes of heavy and prolonged rainfall. ARs are responsible for many of the largest winter floods in the mid-latitudes resulting in major socioeconomic losses; for example, the loss from United Kingdom (UK) flooding in summer/winter 2012 is estimated to be about $1.6 billion in damages. Given the well-established link between ARs and peak river flows for the present day, assessing how ARs could respond under future climate projections is of importance in gauging future impacts from flooding. We show that North Atlantic ARs are projected to become stronger and more numerous in the future scenarios of multiple simulations from five state-of-the-art global climate models (GCMs) in the fifth Climate Model Intercomparison Project (CMIP5). The increased water vapour transport in projected ARs implies a greater risk of higher rainfall totals and therefore larger winter floods in Britain, with increased AR frequency leading to more flood episodes. In the high emissions scenario (RCP8.5) for 2074–2099 there is an approximate doubling of AR frequency in the five GCMs. Our results suggest that the projected change in ARs is predominantly a thermodynamic response to warming resulting from anthropogenic radiative forcing.