Variability of surface climate in simulations of past and future

Abstract

Abstract. It is virtually certain that the mean surface temperature of the Earth will continue to increase under realistic emission scenarios, yet comparatively little is known about future changes in climate variability. This study explores changes in climate variability over the large range of climates simulated by the Coupled Model Intercomparison Project Phase 5 and 6 (CMIP5/6) and the Paleoclimate Modeling Intercomparison Project Phase 3 (PMIP3), including time slices of the Last Glacial Maximum, the mid-Holocene, and idealized experiments (1 % CO2 and abrupt4×CO2). These states encompass climates within a range of 12 ∘C in global mean temperature change. We examine climate variability from the perspectives of local interannual change, coherent climate modes, and through compositing extremes. The change in the interannual variability of precipitation is strongly dependent upon the local change in the total amount of precipitation. At the global scale, temperature variability is inversely related to mean temperature change on intra-seasonal to multidecadal timescales. This decrease is stronger over the oceans, while there is increased temperature variability over subtropical land areas (40∘ S–40∘ N) in warmer simulations. We systematically investigate changes in the standard deviation of modes of climate variability, including the North Atlantic Oscillation, the El Niño–Southern Oscillation, and the Southern Annular Mode, with global mean temperature change. While several climate modes do show consistent relationships (most notably the Atlantic Zonal Mode), no generalizable pattern emerges. By compositing extreme precipitation years across the ensemble, we demonstrate that the same large-scale modes influencing rainfall variability in Mediterranean climates persist throughout paleoclimate and future simulations. The robust nature of the response of climate variability, between cold and warm climates as well as across multiple timescales, suggests that observations and proxy reconstructions could provide a meaningful constraint on climate variability in future projections.

The impact of climate change on astronomical observations

“…Strong IWV events are primarily related to high central equatorial sea surface temperature during El Niño events. Recent climate studies give a hint that an increased CO 2 concentration in the atmosphere will give rise to a global increase in humidity [17] and to more violent El Niño events [13], which means more frequent severe flooding in the Atacama region, as witnessed in March 2015 [18]. Here, we extracted the IWV at the Paranal observatory location from the ERA5 reanalysis, complemented by the ERA20C reanalysis that covers a full century (to probe the 4-year ENSO cycle) [19], which we compared to on-site measurements using the Low Humidity and Atmospheric Temperature PROfiler (LHATPRO), a radiometer installed at the Paranal observatory [20] (Fig.…”

The PMIP4 Last Glacial Maximum experiments: preliminary results and comparison with the PMIP3 simulations

“…Changes in model configuration have resulted in several of the PMIP4-CMIP6 models having substantially higher climate sensitivity than the PMIP3-CMIP5 versions of the same models, and thus the range of climate sensitivity sampled by the PMIP4-CMIP6 models is much wider. This provides an opportunity to re-examine whether the LGM could provide a strong constraint on climate sensitivity (Renoult et al, 2020;Zhu et al, 2021).…”

Weakened impact of the Atlantic Niño on the future equatorial Atlantic and Guinean Coast rainfall

“…The signal in areas where at least 50 % of the models show a significant regression (two-sided Student test at 95 % confidence level) and where at least 80 % of the models agree on the sign of the mean is considered as robust. When we consider the change between two periods, the sign of the change averaged over a subset of models is considered as robust if at least two thirds of the models agree on the multimodel mean (Rehfeld et al, 2020), and at least the change of the multimodel mean is significant at 95 % confidence level according to a two-sided Welch t-test. Finally, coordinates of different geographic areas used in the present study are defined in Table 2.…”