skip to main content


Title: The Pliocene Model Intercomparison Project Phase 2: large-scale climate features and climate sensitivity
Abstract. The Pliocene epoch has great potential to improve ourunderstanding of the long-term climatic and environmental consequences of an atmospheric CO2 concentration near ∼400 parts permillion by volume. Here we present the large-scale features of Plioceneclimate as simulated by a new ensemble of climate models of varyingcomplexity and spatial resolution based on new reconstructions ofboundary conditions (the Pliocene Model Intercomparison Project Phase 2;PlioMIP2). As a global annual average, modelled surface air temperaturesincrease by between 1.7 and 5.2 ∘C relative to the pre-industrial erawith a multi-model mean value of 3.2 ∘C. Annual mean totalprecipitation rates increase by 7 % (range: 2 %–13 %). On average, surface air temperature (SAT) increases by 4.3 ∘C over land and 2.8 ∘C over the oceans. There is a clear pattern of polar amplification with warming polewards of 60∘ N and 60∘ S exceeding the global mean warming by a factor of 2.3. In the Atlantic and Pacific oceans, meridional temperature gradients are reduced, while tropical zonal gradients remain largely unchanged. There is a statistically significant relationship between a model's climate response associated with a doubling in CO2 (equilibrium climate sensitivity; ECS) and its simulated Pliocene surface temperature response. The mean ensemble Earth system response to a doubling of CO2 (including ice sheet feedbacks) is 67 % greater than ECS; this is larger than the increase of 47 % obtained from the PlioMIP1 ensemble. Proxy-derived estimates of Pliocene sea surface temperatures are used to assess model estimates of ECS and give an ECS range of 2.6–4.8 ∘C. This result is in general accord with the ECS range presented by previous Intergovernmental Panel on Climate Change (IPCC) Assessment Reports.  more » « less
Award ID(s):
1903650
NSF-PAR ID:
10280510
Author(s) / Creator(s):
; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; more » ; ; ; ; ; ; ; ; ; ; ; « less
Date Published:
Journal Name:
Climate of the Past
Volume:
16
Issue:
6
ISSN:
1814-9332
Page Range / eLocation ID:
2095 to 2123
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. null (Ed.)
    Abstract. The Arctic is warming 2 to 3 times faster than the global average, partly due to changes in short-lived climate forcers (SLCFs) including aerosols. In order to study the effects of atmospheric aerosols in this warming, recent past (1990–2014) and future (2015–2050) simulations have been carried out using the GISS-E2.1 Earth system model to study the aerosol burdens and their radiative and climate impacts over the Arctic (>60∘ N), using anthropogenic emissions from the Eclipse V6b and the Coupled Model Intercomparison Project Phase 6 (CMIP6) databases, while global annual mean greenhouse gas concentrations were prescribed and kept fixed in all simulations. Results showed that the simulations have underestimated observed surface aerosol levels, in particular black carbon (BC) and sulfate (SO42-), by more than 50 %, with the smallest biases calculated for the atmosphere-only simulations, where winds are nudged to reanalysis data. CMIP6 simulations performed slightly better in reproducing the observed surface aerosol concentrations and climate parameters, compared to the Eclipse simulations. In addition, simulations where atmosphere and ocean are fully coupled had slightly smaller biases in aerosol levels compared to atmosphere-only simulations without nudging. Arctic BC, organic aerosol (OA), and SO42- burdens decrease significantly in all simulations by 10 %–60 % following the reductions of 7 %–78 % in emission projections, with the Eclipse ensemble showing larger reductions in Arctic aerosol burdens compared to the CMIP6 ensemble. For the 2030–2050 period, the Eclipse ensemble simulated a radiative forcing due to aerosol–radiation interactions (RFARI) of -0.39±0.01 W m−2, which is −0.08 W m−2 larger than the 1990–2010 mean forcing (−0.32 W m−2), of which -0.24±0.01 W m−2 was attributed to the anthropogenic aerosols. The CMIP6 ensemble simulated a RFARI of −0.35 to −0.40 W m−2 for the same period, which is −0.01 to −0.06 W m−2 larger than the 1990–2010 mean forcing of −0.35 W m−2. The scenarios with little to no mitigation (worst-case scenarios) led to very small changes in the RFARI, while scenarios with medium to large emission mitigations led to increases in the negative RFARI, mainly due to the decrease in the positive BC forcing and the decrease in the negative SO42- forcing. The anthropogenic aerosols accounted for −0.24 to −0.26 W m−2 of the net RFARI in 2030–2050 period, in Eclipse and CMIP6 ensembles, respectively. Finally, all simulations showed an increase in the Arctic surface air temperatures throughout the simulation period. By 2050, surface air temperatures are projected to increase by 2.4 to 2.6 ∘C in the Eclipse ensemble and 1.9 to 2.6 ∘C in the CMIP6 ensemble, compared to the 1990–2010 mean. Overall, results show that even the scenarios with largest emission reductions leads to similar impact on the future Arctic surface air temperatures and sea-ice extent compared to scenarios with smaller emission reductions, implying reductions of greenhouse emissions are still necessary to mitigate climate change. 
    more » « less
  2. null (Ed.)
    Abstract. Accurate estimates of past global mean surface temperature (GMST) help tocontextualise future climate change and are required to estimate thesensitivity of the climate system to CO2 forcing through Earth's history.Previous GMST estimates for the latest Paleocene and early Eocene(∼57 to 48 million years ago) span a wide range(∼9 to 23 ∘C higher than pre-industrial) andprevent an accurate assessment of climate sensitivity during this extremegreenhouse climate interval. Using the most recent data compilations, weemploy a multi-method experimental framework to calculate GMST during thethree DeepMIP target intervals: (1) the latest Paleocene (∼57 Ma), (2) the Paleocene–Eocene Thermal Maximum (PETM; 56 Ma), and (3) the earlyEocene Climatic Optimum (EECO; 53.3 to 49.1 Ma). Using six differentmethodologies, we find that the average GMST estimate (66 % confidence)during the latest Paleocene, PETM, and EECO was 26.3 ∘C (22.3 to28.3 ∘C), 31.6 ∘C (27.2 to 34.5 ∘C), and27.0 ∘C (23.2 to 29.7 ∘C), respectively. GMST estimatesfrom the EECO are ∼10 to 16 ∘C warmer thanpre-industrial, higher than the estimate given by the Intergovernmental Panel on Climate Change (IPCC) 5thAssessment Report (9 to 14 ∘C higher than pre-industrial).Leveraging the large “signal” associated with these extreme warm climates,we combine estimates of GMST and CO2 from the latest Paleocene, PETM,and EECO to calculate gross estimates of the average climate sensitivitybetween the early Paleogene and today. We demonstrate that “bulk”equilibrium climate sensitivity (ECS; 66 % confidence) during the latestPaleocene, PETM, and EECO is 4.5 ∘C (2.4 to 6.8 ∘C),3.6 ∘C (2.3 to 4.7 ∘C), and 3.1 ∘C (1.8 to4.4 ∘C) per doubling of CO2. These values are generallysimilar to those assessed by the IPCC (1.5 to 4.5 ∘C per doublingCO2) but appear incompatible with low ECS values (<1.5 perdoubling CO2). 
    more » « less
  3. null (Ed.)
    Abstract. Palaeoclimate simulations improve our understanding ofthe climate, inform us about the performance of climate models in adifferent climate scenario, and help to identify robust features of theclimate system. Here, we analyse Arctic warming in an ensemble of 16simulations of the mid-Pliocene Warm Period (mPWP), derived from thePliocene Model Intercomparison Project Phase 2 (PlioMIP2). The PlioMIP2 ensemble simulates Arctic (60–90∘ N) annual meansurface air temperature (SAT) increases of 3.7 to 11.6 ∘Ccompared to the pre-industrial period, with a multi-model mean (MMM) increase of7.2 ∘C. The Arctic warming amplification ratio relative to globalSAT anomalies in the ensemble ranges from 1.8 to 3.1 (MMM is 2.3). Sea iceextent anomalies range from −3.0 to -10.4×106 km2, with a MMManomaly of -5.6×106 km2, which constitutes a decrease of 53 %compared to the pre-industrial period. The majority (11 out of 16) of models simulatesummer sea-ice-free conditions (≤1×106 km2) in their mPWPsimulation. The ensemble tends to underestimate SAT in the Arctic whencompared to available reconstructions, although the degree of underestimationvaries strongly between the simulations. The simulations with the highestArctic SAT anomalies tend to match the proxy dataset in its current formbetter. The ensemble shows some agreement with reconstructions of sea ice,particularly with regard to seasonal sea ice. Large uncertainties limit theconfidence that can be placed in the findings and the compatibility of thedifferent proxy datasets. We show that while reducing uncertainties in thereconstructions could decrease the SAT data–model discord substantially,further improvements are likely to be found in enhanced boundary conditionsor model physics. Lastly, we compare the Arctic warming in the mPWP toprojections of future Arctic warming and find that the PlioMIP2 ensemblesimulates greater Arctic amplification than CMIP5 future climate simulationsand an increase instead of a decrease in Atlantic Meridional OverturningCirculation (AMOC) strength compared topre-industrial period. The results highlight the importance of slow feedbacks inequilibrium climate simulations, and that caution must be taken when usingsimulations of the mPWP as an analogue for future climate change. 
    more » « less
  4. null (Ed.)
    Abstract. The early and late Eocene have both been the subject of many modelling studies, but few have focused on the middle Eocene. The latter still holds many challenges for climate modellers but is also key to understanding the events leading towards the conditions needed for Antarctic glaciation at the Eocene–Oligocene transition. Here, we present the results of CMIP5-like coupled climate simulations using the Community Earth System Model (CESM) version 1. Using a new detailed 38 Ma geography reconstruction and higher model resolution compared to most previous modelling studies and sufficiently long equilibration times, these simulations will help to further understand the middle to late Eocene climate. At realistic levels of atmospheric greenhouse gases, the model is able to show overall good agreement with proxy records and capture the important aspects of a warm greenhouse climate during the Eocene. With a quadrupling of pre-industrial concentrations of both CO2 and CH4 (i.e. 1120 ppm and ∼2700 ppb, respectively, or 4 × PIC; pre-industrial carbon), sea surface temperatures correspond well to the available late middle Eocene (42–38 Ma; ∼ Bartonian) proxies. Being generally cooler, the simulated climate under 2 × PIC forcing is a good analogue for that of the late Eocene (38–34 Ma; ∼ Priabonian). Terrestrial temperature proxies, although their geographical coverage is sparse, also indicate that the results presented here are in agreement with the available information. Our simulated middle to late Eocene climate has a reduced Equator-to-pole temperature gradient and a more symmetric meridional heat distribution compared to the pre-industrial reference. The collective effects of geography, vegetation, and ice account for a global average 5–7 ∘C difference between pre-industrial and 38 Ma Eocene boundary conditions, with important contributions from cloud and water vapour feedbacks. This helps to explain Eocene warmth in general, without the need for greenhouse gas levels much higher than indicated by proxy estimates (i.e. ∼500–1200 ppm CO2) or low-latitude regions becoming unreasonably warm. High-latitude warmth supports the idea of mostly ice-free polar regions, even at 2 × PIC, with Antarctica experiencing particularly warm summers. An overall wet climate is seen in the simulated Eocene climate, which has a strongly monsoonal character. Equilibrium climate sensitivity is reduced (0.62 ∘C W−1 m2; 3.21 ∘C warming between 38 Ma 2 × PIC and 4 × PIC) compared to that of the present-day climate (0.80 ∘C W−1 m2; 3.17 ∘C per CO2 doubling). While the actual warming is similar, we see mainly a higher radiative forcing from the second PIC doubling. A more detailed analysis of energy fluxes shows that the regional radiative balance is mainly responsible for sustaining a low meridional temperature gradient in the Eocene climate, as well as the polar amplification seen towards even warmer conditions. These model results may be useful to reconsider the drivers of Eocene warmth and the Eocene–Oligocene transition (EOT) but can also be a base for more detailed comparisons to future proxy estimates. 
    more » « less
  5. null (Ed.)
    Abstract. Equilibrium climate sensitivity (ECS) has been directly estimated using reconstructions of past climates that are different than today's. A challenge to this approach is that temperature proxies integrate over the timescales of the fast feedback processes (e.g., changes in water vapor, snow, and clouds) that are captured in ECS as well as the slower feedback processes (e.g., changes in ice sheets and ocean circulation) that are not. A way around this issue is to treat the slow feedbacks as climate forcings and independently account for their impact on global temperature. Here we conduct a suite of Last Glacial Maximum (LGM) simulations using the Community Earth System Model version 1.2 (CESM1.2) to quantify the forcingand efficacy of land ice sheets (LISs) and greenhouse gases (GHGs) in order to estimate ECS. Our forcing and efficacy quantification adopts the effective radiative forcing (ERF) and adjustment framework and provides a complete accounting for the radiative, topographic, and dynamical impacts of LIS on surface temperatures. ERF and efficacy of LGM LIS are −3.2 W m−2 and 1.1, respectively. The larger-than-unity efficacy is caused by the temperature changes over land and the Northern Hemisphere subtropical oceans which are relatively larger than those in response to a doubling of atmospheric CO2. The subtropical sea-surface temperature (SST) response is linked to LIS-induced wind changes and feedbacks in ocean–atmosphere coupling and clouds. ERF and efficacy of LGM GHG are −2.8 W m−2 and 0.9, respectively. The lower efficacy is primarily attributed to a smaller cloud feedback at colder temperatures. Our simulations further demonstrate that the direct ECS calculation using the forcing, efficacy, and temperature response in CESM1.2 overestimates the true value in the model by approximately 25 % due to the neglect of slow ocean dynamical feedback. This is supported by the greater cooling (6.8 ∘C) in a fully coupled LGM simulation than that (5.3 ∘C) in a slab ocean model simulation with ocean dynamics disabled. The majority (67 %) of the ocean dynamical feedback is attributed to dynamical changes in the Southern Ocean, where interactions between upper-ocean stratification, heat transport, and sea-ice cover are found to amplify the LGM cooling. Our study demonstrates the value of climate models in the quantification of climate forcings and the ocean dynamical feedback, which is necessary for an accurate direct ECS estimation. 
    more » « less