Can a global mean sea-level rise reduce the Last Interglacial model–data mismatch in East Asia?

Zhiqi Qin ,Tino Xu ,Zhongshi Zhng,c,d,＊ ,Chunju Hung

a State Key Laboratory for Biogeology and Environmental Geology, Hubei Key Laboratory of Critical Zone Evolution, School of Earth Sciences, China University of Geosciences, Wuhan, China

b Department of Atmospheric Science, School of Environmental Studies, China University of Geosciences, Wuhan, China

c School of Geographic Science, Nantong University, Nantong, China

d Centre for Severe Weather and Climate and Hydro-geological Hazards, Wuhan, China

Keywords: Last Interglacial Sea-level rise Model–data mismatch

ABSTRACT The Last Interglacial (LIG),with its many reconstructions and simulations,provides an ideal analog for investigating the future warmer climate.However,there has been a persistent mismatch between simulated and reconstructed LIG climates in East Asia,with simulations generally indicating a colder and drier climate than reconstructions.In this study,utilizing the Norwegian Earth System Model (NorESM1-F),the authors investigated whether incorporating the global mean sea-level rise in LIG simulation experiments can reduce the model–data mismatch.The new experiments reveal a discernible,yet insufficient,warming and wetting effect in East Asia resulting from the sea-level rise.Therefore,the model–data mismatch remains unresolved.Based on these results,the authors explore alternative factors that may contribute to this mismatch,offering insights for future studies.

1.Introduction

The Last Interglacial(LIG),which occurred between 129 and 116 ka,also known as the Eemian or MIS5e,was one of the warmest periods since the middle Pleistocene (Past Interglacials Working Group of PAGES,2016;Hoffman et al.,2017;Turney et al.,2020).Global temperatures during the LIG are estimated to have been 0.1?C to more than 2?C higher than present-day levels (Turney et al.,2020),while the atmospheric CO2concentration was approximately 275 ppm (Petit et al.,1999),slightly lower than the preindustrial level.Although the naturally driven warming mechanisms during the LIG differ from the current global warming caused by greenhouse gas emissions (Pedersen et al.,2017),the LIG remains an ideal analog for investigating Earth system feedback in the future due to its relatively recent occurrence and similar geographical conditions.

In China,ice cores,loess deposits,cave stalagmites,pollen,lakes,and deep-sea sediments have provided rich records of the environment of the LIG (Table 1).Although there are some differences among these records,most indicate a warmer and more humid climate during the LIG(Leng et al.,2019).However,results from phase 4 of the Paleoclimate Modeling Intercomparison Project (PMIP4) multimodel ensemble exhibit colder and drier conditions compared to reconstructions,highlighting a model–data mismatch(Otto-Bliesner et al.,2020,2021;Jiang et al.,2022).These PMIP4 LIG simulations overall reveal the effects of orbital parameters on summer warming and humidification,winter cooling and drying,and annual cooling in China.

Table 1 Reconstructed temperature and precipitation changes during the LIG (modified from Leng et al.(2019)).

The PMIP4 LIG experimental protocol mainly focuses on orbital parameters and greenhouse gas concentrations,without specific constraints on the global mean sea level(GMSL)(Otto-Bliesner et al.,2021).However,previous studies have shown that during the LIG,the GMSL was about 5–10 m higher than today (Fox-Kemper et al.,2021).The higher GMSL may potentially have created extensive and significant impacts on a global scale (Zhang et al.,2023).Growth of the GMSL elevates the sea-level datum (the reference surface between the topography of the land and the ocean bathymetry),leading to a deepening of ocean gateways,expansion of ocean surfaces(Farnsworth et al.,2019),a reshaping of regional relative sea levels (Church et al.,2004;Richter et al.,2020),and reorganization of ocean density structures and dynamics.Moreover,GMSL changes in the glacial–interglacial cycles exert substantial influences on global and regional climates.For instance,the opening and closing of the Bering Strait affected the Atlantic Meridional Ocean Circulation and the transport of oceanic heat (Hu et al.,2010).The landmass configuration of the Maritime Continent plays a crucial role in shaping regional atmospheric circulation and rainfall (Di Nezio et al.,2016).

Here,we further diagnose LIG snapshot experiments with the fast version of the Norwegian Earth System Model (NorESM1-F).We incorporate the GMSL rise into the LIG experiments to investigate whether considering it can reduce the long-standing model–data mismatch.

2.Model and experiments

NorESM1-F is a computationally efficient model within the Norwegian Earth System Model family that simulates the global climate well(Guo et al.,2019).It was built on the Community Climate System Model,version 4(Gent et al.,2011).NorESM1-F utilizes a grid with a horizontal resolution of 2?and 26 vertical levels in the atmospheric component and a tripolar grid with a nominal 1?horizontal resolution and 53 vertical layers in the oceanic component.A detailed description and evaluation of NorESM1-F can be found in the model documentation by Guo et al.(2019).

In addition to the preindustrial control experiment(piControl),three other LIG experiments have been conducted.First,the lig126 experiment only considers changes in Earth’s orbital configuration and greenhouse gas levels,using the orbital parameters of 126 ka (Berger and Loutre,1991),an atmospheric CO2level of 274.99 ppm,and an atmospheric CH4level of 652.52 ppb (Siegenthaler et al.,2005).Subsequently,the lig126sl5m and lig126sl10m experiments incorporate additional GMSL uplifts of 5 and 10 m,respectively.In these experiments,the ocean bathymetry is adjusted by adding 5 m(10 m),while the land topography is reduced by the same amount.For a more detailed experimental design,please refer to Zhang et al.(2023).

3.Simulation results

Compared to the lig126 experiment,the lig126sl5m and lig126sl10m experiments show a slight overall increase in annual mean surface air temperature (SAT) in China,with most regions experiencing an insignificant temperature rise of less than 0.2?C.Notably,the Japan Sea and eastern coastal areas of China exhibit a more pronounced warming,with temperature increases of up to 1?C(Fig.1(a,d)).This warming can be attributed to increased warm water inflow from the Northwest Pacific into the relatively enclosed Japan Sea,leading to a concentration of warm water and resulting in higher temperatures in these regions compared to the surrounding areas.The change in seasonal SAT follows a similar pattern to the annual mean SAT.In summer,apart from the Bohai Sea region,the SAT over China slightly increases(Fig.1(b,e)).In winter,the eastern part of China experiences a slight cooling,whereas other regions show a slight warming.The significant seasonal warming is also mostly confined to the Japan Sea (Fig.1(c,f)).

Fig.1.Monthly temperature changes due to GMSL rise in interglacial sensitivity experiments:(a)comparison between the lig126sl5m and lig126 experiments in the response of monthly temperature(units:℃)for the whole year;(b,c)as in(a)but in winter(DJF)and summer(JJA),respectively;(d–f)as in(a–c)but the sea-level uplifts are 10 m.The gridded area passed the 0.05 significance t-test.

Similar to the temperature responses,the change in annual mean precipitation remains insignificant in most of East Asia,except for the increased precipitation in the Japan Sea and decreased precipitation along southern coastal China (Fig.2(a,d)).Compared with the annual precipitation changes,the adjustments in seasonal precipitation are more complex.Significant changes occur in winter in the Japan Sea,with a substantial increase of approximately 20 mm/month.Conversely,winter precipitation changes in the other regions are small (within the range of less than 3 mm/month).This reduces slightly over the South China Sea(Fig.2(b,e)).During the summer,a drying anomaly is evident in southern China,accompanied by increased precipitation in northern China,with amplitudes of approximately 10 mm/month (Fig.2(c,f)).

Fig.2.Monthly precipitation changes due to GMSL rise in interglacial sensitivity experiments:(a)comparison between the lig126sl5m and lig126 experiments in the response of monthly precipitation(units:mm/month)for the whole year;(b,c)as in(a)but in winter(DJF)and summer(JJA),respectively;(d–f)as in(a–c)but the sea-level uplifts are 10 m.The gridded area passed the 0.05 significance t-test.

The changes in low-level wind fields (Fig.3),combined with the spatial distribution of water vapor (Fig.4),contribute to the simulated precipitation patterns.In winter,cyclonic anomalies appear over the Japan Sea,indicating a weakening of the western Pacific high and a strengthening of the East Asian trough.Along the western side of these anomalies,anomalous northeasterlies flow over northern China,while anomalous northwesterlies flow over southern China(Fig.3(b,e)).Such circulation anomalies induce moisture convergence over the Japan Sea and subsequent transport from north to south over eastern China,leading to a slight increase in winter precipitation in northern China but a notable reduction in precipitation over southern China and the South China Sea (Figs.2(b,e) and 4(b,e)).The mid–high latitude inland regions exhibit an easterly or northeasterly anomaly (Fig.3(b,e)),indicating a weakening of the westerly jet stream.The significant anomalous areas are associated with a slight decrease in precipitation(Fig.2(b,e)).In summer,the rise in sea level leads to a strengthening of the western Pacific high and weakening of the monsoon circulation (Fig.3(c,f)),resulting in increased moisture transport from the Pacific Ocean to northern China,whereas the moisture flux to southern China decreases.

But do these changes help rectify the model–data mismatch? Comparison between the lig126 and piControl experiments shows a winter cooling and summer warming of up to 5?C (Jiang et al.,2022),but a general cooling in the annual mean (Fig.5(a)).In terms of annual precipitation,the lig126 experiment shows that western China is wetter,whereas eastern China is drier,compared to the pre-industrial experiment (Fig.5(d)).Even when considering high sea-levels in the lig126sl5m and lig126sl10m experiments,the patterns of annual mean temperature and precipitation responses do not change significantly.Compared with the reconstruction records (Table 1) collected in a previous study (Leng et al.,2019),the simulated annual temperature and precipitation still exhibit a remarkable mismatch(Fig.5(b,c,e,f)).This comparison demonstrates that the direct effect of sea-level rise is insufficient to explain the model–data mismatch in China during the LIG.

4.Discussion and conclusions

In summary,the lig126sl5m and lig126sl10m experiments demonstrate a slight increase in annual mean SAT in China,particularly in the Japan Sea and eastern coastal regions.The changes in seasonal temperature and precipitation also align with the overall patterns observed in the annual mean.However,these simulated results still exhibit a notable mismatch when compared with reconstruction records,indicating the need for further investigation and refinement in the model simulations.

Although the consideration of GMSL rise has successfully reduced the model–data mismatch in the Southern Hemisphere(Zhang et al.,2023),its direct effect on reducing the mismatch in East Asia is challenging.The discrepancy is likely due to uncertainties in both reconstructions and simulations.

Some studies suggest that the temperature proxies used for the LIG may exhibit a bias towards the warm season,reflecting summer temperatures rather than annual mean temperatures.This bias can lead to generally warmer reconstruction results(Bakker and Renssen,2014).In addition,climate proxies often capture combined climate signals,making extracting a single climate signal (for example,pure precipitation)challenging(e.g.,Miao et al.,2015;Thomas et al.,2016).Finally,uncertainties in dating techniques and differences in methods for analyzing proxy indicators (Jiang et al.,2012) also contribute to the overall uncertainties in reconstructions.

On the simulation side,our experiments remain idealized by uniformly lifting sea-level spatially(Zhang et al.,2023).In reality,a GMSL rise triggers numerous complex processes that cannot be fully captured in climate models.For instance,the fixed vegetation and ice cover conditions in our experiments may have biased the simulated temperature,surface evaporation,and water budget (Capron et al.,2014;Otto-Bliesner et al.,2020).A GMSL rise can potentially induce changes in vegetation feedback,although the extent of this effect remains uncertain.Cloud feedback in simulations also introduces uncertainty(Stephens,2005;Liu et al.,2014),as it is not yet clearly understood how a GMSL rise influences cloud feedback.The coarse horizontal resolution may not capture the localized environmental influences on East Asian precipitation (Jiang et al.,2022).Given that GMSL rise is a slow feedback process,future investigations must delve deeper into the various slow feedback processes related to sea-level rise.Additionally,the potential influence of the calendar effect should be considered,though it is small.

Future research should focus on improving climate models by incorporating dynamic bathymetry,topography changes,and feedback mechanisms related to relative sea-level changes to address these issues and reduce the model–data mismatch.Meanwhile,more comprehensive and high-resolution reconstructions,using multiple proxies and reducing dating uncertainties,are also crucial.All these new approaches will enhance our understanding of the LIG and the future warmer climate.

Funding

This study was jointly supported by the National Natural Science Foundation of China [grants numbers 42125502,42230208,and 42172039].