Impact of mountains in Southern China on the Eocene climates of East Asia

in


Introduction
The early Eocene is the warmest interval of the Cenozoic (Cramer et al., 2009;Hollis et al., 2019;Westerhold et al., 2020), and is often thought to be an analogy for the future climate (Burke et al., 2018;Tierney et al., 2020), in particular under the SSP5-8.5 scenario (Riahi et al., 2017).Recently, the Deep Time Model Intercomparison Project (DeepMIP) (Hollis et al., 2019;Lunt et al., 2017) was initiated to improve our understanding of the warm early Eocene climate.
Several regional climates during the Eocene are still under debate, and the Eocene environment of China is a particular focus of the discussion.One view, based on the widely distributed pollen of Ephedripites, as well as red beds and evaporite sediments, suggests that a dry zonal belt (Figure 1a) dominated mid-latitude China from west to east (Guo et al., 2002,;Sun & Wang, 2005;Wang, 1990).To the north and the south of the arid zonal belt, humid climates appeared in Northeast and South China, where pollen assemblages indicative of humid conditions, as well as coal and oil shales, occur in Eocene sediments (Guo et al., 2002,;Sun & Wang, 2005; (Li & Zeng, 2003), with index values >0.(d) Monsoon domains defined by precipitation seasonality (Wang & Ding, 2008;Wang et al., 2012).(e) Arid index (Precipitation/Potential evapotranspiration).
The cause of the divergence of the simulations of the Eocene climate in China is also poorly understood.Many factors, such as atmospheric greenhouse gas concentrations, Earth's orbital configuration, land-sea distribution, and topography, may be the cause of the model spread (Farnsworth et al., 2019;X. Li et al., 2018;Licht et al., 2014;Lunt et al., 2010;Zhang et al., 2012).Some extreme orbital settings can cause wet Eocene climates in East Asia (Tardif et al., 2021;Zhang et al., 2012), but that cannot explain the long-term drought reflected in the records.In addition, many previous Eocene simulations have tested multiple combinations of greenhouse gas levels and several land-sea distribution conditions (X.Li et al., 2018;Licht et al., 2014;Zhang et al., 2012), with minimal to no influence of these factors.However, uncertainties regarding paleogeography are a particular problem for Eocene climate modeling (Huber & Goldner, 2012;Zhang et al., 2012,).
In the past two decades, several Eocene paleogeographic maps have been produced, such as those of Sewall et al. (2000), Bice and Marotzke (2001), Markwick (2007), Herold et al. (2014), Scotese (2014), He et al. (2019), andPoblete et al. (2021).These maps show large differences in the topography of East Asia, particularly the mountains in Southern China, including the Gangdese Mountains, the Nanling Mountains, and the Zhe-Min Highlands.The Nanling Mountains and the Zhe-Min Highlands together are called the Southeast Mountains.
The height of the Gangdese Mountains in different reconstructions varies from 1,000 to 4,000 m, and those of the Southeast Mountains varies from 0 to 2,500 m, and they also have diverse morphologies in the maps.These mountains may significantly impact the simulated precipitation in China, because they are on the main path of vapor transport to East Asia during the Eocene (X.Li et al., 2018;Zhang et al., 2012); however, their climatic impact has not been comprehensively investigated.
Here, we use the Norwegian Earth System Model (NorESM1-F) to investigate the impact of the mountains in Southern China on the Eocene climate of East Asia.We use the paleogeographic map reconstructed by He et al. (2019), which synthesized the latest Early Eocene Asian geological records.The paper comprises five sections.First, we introduce the model and the experimental design in Section 2, and present the simulation results in Section 3. We then further discuss the simulated precipitation under different topographic configurations and make a summary in Section 4.

Model Description
NorESM1-F (Guo, Bentsen, et al., 2019) is a computationally efficient coupled Earth system model, based on the Community Climate System Model, version 4 (CCSM4; Gent et al., 2011).In line with CCSM4, NorESM1-F uses the Community Atmosphere Model version 4 (CAM4) as the atmosphere component model.The ocean component is MICOM (Bleck et al., 1992;Bleck & Smith, 1990).NorESM1-F uses a horizontal resolution of f19 (∼2°) in the atmospheric model and 26 vertical layers.The horizontal resolution of the ocean component is ∼1° and there are 53 vertical layers.Compared to the CCSM4, it uses an advanced chemistry-aerosol-cloud-radiation scheme, thus improving the simulation of precipitation seasonality (Guo, Bentsen, et al., 2019).It has been used to simulate the paleoclimate of the Pliocene (Haywood et al., 2020;Li et al., 2020), the Last Interglacial (127 ka) (Kageyama et al., 2021), and Marine Isotope Stage 3 (Guo, Nisancioglu, et al., 2019).Guo, Bentsen, et al. (2019) provide a detailed introduction and assessment of NorESM1-F.Our pre-industrial control experiment (Table 1 MD_280) shows that NorESM1-F realistically simulates the mean climate state in East Asia (Figure 1).

Experimental Design
With NorESM1-F, we first employed the paleogeographic reconstruction of He et al. (2019) to simulate the early Eocene climate (Table 1).We set CO 2 to 1,120 ppm (Anagnostou et al., 2016;Hollis et al., 2019;Jagniecki et al., 2015;Witkowski et al., 2018), and used modern orbital settings and an idealized vegetation cover (forest between 30°N and 30°S, and shrubs in the remainder of the area).No ice sheets were present in Greenland and Antarctica.Since there is no reconstruction of the early Eocene atmospheric methane and N 2 O concentration, we set the methane and N 2 O to pre-industrial levels.This fully coupled experiment (He_1120) was integrated for 1,500 model years to reach a quasi-equilibrium state.Finally, we analyzed the model outputs for the last 100 years.
We then compared this fully coupled experiment to our previous simulations that contribute to DeepMIP.The DeepMIP simulations used the paleogeographic maps of Herold et al. (2014) and Lunt et al. (2017).In these maps, the Gangdese Mountains and the Southeast Mountains are coterminous with a long extension, and both are ∼1,500 m high (Figures S1a and S1c in Supporting Information S1).In contrast, in the boundary conditions of He et al. (2019), the Zhe-Min Highlands range is relatively small, and the Nanling Mountains are absent (Figure S1b in Supporting Information S1).
Based on the boundary conditions reconstructed by He et al. (2019), we modified the topography of the mountains in South China to conduct atmosphere-only sensitivity experiments.The horizontal resolution of the atmosphere model CAM4 is f19 (∼2°).We considered a low (∼1,500 m) and high (∼4,000 m) Gangdese Mountains as suggested in the reconstruction of He et al. (2019).The Gangdese Mountains occupy ∼8 model grids.Since the recent Eocene paleogeographic maps (Baatsen et al., 2020;He et al., 2019;Herold et al., 2014;Lunt et al., 2017) suggested that the elevation range of the Southeast Mountains was ∼500-2,500 m, we used four conditions (∼1,000, ∼1,500, ∼2,000, ∼2,500 m) for the Southeast Mountains.The Southeast Mountains occupy ∼10 model grids.Combining these topographic conditions (Figure 2), we designed six sensitivity experiments (Table 1).
In the sensitivity experiments we used the same land-sea configuration, atmospheric CO 2 level, vegetation and orbital configuration as were used in coupled experiment He_1120.We used the monthly SST fields averaged over the last 100 years of the coupled experiment to force the atmosphere-only sensitivity experiments.The simulated mean annual SSTs for ∼55 Ma are generally more than ∼8°C warmer than the simulated modern SSTs in the western Pacific Ocean (Figure S2 in Supporting Information S1).All of these atmosphere-only experiments were run for 30 model years and we analyzed the outputs for the last 25 model years.

NorESM1-F Eocene Simulations Contributed to DeepMIP
Two NorESM1-F Eocene simulations using different paleogeographic maps have contributed to DeepMIP (Table 1).Experiment Herold_1120 uses the paleogeographic map in Herold et al. (2014) (H14), which is in a mantle reference frame and applied in all other DeepMIP experiments.In contrast, experiment Pmag_1120 uses the same H14 paleogeography, but rotated to a paleomagnetic reference frame (Lunt et al., 2017;Torsvik et al., 2012).Other boundary conditions, such as the atmospheric CO 2 level, are consistent with those used in

Comparison to DeepMIP Simulations
From the DeepMIP ensemble, we chose the simulations forced with three or four times the pre-industrial atmospheric CO 2 level.They include simulations with CESM1.2,HadCM3, IPSL, MIROC, COSMOS, GFDL, and NorESM1-F (Table 2).
Compared to other DeepMIP simulations, NorESM1-F tends to simulate a wetter climate in Eastern China.For example, the simulations with HadCM3, IPSL, COSMOS and GFDL show that an arid zonal climate belt (BWh and BSh) controls almost the whole of China (Figures 3b,3c,3e,and 3f).However, forced with the topographic conditions from DeepMIP, NorESM1-F simulates temperate humid climates (Cwa, Csa and Cfa) in some parts of Eastern China (Figures 3g and 3h).These wet climates also appear in the simulations with CESM1.2 and MIROC (Figures 3a and 3d).
Moreover, when the paleogeographic map from He et al. (2019) was used, our simulation shows that temperate humid climates (Cwa, Csa, and Cfa) occupy the whole of Eastern China (Figure 3i).Furthermore, the simulation

Köppen Climate Classification and Precipitation
In our sensitivity experiments, the Köppen climate classification demonstrates that an arid desert (BWh) and steppe (BSh) climate always dominates Western China (over 65-94°E and 24-65°N), irrespective of the height of Gangdese Mountains and the Southeast Mountains (Figure 4).The simulated annual precipitation is <1.5 mm/day.Under the scenario with low Gangdese Mountains and high Southeast Mountains (experiment G1500_S2500), the simulated arid area (Figure 4b) in East Asia is a continuous zonal band, agreeing best with early Eocene reconstructions (Guo et al., 2008;Sun & Wang, 2005).ZHANG ET AL. 10.1029/2022JD036510 9 of 18

Atmospheric Dynamics and Water Vapor Transport
In our sensitivity experiments, most of the precipitation in Eastern China is in summer (Figure 6), when the land-sea thermal contrast leads to a low-pressure system over the Eurasian continent and high-pressure systems over the oceans.This low-pressure system results in cyclonic circulations.Summer northwesterly and northerly winds transport limited amounts of moisture along the western margin of the cyclonic circulation into Western China (Figures 7 and 8); and along the eastern margin of the cyclonic circulations, southwesterly winds transport moisture along the eastern coast region of East Asia (Figures 7 and 8).
However, the topography of the Southeast Mountains significantly affects the simulated summer southwesterly winds.The uplift of the Southeast Mountains (e.g., G1500_S2500 vs. G1500_S1000 and G4000_S2500 vs. G4000_S1000) leads to a low-pressure anomaly in Eastern China (Figures 7c  and 7d).The blocking effect of the Southeast Mountains compresses the moisture transport path, causing the marginal seas in East Asia to receives more moisture while Eastern China receives less (Figures 8e and 8f).As a result, the simulated arid climate extends to Eastern China, enabling an arid zonal belt to form in East Asia.Compared to the Southeast Mountains, the Gangdese Mountains play a less critical role in reorganizing the atmospheric circulation in Eastern China; however, the Gangdese Mountains have more significant impacts in South Asia (Figures 7e and 7f).

Drought and the Wind/Precipitation Seasonality
Two monsoon indices are widely used in previous paleoclimate studies (Huber & Goldner, 2012;X. Li et al., 2018;Zhang et al., 2012,), based on precipitation and wind seasonality, respectively.According to the precipitation seasonality, the monsoon zone is identified by two criteria: (a) the local summer (May-September) minus winter (November-March) mean precipitation rate is >2 mm/day; (b) the ratio of local summer precipitation to annual total precipitation >55% (Wang et al., 2012).The other monsoon index, the SNS Index, defined by wind seasonality, indicates the monsoon zone where the values are >0.The SNS Index value is calculated as follows: where F 1 and F 7 are respectively the January and the July climatological winds, and   is the annual average climatological winds.The norm  ‖‖ is defined as , where S denotes the domain of integration.Details can be found in Li and Zeng (2003).
Our sensitivity experiments demonstrate that the simulated early Eocene precipitation in Eastern China is strongly seasonal (Figure 9).The ratio indicating seasonal precipitation is comparable to the present-day monsoonal precipitation seasonality in Eastern China (Figure 1c).However, the simulated early Eocene climate in Eastern China is not a typical monsoon climate because these simulated southwesterly summer winds cannot be classified as a summer monsoon.Even in the experiments that show humid climates (Cwa and Cfa) occupying the whole of Eastern China (Figures 4a and 4c), the simulated annual precipitation is much less than the current monsoonal precipitation (Figure 6).Moreover, the aridity index (AI)-the ratio between annual rainfall and potential evapotranspiration (PET) (Zhang & Jiang, 2019)-is always <0.5 in most of China in these early Eocene sensitivity experiments (Figure 10).Today, index values <0.5 only occurs in semi-arid and arid inland China (Figure 1d).Furthermore, the wind seasonality defined by the SNS index (Li & Zeng, 2003) does not support a typical early Eocene monsoon in Eastern China either.Today, the index is >0 in the East Asian monsoon region (Figure 1b).

Modeling Uncertainties
Although most previous studies have focused on the (both thermal and mechanical) climate impact of the large and high Tibet Plateau (An et al., 2001;Molnar et al., 2010;Wu et al., 2014), several studies also revealed the important roles of other mountains, for example, Tian Shan (Baldwin & Vecchi, 2016), Yunnan-Guizhou and Mongolian Plateau (Sha et al., 2015;Shi et al., 2019), mountains in India and Indochina (Wang et al., 2006), and Zagros Mountains (Tang et al., 2012), in shaping Asian climate.In agreement with these studies, our simulations further demonstrate that the Southeast Mountains were also critical in shaping the East Asian climate during the Eocene.The Southeast Mountains cause the mechanical effect by blocking water transport into Eastern China  (c-f) Anomaly between two experiments, (c) G1500_S2500 and G1500_S1000, (d) G4000_S2500 and G4000_S1000, (e) G4000_S1000 and G1500_ S1000, (f) G4000_S2500, and G1500_S2500.during the Eocene.In contrast, the high Gangdese Mountains are less critical in influencing climate in Eastern China but cause a thermal effect in South Asia and a mechanical effect in Western China.
Here, our sensitivity experiments are still not exhaustive in some aspects.The resolution of the atmospheric model is ∼2°, and the moisture transport through narrow mountain valleys can not be well simulated.Due to the potential higher precipitation sensitivity to topography (relative to other DeepMIP models) as well as the idealized vegetation, NorESM1-F may have overestimated the Eocene precipitation in East Asia in these experiments.Moreover, our atmosphere-only sensitivity experiments do not resolve feedbacks due to SST variabilities.Associated with the warm SST, in particular in the Pacific, the variability of the subtropical high-pressure system might cause severe aridity in East Asia during the Eocene.Therefore, it remains valuable to repeat these experiments with high-resolution coupled climate models.

The Gangdese and Southeast Mountains
The height of the Gangdese Mountains during the Eocene is debated.Several studies (Deng et al., 2019) suggest that it was still low (<2,000 m) during the Eocene.Low Gangdese Mountains enable moisture from the Indian  Ocean to penetrate into the humid lowland central Tibetan Plateau (Cai et al., 2019;Deng et al., 2019;Su et al., 2019;Wu et al., 2017).However, other studies based on oxygen isotopes and clumped-isotopes (Ding et al., 2014;Ingalls et al., 2018) suggest that several basins in the Gangdese Mountains were already high during the Eocene.For example, the Linzhou Basin, Penbo Basin, and Oiyug Basin (Figure 1a) may have reached heights of 4,500 ± 400, 4,400 + (1,300-1,700), and 4,100-6,500 m during the Eocene.Our simulations demonstrate that the height of the Gangdese Mountains has relatively strong effects on Western China and South Asia.The thermal effect of high Gangdese Mountains leads to the convergence of low-level summer winds, and strengthens the Somali jet (Figures 7e and 7f).Due to the high Gangdese Mountains, the rain shadow effect significantly reduces the precipitation in Western China (Figures 4 and 10).However, the height of the Gangdese Mountains influences the simulated early Eocene climate weakly in Eastern China (Figure 4).Forced with low Gangdese Mountains in experiment G1500_S2500, the simulated arid climate pattern fits best with the reconstructed zonal dry climate configuration during the Eocene in China (Figure 4b).
Due to the lack of robust evidence, it is difficult to constrain the height of the Southeast Mountains (the Zhe-Min Highlands and the Nanling Mountains), although our simulations suggest that they should be higher than 2,000-2,500 m during the Eocene.The molasse deposits that are widely distributed in Zhejiang, Fujian, Jiangxi and Guangdong (Figure 1a) suggest that the Zhe-Min Highlands were at heights of up to 3,500-4,000 m in the Late Cretaceous, and this paleogeographic configuration did not change until the Early Paleogene (Chen, 1997).However, in the Nanling region, no paleo-elevation reconstructions are currently available, and therefore, following He et al. (2019), we used the deposition rate of the Pearl River to provide an estimate.Yan et al. (2018) have shown that the entire volume of eroded rock in the Pearl River Basin from 45 Ma to the present is ∼1,122 × 10 3 km 3 .Since the area of the Pearl River Basin area is 368 × 10 3 km 2 , the total thickness of the denuded area reached ∼2,900 m since the Middle Eocene.Moreover, modern observations (Dai et al., 2008) show that the annual mean sediment flux from the Pearl River is 75 Mt/yr.During the entire Cenozoic, even if the sediment flux from the Pearl River was much less than the present value (e.g., only half, or 37.5 Mt/yr), the height of the eroded rock adjusted for isostatic balance is ∼2,000 m.Therefore, an estimated height of the Nanling Mountains of 2,000-3,000 m during the Early to Middle Eocene appears reasonable.
In addition to our sensitivity experiments, the height of the Southeast Mountains also determines the precipitation patterns in Eastern China in other Eocene simulations.In the DeepMIP experiments (Figures 3a-3h) and some Eocene simulations (X.Li et al., 2018;X. Y. Li et al., 2018;Shields et al., 2021;Zhang et al., 2012), which contain high and long Southeast Mountains (at least 1,500 m), an arid zonal band appears in China.In contrast, other Eocene experiments without high Southeast Mountains (Baatsen et al., 2020;Huber & Goldner, 2012;X. Li et al., 2018;Licht et al., 2014;Tardif et al., 2020) or with short Zhe-Min Highlands alone (Figure 3i), simulate a wet pattern (annual mean precipitation >500 mm) in Eastern China.Therefore, we suggest that the uncertainty regarding the topography of the Southeast Mountains is the main cause of the divergence of the results in the previous model simulations of the Eocene climate in China.

Comparison Between Eocene and Present-Day Climate in Eastern China
Currently, monsoonal climate leads to pronounced seasonality in winds and precipitation in China.In summer, the southwest winds (also called the southwest monsoon) and the southeast winds (also called the southeast monsoon), which transport water vapor from the Indian and Pacific Oceans, supply most of the annual rainfall in China.On the other hand, cold and dry northwest winds from inland Asia blow over North China in winter, and northeast winds blow over Southeast China (Chen et al., 1991).
Our simulations demonstrate that the simulated early Eocene climate shares some similarities with the present-day monsoon climate in East Asia.For example, the experiments show a strong precipitation seasonality similar to today (Figures 1c and 9), which is also evident in other Eocene simulations (Huber & Goldner, 2012;X. Li et al., 2018;Tardif et al., 2020).Furthermore, the simulated precipitation seasonality agrees with the seasonal cycles revealed in Eocene lake sediment and pollen records from Eastern China (Quan et al., 2012,;Tong, 1996;Wang et al., 2013;Zhang et al., 2005).However, our simulations also demonstrate that the simulated early Eocene climate differs from the present-day monsoon climate in China.During the early Eocene, the southwesterly summer winds are more dominant in Eastern China, and the southeasterly winds are weaker than today (Figures 1a and 7).In addition, the wind seasonality remains weak in most parts of Eastern China (Figure 11).This weak wind seasonality also appears in previous Eocene simulations (X.Li et al., 2018;Tardif et al., 2020;Zhang et al., 2012,).
Furthermore, the simulated early Eocene climate in Eastern China is much drier than today, consistent with plant fossil evidence, indicating an arid Eocene climate in Eastern China.Even the experiment forced with low Southeast Mountains (Figures 4 and 5), which simulates a relatively wet environment in Eastern China, reproduces less annual precipitation than today (Figures 4 and 6).Together with the higher evapotranspiration associated with high temperatures during the Eocene, the low precipitation results in a semi-arid and arid climate across almost the whole of China (Figure 1), in good agreement with the zonal dry band pattern revealed by integrated geological records (Guo et al., 2008;Sun & Wang, 2005).Sporadic indicators of humid climate also occur in Eocene sediments in Eastern China (Quan et al., 2012,), although the temporal resolution of these records is relatively coarse.As revealed in previous simulations (Tardif et al., 2021;Zhang et al., 2012)  Therefore, our early Eocene experiments are a convincing demonstration that seasonality in precipitation is not a good proxy for indicating a monsoon climate in deep time.Even though rainfall seasonality appears similar to current conditions, other features (including the seasonality of winds and the mean climatology) may differ significantly.

Summary
We have evaluated the impact of the topography of the Gangdese Mountains and the Southeast Mountains on the simulated Eocene climate of China.Our results show that an arid zonal band appeared in China during the Eocene when the Southeast Mountains are relatively high (>∼2,000 m with NorESM1-F).The blocking effect of the high Southeast Mountains compress the moisture transport path, causing a reduction in precipitation in Eastern China.Furthermore, our simulations show that the early Eocene climate in East Asia is distinct from the present-day monsoon climate, and that seasonality in precipitation alone does not represent a monsoon climate.
simulated Eocene precipitation in Eastern China • When the Southeast Mountains are high, an arid zonal band appears in mid-latitude China • The early Eocene climate in East Asia is not monsoonal climate Supporting Information: Supporting Information may be found in the online version of this article.

Figure 1 .
Figure 1.(a) Modern map of China showing the geographic locations referenced in the text.The two sky blue lines show the Eocene arid belt between ∼30 and 40°N in Guo et al. (2008).The modern topography of the Nanling Mountains and the Zhe-Min Mountains are shown in southeastern China.On the south side of these mountains, the dark blue line shows the Pearl River.The purple shadows show the rough ranges of molasse depositions (Chen, 1997) in the Nanling-Zhe-Min Mountain region.The modern topography of Tibet Plateau and the location of Gangdese Mountains are shown in western China.The green dot is the Linzhou-Penbo Basin, and the blue dot is Oiyug Basin.(b-e), Norwegian Earth System Model 1-F simulated pre-industrial climate (MD_280) in East Asia.(b) JJA sea level pressure (shading, units: hPa) and 850 hPa wind directions (arrow, units: m/s).(c) Monsoon domains defined by the SNS index(Li & Zeng, 2003), with index values >0.(d) Monsoon domains defined by precipitation seasonality(Wang & Ding, 2008;Wang et al., 2012).(e) Arid index (Precipitation/Potential evapotranspiration).
al. (2014) 1,120 Idealized vegetation for experiment He_1120, Herold_1120, Pmag_1120; Modern vegetation for experiment MD_280 experiment.The NorESM1-F experiments share the same solar constant, orbital configuration, and non-CO 2 greenhouse gas concentrations with the other DeepMIP experiments.

Figure 2 .
Figure 2. The topography of Eastern Asia used in the sensitivity experiments.The shading is altitude (units: m).The blue dashed line indicates 1,000 m contour, and the green dashed line the 2,000 m contour.The black lines show the boundary of China (Figure 1a) rotated to 55 Ma by Gplates.

Figure 3 .
Figure 3. Köppen climate classification in DeepMIP experiments with atmospheric CO 2 levels three to four times higher than pre-industrial.Af, tropical rainforest climate; Am, tropical monsoon climate; Aw, tropical savanna climate; BWh, hot desert climate; BSh, hot steppe climate; Cwa and Cfa, humid subtropical climate; Cwb and Cfb, oceanic climate; Csa and Csb, Mediterranean climate; Dsa, and Dfa, hot summer continental climate; Dsb, and Dfb, warm summer continental climate; Dsc, and Dfc, cold summer continental climate.(a-g) With the paleogeographic map in Herold et al. (2014) (H14); (h) with the H14 map using a paleomagnetic reference frame in Torsvik et al. (2012), provided in Lunt et al. (2017); (i) with the map in He et al. (2019).
, changes in Earth's orbital configuration can cause large climatic oscillations in Eastern China during the Eocene, consistent with the occurrence of both wet and dry climatic indicators in Eocene sediments.If a modern-like monsoon climate ZHANG ET AL.
the Eocene, these arid climate indicators should be completely absent from the Eocene sedimentary record.

Table 2
Summary of the DeepMIP Model Simulations, With Atmospheric CO 2 Levels Three or Four Times Higher Than Pre-Industrial, Analyzed in This Study