The Stalagmite Record of Southern Arabia: Climatic Extremes, Human Evolution and Societal Development

The fluctuating climatic conditions of the Saharo-Arabian deserts are increasingly linked to human evolutionary events and societal developments. On orbital timescales, the African and Indian Summer Monsoons were displaced northward and increased precipitation to the Arabian Peninsula which led to favorable periods for human occupation in the now arid interior. At least four periods of climatic optima occurred within the last 130,000 years, related to Marine Isotope Stages (MIS) 5e (128–121 ka BP), 5c (104–97 ka BP), 5a (81–74 ka BP) and 1 (10.5–6.2 ka BP), and potentially early MIS 3 (60–50 ka BP). Stalagmites from Southern Arabia have been key to understanding climatic fluctuations and human-environmental interactions; their precise and high-resolution chronologies can be linked to evidence for changes in human distribution and climate/environment induced societal developments. Here, we review the most recent advances in the Southern Arabian Late Pleistocene and Early Holocene stalagmite records. We compare and contrast MIS 5e and Early Holocene climates to understand how these differed, benchmark the extremes of climatic variability and summarize the impacts on human societal development. We suggest that, while the extreme of MIS 5e was important for H. sapeins dispersal, subsequent, less intense, wet phases mitigate against a simplistic narrative. We highlight that while climate can be a limiting and important factor, there is also the potential of human adaptability and resilience. Further studies will be needed to understand spatio-temporal difference in human-environment interactions in a climatically variable region.

Differing rainfall amounts between SAHPs are confirmed by stalagmite δ 18 O ca values, which are influenced by the intensity of ASM (Mukallah) and ISM (Hoti) rainfall Nicholson et al., 2020). SAHP 4 (MIS 5e) has the most negative δ 18 O ca values (increased rainfall), whereas SAHP 1 (Holocene) has the most positive δ 18 O ca values (drier conditions) (Nicholson et al., 2020). The competing effects of high-latitude glacial-boundary conditions and low-latitude insolation are both considered to control the expansion, contraction and intensity of the monsoon domain (Burns et al., 2003;Cheng et al., 2009a;Beck et al., 2018) and are key differentiating factors of SAHPs (Nicholson et al., 2020). While precipitation intensities of SAHP 4, 3 and 2 follow the declining intensity of glacial-boundary minima, SAHP 1 contradicts this trend, as positive δ 18 O ca occurred during an interglacial maximum. Instead, SAHP δ 18 O ca values consistently follow the pattern of declining lowlatitude summer Northern Hemisphere Insolation (NHI) maxima, which are regulated on orbital eccentricity (100 kyr) and precession (21 kyr) cycles ( Figure 1B). Low-latitude insolation is a key control on the interhemispheric pressure gradient, whereby greater solar heating of the Tibetan Plateau and northern Indian Ocean results in enhanced low pressure and intensification of northern hemisphere cyclones (Burns et al., . Blue lines for H5 and H12 represent means before and after the change-point at ∼6.2 ka BP. The change-point was identified using the changepoint package for R (Killick et al., 2016). 230 Th ages (cricles) and uncertainty (2σ) bars are given above their respective curves.  Bar-Matthews et al., 2003). 2) There is a clear relationship to the July 30 o N isolation curve, demonstrating rainfall intensity was modulated by low-latitude insolation ( Figures 1B, 2B). 3) While there is considerable variability, δ 18 O ca values are consistently more negative (wetter conditions) than succeeding wet phases. 4) There is an abrupt increase in δ 18 O ca and δ 13 C ca (drier conditions) at the termination of the wet period as the tropical rain-belt retreated southwards and annual rainfall fell below the threshold for large stalagmite formation (Nicholson et al., 2020). Additionally, sub-annually resolved H13 (Hoti) δ 18 O ca and δ 13 C ca records shows MIS 5e was characterised by increased seasonality (wetter summers and drier winters) dominated by a monsoon-driven precipitation regime (Nicholson et al., 2020). This was likely echoed by a seasonal vegetation response, as indicated by the presence of C4 plants (Bretzke et al., 2013;Nicholson et al., 2020), with potentially significant implications for animals and human hunter-gatherers.

Early-Mid Holocene
The Early-Mid Holocene is characterized by another period of increased rainfall in Arabia, known as the Holocene Humid Period (HHP), or SAHP 1 in Southern Arabia (Burns et al., 1998;Burns et al., 2001;Fleitmann et al., 2003aFleitmann and Matter, 2009;Lézine, 2009;Rosenberg et al., 2011;Engel et al., 2012;Rosenberg et al., 2013).  Burns et al., 2003;Fleitmann et al., 2007;Shakun et al., 2007). This is coeval to more negative δ 18 O ca values at Qunf Cave which indicate an intensification of the ISM. At both caves δ 18 O ca values show: 1) Intensification of summer precipitation between 10.6 and 9.4 ka BP, which slightly lags low-latitude insolation due to comparatively high glacial-boundary forcing . 2) Considerable multi-decadal variability within both H5 and Q5, displaying clear relationships with GRIP, NGRIP and DYE-3 ice-core δ 18 O records (Johnsen et al., 2001;Neff et al., 2001;Fleitmann et al., 2003a;Fleitmann et al., 2007;Fleitmann and Matter, 2009). More negative ice-core δ 18 O (colder northernhemisphere conditions) were reflected by more positive (drier conditions) stalagmite δ 18 O ca values. 3) A distinct increase of δ 18 O ca values (drier conditions) is observed between ∼8.2-8.0 ka BP and is related to the so-called "8.2-kyr event"; a global climatic event caused by the collapse of Atlantic Overturning Meridional Circulation (AMOC) due to draining of Hudson Bay glacial lakes and freshwater influx into the Atlantic (Barber et al., 1999;Kobashi et al., 2007). δ 18 O ca values of H14 and H5 (Hoti Cave) show this period was characterised by a weakening of rainfall and led to a hiatus of H14 growth (Cheng et al., 2009b). 4) Summer precipitation declines at ∼6.2 ka BP. At Qunf Cave, this decline is gradual and closely follows the 30°N isolation-curve (for an extended discussion, see Fleitmann et al., 2007). At Hoti Cave, this precipitation decline is more abrupt (identifiable by change point analysis; Figure 2A) and related to winter rainfall becoming the dominant source of precipitation in northern Oman . As Hoti Cave provides solid timing on the shifting dominance of winter vs. summer precipitation, the H5 record has been used to define the duration of SAHP 1 and is consistent with the 230 Th ages of Holocene stalagmites from Mukallah Cave Nicholson et al., 2020). Whereas the Y99 record indicates SAHP 4 (during MIS 5e) persisted for ∼6.5 kyrs, the Hoti Cave composite record indicates SAHP 1 lasted for a shorter period of ∼4 kyrs.
These patterns follow established conditions during SAHP 1, which in Southern Arabia are also evidenced by vegetation expansion (Fuchs and Buerkert, 2008), vegetation that requires adequate precipitation (Parker et al., 2004), and palaeolake and river formation (Farraj and Harvey, 2004;Preston, 2011;Berger et al., 2012). Across Arabia, these changes are asynchronous (Preston and Parker, 2013;Preston et al., 2015), with northern Arabia experiencing a truncated period of increased rainfall compared to the south.

MIS 5e
What do the varied conditions between SAHPs mean for discussions of human populations and climatic extremes? There is a growing body of evidence which relates Pleistocene human movements between Arabia and Africa to periods of enhanced precipitation. Archaeological remains at Jebel Faya were dated to MIS 5e and may evidence the earliest instance of H. sapiens in the region (Armitage et al., 2011). Outside of Frontiers in Earth Science | www.frontiersin.org November 2021 | Volume 9 | Article 749488 Arabia, MIS 5 H. sapiens fossils uncovered at Skhul, Qafzeh (Israel, Millard, 2008) and Fuyan Cave (≥80 ka BP, China; Liu et al., 2015) represent some of the earliest instances of Late Pleistocene humans outside of Africa. MIS 5e saw the most intense enhancement of precipitation, highlighting that this period may have been particularly favorable for hominin occupation and dispersal across the Saharo-Arabian deserts (Larrasoaña et al., 2013;Nicholson et al., 2021b). Such a large increase of precipitation was likely echoed by a greater vegetation response than later SAHPs, as evidenced by Mukallah Cave δ 13 C ca values (Nicholson et al., 2020) and the Jebel Faya phytolith record (Bretzke et al., 2013). It is thus likely that the carrying capacity of the Arabian Peninsula was greater during SAHP 4 compared to subsequent SAHPs, meaning population expansions and/or dispersals could have been rapid (Nicholson et al., 2021b). Additionally, the longer duration of SAHP 4 indicates that "green" environments were longer-lived than in SAHP 1, offering potentially longer-term occupation of the now arid interior. In this sense, climatic conditions during MIS 5e were at one extreme of Southern Arabia climatic variability and should not be understated as an optimal period for human dispersal. However, it must be noted that archaeological finds are also dated to MIS 5c, 5a and 3 (Petraglia et al., 2011;Rose et al., 2011;Delagnes et al., 2012;Groucutt et al., 2018). While MIS 5e may therefore be the most extreme period of increased rainfall, other periods were still able to support human populations despite being "less favorable". Do these climatic differences suggest that strategies of survival differed between SAHPs (e.g., Bretzke and Conard, 2017)? Were subsequent dispersals more limited in terms of numbers of people and other animals? What do statistically significant differences in δ 18 O ca values translate to in terms of annual rainfall differences, as well as spatio-temporal variance on long (e.g., millennial) and short (e.g., annual) timescales? Or were the additional benefits of SAHP 4 compared to other SAHPs simply not that important for human occupation (i.e., humans could make do with less)? Additionally, the presence of H. sapiens in Arabia within MIS 3 suggests either occupation throughout the MIS 4 glacial or reentry despite a "drier" climate (Armitage et al., 2011;Delagnes et al., 2012). While these are questions for future research, one message we may take from this is resilience/adaptation despite climatic differences, and that -while the stalagmite record provides useful information on the timing of major climate changes and major H. sapiens biogeographic shifts -providing a climatic "bench-mark" for Late Pleistocene occupations from the stalagmite record is too deterministic and overlooks taphonomical biases and dating uncertainties within the archaeological record.
One thing that is perhaps clearer is that the termination of these wet periods saw a substantial change in environmental conditions. The termination of SAHP 4 likely meant annual rainfall declined to <300 mm yr −1 and was echoed by a decline in vegetation resources. In terms of the "lived" experiences of humans, such a decline would have likely required a shift in survival strategies (Nicholson et al., 2021b). This may have included increased home-range foraging size and mobility patterns, retraction to high-resource retaining areas (such as the Yemeni Highlands; Delagnes et al., 2012;Delagnes et al., 2013) or in some cases dispersal out of Arabia (Nicholson et al., 2021b). Such responses to declining precipitation were also likely variable and not simplistic. Recent archaeological finds in Northern Arabia hint at techno-cultural continuity between Mid-Pleistocene wetter phases , perhaps suggesting human resilience to increasingly unfavorable climatic conditions.
Reduced rainfall following ∼6.2 ka BP led to a temporary end of Neolithic herding in the desert interiors, shrinking population numbers, and migration to areas with greater ecological diversity, perhaps suggesting a minimum amount of precipitation is required for human occupation in these marginal environments (Uerpmann, 1992;Vogt, 1994;Uerpmann, 2002;Potts et al., 2003;Goudie and Parker, 2010). However, human communities returned to the interior of Southern Arabia from ∼5.2 ka BP without amelioration of climate, which even aridified further; varied occupation continues until the modern day (Magee, 2014;Petraglia et al., 2020). Therefore, it seems likely that drier climates create challenges for human populations, but these can be overcome by technological (e.g., mustatils, pottery, water management; camels domestication) and strategic (e.g., mobility, pastoralism) adaptations (Petraglia et al., 2020).
Finally, stability and variance of precipitation (which can be hard to detect in palaeoclimate records) may have been more influential to humans than long-term changes in amounts (Thornton et al., 2014). A temporary transition to herding practices occurred in some parts of Arabia during the 8.2 ka event (Drechsler, 2009;Crassard and Drechsler, 2013), whilst more positive δ 18 O ca values (drier conditions) are observed at Hoti cave ( Figure 2A). Conversely, Cremaschi et al. (2015) suggested that-although increasing precipitation ∼10.5-9.5 ka Frontiers in Earth Science | www.frontiersin.org November 2021 | Volume 9 | Article 749488 BP facilitated occupation-overly "wet" landscapes at Jebel Qara after 9.5 ka BP led to site abandonment and a preference for coastal settings, hinting at the varied human responses to fluctuating climatic conditions.

CONCLUSION
Overall, when compared to other periods, stalagmite climate records indicate that the African and Indian Summer Monsoons were most intense during MIS 5e, which was one extreme of climatic variability in the last 130 kyrs. This was likely an important period for the dispersal of H. sapiens from Africa, as well as occupation in the now desert interiors of Arabia, and the subsequent decline back to more arid conditions likely impacted survival strategies in Southern Arabia. A comparably weaker intensification of precipitation (yet long-term trends are comparable) occurred during the Early Holocene. The expansion of human populations into the now arid interior was similar to MIS 5e but the responses to climatic variability and subsequent aridification differed. We emphasize that evidence for human occupation during all periods of insolation maxima, and the varying climates of these drier periods, highlights human resilience/adaptation despite climatic differences and mitigates against a simplistic narrative. Future research will benefit from the addition of climate/ environmental proxies (trace-element and perhaps aDNA), increased surveys to advance the spatial-temporal coverage of the speleothem record and development of continuous climate records for SAHP 3 and 2. Understanding the shifting survival strategies in the context of declining rainfall and aridification, as Arabia transitioned from one extreme to another, will be of key importance to future debates of H. sapiens biogeography, behavioral flexibility, and both past and future climate-induced socio-political change.

AUTHOR CONTRIBUTIONS
SLN acted as primary author for the article, conceptualizing the manuscript, producing the initial draft and figures and ongoing editing of the manuscript. MJ acted as secondary author, assisting with the initial draft, and editing of the manuscript. RH and DF supervised and edited the manuscript.

FUNDING
This work was supported by the AHRC South, West and Wales Doctoral Training Partnership (Grant AH/L503939/1) the Swiss National Science Foundation (Grant PP002-110554/1 to DF).

ACKNOWLEDGMENTS
We thank two reviewers for their constructive feedback on our manuscript.