Journal of Human Evolution 59 (2010) 139e142

Contents lists available at ScienceDirect

Journal of Human Evolution journal homepage: www.elsevier.com/locate/jhevol

News and Views

Reply to “Is the Pliocene Ethiopian Monsoon extinct? A comment on Aronson et al. (2008)” James L. Aronson a, *, Million Hailemichael b a b

Earth Sciences, Dartmouth College, 6105 Fairchild Hall, Hanover, NH 03755, USA Center for Archaeology, Materials, and Applied Spectroscopy, Idaho State University, Pocatello, ID 83209, USA

a r t i c l e i n f o Article history: Received 18 October 2009 Accepted 13 January 2010 Keywords: Ethiopian Monsoon Afar Hadar Paleosols Hominid environments Ethiopian climate change Oxygen isotopes African paleoclimatology

Background The answer to Wynn and Bedaso’s (2010) title question is no, the summer Ethiopian Monsoon will continue dominating Ethiopia’s meteorology, waxing and waning in strength, until the tectonic plates rearrange themselves. But with our limited knowledge of how the d18OMRW correlates to amount of rainfall, their new improved estimate of the mean value of the d18 of today’s rain in the Hadar area (d18OMRW) is insufficient to overcome our paper’s tenet that the Ethiopian Monsoon was much stronger, and Hadar much more pluvial, during the late Pliocene than today. The annual rainfall and its seasonal distribution determine the structure, diversity, and productivity of East African ecosystems, including their carrying capacity for hominid populations. Although it is only one proxy of a few for paleo-rainfall, the d18OPDB of paleosol carbonates is important because it can be translated into the d18O of the paleo-soil water and, by inference, into the d18OMRW of the ancient rainwater. Of the many meteorological factors that can interact to change the value of d18OMRW, all the ones we indicated in our paper as associated with a stronger Monsoon and more pluvial Ethiopia tend to shift the d18OMRW in the direction of a more negative

* Corresponding author. E-mail address: [email protected] (J.L. Aronson). 0047-2484/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.jhevol.2010.04.001

value. By example, Wynn and Bedaso’s two storms at w0& and e6& show that such a general tendency can be entirely masked or reversed by the high inter-storm variability in d18O when measurements of only a few storms are used to represent an area’s d18OMRW , as we did at Hadar. Wynn and Bedaso (2010) have produced a creditable value for the d18OMRW of modern rain at Hadar that has overcome the interstorm variability issue. We accept their value of
1.6& as the best running benchmark of Hadar’s d18OMRW, and superior to our few Monsoon rain measurements of þ1.4&. However, the question at hand is how sensitive is the amount of change in d18OMRW since the Pliocene to the change in amount of rainfall. Wynn and Bedaso (2010) implicitly presume this relationship has been calibrated for Hadar when they use a reduction of half of the amount of change in d18OMRW to conclude that the current Ethiopian Monsoon is not strongly diminished compared to the Pliocene. Wynn and Bedaso’s (2010) new value of d18OMRW for modernday Hadar Two bodies of data were available when we wrote our paper: (1) our five measurements of intense Spring and Monsoon storms; and (2) the contradictory wadi shallow aquifer measurements by Levin et al. (2004) at Gona. We did not use the wadi aquifer data from Gona at the edge of the escarpment because we thought they were hydrologically too prone to tapping spring inputs of the isotopically negative abundant Plateau groundwater as well as local Afar rainwater. Wynn and Bedaso (2010) creatively selected wadi aquifers from East Dikika, across the Awash from the escarpment, to measure d18OMRW. We agree these should be free of inputs from the escarpment and accept their evaporation-corrected extrapolation of
1.6& as the current best benchmark for the d18OMRW of modern rain at Hadar. But the considerable spread along “the meteoric line” of four of their eight Asbole wadi samples in West Dikika (their Fig. 3) proves this wadi, in heading up the escarpment, is significantly contaminated by plateau-derived waters, likely by both present-day (w
2.5&) and fossil groundwaters (w
10&). Therefore the Asbole data cannot be pooled with the unimpeachable East Dikika data to move the e1.6& benchmark to the more negative
2.6&, which the labeled arrow in their Figure 3 suggests as a viable alternative. Furthermore, the Asbole results permit that the wadis at Gona are also contaminated. In accepting Wynn and Bedaso’s (2010) modern d18OMRW, the

140

J.L. Aronson, M. Hailemichael / Journal of Human Evolution 59 (2010) 139e142

isotopic shift in going from the Pliocene d18OMRW to that of the present is only þ2.5&, not as striking a shift as the þ5.5& we projected in our paper. Our climate change model of the Ethiopian Summer Monsoon has the Afar of the past one million years predictably cycling from two dramatically different rainfall regimes: a long w100 kyr arid one which we argued includes today, cyclically interrupted by brief pluvials like the latest one, the African Humid Period (AHP) of w9e6 ka ago. We held that the distal Awash River flood and delta plain where it met the expanded Lake Gamari in the central Afar during the AHP was both a near-modern depositional and a climatological analog of our persistently wetter late Pliocene Hadar. They claim we based our paleoclimatic argument that Pliocene Ethiopia was more pluvial than today on just a comparison of our d18OMRW for late Pliocene Hadar to, what turns out to be our too positive an estimate of, the d18OMRW “fingerprint’” for modern-day Hadar rain. This is incorrect. As reflected by the title of our paper our analysis and derivation of the paleo-rainfall was comprehensive and marshaled non-isotopic sedimentological and fossil paleoenvironmental evidence together with the paleosol d13C data on paleovegetation in addition to the d18O on the paleosol carbonates (Aronson et al., 2008). Among this non-isotopic evidence, we cited the abundance and diversity of the fossil vertebrates that imply a considerably more wooded ecosystem than today (Reed, 1997). The depth of the carbonate horizon of the paleosols, particularly at A.L. 333, using the conservative criteria of Royer (1999), suggested about 65 cm/yr of annual rainfall, which is at the low end of the estimated range of 80e120 cm/yr in the comparative palynology study by Bonnefille et al. (2004). Wynn and Bedaso (2010) seem to have ignored this body of non-isotopic evidence that the rainfall at paleo-Hadar was likely significantly greater than today.

A reduced d18OMRW of modern rain by even half does not refute a stronger Ethiopian Monsoon in the Pliocene than today In Figure 1 we plot our 64 d18OCarbonate values for Hadar paleosols as a histogram. The carbonate-water fractionation equation calculates the mean d18OMRW value for Pliocene rain at Hadar as
4.1& from the mean paleosol d18OCarbonate of
6.37&. We smoothed the histogram into a curve and colored it green to signify our view of a wetter and greener Hadar in the Pliocene. To illustrate the impact of Wynn and Bedaso’s (2010) new improved
1.6& value of d18OMRW in Figure 1 we replicate and shift the Hadar carbonate histogram toward the left to center it on two less negative d18OPDB values. These two, in yellow and red, would be the hypothetically identically shaped histograms for w60 comparable modern-day carbonates, should they have been forming over recent times in the immature soils surrounding Hadar. The yellow one is centered about the
3.9  1& carbonate-equivalent of Wynn and Bedaso’s (2010)
1.6& modern rain; while the red one (only partially shown) is centered about the carbonate-equivalent of our now less favored þ1.4& d18MRW. The isotope distance in going from the late Pliocene (green) histogram to the yellow one is about half what it was in going to the red one. Even though it partially overlaps the Pliocene green histogram, the hypothetical yellow histogram is statistically distinct from it at n ¼ w60. Wynn and Bedaso’s (2010) “isotope difference” in going from some past value of d18OMRW to the present value is a vector with direction and length, here called the d18OMRW Shift as shown in Figure 1. Of key importance, the observed direction of change in the d18OMRW Shift from the past to the present d18OMRW is toward more positive values. This plausibly suggests a shift to a diminished Monsoon and less rainfall today than in the late Pliocene, no matter if the þ1.4& we used is replaced by the likely

The Hadar Paleosol Histogram in equlibrium with hypothetical δ18OSMOW waters

Hadar Paleosol Hadar Paleosol Histogram in eqil. w/ Histogram in eqil. w/ -1.6 rain water +1.4 rain water δ18OSMOW @ 25°C δ18OSMOW @ 25°C (Aronson et al., 2008)

(Wynn and Bedaso, 2009)

Mean Hadar & L. Busidima Fms. -6.37±1.03 (n=64) = -4.1 soil water δ18OSMOW @ 25°C The Hadar δ18OPDB Paleosol Histogram

18

the δ O Shift

(Aronson et al., 2008) Above Disconformity Below Disconformity

-1.0

-2.0

-3.0

-4.0

-5.0

-6.0

-7.0

-8.0

-9.0

-10.0

18

Hypothetical and Actual Oxygen δ OPDB in Hadar Paleosol calcite Fig. 1. In green, to the far right, is plotted the actual histogram of d18OPC(Paleosol Carbonate) measured relative to the PDB standard on all 64 samples by Hailemichael (2000) from the Hadar and Lower Busidima Fms. Also shown are two hypothetically identically shaped histograms in yellow, in the middle, and in red to the far left as if the relatively identical spectrum of PC results was precipitated from isotopically less negative rain, that is, in the direction associated with less rainfall. The yellow histogram centered at
3.9& PDB would theoretically be forming in Hadar’s immature soils from soil water of Wynn and Bedaso’s (2010) newly determined d18OMRW of
1.6&. The red histogram, partially shown and only for illustrative purposes, would be forming if our previously estimated d18OMRW of þ1.4& were right. At
3.9  1&, the hypothetical modern yellow carbonate histogram is identifiable statistically (n ¼ w60) as distinct from the Pliocene histogram at
6.4  1&, and is displaced by the D shift in d18OMRW arrow (in blue) from the Pliocene value of
4.1& in the Pliocene to the newly determined
1.6& d18OMRW of Wynn and Bedaso (2010) in the direction of a diminished Ethiopian Monsoon. The þ2.5& amount of shift is about half the þ5.5& change we projected in our paper. We emphasize here that the direction of the isotope shift is away from more pluvial conditions and away from a stronger Summer Monsoon in the Pliocene but that the amount of shift has never been calibrated for amount of rainfall along the western Afar gradient in modern rainfall.

J.L. Aronson, M. Hailemichael / Journal of Human Evolution 59 (2010) 139e142

superior
1.6&. Until the magnitude (length) of the isotope shift is calibrated for rainfall amount in the western Afar, it is premature of Wynn and Bedaso (2010) to use the reduced length of the shift in going from the paleo-Hadar green histogram to the yellow (as opposed to our red histogram) histogram to negate a much more pluvial Ethiopian Monsoon in the Pliocene than at present. A tentative correlation of the length of the shift in d18OMRW with rainfall Until future comparative research of modern environmental features with increasing rainfall refines or rejects the estimates of 65 cm/yr of
4.1& paleo-rainfall and makes a longer term verification of Wynn and Bedaso’s (2010)
1.6& d18OMRW for the 40 cm/ yr of modern rainfall, the magnitude of the shift in d18OMRW sits crudely at w
1& for every 10 cm/yr increase in annual rainfall ([Dd18OMRW ¼
4.1& minus
1.6& ¼
2.5&] divided by [D rainfall of 65e40 ¼ 25 cm/yr]). This is roughly half of the more sensitive
2& for every 10 cm/yr increase in annual rainfall it would have been with the d18OMRW of modern rain we used in our paper. A modern example from the Afar suggests such a reduced, lower

141

sensitivity of d18OMRW for rainfall may not be unreasonable for Afar settings. Schoel and Faber (1976) intensely sampled shallow wadi aquifers to estimate a d18OMRW south of the city of Djibouti, at the eastern edge of the Afar, 300 km east of Hadar. Unlike the rest of Ethiopia, the narrow coastal strip of the Afar has a “Mediterranean” dry-summer climate at the arid extreme, with only 15 cm of annual rain in winter and spring. Four of nine of their samples have dDs, which with the d18O, fall on the meteoric line and indicate the sampled waters were unevaporated. The d18O’s are extremely precise at
1.4  0.14&, and surprisingly close to the
1.6& at Dikika. Any rains during Djibouti’s rainy season must come from the Indian Ocean. This is because only the heating produced by a summer sun, positioned well north of the equator can produce the strong atmospheric lows capable of pulling Atlantic air-masses across Africa to Ethiopia (and especially Djibouti). The negative d18O character of the Djibouti wadi aquifer waters and their value close to Wynn and Bedaso’s (2010) new d18OMRW value for modern rain at the western escarpment boundary of the Afar open up the reasonable meteorological possibility that all the rains in the western Afar, not only today but perhaps also in the humid episodes of the past, have also been largely derived from the Indian

Fig. 2. In African biomes with pronounced dry seasons the amount of Summer Monsoon rainfall controls the vegetative structure. We pose two modern African biomes as rival analogs of Pliocene Hadar in contrast with today’s Hadar and its w40 cm/yr of annual rain (interpolating from the two Ethiopian Meteorological Services Agency [EMSA] stations in the Western Afar, at Mile [with 30 cm/yr] and at Gewani [with 48 cm/yr; Ayenew et al., 2007]). With 65 cm/yr of annual rainfall, biome (A), here only 15 km from the 40 year operational rain station at Kasane, Botswani (Jain et al., 2006), exemplifies the Zambezi belt of grassy woodland with its diverse, spaced, sizeable C3 trees and tall C4 grasses. At w50 cm/yr annual rainfall as interpolated between the EMSA rain stations at Gewani (48 cm/yr) and at Metehara (51 cm/yr; Ayenew et al., 2007), biome (B) is a much less diverse low tree, scrub-thorn savanna on a well-developed soil in the slightly wetter southern Afar, 200 km south of Hadar. In the distance beneath the 2500 m high Asebot volcano these plains can be seen to grade into a savanna with only scattered low acacia. Photo (C) is at today’s Hadar and emphasizes the harsh limitations imposed on any hominid population during the pronounced Afar dry season, especially today when so little of the smaller wet-season rainfall is stored by nature over the highly evaporative dry season (photos by Aronson).

142

J.L. Aronson, M. Hailemichael / Journal of Human Evolution 59 (2010) 139e142

Ocean. The significant increase in amount of rainfall from Djibouti’s 15 cm/yr to Hadar’s 40 cm/yr from Indian Ocean air-masses comes with hardly any decrease in d18OMRW. This modern Afar example of an even lower sensitivity of Dd18OMRW for amount of rainfall, albeit for different seasons, makes it plausible that the yellow (center) histogram of Figure 1 has not moved so close to the green one to invalidate our tenet that the Ethiopian Monsoon has “strongly diminished” from the Pliocene to the present. A more relevant calibration of Dd18OMRW with annual rainfall could come from determining d18OMRW along the North-South 30e50 cm/yr rainfall gradient in the western Afar from Mille to Odeitu (near Hadar) to Gewane (data in caption of Figure 2). Was the AHP more pluvial than the Pliocene? In following the conclusion of Wynn and Bedaso (2010; 133-138), which we hesitate to accept, that the present-day strength of the Ethiopian Monsoon is not that diminished from Pliocene Hadar, and in accepting our identical conclusions that the African Humid Period (AHP) of 9e6 ka was clearly more pluvial than the present, Wynn and Bedaso (2010) end up concluding that the AHP was too pluvial to be a modern analog for the climatology of Pliocene Hadar. The evidence in our paper is insufficient to refute this moot point. But the histograms in Figure 1 point to how to test which hypothesis is correct. One should be able to find near-modern paleosols, near Hadar, say in the Gamari Plain, and verify if they formed during the AHP by 14C dates. If we are right, the histogram of such AHP d18OCarbonate results will overlie the green paleo-Hadar histogram; if they are right such a “Gamari histogram” will center statistically distinctly to the right of the green paleo-Hadar histogram (isotopically more negative) and could be colored dark green to signify even wetter conditions than during Pliocene Hadar. Either outcome will sharpen our image of the Pliocene environmental setting along the western Afar by setting a probable upper limit, either way, to the range in annual rainfall that has affected the western Afar from the Pliocene to the present. Conclusions We applaud Wynn and Bedaso (2010: 133-138) for their many important discoveries, which required quick, strategic responses to serendipitous circumstances arising in the field. We fully subscribe to their strong caveat that “without this knowledge of the present (controls on the d18O of Monsoon rainfall), we fall short at interpreting the past.” Their findings and creative strategies put isotopic meteorology further along the path to unraveling the intricate nature of the much neglected Ethiopian Monsoon, so it can be used as a modern analog by which to judge its past history of possible significant periodic changes in rainfall. We pose two modern ecosystems (A and B) as rival analogs of ancient Hadar to compare with modern Hadar (C) in Figure 2. The photos illustrate how important it is to refine the d18O proxy for paleo-rainfall by documenting the present as Wynn and Bedaso (2010) have done nicely here at Dikiki. In African seasonal climates with a pronounced dry season small differences in annual rainfall result in major changes in the structure of ecosystems. (B) is

a low tree scrub-thorn savanna in the Afar with the 50 cm/yr of rainfall we assign to Wynn and Bedaso’s (2010) paleo-Hadar from which the present-day 40 cm/yr Ethiopian Monsoon is said to have not diminished that much. We compare it to (A), a grassy woodland with diverse sizeable spaced trees developed under the 65 cm/yr of rain we project for paleo-Hadar (from non-isotopic considerations). Photo C, at modern-day Hadar, with a rainfall of w40 cm/yr emphasizes the harsh limitations that the long pronounced Ethiopian dry season imposes on populations of any hominid, here a geologist, at midday limited by the water he can carry. Though the steep badlands exposure exaggerates the ecologic bleakness at today’s Hadar, the adjacent undissected plains are largely bare, immature, gravelly Aridosol with an ecologically much less productive open cover of scattered low grasses and shrub-acacia than in B. How one interprets the physical and cultural evolution of Afar hominids of the past differs between the two analogs. Just 15 cm/ yr more rainfall transforms B into A, where Tattersall (1998: 117) projected early hominids could “make the most of their varied locomotor abilities” for gathering food and avoiding predators.

Acknowledgements We thank Ed Meyer for helping produce the figures, and a thorough anonymous reviewer whose detailed comments influenced the structure of our presentation. We especially thank Susan Antón for staying with JHE beyond her term to continue as Editor for this interchange. With her welcoming and efficient manner she shepherded both parties to sharpen our presentations and focus on the testable issues in this debate.

References Aronson, J.L., Hailemichael, M., Savin, S.M., 2008. Hominid environments at Hadar from paleosol studies in a framework of Ethiopian climate change. J. Hum. Evol. 55, 532e550. Ayenew, T., Kebede, S., Alemyahu, T., 2007. Environmental isotopes and hydrochemical study applied to surface water and groundwater interaction in the Awash River basin. Hydrol. Process 22, 1548e1563. Bonnefille, R., Potts, R., Chalie, F., Jolly, D., Peyron, O., 2004. High resolution vegetation and climate change associated with Pliocene A. afarensis. Proc. Natl. Acad. Sci. U.S.A. 96, 12125e12129. Hailemichael, M., 2000. The Pliocene environment of Hadar, Ethiopia: a comparative isotopic study of paleosol carbonates and lacustrine mollusk shells of the Hadar Formation and of modern analogs. Ph.D. Dissertation, Case Western Reserve University. Jain, P.K., Prakash, J., Lungu, E.M., 2006. Climatic characteristics of Botswana. In: Proceedings of the Sixth IASTED International Conference on Modelling, Simulation and Optimization. pp. 96e105. Levin, N.E., Quade, J., Simpson, S.W., Semaw, S., Rogers, M., 2004. Isotopic evidence for Plio-Pleistocene environmental change at Gona, Ethiopia. Earth Planet. Sci. Lett. 219, 93e110. Reed, K.E., 1997. Early hominid evolution and ecological change through the African Plio-Pleistocene. J. Hum. Evol. 32, 289e322. Royer, D.L., 1999. Depth to pedogenic carbonate horizon as a paleoprecipitation indicator? Geology 27, 1123e1126. Schoel, M., Faber, E., 1976. Survey on the isotopic composition of waters from NE Africa. Geol. Jahrb. D17, 197e213. Tattersall, I., 1998. Becoming Human: Evolution and Human Uniqueness. Harcourt Brace, New York (First Harvest, pbk) 258 pp. Wynn, J.G., Bedaso, Z.K., 2010. Is the Pliocene Ethiopian Monsoon extinct? A comment on Aronson et al. (2008). J. Hum. Evol 59 (1), 133e138.