| 22 May 2026
Late Pleistocene temperature reconstruction of the interior Kalahari Plateau of Northwestern Botswana
Abstract. Late Quaternary climate reconstructions for southern Africa remain poorly constrained due to the scarcity of paleotemperature data, particularly across the interior, highly elevated Kalahari Plateau. This study addresses the gap by focusing on the Late Pleistocene period in Ngamiland of NW Botswana. We conducted an integrated and multi-proxy mineralogical and isotope geochemical study of marlstones, fossil gastropods, soil concretions, calcretes, and pan and lake sediments from short cores and outcrops of the Kalahari Group near the Tsodilo Hills. Our approach was applied to address the following objectives: (i) identify primary carbonates using petrographic screening, (ii) constrain the hydroclimate using δ13C and δ18O stable isotope signatures and processes, (iii) provide new age constraints of deposition using radiocarbon 14C dating, and (iv) interpret paleotemperatures by analysing clumped isotopes (∆47) and dual clumped isotopes (∆47-∆48). This multi-proxy approach aimed at defining a robust proxy to reconstruct the Late Pleistocene climate in the interior Kalahari. Our studies revealed that the sediments were deposited between 49.05 ± 1.98 and 24.89 ± 0.13 ka BP ago based on 14C calibrated ages (SHCal20). The marlstones, calcretes and soil concretions are micritic and sparitic with evidence of bioturbation, whereas fossil gastropods are devoid of reworking and suited for paleoclimate studies. The δ13C and δ18O values range between 0.1 and 4.2 ‰ VPDB and -0.3 and 9.4 ‰ VPDB, respectively while measured ∆47 temperatures range between -3.1 and 47.5 oC. Stable isotope variations suggest a strong biological control in the soil concretions, whereas the groundwater signature is likely overprinted in the calcretes, and a combination of meteoric and groundwater has influenced the texture of marlstones. The extreme temperatures are attributed to organic contaminants in the sediments. It is interpreted that fossil gastropods retain the primary geochemical signatures based on the stable and clumped isotope variations and therefore represent the most reliable material for climate reconstructions in the region. Our study has provided the first direct temperature estimates of the Late Pleistocene for the interior Kalahari region from these gastropods which appear to be cooler than the present at ~15 oC. This study therefore addresses a major data gap in southern African paleoclimate records, which has implication for global climate models.
This is an interesting study in confirming, using clumped isotope data from fossil gastropods, that temperatures in NW Botswana were lower during the Late Pleistocene relative to the present. The mean annual temperature estimate of ~15 deg C aligns well with previous palaeotemperature estimates from Nobel gas and δ¹⁸O isotope analysis of groundwater in Stampriet, Namibia, which suggest that glacial temperatures were roughly 5°C to 6°C cooler than today. The paper is well constructed and generally well written but of variable quality. In particular, the literature elements of section 2 suggest a lack of engagement with the wider literature on Kalahari geology and palaeoenvironments.
I cannot comment on the robustness of the various analytical results. However, what is key is how the samples on which the analyses were conducted originated. I’m not entirely clear from the text that the authors have fully got to grips with how the sediment sequences formed – and hence what the palaeotemperatures from the cores actually mean. Are we dealing with a progressively accumulating sediment pile, with calcretes representing periods of landscape stability and pedogenesis? Or, as the presence of sparry calcretes might suggest, periods when calcretes were forming at depth close to a palaeo-water table within an already deposited sediment sequence. Or both? The implications of the two contexts for the interpretation of palaeotemperature results are very different. As the authors point out, temperature estimates from soil concretions and Bk concretions are generally higher and presumably reflect near-surface temperature. Could groundwater affected measurements represent temperature at depth? Regardless, I agree that fossil gastropods appear to be the best way forward for future investigations.
Specific comments
Lines 50-54 – the word ‘region’ is used 4 times here – I suggest an edit.
Line 58-60 – this is a big claim. What about the palaeotemperature data from the Stampriet region by Stute & Talma (1998)?
Stute, M., & Talma, A. S. (1998). Glacial temperatures and moisture transport regimes reconstructed from noble gases and δ18O, Stampriet aquifer, Namibia. IAEA (pp. 307-318)
Line 71 – Nicholson (2108) has argued against the use of the term ‘ITCZ’ for studies of African rainfall – you would be better referring to ‘shifts in the tropical rainbelt’.
Nicholson, S.E. (2018). The ITCZ and the seasonal cycle over equatorial Africa. Bulletin of the American Meteorological Society 295: 337-348.
Line 73-75 – surely you mean ‘mean monthly temperatures’? Also, why the two different definitions for the austral summer on adjacent lines?
Line 75-78 – this feels like a big jump moving from discussing the climate of Ngamiland to the Okavango River. A bridging sentence is needed.
Line 80 – the five different shades of green are almost impossible to distinguish on this figure.
Line 85-86 – this is a vast over-simplification of the age and hydroclimatic regime of the Makgadikgadi Pans – see the numerous other papers by Burrough, Shaw, Thomas and others, plus studies of the earlier basin evolution by Moore.
Line 89 – again, what about Stute & Talma (1998)?
Line 90 – should be 2.2.
Lines 91-92 – this is a vast over-simplification of the basement geology of NW Botswana. There are numerous rock types in addition to granite and gneiss beneath the Kalahari Group cover.
Lines 100-101 – needs a supporting citation.
Line 130 – this whole section needs a lot more detail – see below.
Lines 131-135 – it would be helpful to include a little more detail about the KGP, TSO1 and NGA2 cores. How long are these cores? At what depths were they sampled? Simply the giving the borehole head elevation in Table 1 is not helpful. I know more detail is given in section 4.1 but more contextual information here would be useful.
Line 135-136 – how do you know the outcrop samples from Tsodilo are of Pleistocene age (they almost certainly are, but you need to justify this)?
Line 140 – how representative are two modern shells as analogues?
Lines 143-145 – for purposes of replicability, how did you identify preservation states?
Lines 252-257 – It would be useful to provide some reflection on the veracity of the radiocarbon ages. They include measurements on carbonate within marlstones at depths of tens of metres below ground level. Presumably, groundwater bearing carbonate has percolated through these sediment and – potentially – introduced younger carbon into the samples. This could make them appear younger than their true age.
Line 260-265 – can you say a little more about how you think the stack of sediment in the cores formed from these data? Presumably, if there is evidence of pedogenesis then there may have been cycles of surface stability followed by phases of burial to build up the profile. How might this affect the interpretation of the isotopic record? This feeds into section 5.1.
Line 291 – it would be helpful to explain why the low temperature value reflects contamination. It is perfectly feasible for temperatures in NW Botswana to fall as low as -3.1 deg C.
Line 314-323 – I would like to see more discussion of the validity of using two shell samples for validation of the clumped isotope data in this section. Is this standard practice?
Line 346 – a sparry calcrete will almost certainly have formed as a groundwater calcrete, close to a palaeo-water table. It would be worth exploring the implications of this for the interpretation of palaeotemperature data.