|The authors have made considerable effort to improve this manuscript. The organization and writing are now better suited for a Climate of the Past audience rather than an Organic Geochemistry or GCA audience. The authors have provided important additional information on the age model and other issues that were previously absent. They have also provided a new/revised discussion of the Challa BIT record over the past 2 millennia in comparison with the varve thickness record and other records from the region, which will be of broad interest to Climate of the Past readers. The manuscript has overall been improved by these additions and revisions. As in my first review, I feel as though this very detailed modern time series is unprecedented and worthy of publication in Climate of the Past. However, three substantial issues still need to be addressed in order for this paper to be publishable.|
Issue #1.) The discussion of how BIT works is now more clear and streamlined in the paper. However, some confusion still remains about how BIT relates to climatic parameters. The authors argue that seasonal Thaumarchaeota blooms during austral summer determine the amount of Crenarchaeol deposited in the sediment. The BIT index strongly correlates with Crenarchaeol. Therefore, the BIT index is sensitive to seasonal Thaumarchaeota blooms, which are suppressed during periods of high soil erosion, which themselves occur when a heavy rainfall event follows a drought. The chain of events that occurs between climate parameter and GDGT production is now more clear in this revised manuscript than in the previous manuscript. The authors’ new statement “The BIT index can thus be considered to reflect the frequency of ‘extreme’ soil-erosion events, which in this semi-arid region have a threshold relationship with rainfall extremes” (lines 462-464) is now well-argued in the paper.
However, the authors provide several other mechanisms that are either confusingly worded, or contradictory to each other. They state several mechanisms by which increased rainfall can lead to high BIT (either short time scales relevant to erosional events, or longer timescales integrated by lake level) by mobilizing (micro?)nutrients from the soils and promoting productivity of planktonic and microbial communities that outcompete Thaumarcheota, leading to relatively less Crenarchaeol in the sediments, increasing BIT values. A different mechanism describes the dry conditions/low BIT correspondence, namely that Thaumarchaeota thrive after the diatom blooms that follow windy conditions, hence promoting Crenarchaeol production and decreasing BIT values. These mechanisms are mainly discussed in section 4.4.
However the authors also provide a mechanism in lines 422-440 by which prolonged/strong windy conditions (and inferred dryness) can also cause high BIT values, via outcompetition by GDGT0-producing archaea at high ammonium levels. It appears that prolonged windiness/mixing and high precipitation can both lead to high BIT. In a sense, the authors’ logic is consistent in that the production of Crenarchaeol is strongly tied to fluxes of nutrients into the suboxic zone. However this can be accomplished by dry conditions (windy→more/deeper mixing→mobilization of nutrients from deeper in the water column) or by wet conditions (erosion events→transport of nutrients from the soils into the lake).
Therefore I am still unclear on the mechanisms by which BIT reflects wet or dry conditions. As the stated objective of this paper is to describe the mechanisms by which climatic conditions are translated into BIT at Challa, the discussion of these mechanisms needs to be crystal clear, especially to a non-organic geochemistry audience (inorganic chemists, climate modelers, etc etc etc).
Once those mechanisms have been clarified in the text, it would be helpful for the authors to include an illustration with flow charts that better summarize the processes that indirectly tie rainfall to BIT. Such an illustration would very much aid non-organic geochemists in following the logic in the paper.
1. BIT and dry conditions:
Prolonged windiness → deeper mixing → remobilization of nutrients from lower water column to the photic zone → diatom blooms → ammonium released by biomass degradation →
1a.) EITHER: → nitrification by Thaumarchaeota → Thaumarchaeota bloom → Crenarchaeol production → low BIT
1b.) OR: → outcompetition by GDGT-0 producing archea → less Thaumarchaeota → less Crenarchaeol → high BIT
... perhaps depending on the level of ammonium, which influences competition of nitrifying archaea vs bacteria?
2. BIT and wet conditions:
2a.) Increased rainfall → high soil erosion → extra nutrients → change in planktonic and microbial communities → outcompetition of Thaumarchaeota → reduced crenarchaeol → high BIT
2b.) High lake level → increase in accommodation space for Thaumarchaeota in suboxic zone → lacustrine brGDGT production rather than Thaumarchaeota → reduced crenarchaeol → high BIT
Issue #2.) Of all the mechanisms described in the paper (1a,1b,2a,2b above), the ones that apply to most of the paleoclimatic interpretations are related to dry, windy conditions (#1a especially). There is indeed correspondence between windier conditions and reduced rainfall on multi-month to annual timescales, as described in the supplement of Wolff et al 2011. However, it is incorrect to say that BIT (or varve thickness) is a record of monsoon precipitation. From the discussion, it is clear that BIT is a record of dry season strength/length. The difference may just appear semantic to the authors but it is of great importantance to the wider palaeoclimate community. Changes in the annual cycle of precipitation in the past, such as the balance between the two rainy seasons (eg during the early Holocene), will affect the degree to which dry season length co-varies with wet season rainfall. Moreover, “monsoon precipitation,” “ITCZ,” “ENSO driven rainfall,” and “extreme soil erosion events” are often used interchangeably in palaeoclimate literature, but this is a problem for climate dynamicists, since these terms are all quite different from a climatological standpoint. Since the audience of Climate of the Past includes climate modelers and observationalists, this needs to be made very clear in the paper. In addition, the subject heading in 4.5 should be changed accordingly.
Relevant from the authors’ Response to Referees: Wolff et al do infer wetter or drier conditions based on varve thickness because windiness is so strongly related to ITCZ position, which itself is modulated in East Africa by ENSO. However they are careful with their wording. They state that varve thickness represents regional windiness, but they only bring in a more direct discussion of precipitation itself once they compare the varve thickness record with the BIT index. Therefore, to claim that the BIT index discussed in the current manuscript is a record of precipitation because the varve thickness record is a record of precipitation is rather circular reasoning.
Issue #3.) Throughout the paper, more clarification needs to be made regarding the authors’ meaning of “high resolution” and “low frequency” etc. Most of the palaeoclimate interpretations seem to point toward the new BIT record as being a better recorder of high frequency variability than the varve thickness record, but the discussion in lines 509-519 seem to point toward the varve thickness record also not being a good recorder of low frequency variability. The authors discuss event, seasonal, interannual, (multi)decadal, and centennial timescales in this paper. They should clearly define which ones they mean to be high frequency/high resolution/low frequency/low resolution.
Line by line comments:
The use of GPCP precipitation is an excellent improvement over the previous version of the manuscript, which intermingled rain events from Taveta and Challa on one plot.
Line 50: The addition of an appendix with structures and nomenclature is very helpful.
Line 55: The first time Challa is introduced, a latitude/longitude should be given
The figures need to be re-numbered such that they appear sequentially in the text. Currently, Figs 2, 5, 7 are out of order.
Lines 536-7: The discussion of the more recent BIT minima is a very interesting addition considering all the author studies that find dry events at these times. The authors should add that the drought around 1870-1890 AD also occurs in the Lake Challa dD record (Tierney et al., 2011, QSR) and on Mount Kenya (Konecky et al 2013 Palaeo-3). Comparison of these records will bolster the authors’ point that these droughts were widespread. Otherwise, the comparison is lacking since Ethiopia has a different seasonal rainfall pattern than the more equatorial sites. Additionally the presence of this drought in multiple locations lends more credibility to the authors’ claim that Challa BIT reflects annual scale rainfall/drought, rather than just patterns that are specific to the March-May rains.
Line 576: Tierney et al. 2013 should also be cited here.