Interactive comment on “Enhanced Mediterranean water cycle explains increased humidity during MIS 3 in North Africa” by Mike Rogerson et al

We thank the two anonymous Reviewers of our draft manuscript for their detailed and constructive reviews, and are extremely pleased that they find our work both interesting and worthy of publication. We fully concur that the interpretation of the data we present is complicated by structure of the dataset, and we are happy both reviewers agree with us that the data itself is so unique as to make a pressing case for publication and, that our analysis of it is fair, balanced and reasonable.

Response: We concur with the reviewer's sentiment that difficulties in the correlation of the timeseries are "uncomfortable", and this is why we approach the data by analysing large groups of datapoints rather than using a time-series approach. We are extremely pleased that this reviewer also feels that collection of the fluid inclusion dataset is "technically sound" (C1, 2, 7), and that the general accuracy of the dataset is supported by our quantitative analysis of it compared to modern rainfall isotopes (C2, 1, 1-2). It can only be concluded that these data are a good representation of this isotopic system, even though they look unusual.
We suggest that the apparently poor correlation of time-series likely arises from aliasing of a complicated signal in the fluid inclusions, and emphasise that it is unlikely the data presented are sufficiently resolved to demonstrate the two datasets actually have different structure. The correlation of the datasets is therefore ambiguous, rather than disproven. Sadly, the at least orderof-magnitude increase in the size of the fluid inclusion dataset needed to resolve this point is not realistic: indeed, this Reviewer notes that this dataset is already "comparatively large" (C1, 2, 1). The most appropriate way forward in this situation is to minimise the interpretation of the temporal structure of the fluid inclusion data we present, and this is what we have done. We are pleased that despite their discomfort, this Reviewer supports publication of this "interesting study" (C3, 3, 13-17).
A conventional equilibrium test is difficult to perform for this dataset, as each measurement comprises a mixture of inclusions with different compositions from each layer, and therefore an unknown position on the mixing line between these end members. Should we compute mean values (arithmetic or volume weighted); or extrapolate end members, and test for equilibrium of both? All these judgements require assumptions we are not in a position to make. Consequently, we are only able to test for equilibrium in the subset of samples where the end members are sufficiently close together for the analysis to be fully "duplicable" (the subset shown in Figure 6). Modern mean winter temperature in Dernah (the nearest city to Susah Cave) is 11.9 o C, with maximum 17.7 o C and minimum 7.1 o C. These fluid inclusions are therefore certainly at least close to isotopic equilibrium with the carbonate hosting them. C2, 2, 17-19: "It would perhaps be useful if the authors discuss that a bit more in the context of the interpretation of their record".
C3, 1, 5-10: "Bringing a third water source in, as is suggested in the ms, cannot really be supported by the data from my perspective………… One could perhaps argue that slight isotope changes within each of these moisture sources can cause similar isotope patterns?" Response:, The Reviewer agrees with our analysis that the data does show a mixing pattern of western and eastern Mediterranean sources (C3, 1, 4-5). So, we assume the 'third source' mentioned above is therefore the Atlantic external water we argue for, and find in relatively small amounts. We happily agree this is the most speculative part of our analysis. However, we also note that finding no Atlantic moisture at all in this dataset is a rather more startling interpretation than our suggestion that we find only a little. Atlantic-sourced moisture contributes to rainfall in central northern Africa today, and this mode of rainfall has previously been argued to be greater in past humid phases . We therefore find this point of speculation actually rather conservative in its nature.
C4, 2, 1-2: "I'd like to know where your duplicable samples from Fig. 6 are located in the stalagmite (stratigraphically). All in one period, or distributed all over?" Response: See Table 1. All three Growth Phases are represented by at least one fully duplicated sample.
C4, 2, 2-4: "Do you have a better correlation with the d18O values of the carbonate when you consider the duplicable dataset only?" Response: Beyond the differences between the three phases (see next response), it is difficult to judge whether there is true correlation between the reduced fluid inclusion dataset and the calcite isotope dataset, because the former is rather small. To make interpretations based on such a "correlation" would seem to us rather speculative. We are safer limiting the discussion of the time series. Atmospheric latent heat is a major component of global and regional climate energy budgets and 35 changes in its amount and distribution are key aspects of the climate system . 36 Equally, in mid-and low-latitude regions, changes in the water cycle have more impact on landscapes 37 and ecosystems than changes in sensible heat . Rainfall in semi-arid regions is thus 38 one of the key climate parameters that understanding future impact on human societies depends upon 39 (IPCC, 2014), making constraining of mid-latitude hydrology a globally significant research priority. 40 These regions, however, have a particularly sparse record of palaeoclimate due to typically poor 41 preservation of surface sedimentary archives . North Africa is a region that fully exhibits these limitations, and large areas present either no pre-Holocene record or else they present 43 highly discontinuous deposits indicating major reorganisation of the hydroclimate, which are 44 challenging to date (Armitage et al., 2007). North Africa also fully exhibits the progress 45 palaeoclimatologists have made in understanding continental hydrological change from its impact on 46  where large tufa deposits attest to higher past groundwater tables . 91 An emerging picture of MIS 3 as a humid period within the Mediterranean basin is developing 92 (Langgut et al., 2018), and the current study focusses on this time period. However, MIS3 is not well 93 expressed in the Sahara region. The Libyan interior is considered arid or hyperarid throughout the last 94 glaciation . Recent re-evaluation of lake levels in southwest Egypt indicates a groundwater fed system was active around 41 ka (Nicoll, 2018), which is similar to dates 96 for springline tufa systems at Kharga Oasis (Smith et al., 2007). We are not aware of continental 97 MIS3 pollen records from the region, but marine pollen from Tunisia indicates more arid conditions 98 through the last glacial than during the Holocene (Brun, 1991). There is a triple peak in runoff from 99 the Nile recorded in the marine sediment record, with maxima at ~60, ~55 and ~35ka, indicating 100 higher rainfall within the upper Nile catchment (Revel et al., 2010). 101 It is unlikely that significant further progress will be made in understanding the palaeoclimate of 102 North Africa without new empirical evidence of regional hydrological changes from which 103 atmospheric dynamics can be delineated. 104

The central North African speleothem record 105
Speleothem palaeoclimatology has high potential for North Africa, but is only recently becoming 106 light early Holocene. The  13 C record indicates cool periods exhibiting higher carbon isotope values, 113 more clearly delineating the Bølling-Allerød / Younger Dryas oscillation than  18 O. This is assumed 114 to reflect higher soil respiration during warm periods (Genty et al., 2006). A major change in the 115 carbon isotopic composition occurred across the transition from the relatively arid glacial to the more 116 humid Early Holocene, and indicates a significant reorganisation of the regional hydroclimate. 117 However, it is difficult to interpret these data in isolation. A recently reported speleothem record (SC-118 06-01) indicates that conditions in northern Libya during Marine Isotope Stage 3 (MIS 3) were more 119 humid than today, and shows isotopic evidence of a teleconnection between temperature in Greenland 120 and rainfall at the southern Mediterranean margin (Hoffmann et al., 2016). The oxygen isotope record 121 indicates that the water dripping into the cave during MIS 3 was isotopically too heavy for the 122 moisture to be sourced from within the monsoon system (Hoffmann et al., 2016). However, beyond 123 ruling out a southern source  18 Occ values alone are not sufficient to determine the origin of 124 atmospheric vapour. Three distinct humid phases within MIS3 are reported from this speleothem: 65-125 61 ka, 52.5-50.5 ka and 37.5-33 ka. Phases I and III occur during times of low precession parameter, 126 when summer insolation on the northern hemisphere is relatively increased. Phase II represents the 127 first evidence for high obliquity being able to cause a pluvial period in the north African subtropics in 128 the same manner as precession (Hoffmann et al., 2016). In SC06-01, all three growth phases are 129 fractured into multiple short periods of growth, and show a marked temporal coherence with 130 Greenland Dansgaard-Oeschger interstadials (Hoffmann et al., 2016). Here, we report fluid inclusion 131 data from this speleothem and discuss how this helps resolve some of the issues discussed above. 132

Fluid Inclusions 133
Speleothem fluid inclusions are small volumes of water that were enclosed between or within calcite 134 crystals as they grew, ranging in size from less than 1 m to hundreds of m . 135 This water represents quantities of ancient drip-water that can be interrogated directly to ascertain the

Fluid inclusions 202
Petrographic analysis of the thick sections indicates that the distribution of fluid inclusions is highly 203 variable, with macroscopically opaque "milky" calcite typical of rapidly growing intervals containing 204 sometimes very abundant inclusions and the discoloured, translucent calcite of the slowly growing 205 intervals being almost inclusion-free (Fig. 2). In most samples, two distinct populations of inclusions 206 were identified with numerous small intra-crystalline inclusions and larger, but less frequent, inter-207 crystalline inclusions. Consequently, the volume of water analysed per sample was very variable (Fig.  208 3). Indeed, a significant proportion of individual fluid inclusion measurements had analyte volumes 209 too small (<0.1 L) to have confidence in the isotope results. A small number of analyses failed due 210 to excessive water saturating the detector, and these have not been included in the datasets presented 211 here. The major impact of the highly variable availability of inclusions in the speleothem is a 212 significant bias in the analyses towards the most rapidly growing, and therefore probably humid, time isotopically similar (with  18 OFI ranging from -7.5 ‰ to -3.8 ‰ and from -8.5 ‰ to -3.2‰ 216 respectively and  2 HFI ranging from -26.7 ‰ to -18.6 ‰ and from -29.4 ‰ to -16.1 ‰ respectively). 217 However, compositions for Phase II are different, particularly with respect to deuterium ( 18 OFI 218 ranging from -8.9 ‰ to -4.5 ‰ and  2 HFI ranging from -38.3 ‰ to -25.1 ‰). 219 In most samples, achieving within-error replication ( 2 H ±1.5‰,  18 O: ±0.5‰) of both  18 Ofi and 220  2 Hfi was difficult. This must reflect more than one population of inclusions with different properties 221 being present within at least some samples, and each replicate analysis represents some proportion of 222 mixing between these populations. This suggests significant short-term variability in the composition 223 of the water stored in the presumably rather small soil/epikarst zone overlying the cave. 224 Consequently, any given time interval risks being under-sampled with regard to variability at that 225 time. Although there is some visual correspondence between the  18 Ofi,   Hfi and  18 Occ data series 226 (Fig. 4),4), it seems that the fluid inclusion time series risks aliasing changes seen in the calcite 227 isotope time series. Consequently, the usefulness of interpretation that can be drawn from the episodic 228 SC-01-06 fluid inclusion dataset when arranged as a time series is limited. We and we therefore 229 largely limitfocus our discussion to the properties of the population of waters as a full dataset. This 230 approach minimises the impact the different populations can have on interpretation. 231 precipitation values, and that surface re-evaporation was minor at least during humid phases. 246 However, the range of fluid inclusion values is inconsistent with either an exclusively Tunis-type or 247 an exclusively Bet Dagan-type moisture source for precipitation in Cyrenaica during MIS 3. Even 248 when all but the subset of fluid inclusion analyses who replicates are similar are excluded (Fig. 6), the 249 population is split between the Tunisian and Israeli precipitation end-members. 250

Strontium isotopes 252
The 87 Sr/ 86 Sr signal in the SC-06-01 record is rather invariable (Fig. 7), with all analyses indicating 253 values within analytical error. Mean values vary between 0.708275 and 0.708524 and although there 254 is an apparent trend from maxima at 34 and 64 ka BP with a minimum at 52 ka BP, which mimics the 255 precession history, this is too weak to be significant relative to the error. 256

Calcite carbon isotopes 257
Both  13 Ccc and  18 Occ show similar trends throughout the record (Fig. 8)  The weighted mean value for Atlantic-sourced precipitation events in Sfax ( 18 Oppt = -6.7 ‰,  2 Hppt = 276 -37.7 ‰) is distant from any observed fluid inclusion value (Fig. 9). Likewise, compositions similar to 277 the high amount Atlantic-sourced rainfall events in Sfax ( 18 Oppt = -8 ‰,  2 Hppt = -46 ‰) are not 278 reflected in the fluid inclusion data in Figure 9 suggesting a relatively low admixture of water from 279 this source.9). A simple 3-end-member unmixing of fluid inclusion isotope values using the The simplest interpretation of the Susah Cave fluid inclusion data is therefore that they reflect a 286 dynamic balance of moisture sources contributing to rainfall in Cyrenaica which resembles modern 287 precipitation in Tunisia and Israel in roughly equal proportions. An alternative way to explain the 288 trend of some points towards enriched  18 O values on the GMWL would be the temperature-289 dependent fractionation that would be caused by a shift to summertime precipitation. We do not 290 favour this explanation, as it requires a more fundamental reorganisation of regional atmospheric 291 circulation than our suggestion that the winter storms observed today penetrated further east in the 292

past. 293
Within the data presented in Figure 9, the Phase II fluid inclusions are exceptional, because none 294 show compositions consistent with a Bet Dagan source. Indeed, all the measurements for this period 295 resemble GMWL compositions. This seems to reflect a fundamental difference between this period 296 and Phases I and III, where all precipitation is drawn from synoptic westerly storms in the winter. 297 Consequently, it would seem that during the Obliquity-forced period of humidity the Israeli-mode 298 precipitation did not occur in the manner that it did during both Precession-forced periods of 299 humidity.
Although the isotopic composition of Mediterranean water will have been more enriched during MIS 301 3 due to ice-volume effects and increased Mediterranean water residence time (Rohling and Bryden, 302 1994), the similar mean values of the SC-06-01 fluid inclusion waters compared to modern 303 precipitation indicates the meteoric waterline at this time was not displaced to more enriched isotope 304 values. This could reflect balancing of source water effects by changes in kinetic fractionation during 305 evaporation (Goldsmith et al., 2017), which is controlled by normalised relative humidity. This would 306 imply that the Mediterranean air masses were less saturated with moisture than today during MIS 3, 307 which is consistent with the high deuterium excess  2 Hexcess values found in some fluid inclusion 308 samples (Fig. 10), but is difficult to reconcile with the increased precipitation recorded in SC-06-01. 309 In addition, changes in cloud height and cloud formation processes could possibly alter the isotopic 310 fractionation in the atmosphere. Alternatively, the source water effect may be countered by increased 311 runoff from the margins of the Mediterranean supplying isotopically depleted water to evaporating 312 surface water. Isotopic "residuals" consistent with this argument are identified throughout MIS 3 in 313 the eastern Mediterranean marine core LC21 (Grant et al., 2016), and this is also consistent with 314 higher rainfall in Cyrenaica. We therefore favour the latter explanation. 315 We concludeAlthough we find that our results likely reflect patterns of atmospheric transport in MIS3 316 comparable to today, it is possible that some moisture was drawn from re-evaporation of monsoon 317 rain falling further south, with no modern analogue in the region (Aggarwal et al., 2016). This water 318 would likely be extremely isotopically light, reflecting both monsoon-type compositions and further 319 fractionation during secondary evaporation. Moreover, a shift to more southerly-sourced regions is 320 inconsistent with Sr-isotope data from Susah Cave. Sr-isotopes are known to be sensitive to changes 321 in transport of Saharan dust (Frumkin and Stein, 2004), but even considering the most slow-growing 322 and most rapidly-growing parts of SC-06-01, no significant difference in 87 Sr/ 86 Sr has been identified. 323 Although at times of extreme rainfall in the region, Saharan / Sahellian dust production is suppressed, 324 this is not true during MIS3 (Collins et al., 2013). It seems that despite changes in the intensity of 325 moisture transport during the period 65-30 ka BP, there is no large-scale change in atmospheric dust 326 transport direction. This further supports our conclusion from the fluid inclusions that the Eastern convection rather than transport of moisture from the east or south with an atmospheric circulation 329 pattern that prevails today. 330

Different sources at different times? 331
Phase II fluid inclusions are exceptional, because none show compositions consistent with a Bet 332 Dagan source. This is most clearly reflected in the  2 Hexcess values (Fig. 10), which show consistently 333 low values across Phase II comparing well to the Western water end-member (~10 ‰) and not the 334 Eastern water end-member (~30 ‰.). The lack of Eastern water during Phase II seems to reflect a 335 fundamental difference between this period and Phases I and III, as during this time all precipitation 336 was drawn from synoptic westerly storms in the winter. Consequently, it would seem that during the 337 Obliquity-forced period of humidity, the Israeli-mode precipitation did not occur in the manner that it 338 did during both Precession-forced periods of humidity. This difference in the origin of the moisture 339 feeding rainfall may explain the difference in average  18 Occ during these different phases (Hoffmann 340 et al., 2016), and why during some periods in Susah Cave show strong correlation with North Atlantic 341 temperature, whereas others do not (Hoffmann et al., 2016). 342 Palaeoclimatological significance 343 mostMost of the precipitation supplied to Cyrenaica during MIS 3 was sourced from within the 344 Mediterranean basin, which exhibited a similar meteoric water cycle to that observed today, albeit 345 with more freshwater influence. This is a critical observation, as internally-cycled water cannot alter 346 the basin-scale hydrological balance and therefore is a minor influence on deep convection in the 347 Mediterranean Sea . Thethe precipitation feeding runoff must be 348 externally sourced if it is to materially change Mediterranean functioning, as is observed during 349 sapropel events . As most of the precipitation identified in SC-06-01 is sourced 350 internally to the Mediterranean, only the small, Atlantic-sourced portion of this water can be assumed 351 to play a role inThe internally-cycled water we report from Susah Cave cannot alter the basin-scale 352 hydrological balance, and therefore is a minor influence on deep convection in the Mediterranean Sea 353 : put simply, this means evidence of increased rainfall in the coastal of the Mediterranean.does not provide evidence for decreased net evaporation in the marine system. 356 This observation is critical, as it decouples the processes of precipitation on the Mediterranean 357 margins with sapropel formation, and consequent changes in moemtummomentum transfer to the 358 North Atlantic (Rogerson et al., 2012). Consequently, we recommend that great care is taken to 359 determine whether past precipitation peaks reflect significantly enhanced external water advection 360 before any continental record can be used as a basis for inferring Mediterranean freshening. 361

Palaeoclimatological significance 362
The consistency of MIS 3 and modern precipitation isotope values permits comparison of fluid 363 inclusion values and precipitation magnitude records at Sfax and Bet Dagan. Most of the water 364 reaching Susah Cave seems to have been derived from large-magnitude rainfall sourced from the 365 Western or Eastern Mediterranean surface water. The primary difference between these end-members 366 is the level of Dexcess, with the Western water ~10 ‰ and Eastern water ~30 ‰. This difference allows 367 the influence of these two sources to be compared between the three major humid phases (Hoffmann 368 et al., 2016) recorded in SC-06-01 (Fig. 10). These phases reflect changes in the distribution of 369 insolation as a consequence of changes in orbital tilt, with Phase I (65 to 61 ka BP) and Phase III 370 (37.5 to 33.5 ka BP) associated with reflecting Northern Hemisphere heating during precession 371 minima and Phase II (52.5 to 50.5 ka BP) which has been associated with a change in obliquity. In all 372 cases, the peak in rainfall recorded by the speleothem leads the orbital peak by ~3 ka. Phases I and III 373 both show very elevated Dexcess, whereas no such values were found in Phase II. This provides further 374 support to our conclusion that the Eastern Mediterranean source contributed significant moisture to 375 Cyrenaica during precession-related humid events, but that it did not during the obliquity-related 376 humid event. This difference in the origin of the moisture feeding rainfall may explain the difference 377 in average  18 Occ during these different phases (Hoffmann et al., 2016). 378 The varying balance between Eastern and Western precipitation is diagnostic of changing basin-scale 379 atmospheric structure during the pastDespite the low level of Atlantic moisture contributing to rainfall 380 in Libya in MIS 3. Eastern-sourced rainfall may occasionally relate to wintertime storms, as today (Gat et al., 2003), but essentially reflects convective rainfall with relatively small advection distances. 382 The significant enhancement of the magnitude and regional significance of this convective rainfall 383 observed at Susah Cave must reflect greater atmospheric convergence due to northward displacement 384 of the annual average position of the ITCZ . Contrary to this, the Western-385 sourced moisture is transported ~1500 km eastwards to reach Cyrenaica, which must reflect the mid-386 latitude storm track (Brayshaw et al., 2009). Consequently, although it does not seem that Atlantic 387 moisture is important to the climatology of Cyrenaica, the momentum derived from Atlantic winter 388 storms predicted by regional climate modelling (Brayshaw et al., 2009)  Further constraint on large-scale atmospheric advection can be provided by Sr-isotopes, which are 404 known to be sensitive to changes in transport of Saharan dust (Frumkin and Stein, 2004). Even 405 considering the most slow-growing and most rapidly-growing parts of SC-06-01, no significant 406 difference in 87 Sr/ 86 Sr was identified. This is unexpected and significant, as climate-driven changes in 407 87 Sr/ 86 Sr have previously been reported from speleothems in the Mediterranean region (Frumkin and 408 insolation peaks reflecting Precession, coastal Libya experiences greater westerly advection of water 410 due to an increase in Atlantic heat and greater convective rainfall due to migration of the ITCZ 411 whereas 2) insolation peaks reflecting obliquity show increased Atlantic heat and westerlies, but no 412 comparable change in the ITCZ position. 413 It seems that despite changes in the intensity of moisture transport during the period 65-30 ka BP, 414 there is no large-scale change in atmospheric dust transport direction. This further supports our 415 conclusion from the fluid inclusions that the Eastern Mediterranean rainfall operating during 416 precession minima reflects enhanced internal convection rather than transport of moisture from the 417 east or south with an atmospheric circulation pattern that prevails today. 418

419
Aside from those data with high deuterium excess, which reflect influence from the Eastern 420 Mediterranean source, much of the variance in the fluid inclusion dataset is captured by a two end-421 https://nucleus.iaea.org/wiser/gnip.php). Furthermore, addition of heavy rain events derived from the 431 Eastern Mediterranean aliases the tendency towards depleted  18 Odripwater, as this water is also more 432 depleted than modern Western Mediterranean precipitation. In the Bet Dagan data, there is also a 433 tendency to lower  18 Oppt with higher precipitation amount, but the relationship between rainfall 434 amount and rainfall isotope composition is not identical to Tunis. Ultimately, it seems likely that rainfall amount changes at Susah Cave do cause depleted (enriched)  18 Occ values to be associated 436 with high (low) rainfall, but this is too complicated by independent changes increases (decreases) in 437 westerly moisture advection and increases (decreases) in convergence. Qualitatively, all these 438 parameters are expected symptoms of North African humid phases and so these trends remain a 439 valuable expression of climatic variability. Quantitatively, more information is required to translate 440 the trends into fully-functional palaeoclimatologies, and this analysis pivots on whether  18 Occ trends 441 reflect changes in water deficit / surplus in Cyrenaica. 442 Although it is likely the oxygen isotope fractionation during calcite precipitation occurred close to 443 isotope equilibrium (Hoffmann et al., 2016), there is a good degree of correspondence between 444 positive and negative phases in  18 Occ and  13 Ccc, indicating a shared control. Indeed,  13 Ccc has a 445 markedly higher amplitude variability than  18 Occ. More isotopically depleted carbon may represent 446 increased incorporation of respired soil carbon, increased dominance of C3 over C4 plants, and/or 447 decreased degassing of aquifer water . Today, the Susah Cave location on Jebel 448 Malh has very thin soil cover, colonised by shrubby maquis vegetation. Soil respiration and 449 colonisation by C3 plants is limited by the strong water deficit of the region, and aquifer water 450 outgassing is enhanced by long residence times due to low water infiltration. Increased water 451 availability will progressively deplete the  13 C of dripwater by all three mechanisms described above. 452 Consequently, all three of these processes promote correlation between  13 Ccc and precipitation 453 amount. Within the  18 Occ data series, peak growth rates occur both during relatively enriched and 454 relatively depleted isotope stages. This is not the case for  13 Ccc, which more consistently shows 455 depleted values during times of rapid growth (SC-06-01 growth phases shown in Fig. 11). We 456 therefore consider it likely that  13 Ccc indeed more accurately records rainfall amount than  18 Occ does. 457

458
A key feature of this combined dataset is the long-term sinusoidal trend in both the  18 Occ and  2 Hfi, 459 reflecting the differing rainfall regimes dominant between Humid Phases I and III compared to Phase 460 II. This is not developed in  13 Ccc, implying that the process forcing the long-term cycle in moisture 461 source is not impacting on carbon dynamics in the soil and epikarst. We therefore conclude that there 462 is a mixed amount and source control on  18 O and  2 H in the SC-01-06 record, whereas  13 C is 463 dominantly controlled by water availability. 464 The fluid inclusions from SC-06-01 show that rainfall compositions in the southeast Mediterranean 465 region during MIS 3 were comparable to modern rainfall compositions recorded in regional GNIP 466 datasets. However, the diversity of compositions is impossible to explain with a single rainfall source, 467 rather indicating that moisture derived from the Atlantic, the Western Mediterranean and the Eastern 468 Mediterranean basins have all contributed to MIS 3 precipitation in Libya. This requires both 469 enhanced westerly advection of moisture to this region, reflecting the Atlantic storm track, and 470 enhanced convective rainfall within the Eastern Mediterranean basin. There is some indication that 471 these two mechanisms differ in terms of their response to orbital forcing, with precession parameter 472 minima enhancing westerly advection and internal convection, whereas obliquity minima enhance 473 westerly advection without significantly altering internal convection. 474 Crucially, this picture is most consistent with atmospheric circulation over the Mediterranean 475 remaining essentially unchanged during precession cycles. This is consistent with regional climate 476 model experiments showing major enhancement of winter westerly storm activity, but it not 477 consistent with the extreme migration of the ITCZ, where the monsoon belt approaches the North 478 African coast. The strong implication is that a significant arid belt is retained between the 479 Mediterranean and the ITCZ, even when northernmost Africa is experiencing significantly enhanced 480

rainfall. 481
It is likely that rainfall amount played a role in controlling the isotopic composition of the calcite in 482 this speleothem ( 18 Occ). However, the more depleted values reflecting higher rainfall are also 483 consistent with different mixing between the end members identified by the fluid inclusion analysis. 484 The structure of the  13 Ccc record provides an independent means of assessing changes in water 485 surplus / deficit, as more depleted values will reflect lower aquifer residence times, enhanced soil 486 respiration and changes in vegetation structure, all of which are limited by water availability in this semi-arid environment. Combined analysis of the proxies provides a powerful new demonstration that 488 the northeast Libyan climate was more humid during millennial-scale warm periods in the North 489 Atlantic realm, but quantification will be dependent on generating unambiguous independent evidence 490 for water availability in the soil and epikarst.   I, II and II  911 respectively. 912 Figure 6) Double-replicated fluid inclusion measurements from SC-06-01, and regional precipitation 913 isotope trends. 914 High Precipitation measurements from an Atlantic moisture source (as described in Discussion) 924 respectively. "Mean Bet Dagan" is the mean of GNIP measurements from this location, and "High 925 Precip Bet Dagan" is the subset of high precipitation measurements as described in the Discussion. 926 Figure 10) Fluid inclusion deuterium excess ( 2 Hexcess-FI) relative to calcite  18 Occ. Note some fluid 927 inclusions (70 to 60 ka BP and 40 to 30 ka BP) show high Dexcess( 2 Hexcess -Fi indicative of an Eastern 928 Mediterranean source. Growth Phases I, II and III are shown as grey areas. 929