We thank the referee for their constructive comments on this article. The referee. was concerned about the quality of the figures. We produced the figures for this article in high resolution (300 dpi) but their quality appears to have been reduced upon upload. We have attached the high-resolution figures in this comment, though we will also re-work figures 3 and 5 to address legibility concerns (see below).

Overall comments:

• The authors describe changes between the two time periods as anomalies. In my understanding, anomalies are deviations from the long-term mean. Since we are comparing two time periods, it is not really clear what an anomaly is and which "long-term mean" they are referring to; in fact, I think it is just the absolute difference that is being referred to here. Please clarify this at the beginning of the methods section.

Within the paleoclimate community it is standard usage to employ the term anomaly when referring to the absolute difference between modern climatological averages of a variable and the climatological averages of a simulated episode of past climate for the same variable. In this study, for each grid cell and each climate variable, the long-term LGM climatology simulated by each climate model was subtracted from the long-term pre-industrial climatology simulated by the same climate model and then added to modern climate values in order to obtain a bias-corrected LGM climatology. What is referred to as “anomalies” in the resulting discussion is the difference between the simulated burnt area, fire size and fire intensity under the MOD climate/MOD CO₂ experiment and the other four experiments. Since we are representing the average spatial patterns of each fire properties under each experimental condition, we believe this term to be appropriate. However, we will add text to the methods section to clarify these definitions (and the difference between the climate and the fire anomalies) as well as referring the fire anomalies as BA, FS or FI anomaly in the results and discussion section:

Line 102: “Figure 2. Latitudinal distribution of the LGM-MOD climate anomalies”

Line 115: “difference between the PI and LGM values (LGM-PI climate anomalies) were calculated and added to the MOD climatology (LGM-MOD climate anomalies) (see Figure 1). We use the term climate anomalies to refer to the difference between the MOD climatology for each climate variable and the computed bias-adjusted LGM climatology for the same variable, consistent with the PMIP4 protocol (Kageyama et al., 2017). The use of anomalies is designed to minimise the impact of systematic model biases on the derived climate. This approach provided three LGM climate scenarios, resulting in nine experiments for BA, FS and FI respectively.”

And Line 160 (break and create a new paragraph): “The resulting BA, FS and FI anomalies refer to the difference between the MOD climate/MOD CO₂ experiment and the three other experiments since each experiment is considered to represent the long-term average spatial pattern for each fire property under the set experimental conditions. We used the sensitivity experiments to quantify the separate effects of CO₂ and climate on BA, FS and FI independently. We then used the realistic experiments to identify which predictors were driving the largest change between MOD and the three LGM scenarios by excluding one predictor at a time from the GLM models, re-running the LGM experiments and identifying which excluded variable caused the greatest change in the BA, FS and FI MOD-LGM anomalies in each grid-cell. Comparing these results to the BA, FS and FI MOD-LGM anomalies of the full GLM models allowed us to determine if the predictor was responsible for an increase or a decrease in BA, FS and FI.”

2) The figures are quite small and hard to read (especially figure 3). They should be reworked (e.g. using BA, FI, FS as columns instead of lines) and provided at a higher resolution.

Figures were produced at higher resolution for this article but were not posted (only the figures as part of the embedded word document were uploaded). We will upload the high-resolution figures in this comment. However, we also agree with the referee that the figures could be improved and will re-work them for the revised manuscript. We will use stronger colours in Figure 2 (and move key into bottom right-hand corner to give a bit more space), we will create columns instead of rows (as suggested) for Figure 3 and 6 as well as remove Antarctica.

3) The discussion section (323-337) lacks a literature-based discussion of how VPD, DD and DTR have been observed/modeled in other studies. For example, the authors could already refer to the reference to Diffenbaugh in lines 372-374: "This work also highlights the role of VPD in promoting fuel loads and limiting fire ignition and spread, a climatic variable that has been linked to wildfire occurrence (Diffenbaugh et al., 2021)."   

The calculation of all three of these variables are standard therefore we do not believe there is a need for a literature-based discussion of how VPD, DD and DTR have been observed/modeled in other studies. We believe the most crucial point is that the calculations are consistent with the methods used in Haas et al., 2022 since this was the modern data that the GLM models were built on. However, we agree with the referee that a discussion on how these variables have been shown to be important in previous observation-based studies would be beneficial. We suggest adding the following:

Line 362: “These results add to a growing body of literature highlighting the importance of considering not only changes in wildfire weather but also vegetation properties in projections of future wildfire regimes (e.g. Harrison et al., 2021; Kuhn-Régnier et al., 2021; Pausas & Keeley, 2021). The impact of rising CO₂ levels will most likely enhance vegetation growth and litter accumulation, which are important controls on fuel availability, continuity, and load. However, climate and specifically VPD may have opposing effects to that of rising CO₂ levels. Since VPD controls plant growth, increasing VPD can limit ecosystem productivity and tree growth, in turn reducing fuel loads (Williams et al. 2013). Nevertheless, VPD has also been shown to increase litter fall, thus increasing available dead fuel (Resco de Dios 2020, De Faria et al. 2017). As such, it is important to consider how temporal and spatial scales affect the response of vegetation to changing VPD (Grossiord et al., 2020). Although the trade-offs between future increases in CO₂ and reductions in productivity due to higher temperatures and atmospheric dryness are not fully understood, this work highlights the importance of considering both. These effects will most likely not be evenly distributed across the globe (Gonsamo et al., 2021; Piao et al., 2020; van der Sleen et al., 2015) and CO₂ effects may be more important in some regions than others. In fuel-limited ecosystems, CO₂ fertilization could increase fuel loads and fuel continuity, increasing overall burnt area but also the potential for larger and more intense wildfires. This is particularly worrying in regions with anticipated decreases in atmospheric moisture, especially since evidence suggests rising VPD may only counteract a small proportion of CO₂-induced plant growth (Y. Song et al., 2022). Increased woody thickening, for example in tropical South Asia (Kumar et al., 2021; Scheiter et al., 2020), may also alter fuel loads in regions that are likely to be vulnerable to ignition under a drier and warmer atmosphere (Clarke et al., 2022). Whilst climate variables such as DD and DTR have also shown to be strong controls of global wildfires regimes (e.g. Bistinas et al., 2014; Forkel et al., 2019; Kuhn-Régnier et al., 2021), this study highlights the importance of VPD relative to other climate variables in driving spatial patterns of BA, FS and FI. This is in line with previous studies that have highlighted the important role of VPD in promoting fuel loads and fire spread (e.g. Diffenbaugh et al., 2021; Grillakis et al., 2022; Duane et al., 2021; Balch et al., 2022).”

Balch, J.K., Abatzoglou, J.T., Joseph, M.B., Koontz, M.J., Mahood, A.L., McGlinchy, J., Cattau, M.E. and Williams, A.P., 2022. Warming weakens the night-time barrier to global fire. Nature, 602(7897), pp.442-448.

De Dios, V.R., Hedo, J., Camprubí, À.C., Thapa, P., Del Castillo, E.M., de Aragón, J.M., Bonet, J.A., Balaguer-Romano, R., Díaz-Sierra, R., Yebra, M. and Boer, M.M., 2021. Climate change induced declines in fuel moisture may turn currently fire-free Pyrenean mountain forests into fire-prone ecosystems. Science of The Total Environment, 797, p.149104.

De Faria, B.L., Brando, P.M., Macedo, M.N., Panday, P.K., Soares-Filho, B.S. and Coe, M.T., 2017. Current and future patterns of fire-induced forest degradation in Amazonia. Environmental Research Letters, 12(9), p.095005.

Duane, A., Castellnou, M. and Brotons, L., 2021. Towards a comprehensive look at global drivers of novel extreme wildfire events. Climatic Change, 165(3-4), p.43.

Grillakis, M., Voulgarakis, A., Rovithakis, A., Seiradakis, K.D., Koutroulis, A., Field, R.D., Kasoar, M., Papadopoulos, A. and Lazaridis, M., 2022. Climate drivers of global wildfire burned area. Environmental Research Letters, 17(4), p.045021.

Grossiord, C., Buckley, T.N., Cernusak, L.A., Novick, K.A., Poulter, B., Siegwolf, R.T., Sperry, J.S. and McDowell, N.G., 2020. Plant responses to rising vapor pressure deficit. New Phytologist, 226(6), pp.1550-1566.

Williams AP, Allen CD, Macalady AK, Griffin D, Woodhouse CA, Meko DM, Swetnam TW, Rauscher SA, Seager R, Grissino-Mayer HD et al. 2013. Temperature as a potent driver of regional forest drought stress and tree mortality. Nature Climate Change Change 3: 292–297.

We also suggest adding the following to the methods to make the calculations clearer:

Line 91: “The number of monthly dry days (DD) (days with ≤ 1mm of precipitation), monthly diurnal temperature range (DTR) (daily maximum temperature – daily minimum temperature) and monthly vapour pressure deficit (VPD), a function of specific humidity, temperature and pressure were all calculated following the methodology in Haas et al. (2022).”

 Minor corrections:

Line 26: missing e for emphasis

 This has been corrected.

Line 66: missing space after Haas et al.

This has been corrected.

Line 79: please rephrase, it sounds like Haas was an study providing observations

Line 79: “Haas et al (2002) developed empirical models of the global spatial patterns of burnt area (BA), fire size (FS) and fire intensity (FI) using generalised linear modelling (GLM) of modern observations. Here we use these models to simulate the global spatial patterns of burnt area (BA), fire size (FS) and fire intensity (FI) under four climate/CO2 scenarios (Figure 1).”

Line 141: "did not change dramatically". In the framework it is stated that it did not changed at all, please clarify

We have cut the word “dramatically”, there was no significant change.

Figure 3: too small, legends unreadable

(see above)

Line 300: rephrase " somewhat worse ".

Line 300: We have substituted “not as good”

Line 374 “Although the effect of human activity was not considered in this analysis, if reductions in burnt area do contribute to greater fuel loads, suppression policies may artificially increase fuel loads in the same way reduced burnt area increased fuel loads under LGM conditions, suggesting resulting wildfires may be larger and more intense. “This statement is highly speculative and oversimplifies human influence. It is also not clear how results from "past anomalies" can be extrapolated to "future anomalies". Please rephrase and elaborate on these two points or delete the statement.

We agree with the referee and have deleted this statement.

We thank the reviewer for their comments. Below we have addressed the concerns raised by the reviewer on the assumptions of this work regarding human activity, as well as the legibility of the figures.

"Lack of human activity data: the model does not consider the influence of humans in modulating wildfire regimes during the LGM. This is a major concern for areas that were cared for by First Nations during this period (e.g. Africa and Australia). The most appropriate approach in this case would be to focus your study on the regions that were NOT populated by humans at 21kya. This would completely remove the confounding effects deriving from land management and cultural burning adopted by First Nations peoples. I think this selection would also ‘clean up’ the images and make the delivery of key points more straightforward. The removal of populated regions will sound much better than the justification provided in LL144-147. Agriculture (whatever definition you use) is also not the only way people have managed landscapes with."

Although there are reconstructions of human population density for the Holocene (e.g. HYDE, KK10), these are poorly constrained before the past 2-3 thousand years, as is shown by the marked differences between existing global maps. Uncertainty is even greater regarding population densities at the LGM. Nonetheless, to confine our analyses to areas where human populations were unequivocally absent during the LGM would remove much of the world. Therefore, we think this suggestion by the reviewer is not a viable option.

In our modelling of modern wildfires, we used road density and cropland cover as measures of landscape fragmentation, and population density as a measure of potential human ignitions. Including these anthropogenic predictors in the GLM models was found to be essential to capture the global drivers of the observed spatial patterns of wildfires, as modern fire regimes are influenced by human activity at a global scale (e.g. Marlon et al., 2008; Bowman et al., 2020; Harrison et al., 2021). In the absence of agriculture and roads at the LGM, we naturally set these predictors to zero in our simulations of the LGM. We also set population density to zero because of the lack of quantitative information about population density globally at the LGM. This is why we re-ran all the experiments, including the modern, with the human predictors set to zero. This approach highlighted (as expected) the significant effect of anthropogenic predictors on modern burnt area and fire size. However, it did not affect the large-scale spatial trends that we comment on, nor the anomalies between the different experiments as reported here.

We recognize that pre-agricultural populations, for example in Australia, set fires both to facilitate hunting and to promote the local abundance of food plants (see e.g. Gott, 2005). However, work in Australia indicates that Aboriginal populations also managed these fires to ensure that they did not become large or catastrophic, by ensuring that the fires were set at appropriate times. These so-called "cool burning" practices had the added benefit of reducing wildfires (see e.g. Constantine IV et al., 2023). Be-derived erosion rates on the Southern Tablelands (Portenga et al., 2016) indicate that Aboriginal burning there did not become important until the late Holocene – a conclusion supported by other lines of evidence (see e.g. Black et al, 2007). Indeed, these authors stated: "This is not to say that Aboriginal Australians never used fire to alter landscapes prior to the late Holocene, but that Aboriginal burns prior to this time were too infrequent, localized, or low impact to have altered erosion rates". This, and similar evidence from other regions, supports our statement that the human impact on the landscape was slight and relatively localised. Additionally, previous studies have shown a weak influence of population and land-use change on driving global wildfire trends prior to the 18^th century (e.g Pechony and Shindell, 2010; Bowman et al., 2020). We will modify our justification in the Methods section to make the basis for excluding predictors associated with human activity clearer, as follows:

LL147: The original GLM models (Haas et al., 2022) included predictors associated with human activity, specifically road density, cropland cover and population density. However, there was no agriculture (or modern roads) at the LGM, and information about pre-agricultural population sizes is limited and highly uncertain (see e.g. Williams et al., 2013; Gautney & Holliday, 2015); the human impact on the natural landscape was slight and relatively localised (Black et al, 2007; Fuller et al., 2014; Portenga et al., 2016). Therefore, we excluded these anthropogenic predictors in all the experiments by setting them to zero. This ensured that differences between the experiments were driven solely by climate and CO₂.

We agree with the reviewer that it would be worthwhile to include a fuller discussion of the potential impact of hunter-gatherer populations on fire regimes at the LGM. One important issue is the degree to which the LGM climate was unsuitable for human populations. Both in Africa and Australia, recent work indicates that there was a considerable reduction in suitable habitat and that human populations were confined to suitable refugial areas and that population densities were low even within these areas (Williams et al., 2013; Gautney & Holliday, 2015; Blinkhorn et al., 2022). We suggest adding the following paragraph:

LL360: The effect of human activity was not considered in this analysis. Pre-agricultural hunter-gatherer populations used fire for land management, for example to facilitate hunting and to promote the local abundance of food plants (Bowman, 1998; Gott, 2005), although recent work indicates that the burning regimes, they practiced tended to reduce fire overall compared to the natural state (see e.g. Constantine IV et al., 2023). However, the areas suitable for hunter-gatherer populations was much reduced at the LGM by generally colder and drier climates. It has been estimated that less that 23% of Sahulland (the extended continent of Australia) and less than 58% of Africa was habitable at the LGM (Gautney and Holliday, 2015; note Blinkhorn et al., 2022 estimate the range of habitable area as between 27 and 66%) and that hunter-gatherer populations were confined to climatically suitable refugia (see e.g. Williams et al., 2013; Blinkhorn et al., 2022). Furthermore, although the estimates of population density are highly uncertain, the LGM population of Australia was less than 5% of the modern population and the reduction in Africa was even larger (Gautney and Holliday, 2015). Palaeoecological evidence from Australia suggests that the use of fire by pre-agricultural hunter-gatherers had a low impact on the environment before the late Holocene (e.g. Black et al., 2007; Fuller et al., 2014; Portenga et al., 2016). Thus, it is unlikely that human activities during the LGM would have substantially increased fire or offset the impact of the changes in climate and CO₂ on fire regimes.

"LL374-377: this statement is way too simplistic and unclear. If you decide to retain regions that were populated during the LGM, you should consider the human activity element more fully in your discussion. For example, you should provide proxy-based evidence of this claim."

Reviewer 1 pointed out that the comment about the impact of fire suppression on fuel loads was speculative, and we have now removed this statement. However, we have added a paragraph in the Discussion section on the potential role of pre-agricultural populations on fire regimes and specifically their likely impact during the LGM.

"Figures are too small to be read. The legends are especially not readable, and they contain crucial information to understand/assess this work."

As we have explained in response to Reviewer 1, the figures were produced at higher resolution, but the uploaded version was only at low resolution and therefore less readable. However, in response to Reviewer 1's helpful comments, we have now revised the figures to make them more readable and these have been uploaded along with our comments to that reviewer.

References (to be added to the revised manuscript):

Black, M. P., Mooney, S. D., & Haberle, S. G. (2007). The fire, human and climate nexus in the Sydney Basin, eastern Australia. The Holocene, 17(4), 469-480.

Blinkhorn, J., Timbrell, L., Grove, M., & Scerri, E. M. L. (2022). Evaluating refugia in recent human evolution in Africa. Philosophical Transactions of the Royal Society B, 377(1849), 20200485.

Bowman, D. M. J. S. (1998). The impact of Aboriginal landscape burning on the Australian biota. The New Phytologist, 140(3), 385–410.

Bowman, D. M. J. S., Kolden, C. A., Abatzoglou, J. T., Johnston, F. H., van der Werf, G. R., & Flannigan, M. (2020). Vegetation fires in the Anthropocene. Nature Reviews Earth & Environment, 1(10), 500–515.

Constantine IV, M., Williams, A. N., Francke, A., Cadd, H., Forbes, M., Cohen, T. J., Zhu, X., & Mooney, S. D. (2023). Exploration of the burning question: a long history of fire in eastern Australia with and without people. Fire, 6(4), 152.

Fuller, D. Q., Denham, T., Arroyo-Kalin, M., Lucas, L., Stevens, C. J., Qin, L., Allaby, R. G., & Purugganan, M. D. (2014). Convergent evolution and parallelism in plant domestication revealed by an expanding archaeological record. Proceedings of the National Academy of Sciences, 111(17), 6147–6152.

Gautney, J. R., & Holliday, T. W. (2015). New estimations of habitable land area and human population size at the Last Glacial Maximum. Journal of Archaeological Science, 58, 103–112.

Gott, B. (2005). Aboriginal fire management in south-eastern Australia: aims and frequency. Journal of Biogeography, 1203–1208.

Harrison, S. P., Prentice, I. C., Bloomfield, K. J., Dong, N., Forkel, M., Forrest, M., ... & Simpson, K. J. (2021). Understanding and modelling wildfire regimes: an ecological perspective. Environmental Research Letters, 16(12), 125008.

Marlon, J.R., Bartlein, P.J., Carcaillet, C., Gavin, D.G., Harrison, S.P., Higuera, P.E., Joos, F., Power, M.J. and Prentice, I.C., 2008. Climate and human influences on global biomass burning over the past two millennia. Nature Geoscience, 1(10), pp.697-702.

Pechony, O., & Shindell, D. T. (2010). Driving forces of global wildfires over the past millennium and the forthcoming century. Proceedings of the National Academy of Sciences, 107(45), 19167-19170.

Portenga, E. W., Rood, D. H., Bishop, P., & Bierman, P. R. (2016). A late Holocene onset of Aboriginal burning in southeastern Australia. Geology, 44(2), 131–134.

Williams, A. N., Ulm, S., Cook, A. R., Langley, M. C., & Collard, M. (2013). Human refugia in Australia during the Last Glacial Maximum and terminal Pleistocene: A geospatial analysis of the 25–12 ka Australian archaeological record. Journal of Archaeological Science, 40(12), 4612–4625.