| 16 Jun 2026
Modelling compounding global climate extremes following the Mazama eruption of Crater Lake 7600 years ago
Abstract. The Mount Mazama eruption (Crater Lake, USA) c. 7600 years ago counts among the largest eruptions of the Holocene, yet its impact on contemporaneous climate, environment, and humans remains incompletely understood. Here, we simulate the Mazama eruption using the Max Planck Institute Earth System Model with a volcanic stratospheric sulfur injection of 162 Tg S based on estimates from ice core records, to project potential impacts on global climate and society. The model simulations reveal severe and diverging surface climate anomalies in different regions of the world. We investigate specifically the regions of the Mediterranean and Near East, and southeast Asia distal from the eruption source, for which we project compounding extreme events. We argue that the compound occurrence of severe cooling and precipitation extremes likely had a significant impact on these main regions of human settlements and spread of agricultural practices, with crop failures due to drought, and potential flooding in areas experiencing extreme precipitation increase. Our study illustrates how very large volcanic eruptions can alter surface climate with varying and contrasting compound anomalies affecting much of the land surface. A volcanic event similar to the Mazama eruption may well occur in the next decades to centuries. Today, an eruption of this magnitude would pose substantial risks of multiple breadbasket failures, impacting food security globally, which in turn may lead to further societal upheaval. Studying very large past eruptions is imperative for better understanding the risks associated with low-likelihood, high-impact events – that global society is patently ill-prepared for.
Review of the manuscript
Modelling compounding global climate extremes following the
Mazama eruption of Crater Lake 7600 years ago
by Evelien van Dijk and co-authors
submitted for publication in Climate of the Past
General:
Authors investigate the historical eruption of the Mazamas volcano 7600 years ago on its compound climatic impacts in different regions of the world in conjunction with implications on food security and past societies. The manuscript is very well written, the motivation is clearly outlined and the analysis and presentation of results is concise. An asset of the manuscript is the link and discussion to potential implications of future volcanic eruptions on societies.
The authors used an ensemble approach that is very meaningful, but in the present form of the manuscript this approach is unfortunately not exploited and highlighted to a large degree. Also the wealth of atmospheric and oceanic output data based on the comprehensive simulations allow in-depth investigations on dynamical reasons, giving rise to the perturbed climate in the aftermath of the eruption.
I suggest publication of the manuscript after the comments listed below are addressed and additional analysis is presented in the context of dynamical atmospheric and oceanic processes in the context of the Mazama eruption.
Specific:
Abstract
The abstract motivates and summarizes very well results achieved in the manuscript. For the revised version authors should include additional statements on the importance of the ensemble approach, investigating regional changes and results based on the dynamical analysis of additional atmospheric and (upper-)oceanic variables.
2.2 Model experiments
l. 100: It looks like the authors use a so called physically perturbed parameter experiment to generate a set of ensembles. It would be helpful if this is explicitly stated, in contrast to the perturbed experiments when different states of the initial conditions are used to generate the ensemble.
Also a few words and background information on the physical meaning of the vertical diffusivity and parameter space used for the generation of the ensemble in the context of volcanic eruptions should be added.
2.4 Analysis
Are there any other volcanic eruptions included in the 374 year long period ? I wonder how sensitive results are when using a somewhat shorter, but still climatological meaningful period, i.e. 30 years prior to the period. It is also interesting whether changes in solar activity are used in the simulation branched off from the main simulation.
l. 136: How is the location of the ITCZ calculated ? Is it the maximum of zonal averaged precipitation ?
3.1 Global Climate Response
In general the sub chapter is quite short, although it would be interesting to explore in greater detail continental-to-hemispheric scale climatic changes. Therefore I suggest adding additional variables (i.e. SSTs (ENSO), Sea Ice, Wind, SLP and 500 hPa (NAO) atmospheric circulation) that could also provide further indications on respective changes in the aftermath of the volcanic eruption. The atmospheric variables could also help to explore and display changes in important modes of natural variability (North Atlantic Oscillation, Antarctic Oscillation, Indian and African Monsoon, Antarctic Oscillation).
Since the authors carried out a quite large number of ensemble simulations it is also imparative exploring the potential spread of the individual simulations for a selection of phenomena (i.e. response on evolution of ENSO and DJF NAO indices), depending on the individual setting of vertical diffusivity.
l. 130 ff: Why are changes in Arctic and Antarctica in 2m temperatures so small ?(Antarctica is even excluded in the „global“ map). Authors should state reasons why no or little (stat. significant) changes are evident (i.e. low signal-to-noise ratio because of high amount variability, and add reasons if patterns might change when accounting for longer periods after eruptions. There are studies pointing to the effect of ocean-atmosphere response modulated via the subpolar gyre, leading to sustained cooling in the North Atlantic that could also be cited in this context (cf. Moreno-Chamarro et al., 2017)
3.2 Impacts in the Mediterranean and the Near East
The impact on different regions is a very important issue, especially in the context of the climate-society nexus. I encourage authors to add a few words on potential limitations in using global climate models for sub-continental inferences. Although large-scale climate forcing such as shifts in ITCZ or NAO substantially influence regional and local climate, meso-scale circulation features (i.e. cut off lows, Genoa low) that are under-represented in global climate models and their impact on local temperatures, and even more important hydrological variables, should explicitly be mentioned.
l. 163: Here also the ensemble could be used for investigations on the potential spread of flood/droughts using the same magnitude in volcanic forcing.
l 163 ff: The hypotheses proposed here could be nicely tested with the simulations using daily output – in case its available. One could at least address this by looking into changes in high percentiles of (regional) precipitation – again with the notification that results based on spatially coarsely resolved model simulations are afflicted with a high amount of uncertainty on the regional and local scales.
3.3 Impacts in India and Southeast Asia
The whole section would benefit if additional plots on atmospheric circulation (i.e. 850 hPa winds and according plots with difference between volcanic minus reference period for the Indian summer monsoon are presented).
3.4 Compound extremes
l. 195 ff: The basic motivation of the intro paragraph and its links to present-day climate and societies reads perfectly fine. I suggest to include an additional paragraph on the potential and already existing differences between the background climate and the societal differences between the period of the Mazama eruption and present-day conditions. This will help the reader to better put the study in context and that results cannot per se translated into our present day world.
l. 210 ff: The statement is quite hypothetical, although large (tropical), repetitive and long-lasting volcanic eruptions may indeed have a had substantial impacts on past societies. Here authors should clarify how the NH eruption can be put into the concert of other larger tropical volcanic eruptions, specifically on the time span of climatic deterioration in the aftermath of the eruption.
I assume that very short-lived (NH) eruptions affecting two or three harvesting seasons might have been compensated by resilience measures of past societies. At least this point should somehow be addressed within this section. Again, the individual trajectories of each simulation might help to elucidate if especially at regional scales the impact might have been ameliorated through internal processes. This point could be addressed by presenting the evolution of individual simulations for selected regions and seasons (e.g. for temperature and precipitation), including the calculation of the respective 2sigma natural variability levels of the reference simulation.
4 Discussion
4.1 Comparison to other large explosive eruptions
l. 229: I suggest to put this very strong statement a bit into context already at this point, given that the eruption was only over the NH.
l. 245: Authors should address how different the trajectories of the individual simulations compare at selected regional scales, e.g. is the max T2m decrease always located over the same region ?
l. 277: As the authors state this argument is based on speculation rather than on empirical evidence and therefore this should somehow be mentioned more explicitly.
4.2 Uncertainties and limitations
Authors should include the perspective how the trajectories would be affected using point source sulfate emissions, ie. using the exact location of the volcano and not the zonally averaged AOD-time forcing. Are there estimations whether the implementation used might reflect an exaggeration of the real-world impacts of the aerosols because of the simplifications of the specific implementation into the MPI model ?
4.3 A Mazama-like eruption in present or future conditions
In addition to the implications mentioned, I suggest to include differences of paleo eruptions and climatic downturns in comparison to the present-day world, e.g. in terms of increased international trade and storage options counteracting adverse effects of the eruption, opposed to a potentially highly vulnerable infrastructure (complex and interconnected Systems in the IT, energy and traffic sector) that could have substantial impact on the world economy.
5 Summary and conclusion
For this chapter I suggest also to mention the regional varying trajectories that are plausible using the ensemble approach to illustrate the bandwidth of possible climatic evolution on the sub-continental to regional scale in the aftermath of (large) volcanic eruptions.
Tables and Figures:
Fig. 1: The NH and global lines should be referenced in brackets in the Figure caption and/or the label could be added into all diagrams.
The gray dashed lines for 2sigma significance are difficult to recognize besides the grid of the Figures. The question is also for which curve the lines are referring to, i.e. the global or the NH 2sigma levels ?
Fig. 2: Colorbars should be centered at zero with symmetric axes, Lat/Lon information at the axes of the spatial plots is missing
Fig. 3: Lat/Lon information at the axes of the spatial plots is missing. Again, color bars should be symmetric around zero, using same absolute max/min values.
Fig. 4: c.f. comments related to Fig. 3
Additional Reference:
Moreno-Chamarro, E., Zanchettin, D., Lohmann, K. et al. An abrupt weakening of the subpolar gyre as trigger of Little Ice Age-type episodes. Clim Dyn 48, 727–744 (2017). https://doi.org/10.1007/s00382-016-3106-7