By Matt Boyd & Nick Wilson

Update (6 Dec 2022): the research described in this post is now published as a peer-reviewed paper in the journal Risk Analysis. You can read the published version here. Minor revisions through the peer review process mean that some quantitative results were updated and the updated results are contained in the peer-reviewed version. Please cite these results rather than those listed below. However, the key messages and conclusions in the post below remain unchanged.

Purpose: In this long-read post we provide a summary and extended commentary on our recent paper: Island refuges for surviving nuclear winter and other abrupt sun-reducing catastrophes (available as a pre-print via the link). We demonstrate how this research could help direct efforts to reduce the risk from catastrophic sun-blocking scenarios and suggest some next steps to safeguard humanity from these types of existential threat. 

TLDR:

• Abrupt sunlight reducing scenarios (ASRS) such as nuclear winter or volcanic super-eruption are plausible and could have serious consequences for climate and food production.
• It is often thought that some Southern Hemisphere islands might resist the more severe impacts of these winters.
• We find that some locations could likely produce enough food in a nuclear winter to keep feeding their populations, but food supply alone does not guarantee flourishing of technological society if trade is seriously disrupted.
• The potential disruption to industry and society caused by serious trade collapse could be severe and cause deindustrialization of societies.
• It is problematic to assume that locations such as New Zealand and Australia might survive catastrophes such as nuclear winter with their institutions and technology intact – without major upgrades in their levels of resilience.
• This has implications for the future of humanity and human civilisation, given the existing assumptions about Southern Hemisphere islands.
• Many nations might be advised to pursue resilient foods to mitigate ASRS, but New Zealand and Australia might focus on resilience measures for preserving transport, energy, manufacturing, and industrial inputs in the absence of global trade.

Background

Scenario

Sudden reduction in sunlight reaching the earth could have catastrophic impacts. Abrupt sun-reducing scenarios (ASRS) plausibly include nuclear winter, massive volcanic eruption, and asteroid/comet impact. In these scenarios, soot, sulphur dioxide or dust could enter the stratosphere, spread globally, and cause global climate impacts such as a drop in mean temperature and sunlight levels that severely limits food production.

Probability

There are many estimates of the probability of nuclear war, some published on this Forum such as here, and here. Most estimates of total risk (eg, not only US-Russia war) are approximately 1% per annum range (or this order of magnitude, being greater than 0.1% per annum but usually less than 10%). Nuclear war alone would not guarantee nuclear winter, but it is a possibility and vast amounts of soot could be ejected into the stratosphere by a nuclear war, as discussed in this forum here.

We are not going to unpack these likelihoods in this post, however, considering the motivations of the relevant actors and the means at their disposal, nuclear war is plausible. 

The probabilities of major volcanic eruptions were detailed in another Forum post here. Sufficiently concerning eruptions probably have a return period of about 680–2,100 years. Large asteroid/comet impacts are very much less frequent. 

Consequences

As well as uneven but potentially severe reduction in global temperature and sunlight and associated reduction in crop yields, ASRS could impact precipitation, and reduce atmospheric ozone, phytoplankton, and marine food. There would be knock on effects that disrupt human systems such as energy, trade, social cohesion, communications, health, and psychology. These impacts could be magnified if strategic global pinch points are affected. Many of these consequences are yet to be fully understood but we discuss them in more detail in our paper. 

The scale of any climate impact depends on the volume of material that is injected into the stratosphere. Some existing modelling estimates in the literature put this at 5 teragrams (Tg) for a ‘regional’ nuclear war (such as between India and Pakistan), and 150 Tg for an ‘all out NATO-Russia war’. Dissenting from the extreme scenario, a post in this Forum has argued for a middle ground estimate in a US-Russia nuclear war of 30 Tg (90%CI: 14–66), see here.

Much concern about the climate impacts of ASRS (eg, ‘nuclear winter’) has focused on the likely reduction in food supply. The impact of nuclear winter on food production globally has been modelled by Xia et al 2021. Results of the modelling indicate severe global food shortage (~80% reduction under the 150 Tg scenario) with near total food production loss in large continental landmasses (eg China, Russia, Europe, North America). It must be noted that the possibility of nuclear winter, and the resulting impact on food production, remains controversial (see for example Hess 2021 who discusses this). Therefore, the results of Xia et al’s modelling displayed in the following table should be viewed in the likely “worst case” range:

Table 1: Reduction in calories (%) after simulated nuclear conflict (see Xia et al 2021 for details)

Climate impacts would not be homogenous and although food supply might drop precipitously in many Northern Hemisphere regions, crop yields may be impacted less in tropical regions. Islands in the Southern Hemisphere might do better than Northern Hemisphere landmasses because of the thermal moderating influence of the ocean. Integrated atmosphere, ocean and crop modelling supports this claim. However, this is in a context where only 66 of 164 countries in one study were food self-sufficient (and this number declined in recent decades). 

There are of course many effects of ASRS beyond the impact on food production. If the initial event causes widespread destruction (such as nuclear war) then as well as the deaths of perhaps 30–75 million people, we might expect devastation to transport (ports, airports, railways), energy production, manufacturing capacity, and communications. Supply chains and trade could be cut with disastrous feedback effects. 

Despite all these impacts, extinction risk and existential catastrophe by collapse of civilization might seem unlikely to result from nuclear war. This is in part because of the heterogenous climate impacts and relatively better position in which several island nations might find themselves. Toby Ord singles out Australia and New Zealand in The Precipice. These larger islands might sustain ‘nodes of persisting complexity’ thereby ensuring successful continuation of technological civilization in the face of adverse circumstances. 

Aim

It was our hypothesis that island nations, particularly in the Southern Hemisphere, would typically suffer less from ASRS. Complex technological society on such islands might persist, and targeted preparation could increase the probability of a global recovery. We tested this hypothesis with: 

• a food threshold analysis
• profiling islands according to resilience factors
• a more detailed case study (New Zealand)

We aimed to identify the island nations whose societies are most likely to survive a nuclear winter or other ASRS while preserving complex societal and industrial functioning.

What we did

Full methodological details can be found in our pre-print. But basically, we did the following three things: 

1. Food threshold analysis

We obtained estimates of the dietary energy production (DEP; kcal per person per day) for most countries from Schramski’s 2019 study. We then applied the percentage reduction on agricultural yield under 5 Tg, 27 Tg, and 150 Tg nuclear winter scenarios as estimated by Xia et al. We eliminated islands that had resulting DEP <2,200 kcal per capita per day. We used this threshold because it was the same as that used by Xia et al for approximating survival without weight loss. 

2. Profiling islands’ resilience factors

From the literature on island refuges, nodes of persisting complexity, and the impacts of nuclear war and nuclear winter, we identified 13 factors that appeared relevant to an island society surviving with institutions and technology relatively intact. We then identified indicators that we could use to measure these ‘resilience factors’ (see Table 2). 

3. Island case study

We undertook a more in-depth case study for the island nation of New Zealand. In the case study we identified themes covered in the 1987 New Zealand Nuclear Impacts Study, and updated this with new information about New Zealand in 2022, including reference to the resilience factors identified above.

Table 2: Factors plausibly increasing societal resilience to nuclear winter-type scenarios

What we found

We’ll now summarise some highlights of our results, and then follow this with a discussion of what it might mean for human resilience. 

Food

We found that applying the modelled dietary calorie reduction (%) for a 150 Tg nuclear winter to the baseline DEP estimates, only five island nations (of 38 in Xia’s dataset) continued to produce a minimum of 2,200 kcal/capita/day. These were Australia, New Zealand, Iceland, the Solomon Islands, and Vanuatu. In the 5 Tg scenario there were only three more: Mauritius, Indonesia, and the Philippines (given the diverse nature of the archipelagos of Indonesia and the Philippines, it could well be the case that some individual islands are more productive than others, and may reach the threshold of 2,200 kcal/capita/day, but we did not analyse sub-national jurisdictions). 

We also found that Australia produced enough food, even in modelled 150 Tg nuclear war scenario to feed 85 million people, in addition to its own population. New Zealand could possibly feed another 4 million. 

Table 3: Dietary energy production under various modelled nuclear winter scenarios

*below 2,200 kcal/capita/day weight loss threshold

In a separate analysis of major export food data for New Zealand, we found that food normally exported totals 8,150 kcal/capita/day, which might reduce to 3,423 kcal/capita/day under severe nuclear winter conditions (New Zealand’s production is modelled to reduce by about 58% in the 150 Tg scenario). Currently exported milk powder alone could provide approximately 90% of caloric needs in nuclear winter conditions (with all dairy products 142%). So it seems like at least some of these island nations should be able to produce enough food for their populations to survive severe ASRS conditions without such problems as internal conflict caused by food shortages (albeit assuming appropriate fuel supplies to the agricultural sector continue and food is equitably distributed). 

Table 4: Daily dietary energy provided by major food exports (NZ), % of current population intake. 

Island resilience profiles

The following table details the 13 indicators of resilience allowing comparison within the islands in the ‘food secure’ set. Comparing each factor across these islands and attributing a ranking it is possible to calculate a rough mean ranking on all 13 indicators. The table lists the top six islands on this rough metric. What we can say is that across these 13 factors and compared to the other islands that also have food >2,200 kcal/capita/day in modelled nuclear winter conditions, Australia appears to have the most resilient profile, followed by New Zealand (we’ve left the Solomon Islands and Vanuatu off the table for brevity). 

Table 5: Islands and their comparative resilience

Australia’s food supply buffer is gigantic (9,483 kcal/capita/day in 150 Tg scenario), with potential to feed many tens of millions of extra people. Good quality infrastructure, vast energy surplus, the second highest health security in the world, and triple the defence spending of any other island in our analysis, all suggest Australia has the potential to survive and even thrive (comparatively) during an ASRS. 

However:

• Australia may be a military target in a nuclear war due to its AUKUS alliance, and US-aligned intelligence stations (eg, via an electromagnetic pulse to damage the Pine Gap facility).
• Australia may be an accessible destination for refugees (eg via South East Asia) or invaders and would need to plan for this.
• Like all the islands, Australia is at risk of progressive societal degradation if cut off from international trade.

The issue of trade is particularly important. The interdependencies of the global industrial system are immense and convoluted. Any loss of global trade triggered by a nuclear war, or severe climate impacts following an ASRS (necessitating hoarding of food and fuels, or triggering further conflict) could be catastrophic for industry. There is a very accessible 2022 book by Peter Zeihan laying out these cascading interdependencies as they pertain to transport, energy, finance, industrial inputs, manufacturing, and agriculture. In sum, the world is brittle to trade disruption. 

Case study: New Zealand

In our paper we argue that just because a country has plenty of food production, and a favourable resilience profile relative to other locations, this does not mean it is yet resilient to severe ASRS. 

The vulnerability of New Zealand to loss of international trade was noted as far back as 1982 in the Future Contingencies: Nuclear Disaster report. The severity of this risk was further highlighted in the 1987 New Zealand Nuclear Impacts Study. At the time, New Zealand appeared severely at risk from nuclear war due to its dependence on trade, energy imports and the intricate interdependence of societal systems. These concerns remain relevant today – possibly even more so in a more interconnected world.

Unlike countries in Europe or North America that might suffer greater than 90% reduction in crop yield, food does not appear to be the main issue for New Zealand. However, the following factors are of major concern in ASRS (and we present more details in our paper):

• New Zealand is the most isolated temperate land mass in the world, it is therefore highly susceptible to downturn or cessation of long-haul shipping
• No shipping would mean many industrial inputs are unavailable, including commodities such as valves, lubricating oils, tyre rubber, many chemicals and their precursors (eg, fertilizer), not to mention electronics, semiconductors, and critically, diesel and other refined fuels for which NZ is 100% dependent on trade
• Compounding this issue New Zealand has no capacity to refine oil, given the recent closure and decommissioning of its only refinery (and only a tiny biofuel production capacity, which is similarly being wound back)
• New Zealand has a sparse distributed population and transportation of food is almost exclusively by road.
• Only 1.2% of vehicles in New Zealand are electric
• Diesel is also needed for inter-island and coastal shipping
• New Zealand lacks a sophisticated manufacturing sector, so many replacement components would not be available to service breakdowns
• New Zealand lacks local expertise for many industrial systems and processes (eg, international experts are often flown in for repair of essential machinery/infrastructure). Furthermore, much legacy infrastructure needs massive upgrades (eg leaking city water pipe systems).
• Significant operational data is stored offshore in the cloud (eg, some major banks)

These shortcomings, and many others, mean New Zealand could experience catastrophic limitations in transportation, infrastructure, or machinery for growing, harvesting, processing, packaging, transport, and refrigeration of food. Certain failures or unavailable inputs could degrade communications or electricity infrastructure. Such failures could have devastating flow on impacts across almost all sectors impeding communication, coordination and distribution. These difficulties would be amplified in a context of trying to pivot export food to the domestic market. There could be rationing (of food and fuels), hyperinflation, or price collapse. People’s fear of scarcity, psychological trauma, or concern for loved ones, could see an exodus from urban areas, resulting in absenteeism, and further degradation of industrial and societal functioning. A suite of cascading scenarios can be imagined akin to the cascading deindustrialization processes described by Zeihan and others. The 1980s Nuclear Impacts Study even posed the question of whether government control would be possible, or appropriate, under such circumstances. 

The problem is that modern high-income societies are an infrastructure-of-infrastructures. Namely complex adaptive ecological systems embedded within complex adaptive ecological systems. Yet ecological-social resilience theory tells us that the effects of perturbations in such systems are unpredictable and key ‘slow variables’ matter. 

Our case study of New Zealand indicates that slow variables such as diminishing supply of refined fuel, the failure of irreplaceable components, or unravelling social cohesion, may be important drivers of state transitions in the system. In the case of ASRS the new equilibrium may be a non-digital, non-industrial, 18^th century-level-agricultural state, making local recovery much more difficult. 

If New Zealand is supposed to be one of the best placed countries in the world in an ASRS situation, then this has implications for global recovery as well. 

Although superficially, when inspecting macro-indicators (the 13 factors in table 2 above), New Zealand appears well-placed, most systems may not be resilient to major shocks. 

Some earlier work (including our own) identifying New Zealand (and other islands) as favourable ‘refuges’ against extreme pandemics, or as resilient ‘nodes of persisting complexity’ may not have sufficiently emphasised the importance of trade.  

Implications for avoiding global catastrophic and existential risk

In many regions, locally irreparable failures could lead to ungraceful degradation of essential systems, and beyond certain thresholds there could be catastrophic cascading effects in a cycle of decline to around pre-industrial levels if global trade ended. 

Contrary to statements such as, ‘[i]t is hard to see why [New Zealand] wouldn’t make it through with most of their technology (and institutions) intact’ (Ord, 2020), there are a number of paths to societal collapse in ASRS. As identified in the 1980s Nuclear Impacts Study, ‘[f]undamental disruptions to New Zealand society would occur in the absence of direct targets or climatic change’ (Green, 1989).

It had previously been thought that nuclear winter could risk human extinction. However, more recent research on nuclear winter focussing on climate impacts and food production indicates that Southern Hemisphere island nations such as Australia or New Zealand (or the Solomon Islands or Vanuatu) might be relatively spared from starvation and food conflict due to abundance and remoteness. 

Human extinction due to ASRS seems unlikely via direct impacts, given the islands just identified and their agricultural excess. However, existential catastrophe (and limitation of human potential) seems possible. There is a difference between surviving and thriving, and there is a further difference between thriving locally and restoring or ‘rebooting’ a path to a refreshed global industrial and technological civilisation. Places like New Zealand and Australia might be well advised to increase their resilience to avoid the critical transition from industrial to pre-industrial states. 

Resilience Measures

We have identified a range of factors that could be addressed to improve New Zealand’s situation (and many of these might generalise to other island or even non-island locations). 

Many of these suggestions might apply to any period of isolation experienced by New Zealand. Possible scenarios include: extreme pandemics, conventional world war, nuclear war, nuclear winter, large volcanic eruption, asteroid/comet impact, industry disabling solar flare, cascading impacts of extreme climate change, or various other possibilities. We list the main themes below, but a substantial list of specific suggestions can be found in the Supplement to our paper. 

Table 6: Brief version of the list of possible resilience interventions

Importantly, the ideas and examples in the above table are not yet recommendations, and they are not prioritised. Furthermore, the New Zealand Nuclear Impacts Study is substantially out of date. New Zealand (and elsewhere) now needs a detailed research project that describes the key scenario(s), and then gathers information from across all sectors (public services, transport, energy, manufacturing, industry, health, etc) as to what impact ASRS scenarios with trade disruption would have on each sector, and how these impacts would in turn impact other sectors. 

It is this project that we are now undertaking in the New Zealand context with assistance from the Regranting Programme of the FTX Foundation Future Fund. The aim is to eventually produce outputs that to some degree generalise across other islands. 

Conclusions

ASRS are not improbable (nuclear war is plausible, and supervolcanic eruptions are probably inevitable). But much of the thinking outlined above applies to any global catastrophe that causes isolation of industrial populations (conventional war, solar flare, extreme cascading climate impacts). 

However, these scenarios do not appear to be addressed systematically in any national risk assessment of which we are aware (we’ve written about this issue, see our preprint here on national risk assessment and global catastrophic risks). 

New Zealand’s National Risk Register is classified, thereby impeding any communication of a risk and capabilities assessment to the public and industry. Additionally, a summary of public consultation on the topic for New Zealand’s newly mandated ‘National Security Long-term Insights Briefing’ did not mention the words ‘nuclear’ or ‘volcano’ (although ‘food security’ was mentioned in the context of climate change). These deficits need addressing.  

It is unclear what planning exists to manage situations of extreme trade isolation, or to respond to, or ideally prevent, cascading societal failures. In our previous work on island refuges for extreme pandemics we noted that some islands appear better placed than others to survive voluntary isolation, although we were also clear that this observation does not yet mean that any of these ‘best’ islands are yet optimised or capable of doing so. 

Such optimising work should now begin (ideally in several locations around the world) with a view to mitigating a range of global catastrophic scenarios. Development of possible resilience plans could then be followed by implementation prioritisation and cost-effectiveness analyses. 

Next Steps

This New Zealand Catastrophe Resilience Project is the next phase of our research and will deploy the following initial project components:

1. Use recommended national risk assessment methodology to produce relevant risk profiles of national importance
2. Conduct workshops to develop scenarios and identify information gaps so as to elicit the relevant information via industry/sector/public service surveys and interview methodology
3. Survey sector experts in a knowledge brokering process to collate relevant vulnerabilities and interdependencies
4. Interview targeted key informants to develop deeper understanding of key issues
5. Conduct a roundtable/workshops with the survey/interview results to identify priority issues (eg a ‘big 10’ by Delphi processes)
6. Develop a bullet point ‘proto-plan’ that can serve as the basis for a fully worked up National (ASRS/Other) Catastrophe Resilience Plan
7. Engage with government risk coordinators throughout the process

Related links

Our papers on island refuges (in the context of extreme pandemics):

• Optimizing Island Refuges against global Catastrophic and Existential Biological Threats: Priorities and Preparations
• The Prioritization of Island Nations as Refuges from Extreme Pandemics

Blogs we wrote part way through this project:

• Sustained Resilience: the impact of nuclear war on New Zealand and how to mitigate catastrophe
• Putin and the Bomb: Why New Zealand national risk assessments should include planning for the potential impacts of nuclear winter

Popular Media: Radio interview related to the project:

• Do we really need to prepare for nuclear war?

Acknowledgments

This work was kindly supported by a grant from the Centre for Effective Altruism Long-term Future Fund. 

We thank Wren Green, David Denkenberger, Morgan Rivers, Marnie Prickett, and Bill Brander for helpful input at various stages. 

Our ongoing work in this topic area is supported by the Regranting Program of the FTX Future Fund.