COVID-19 Reveals Vulnerabilities of the Food-Energy-Water Nexus to Viral Pandemics.

Food, energy, and water (FEW) sectors are inextricably linked, making one sector vulnerable to disruptions in another. Interactions between FEW systems, viral pandemics, and human health have not been widely studied. We mined scientific and news/media articles for causal relations among FEW and COVID-19 variables and qualitatively characterized system dynamics. Food systems promoted the emergence and spread of COVID-19, leading to illness and death. Major supply-side breakdowns were avoided (likely due to low morbidity/mortality among working-age people). However, COVID-19 and physical distancing disrupted labor and capital inputs and stressed supply chains, while creating economic insecurity among the already vulnerable poor. This led to demand-side FEW insecurities, in turn increasing susceptibility to COVID-19 among people with many comorbidities. COVID-19 revealed trade-offs such as allocation of water to hygiene versus to food production and disease burden avoided by physical distancing versus disease burden from increased FEW insecurities. News/media articles suggest great public interest in FEW insecurities triggered by COVID-19 interventions among individuals with low COVID-19 case-fatality rates. There is virtually no quantitative analysis of any of these trade-offs or feedbacks. Enhanced quantitative FEW and health models are urgently needed as future pandemics are likely and may have greater morbidity and mortality than COVID-19.

Introduction

Modern viral pandemics such as the 2019 coronavirus disease (COVID-19, caused by the virus SARS-CoV-2) are predominantly caused by food systems that expose increasingly interconnected populations to reservoirs of viruses against which humans have little resistance.[1] As of July 2021, over 190 million cases of COVID-19 have been confirmed globally, resulting in over 4 million deaths.[2] Attempts to arrest the spread of the virus have meanwhile resulted in widespread economic and social disruption. For example, in the United States, gross domestic product (GDP) fell by an annualized rate of 32% between the first and second quarters of 2020 as physical distancing and travel restrictions constrained economic activity.[3,4] COVID-19 demonstrates that modern food systems may disrupt global society by introducing novel pathogens, but interactions with tightly coupled water and energy systems have been underexplored. Characterizing how COVID-19 has influenced food–energy–water (FEW) systems is a first step toward building resilience to avoid catastrophic breakdowns in future pandemics.

Modern food systems promote pandemic zoonoses. Worldwide, animal supply chains create novel contacts between animal species and between animals and humans, promoting viral recombination, bacterial adaptation, and interspecies infection.[5−8] Low-latitude, less developed countries are biodiverse (with respect to fauna and pathogens) and are subject to social and ecological pressures that increase harvest and consumption of bushmeat, increasing risks of pathogen emergence.[1,9−11] Virtually all major infectious diseases throughout history such as influenza, measles, and smallpox have been zoonotic and emerged alongside animal husbandry.[12−14] The continual emergence of new diseases such as HIV/AIDS (1980 to present, >33 million deaths, likely chimpanzee origin),[15,16] Ebola (1976 to present, >13 000 deaths over multiple outbreaks, likely bat origin),[17−20] and SARS-CoV-2/COVID-19 (likely bat origin and pangolin intermediary)[21] illustrate that emergent zoonotic disease continues to pose novel human health risks.

Current methods to forecast impacts of pandemics on social systems and to understand how policy responses mediate health and wellbeing are inadequate. State-of-the-science pandemic management and forecasting tools consider health as a single causal chain between infection, disease, and recovery or death.[22] Yet policy responses instituted to slow disease spread have diverse consequences with impacts and trade-offs not anticipated at the start of the pandemic. For example, shocks to economic and social networks associated with responses to COVID-19 have caused a serious countervailing mental health risk. In a recent sample of the U.S. population, Czeisler et al.[23] found that symptoms of anxiety and depression in June 2020 were 3 and 4 times more prevalent, respectively, compared to June 2019, with >25% of respondents aged 18–24 reporting suicidal ideation.

Meanwhile, FEW research is oriented around unsustainable supply-side demand for overexploited resources (especially water) and its impacts on food, energy, and water availability. It is now widely understood that FEW sectors are inextricably linked given the strong dependence of each sector on the other (e.g., agriculture accounts for 71% of global water withdrawals) and that unsustainable practices in one sector can cause volatility and crises in another (e.g., water shortage-induced food price volatility in 2008–2010).[24,25] COVID-19 however reveals that FEW interconnections are vulnerable not just to resource constraints but also to disruptions in human and financial capital on the supply and demand sides.[26] For instance, economic disruptions in the U.S. led to electric and gas arrearages totaling roughly $32 billion by the end of 2020.[27] Existing FEW modeling tools do not consider the human labor and financial capital required to sustain outputs and are therefore poorly suited to evaluating impacts of pandemics on FEW security.

While both pandemic zoonoses[13,28] and FEW systems[29,30] are major determinants of health globally, their combined effect on health is poorly understood. Pandemic research centers on biophysical processes, some of which have negligible if any impact on health. For example, several recent studies characterized wastewater as a vector of SARS-CoV-2 despite this pathway posing at most a negligible risk.[31−36] Meanwhile, larger scale processes such as strains on infrastructure and the water insecurity caused by COVID-19-related economic precarity have been discussed often in the media but rarely in the scientific literature. Other topics of ongoing research on sector-specific impacts of COVID-19 include impacts on nutritional adequacy[37,38] and the disproportionate risk of infection faced by workers in the water[39] and food[40,41] sectors. While some studies recognize intrasector system dynamics (e.g., risk of an outbreak leading to labor and thus food system disruption in the United States[42]), there has been a lack of analysis characterizing cross-sector trade-offs, feedbacks, and other system dynamics. This lack of integrated conceptual understanding hinders prediction and management of trade-offs and unintended consequences.

In this paper, we characterize the interactions between FEW processes, COVID-19, and pre-vaccine mitigation measures such as physical distancing and hygiene protocols. We also identify processes that largely withstood the stresses of COVID-19 but may break down in the context of a pandemic disease with higher mortality and morbidity among the working-age population. We combine a review of the scientific literature with a review of news/media articles to characterize causal pathways and identify potential research gaps. Where possible, we identify processes that are unique to less developed countries and processes that interact with pre-existing social inequities in the United States. We conclude with a summary of urgent needs for FEW-health modeling to improve decision-making for future viral pandemics.

Methods

We conducted a scoping review of the peer-reviewed literature for articles addressing food–energy–water phenomena in the context of COVID-19. A scoping review aims to quantify the nature and extent of available research without critical appraisal of research quality.[43] Article abstracts of journals indexed in the Web of Science (WoS) Core Collection[44] were queried on October 15, 2020 using a syntax that allows for diverse words to qualify as food, energy, or water concepts. This syntax is summarized in Supporting Information (SI) Table S1. The search was confined to publications in 2019 and 2020. We consider it very likely that virtually all articles that focused on food/energy/water interactions in any substantive way in the context of COVID-19 would have used one of the relevant terms in the abstract.

The search returned 179 articles, which were screened manually for relevance. A total of 98 articles were excluded without further review based on the conditions summarized in SI Table S2. The remaining 81 articles were reviewed to extract causal associations asserted or evaluated by authors. SI Table S3 lists all articles returned by the WoS search and excluded (sorted by exclusion reason). SI Table S4 summarizes all articles retained from the WoS search.

Article abstracts were reviewed for assertions about causal relationships. These relationships were recorded by authors (M.J., R.C., and R.H.) and research assistants (G.B. and M.M.). Author R.C. reviewed all coding and consolidated similar concepts using consistent terminology. Causal associations were combined using general terminology to avoid exponential proliferation of causal relationships appearing in a graphical model. Author R.C. synthesized all articles using R.[45] This method was used successfully on a larger scale by Calder et al.[46]

We note that the scientific literature is rapidly evolving, and so our mapping of literature coverage and gaps is subject to ongoing reevaluation. We also note that we have focused on cross-sector FEW nexus relationships rather than intrasector phenomena, where systems thinking is also needed.[47]

We undertook a complementary review of news/media articles to evaluate which food–energy–water concepts are being addressed in connection with COVID-19 outside the peer-reviewed literature. This was done to identify possible gaps between issues of popular interest and the concepts addressed by existing research. This is particularly important due to the rapid nature of evolving COVID-19 and the less rapid pace of academic peer review and publishing.

Using the Nexis Uni database,[48] we extracted news/media articles including in their titles at least two of the three nexus concepts: food + energy (FE), food + water (FW), or energy + water (EW). Article types include press releases and newspaper articles. The distribution of article types is presented in SI Figure S1. The words retained to describe each of the food, energy, and water concepts are the same as those summarized in the WoS search (SI Table S1, rows 3–5). We confined our search to articles with these terms appearing in the title in order to generate a manageable number of results. Searching in article texts returns more than 1.4 million articles of which most are irrelevant (from a cursory screening). Final and rejected search syntaxes are described in SI Table S5. News/media searches were performed on October 4, 2020 and returned 572 total results, split between FW (184), EW (185), and FE (203). No articles included food, energy, and water together in their titles.

All authors and research assistant C.G. then coded causal relationships asserted in news/media articles as was done for the WoS search. Given the large number of news/media articles returned, we manually reviewed a subset of 25% each of FW, EW, and FE. The 25% reviewed for each category were the first 25% as sorted by Nexis Uni in decreasing order of relevance; cursory review of the remaining 75% of articles in each group coded revealed that the majority presented no discernible link to food–energy–water systems and/or included only a tangential reference to COVID-19. Duplicate articles (approximately 22%) were skipped. Roughly 52% of articles were market reports that mentioned COVID-19 that did not make any causal claims.

For all 572 news/media results, we evaluated the distribution of articles by industry and by article type and performed word frequency analysis for each nexus group (EW, FE, and FW). Word frequency analysis is often used as a screening method to identify concepts or associations of particular concern within a large body of literature.[49] This analysis was carried out using the WordCloud package[50] in Python.[51]

Finally, for all 572 news/media results, we investigated distributional similarity measures for a subset of high-frequency words to better understand their contextual significance in news/media articles. The distributional similarity measures the number of contexts shared by words in a given text under the assumption that words that occur in similar contexts have similar meanings.[52,53] This analysis was carried out using the NLTK package for Python.[54]

We note that our findings are based on articles written in English and therefore may not capture phenomena occurring only in non-English-speaking countries. However, we are not aware of FEW/COVID-19 phenomena unique to the non-English-speaking world.

Results and Discussion

Supply-Side Stresses

The most frequently identified relationships correspond to risks associated with work in the food sector (notably meatpacking and processing); food sector workers are at high risk of SARS-CoV-2 infection, posing a risk of worker illness and mortality and necessitating physical distancing interventions, all of which have disrupted a food system dependent on labor inputs. This has resulted in localized food shortages with unclear impacts on nutrition and health.[55] All relationships identified in scientific and news/media articles relevant to supply-side disruptions in FEW sectors are displayed in Table . All causal relations identified are listed in SI Table S6 along with summary statistics and citation information.

Table 1

FEW Supply-Side Relationships Asserted in Scientific and News/Media Articles

	percent reportinga
relationship	scientific articles	news/media articles
economic support → purchasing power (of producers)	22%	16%
food supply → food/nutritional security/choices	6%	20%
physical distancing → food supplyb	9%	16%
food supply → SARS-CoV-2 transmissionc	9%	9%
water availability → food supply	2%	9%
SARS-CoV-2 transmission → food supplyd	4%	4%
physical distancing → water qualitye	4%	0
employment/economic activity → water qualityf	4%	0
water availability → water security	0	1%

“Percent reporting” reflects articles discussing supply-side and demand-side effects.

Physical distancing measures instituted in meat processing and packing plants and elsewhere have resulted in food supply slowdowns.

Meat processing and packing plants are high-risk occupational settings for COVID-19, and there have been many outbreaks there.

COVID-19 outbreaks in meat processing and packing plants and elsewhere lead to absenteeism, slowing down food production.

Increased occupancy of residential buildings as a result of workplace distancing have changed volumes and quantities of wastewater produced worldwide.

Diminished industrial production at the start of the pandemic has changed volumes and quantities of wastewater produced worldwide.

In the United States, black, Hispanic, and Native American individuals face disproportionate COVID-19 risks, partly as a result of food, energy, and water systems. They are highly represented in workplaces with higher COVID-19 risks such as meatpacking[56,57] and are also more likely to be food-insecure and hence vulnerable to disrupted food supply chains.[58] Meanwhile, food insecurity contributes to higher rates of pre-existing health conditions, which in turn increases susceptibility to COVID-19 morbidity and mortality.[59] These system dynamics are represented in Figure and explored further below.

Figure 1

Relations among food, energy, water, health, and economic concepts in the context of COVID-19. Hatched lines are relations not corroborated in a Web of Science (WoS) search and correspond to possible research gaps. All hatched lines correspond to results of the news/media review except energy demand → energy security (limited media predictions not found by news/media review) and energy security → comorbidities (logical inference and media reports that emerged after our formal search). Non-FEW relations occurring only once in scientific articles are omitted for clarity. A complete list of relations with journal citation information and summary statistics for a news/media search included in SI Table S6. Positive/negative associations refer to expected numerical relationships rather than better/worse judgments.

The only cross-sector, supply-side stress identified was the potential impact of water availability on food production. This is consistent with recent observations that, generally, energy systems have weathered COVID-19 with no major disruptions, likely due to the overall suppressive effect the pandemic has had on energy demand (discussed in the Demand-Side Stresses section).[26]

Economic support to agricultural producers was widely discussed, particularly in the U.S. media (between April 2020 and March 2021, roughly $42 billion in public funds was appropriated to support the U.S. food sector).[60−62] However, we did not identify any articles (news/media or scientific) evaluating whether enhanced producer purchasing power affects overall food supply or food security. Supply-side economic support has also generally not translated into support for food workers (e.g., paid sick leave) who are disproportionately susceptible to occupational SARS-CoV-2 exposure, which in turn destabilizes food supplies.[63−65]

The news/media review identified 50 food–energy, 63 food–water, and 129 energy–water articles classified under “Energy & Utilities” (SI Figure S1). Word frequency analysis (SI Figure S2) also revealed a strong industry focus in EW articles (“market” and related words most represented) compared to the consumer focus among FE and FW articles (“COVID,” “people,” and related words most represented). This is also reflected in the results of semantic distribution analysis undertaken for the terms “government” (highly represented in FE and FW articles, SI Table S7) and “market” (highly represented in EW articles, SI Table S8). Many industry-related terms (e.g., “market,” “vendors,” “forecast,” “order”) appeared frequently in a similar context as the term “government,” supporting our finding that EW articles, more than FE and FW articles, focused on challenges, perceptions, and concerns of producers and industries. It was therefore surprising that no FE or EW cross-sector, supply-side relationships were identified from the news/media review.

Close inspection of the energy–water articles however reveals that virtually all are either market reports for cross-sector products (e.g., water heaters) or discuss industry concern about cross-sector trade-offs on the demand side. This includes water and power disconnections and disconnection bans introduced during the pandemic as described in the Demand-Side Stresses section. While supply-side terminology was heavily represented in news/media articles returned by the EW search, the underlying stresses fell virtually exclusively on the demand side.

After our structured searches were completed, in February 2021, a cold wave struck North America. This caused power failures in Texas where the electrical grid depends on poorly winterized generating assets and lacks regional interconnections.[66] This imperiled COVID-19 outpatients dependent on portable oxygen concentrators[67] and forced people in unheated homes to weigh risks of cold exposure against risks of contracting SARS-CoV-2 in congregate shelters.[68] This demonstrates a causal connection between energy security and COVID-19 morbidity and mortality not uncovered in our formal search.

Demand-Side Stresses

The most widely discussed demand-side stresses on FEW security relate to a collapse in purchasing power among lower-income individuals dependent on income from the retail, hospitality, and food services sectors. These sectors were severely affected by physical distancing interventions such as bar and restaurant closures. In the U.S., these disruptions disproportionately affected women and black, Hispanic, and Native American people.[57,69]

Economic impacts from COVID-19 and related interventions have created FEW insecurities on the demand side (summarized in Table ). In the United States, a patchwork of energy and water disconnection moratoria has so far reduced the number of utility disconnections thus representing a type of economic support by utilities to consumers.[70] With energy arrearages by the end of 2020 totaling roughly $32 billion in the U.S. alone and disconnection moratoria now expiring,[27] a utilities access crisis may have been delayed rather than avoided.

Table 2

FEW Demand-Side Relationships Asserted in Scientific and News/Media Articles

	percent reportinga
relationship	scientific articles	news/media articles
economic support → purchasing power (of consumers)	22%	16%
physical distancing → employment/economic activityb	10%	25%
purchasing power → food/nutrition security/choices	9%	25%
food supply → food/nutrition security/choices	6%	20%
physical distancing → food supplyc	9%	16%
purchasing power → water security	0	17%
purchasing power → energy security	0	12%
employment/economic activity → purchasing power	5%	6%
economic support → physical distancingd	0	6%
physical distancing → water demande	4%	4%
physical distancing → food/security nutrition/choicesf	4%	0
water security → physical distancingg	0	4%
water demand → waterborne pathogensh	0	2%
physical distancing → energy demandi	0	1%
physical distancing → hand washingj	0	1%

“Percent reporting” reflects articles discussing supply-side and demand-side effects.

Physical distancing measures heavily disrupted certain economic sectors such as hospitality and tourism leading to unemployment and economic precarity.

Decreased patronage of restaurants and other congregate dining settings (and increased at-home cooking) led to a mismatch in food packaging size supply vs demand.

Enhanced unemployment and other benefits make economically disruptive public health policies more feasible.

Decreased industrial/commercial building occupancy and increased residential building occupancy changes the volume and timing of water demand.

Decreased patronage of restaurants and other congregate dining settings (and increased at-home cooking) led to changes in food choices.

Lack of household drinking water access (e.g., in less developed countries) requires congregation around community spigots and impeding physical distancing

Standing water in the plumbing of commercial/industrial/institutional facilities with reduced occupancy can promote the proliferation of legionella.

Decreased industrial/commercial building occupancy and increased residential building occupancy changes the volume and timing of energy demand.

Physical distancing imperatives reduce the feasibility of hand washing in settings with community spigots (e.g., in less developed countries)

Lower income and/or black, Hispanic, or Native American households are also at higher risks of illness and death from COVID-19. (The Supply-Side Stresses section discusses this in the setting of FEW.) However, we did not identify articles linking COVID-19 morbidity or mortality to FEW disruptions, apart from supply-side impacts on food workers described earlier.

Racial disparities in economic security translate to racial disparities in FEW security during the COVID-19 pandemic. For example, in the U.S., food insecurity increased from roughly 19% to 32% among white households and from 39% to 50% among black households.[58] Even controlling for economic security, racial disparities in FEW security persist. For example, in a nationally representative sample of households within 200% of the federal poverty line, black households had 3 times the odds of having been disconnected from power in the month to May 2020, while households under 100% of the poverty line had no increased odds compared to the study population as a whole.[71]

Physical distancing measures implemented in response to COVID-19 vastly increased person-hours spent in residential neighborhoods, changing food, energy, and water demand patterns. Electrical grid failures were widely prophesied for summer 2020 given the sharp increase in remote work during the COVID-19 pandemic,[72,73] but we did not find evidence that this had occurred as of June 2021. In less developed countries where distributed drinking water access is not universal and water shortages are common, news/media articles suggested a trade-off between routine hand washing and physical distancing[74] and difficulty of following hand washing advice while leaving water available for other uses.[75] We did not identify scientific articles characterizing the physical distancing/hand-washing trade-off or evaluating competing demands for water use in the context of minimizing the burden of disease.

A smaller number of news/media and scientific articles addressed other demand-side disruptions to FEW systems not directly associated with purchasing power. This includes packaging size supply–demand mismatches (associated with the collapse in retail dining), which has led to substantial wastage of food as represented in Table by physical distancing → food supply.[76] The second-order impacts of food waste on energy and water waste are increasingly appreciated, but papers returned by our review did not address this.[77] We are aware of reports of legionella proliferating in stagnant water in pipes and tanks of buildings shut down following physical distancing orders.[78] This is particularly concerning given the risks associated with SARS-CoV-2/legionella coinfection.[79] However, our review did not capture any scientific articles that addressed these phenomena.

FEW/COVID-19 System Dynamics

COVID-19 and associated policy interventions like physical distancing have had profound and uneven impacts on human health. Many of these impacts have been realized through intermediate effects on FEW systems as described above. Aggregating these causal relationships uncovers feedback loops and trade-offs that span traditional areas of research and practice. Below, we discuss these system dynamics. Figure presents our integrated conceptual model of FEW/COVID-19 interactions. (All causal relations identified are listed in SI Table S6 along with summary statistics; FEW supply- and demand-side relations summarized in Tables and 2 above.)

First, SARS-CoV-2 spreads throughout the population as a function of social and virological parameters. Morbidity and mortality are determined by many variables including nutritional status and food security.[80] The conditions of meatpacking and food processing plants promote disease spread, but these facilities are essential components of the food supply chain.[81] Outbreaks there create disease burden among workers and among people who may suffer from supply chain disruptions resulting from worker absenteeism. Physical distancing (e.g., shift reductions) reduces the expected burden of disease among workers but may exacerbate food insecurities in the general population. This represents a direct burden of disease from malnutrition and indirect burden of disease from increased vulnerability to viral disease. There is currently no quantitative decision support for these interacting food–health–virus processes.

Second, economic disruption from COVID-19 (loss of wages among the sick) or policy responses (closure of industries such as retail dining) leads to the burden of disease via FEW insecurities and psychosocial impacts (described in the Demand-Side Stresses section). Physical distancing reduces disease burden, particularly among groups with high case-fatality rates. However, these groups are older and bear fewer of the countervailing economic impacts or second-order FEW insecurity risks. This raises equity issues that have been widely interrogated in the popular press but of which there has been virtually no quantitative analysis. Immediate and long-term impacts of simultaneous food, energy, and water affordability crises among low-income families are also unknown.

Third, food supply (and much else) depends on water and energy inputs, while food, energy, and water supply all depend on labor and capital inputs. COVID-19 morbidity and mortality has been concentrated among older people rather than working-age adults,[82] and large-scale supply-side FEW system breakdowns have so far been avoided. In the context of COVID-19, decision makers have therefore not had to confront trade-offs that might be forced by a pandemic driven by a deadlier pathogen such as the 1918 H1N1 strain. A pandemic that kills or disables a substantial fraction of FEW workers would force decision makers to weigh provision of essential services against risks to workers, where risks to workers would also jeopardize the provision of essential services. Meanwhile, COVID-19 has revealed the lack of modeling tools available to predict the impact of capital or labor shortfalls on integrated FEW systems including risks of cascading FEW crises.

The current evidence base does not facilitate analysis of these feedbacks and trade-offs. Almost all articles coded considered one-way relationships between at most two concepts in series (i.e., A → B and C → D but not B → C). Despite the complexity of coupled human–natural systems, most evidence in environmental and social sciences accounts only for such one-way interactions, limiting the ability of policymakers and others to quantify trade-offs and interactions across sectors.[46] This contrasts markedly to the biomedical sciences where careful statistical analysis is applied to determine the role of environmental exposures and prior disease status on mortality,[83−87] the impact of COVID-19 on comorbidities, and the possible confounding role of treatment for comorbidities on COVID-19 outcomes.[88,89]

While the study of complex causal structures in the FEW sciences has been limited, we did identify six pathways in the literature that reflect interest in second-order impacts, unintended consequences, and system dynamics (see Table for narrative descriptions). Of 81 scientific articles reviewed, 18 evaluated or asserted multiple outcomes from a given intervention or variable (e.g., impacts of physical distancing on nutrition, physical activity, etc.) or, conversely, multiple predictors of a given outcome (e.g., impacts of hand washing and physical distancing on SARS-CoV-2 transmission). This framing allows for analysis of trade-offs and comparative efficacy of alternative interventions. However, no article that we identified explicitly undertook such an analysis.

Table 3

Pathways Longer than Two Relationships in Series Identified by Articles from WoS Search

pathway	narrative explanation	citations
food supply → SARS-CoV-2 transmission → food/nutrition choices/security → environmenta	the perception of certain foods as vectors for SARS-CoV-2 may lead to more sustainable food choices	Yang[90]
physical distancing → employment → purchasing power	physical distancing can create economic precarity by reducing opportunities for certain types of work (e.g., food supply chain)	Orden[91]
physical distancing → food supply → food/nutrition choices/security → comorbidities	physical distancing has second-order effects on health through reduced food security	Akseer et al.[92], Lal[93]
water availability → food supply → food/nutrition choices/security → comorbidities → COVID-19 mortality and morbidity	water availability has second-order effects on health through reduced food security	Woertz[94]
SARS-CoV-2 transmission → COVID-19 mortality and morbidity → comorbidities → burden of disease	without intervention, COVID-19 will leave long-term impacts on maternal and general health	Kapur and Hod[95]
comorbidities → COVID-19 mortality and morbidity → burden of disease	comorbidities exacerbate COVID-19	Gasmi et al.[96]

Yang[90] quantified impacts on food choices in the context of broadly defined environmental “sustainability” but specific environmental end points were outside the scope of their study.

Overall, COVID-19 has exposed trade-offs and feedbacks around FEW systems stressed by viral pandemics, but these dynamics have not yet been well characterized. Because most working-age people who contract SARS-CoV-2/COVID-19 recover, direct impacts on the supply-side FEW systems have been smaller than demand-side impacts of COVID-19 policy responses instituted to protect vulnerable populations (e.g., physical distancing). FEW systems have therefore withstood a number of stresses posed by COVID-19, such as the impact of virus morbidity and mortality on labor inputs, even as COVID-19 has highlighted these vulnerabilities.

Research Needs

The FEW nexus has amplified SARS-CoV-2/COVID-19 spread (e.g., initial outbreak and spread among food sector workers) and has been destabilized by the loss of labor and capital inputs. While widespread food shortages, blackouts, and water shortages have largely been avoided, COVID-19 has demonstrated that FEW models have overlooked purchasing power and labor inputs as factors in assessing system resiliency. Similarly, tools to forecast viral spread across a population do not generally consider how illness and death from viruses contrast with (or interact with) burden of disease introduced by physical distancing measures, particularly among groups with low case-fatality rates. These impacts may be direct (psychosocial impacts) or triggered by FEW perturbations on the supply side (loss of labor inputs to ensure FEW system resiliency) or the demand side (loss of purchasing power to ensure FEW security).

Virtually all research we identified was retrospective, reporting on stresses that had already occurred. Overall, there is a lack of exploratory research including scenario analysis, risk analysis, and numerical simulation of FEW systems. Without these methods, insights into the potential for cascading failures will be limited. For example, as of July 2021, we have avoided large-scale supply-side breakdowns such as utility failures, despite the acknowledged potential of such events.[72,73] We did not identify any research pointing to the conditions (virological or meteorological parameters, etc.) that would lead to such breakdowns, limiting the ability for resiliency planning.

Water stresses in less developed countries have been acknowledged as a constraint on the feasibility of widespread handwashing, potentially increasing COVID-19 risks.[97,98] As of July 2021, inhalation of droplets and aerosols is now understood to drive SARS-CoV-2 transmission,[99,100] reducing the urgency of this constraint. However, hands/fomite transmission may be more important in future pandemics, and we did not identify any quantitative framework to understand how pandemics may change optimal water allocation in densely populated, water-stressed areas.

Recent history has demonstrated that pandemics of greater severity than COVID-19 (e.g., the 1918–1920 H1N1 pandemic) are very possible.[101,102] Major supply-side FEW disruptions were avoided during COVID-19, likely due to relatively low mortality among working-age adults. For example, in the United States, the case fatality rate of COVID-19 was roughly 130 times lower among 18-to-29-year-olds than among 65-to-74-year-olds prior to widespread vaccination.[82] Across the three waves of the 1918–1920 H1N1 pandemic, this was almost exactly reversed with excess deaths among 15-to-44-year-olds 129 times higher than among ≥65-year-olds.[101] Likewise, MERS-CoV and SARS-CoV-1 have case fatality rates of roughly 34% and 9.5%, respectively, compared to ∼1% with SARS-CoV-2.[103] Future risk and resilience simulations developed to enhance preparedness must account for such possibilities.

Quantitative risk trade-off frameworks can improve decision-making even when underlying processes are subject to large uncertainties.[104,105] These approaches are increasingly appreciated as tools to structure complex policy questions[106] and would likely be a useful approach for FEW resiliency planning. It is essential that such characterizations leverage the existing simulation capacity of the supply side of food–energy–water systems.[107]