Search & Screening Results
The study selection process is shown in the PRISMA Flowchart (Fig. 1). We screened 2,754 unique records on title and abstract and 1,326 on full text. 844 studies met the inclusion criteria and were mapped onto our EGM using a fixed framework of climate change and FSNH categories (see HTML file). The cells in the EGM are segmented by climate change strategies: adaptation (orange), mitigation (yellow), both (green) and none (blue).
Climate change and FSNH thematic analyses are presented below in order of the prevalence of reports covering each domain. The results presented here forth are not additive to the total number of reports included, as many studies spanned multiple domains, populations, and settings.
Aspects of Climate Change
Almost two-thirds of reports focused on changes in weather patterns (n = 500), with over half of these (n = 264/500) discussing non-specific weather changes. Of the specific weather phenomena examined, ambient temperature was most prominent (35%, n = 177/500), followed by drought (27%, n = 137/500) and precipitation patterns (16%, n = 78/500).
Half of the reviews examined GHGs and related emissions (n = 427), mostly in general (48%, n = 207/427) but many focused on specific GHGs, particularly carbon (43%, n = 185/427) and nitrogen (25%, n = 107/427). Almost a fifth covered both GHGs and weather (18%, n = 155). Some studies did not specify which aspects of climate change they were discussing (11%, n = 95).
While not explicitly part of our search strategy, many of the included studies examined other anthropogenic drivers indirectly linked to climate change (34%, n = 289). Of these, the most prominent groups were environmental pollutants (35%, n = 101/289), energy (28%, n = 81/289), environmental degradation (25%, n = 72/241), and salinization (24%, n = 70/289).
Aspects of Food Systems, Nutrition and Nutrition-related health (FSNH)
Reports on farm level food production comprised the most saturated of the FSNH domains, consisting of 87% of reviews (n = 734). Within this, crops (65%, n = 480/734)) and Animal Source Foods (ASF; 28%, n = 204/734)) were the most covered, with much of each category focusing on general crop or livestock production, rather than specific crops or animals. Crop and ASF themes were not mutually exclusive, and many papers examined both (14%, n = 117). Aquaculture was the least reviewed category within food production (9%, n = 66/734)).
Reports on agro-environments were the second most prominent FSNH domain, accounting for 57% of reviews (n = 481). Within this, the major themes were soil (52%, n = 251/481), water systems (37%, n = 179/481), biodiversity/agroecology (29%, n = 141/481), land use (20%, n = 96/481), and disease, pathogens, and pests (20%, n = 96/481).
Post-harvest food systems reports were the third most populated FSNH domain (30% of all records, n = 256). Most of these reports focus on food environments (36%, n = 93/256) or food system policy (31%, n = 79/256), but other prominent themes were food waste or loss (24%, n = 62/256) and economics (24%, n = 61/256). Very few reports focused on post-harvest parts of food systems (prior to consumption), namely food supply chains and transport (n = 42), packaging and labelling (n = 19), food safety (n = 19) or food prices and expenditure (n = 17), albeit several of these subdomains are part of larger food environments frameworks (in addition to being commonly discussed in relationship to consumer choice and diets), and thus are doubly coded under food environments, as appropriate.
Food security and diets were a central theme for 17% of records (n = 140), the overwhelming majority of which focused on food security (91%, n = 128/140), followed by diets (47%, n = 66/140). Food security was covered by reviews both as an aspect of food systems (including pillars aligned to food environments such as availability and utilisation) and an aspect of consumption or nutrition. Thus, there was an overlap between these subdomains in the coding. While considered part of diets, research on Infant and Child Feeding (IYCF) made up only 5% of reports (n = 7/140).
Nutrition and nutrition-related health reports comprised the least populous FSNH domain, containing 12% of the records (n = 103). Half of these corresponded to non-communicable diseases (51%, n = 53/103) followed by forms of malnutrition (46%, n = 47/103) and communicable diseases (33% n = 34/103). Only 21 reviews focused on overweight or obesity, while 38 reviews considered underweight, micronutrient malnutrition or growth faltering in early life. The links made between climate change and FSNH domains are illustrated in Fig. 2.
For linkages between parts of food systems, food security and diets, nutrition, and nutrition-related health, 227 reviews only focused on one of these domains. Many of the reviews covered more than one FSNH domain (73%), especially those in ‘adjacent’, or most closely related categories of FSNH (e.g. agriculture environments and on-farm food production). Most covered two FSNH domains (n = 416). Fewer incorporated more than two FSNH domains, especially from the most ‘distant’ or different ends of the FSNH pathway. For instance, 35% (n = 298) of records examined both agro-environments and on farm food production domains, whereas 8% (n = 70) examined these two domains as well as domains from the post-harvest food system. Only four reports included aspects of all five FSNH domains. The links made between climate change and FSNH domains are illustrated in 2, and further broken down in Fig. 3.
Type of Reviews included
We identified several different types of evidence synthesis (Fig. 4). Over half of included reviews were subject or content reviews with no stated methodological approach (n = 494). Just over a quarter of reports were non-systematic literature reviews (n = 228), and 10% were systematic reviews that generally followed PRISMA or ROSES reporting guidelines (n = 89). Four percent were policy or organisational reviews, and another 4% offered a meta-analysis. Scoping reviews and bibliographic analyses were less common.
Settings, populations and trends over time
Most reports drew from diverse global literature (68%, n = 576) without limiting their geographic scope to specific countries or regions (Fig. 5). Some took a regional focus, such as on Sub-Saharan Africa (14%, n = 116), East Asia & Pacific (6%, n = 52), Europe and Central Asia (6%, n = 49), South Asia (5% n = 44), North America (4%, n = 29), Latin America and the Caribbean (LAC; 3%, n = 22) and Middle east and North Africa (MENA; 2%, n = 17). Others specifically examined economic settings, such as LMICs (13%, n = 110) and the European Union (4%, n = 32). A few specifically examined agro-environmental zones, such as arid/semi-arid (5%, n = 40) and tropical (4%, n = 33) regions. Small Island Developing States (SIDS) were the economic grouping with the least focus (1%; n = 9). Of the country specific papers, India, USA, and China were most prominent (3%, n = 23; 2%, n = 21; 2%, n = 18, respectively).
Of the reviews on human populations (n = 608), 77% were general (n = 468/608). Occupation, which mainly consisted of farmers/small-holder populations, accounted for 17% of these reviews (n = 102/608). Reviews focusing on age groups, genders, race/ethnicity/culture, or place of residence were rare, as were those focusing only on children. Almost 40% of reports did not focus on a human population and rather focused on other relevant FSNH components, like environments, animals, or plants (n = 236).
Our analysis shows that the overarching body of evidence syntheses linking climate change to FSNH has steadily grown by almost 200% in the past 5 years, from 89 in 2018 to 259 in 2022 (Fig. 6). As we concluded our search at the beginning of 2023, the number of these studies annually will most likely continue to increase, marking a continued interest in this cross-section of fields.
Equity
Three-quarters of reviews did not critically evaluate equity, meaning that they did not discuss equity as a driver of climate change-FSNH relationships, or equitable outcomes related to either. Of the papers that did engage critically with issues of equity (n = 214), socioeconomic status was the most discussed (62%, n = 132/214), followed by place of residence (32%, n = 68/214), gender/sex (29%, n = 63/214), and occupation (26%, n = 55/214). The focus on occupation mirrors the prevalence of reviews focusing on farmer and smallholder populations.
Directionality: hypothesized relationship between climate change and FSNH
Forty-nine percent of all records examined climate change as an exposure driving FSNH outcomes, 32% examined FSNH as an exposure driving climate change outcomes, and 19% assessed bidirectional relationships. Figure 7 summarises these relationships.
Climate Change Adaptation & Mitigation Strategies
Forty-three percent of reviews discussed adaptation strategies (the process of adapting to a changing climate), whereas 25% evaluated mitigation strategies (the process of reducing the drivers of climate change). Over 23% discussed both adaptation and mitigation strategies, usually through dynamic mechanisms of change, such as Climate-Smart Agriculture (CSA) or dietary decision support. Around 9% of reports discussed neither—these were typically descriptive reviews lacking solution-oriented discussions.
Mechanisms of Change
Most reports (89%) described some mechanism of change to achieve better outcomes, such as policies, technologies, programmes, or interventions (i.e. ways to change outcomes) or understand relationships better through research innovation (e.g. frameworks, methods or metrics, also including technology). Categories of mechanisms of change included policy, technology, Knowledge, Attitudes and Practices (KAP) (i.e. ways to shift knowledge, attitudes, or practices; see Supplemental methods 3A for definitions), and research-focused mechanisms (i.e. tools, methods, metrics and data required to produce better research).
Mechanisms of changing KAP were the most commonly discussed (67%, n = 564), with practices being most prevalent (90%, n = 509/564). The largest of this category were agro-environmental practices (45%, n = 227/564), crop practices (24%, n = 124/564), water practices (23%, n = 115/564), and animal practices (21%, n = 106/564). Technology mechanisms were the second largest grouping (51%, n = 433). They were mostly saturated by biotechnology (79%, n = 344/433), particularly genetics and genetically modified organisms (e.g. CRISPR gene engineering and climate-resilient breeding) (37%, n = 128/433), fertilizers or pesticides (31%, n = 107/433) and microbiomes (17%, n = 58/433). Policy and governance change mechanisms were not often put forth 16% (n = 137).
Informing action through different research approaches (i.e. research-focused mechanisms of change) were proposed less frequently. Of the reviews critically evaluating research approaches, 20% discussed shifts in research methods (n = 172), mostly focused on models (83%, n = 143/172), including prediction, econometric, simulation and systems models. Only 11% of reviews discussed research metrics and measures for either FSNH or climate change (n = 96), and less than 6% discussed changes to data sources and uses (n = 49). Figure 8 shows groupings of different mechanisms of change mediating FSNH and climate change factors.
Recommendations proposed by authors
The majority of authors included recommendations for future work and action, beyond the mechanisms of change they described (91%, n = 770). Some reviews did not make specific recommendations (9%). When we classified author recommendations, including future research, policy and programme changes, we found that changes in practices, including production or supply management, behaviours and systems were the most commonly recommended (58%, n = 449). This could have included strategies to manage heat stress among livestock or types of messaging that influence dietary choices. Changes to research focus, including specific comments on filling gaps and creating new evidence, was the second most common recommendation (38%, n = 296), followed by changes to policy and governance (36%, n = 278). Common recommendations around filling research gaps included things like using certain types of models (consequential LCA), or collecting data to understand contextual factors in interventions or actions. Suggested changes to policy or governance were often broad, such as making financing schemes more accessible, imposing new taxes, tariffs, or supporting certain innovations in production. Changes to technology were recommended for 28% of solution-oriented papers, which often included breeding improvements to seeds or varietals. Changes to knowledge, including decision support and education, changes to research methods and frameworks, and more interdisciplinary and cross-sectoral collaboration were each recommended by about 20% of papers. Changes in resource access or availability (10%, n = 78), and inclusion of equity or changes to equitable approaches were the least mentioned in reports (7%, n = 52). These groups of recommendations can be selected as filters in the EGM and papers assessed by the users regarding their specific recommendations.
Expert consultation
We received 18 responses from the expert consultation, who gave input on a) landmark literature on climate change and FSNH, b) research evidence gaps and recommendations, and c) promising ways forward, especially for research tools, methods and metrics. We summarised responses, and present illustrative quotations of these responses (numbered to maintain anonymity). The full analysis of evidence gaps and areas for future research are provided in Supplemental Results 2–4.
Suggested literature
Some of the key literature recommendations were identified in our literature search (21–24). Our search strategy did not capture certain relevant reviews since they did not mention ‘review’ or ’overview’ in the title or abstract (5, 25–27), did not have an abstract, were not indexed (2, 28–30), or were housed in unsearched grey literature databases (31–36). These reviews were added for data extraction. Some recommended papers did not fit inclusion criteria, (e.g. full text books, primary analyses, or papers not yet published), so they were not included in the EGM.
Evidence Gaps
The expert consultation identified critical research gaps and suggested priorities for advancing understanding of the climate change-FSNH nexus. Broad themes mentioned by many participants included increasing research within post-harvest subsectors of the food system. It also included further understanding of climate change impacts, both on food system functions and on the nutrition and health of understudied and marginalized population groups. Many mentioned interactions of climate impacts with socio-economic and political contexts and cascading risks across systems and sectors. Improving knowledge of food system vulnerabilities and adaptive capacity was highlighted alongside strong calls for evidence on mitigation and adaptation strategies. Several experts noted a need to know about the effects of combining strategies, and the feasibility and translation of strategies for different contexts. They also highlighted a dearth of information on the impacts of these approaches on nutrition and health, including unintended outcomes. They argued a need for: new and accessible data; research approaches that address equity and include indigenous knowledge; advancing analytical models that capture complex systems and dynamic, interacting risks and impacts; and increasing the availability of tools that can inform research and policy. Illustrative quotes are highlighted in Box 1.
Box 1: Illustrative quotes from experts about current evidence gaps on climate change linked to food systems, nutrition and nutrition-related health
“The majority of food retailers believe they can adapt their supply chains very quickly but we actually do not know the stress limits of this or how it might affect nutrition and health in future.” (Respondent 1)
“Middle parts of the chain are a huge research gap, such as commodity structures, processing and transport as most of us eat food-like substances. These parts of the food chain can be modelled in a much more rigorous way, particularly transportation as a contributor to or mitigator of climate change.” (Respondent 5)
“Most of the literature focuses on fairly simple pathways (e.g. heat stress affecting food production) rather than more complex interactions such as climate change compounding other drivers which affect agriculture, nutrition and health, or having unanticipated outcomes.” (Respondent 9)
“Research gaps include understanding which strategies are adaptable to various contexts (particularly those most vulnerable to climate change), their impacts on nutrition and health outcomes, and equity outcomes.” (Respondent 4)
“The basic story of migration from the evidence so far is that the poorest and most vulnerable move very little – for them there will be more climate-poverty traps than they will be climate refugees.” (Respondent 3)
“There is a gap in integrating indigenous and local knowledge into research on climate change adaptation and mitigation (climate-resilient agriculture) and policies. Indigenous communities are particularly vulnerable to the effects of climate change, and also hold important knowledge on climate resilience.” (Respondent 4)
|
Expert-identified directions for future research:
Suggestions for future directions in research on climate change related to FSNH echoed the gaps respondents identified, and the key recommendations made by authors of reports included in the EGM. Broadly, respondents mentioned the need to identify which kinds of data, research methods and approaches would inform food system action. Some questioned whether macro- or micro-systems focus would be most informative for identifying levers of change or be most useful for decision-makers at various levels. Respondents noted that methods are more evolved to measure impacts of acute events or disasters but much less evolved for slower (but no less profound) changes such as pest infestation, non-acute flooding, or temperature changes. The inability to capture the spatial-temporal aspects of changing climates, food systems, and health poses challenges to planning effective interventions. They also raised issues of cost modelling and political economy analysis to figure out where the onus for the transformation of environments, food systems and nutrition should lie. Further integration across disciplines, sectors, and thematic research communities was posed as an important future direction. Illustrative quotations from experts are listed in Box 2.
An overall conclusion from the consultation was that there are two different data-related gaps to fill: a) models and analyses built by experts that include much more data and types of data, and, critically, have better precision and predictive capacity even in the face of changing scenarios; and b) effective tools that are straightforward, easy and free to use by non-experts, that have enough granularity for decision-making at a sub-national level or above. Several approaches or principles were mentioned as essential to ensure equitable progress on both human and planetary health, especially feminism, bottom-up, participatory approaches and co-produced solutions, and biodiversity and agroecology.
Box 2: Illustrative quotes from experts about future directions for research on climate change linked to food systems, nutrition and nutrition-related health
Progress “requires collaboration across lots of disciplines and scales.” (Respondent 8)
Even given proper data and perfect models, there will need to be “supporting capacities to use the tools, realize value around tools” and enable decision-making. (Respondent 6)
“There is a big push to collect metrics at various scales and of different dimensions of food systems…We need to…explore how some of those metrics could be projected into the future and related to climate change. There is a need for funders, policy makers and private sector incumbents to come together to…improve data collection and accessibility.” (Respondent 8)
“True validation of…indicators commonly used to proxy for dimensions of climate, planet, food system, health, economics, justice dimensions is still needed.” (Respondent 10)
“Qualitative and quantitative evaluations of climate risk impacts on FSNH among LMICs should also be developed, particularly regarding agriculture adaption strategies.” (Respondent 4)
“Models and tools that can compare impacts and trade-offs from different policies, programmes and actions” [would be transformative in the field]. (Respondent 2)
“There is a distinct lack of decision tools that are freely accessible, low-input and do not require extensive data knowledge available to policy-makers and government employees. [We need] ways to look at food systems in totality, and triggers for performance. We are not currently building tools that are relevant for governments.” (Respondent 5)
|
Integration with the EGM
Overall, the thematic gaps and the contextual elucidation of gaps and future directions evident in the consultation matched well with the EGM results. Especially the analysis of recommendations for future research directions was strengthened and enriched by the overlaps between the literature presented in the EGM and the expert consultation.
There was a notable overlap between the results of the consultation and the EGM in the lack of reviews on post-farmgate food systems (including processing, packaging, and transport) (mentioned by half of respondents in the consultation). Food and/or nutrient loss and waste, especially issues of perishability and infrastructure, were commonly mentioned by experts and underrepresented in the EGM. Another example of overlap was the lack of reviews focused on food environments and dietary considerations such as accessibility and affordability. Many expert respondents brought up different aspects of equity and vulnerability, albeit from different perspectives. Evidence gaps in sub-domains, such as vector-borne and infectious disease in the health domain, or biodiversity in the agro-environment domain, were also overlapping gaps.
Some gaps mentioned in the expert consultation were, in fact, areas of evidence clustering in the EGM, for instance, the synthesis of evidence on the nutrient quality of crops, or climate responsive practices or programmes. These are examples of where much literature exists, but may not have yet translated into clear, actionable ways forward, especially given that most review types were not systematic reviews or meta-analyses. Experts noted that despite the proliferation of reporting on these topics, the scientific literature falls short of being useful for programme and policy design.