Erratum to “How climate change interacts with inequity to affect nutrition”

Climate change poses a growing threat to the achievement of optimal nutritional status, both directly through affecting food production and indirectly through altering social and economic influences in people's lives. These adverse nutrition outcomes are not evenly distributed across the world, and vulnerable populations are the most impacted. Understanding how different forms of inequity interact with climate change and adverse nutritional outcomes is a novel area of research in today's challenging environment of increased climate change pressures. This article presents the results of a systematic literature search undertaken to identify the connections, trends and pathways between climate change, inequity and nutrition outcomes. Forty‐six peer‐reviewed studies are identified that explore these complex interactions with a specific focus on the extent to which equity is a fundamental component of climate change and nutrition research. The pathways captured in this body of evidence are mapped to current framework thinking to identify trends and gaps. While there is a trend for studies to acknowledge an unfair distribution of vulnerability to adverse nutrition outcomes, there is less attention given to the (lack of) recognition of the social situations which increase these groups' vulnerability and the absence of representation or inclusion of these groups as vital decision‐makers. Studies that do incorporate these core dimensions of equity take mixed‐method and qualitative approaches. This highlights an inherent value in stepping outside the usual scope of empirical climate change research, one that incorporates the voices of those most affected.

Through this decade, a significant amount of work has been done on conceptualizing pathways and links between climate change and nutrition. Though a combination of scoping internet searches, snowballing, and citation tracking we identified many such frameworks (Bryan et al., 2017;Crahay et al., 2010;Fanzo, Mclaren, Davis, & Choufani, 2017;FAO, IFAD, UNICEF, WFP, & WHO, 2018;Mckune et al., 2015;Smith et al., 2014;Thomson & Fanzo, 2015;Tirado et al., 2013;UNICEF, 2011;Watts et al., 2018;WHO, 2016). We found a high degree of overlap between frameworks. For example, many utilized the longstanding UNICEF, 1990 conceptual framework of drivers of malnutrition as the structure onto which climate considerations were applied. What we found to be lacking across these frameworks however was a consideration of inequity as a significant modifying factor between climate change pathways and nutrition outcomes. We sought to build upon such conceptual frameworks by incorporating equity considerations as a core component.
The exercise of identifying frameworks was valuable in itself as we found equity considerations to be lacking in the conceptual literature. In the context of this review, the framework also serves as a template onto which we can map the findings in order to get a sense of which pathways are prominent in the captured research. This is an adapted approach of the "Best Fit Framework Synthesis" model (Carroll, Booth, Leaviss, & Rick, 2013) which can be defined as a common sense approach to identify and a priori framework/frameworks against which to code data extracted from included studies, and then use analysis of that extracted data and/or thematic analysis to improve the framework, identify new themes, unexplored pathways or other data gaps.
We merged three frameworks which we found to accurately represent the climate and nutrition science as well as the broader social science analysis of equity to create our own "meta-framework" (Figure 1). The choice of these frameworks was based on detailed comparison of identified frameworks and discussion among the article authors, which included representation of experts on climate change, public health nutrition, equity, and international development. The first framework we used was the 2018 Lancet Countdown on health and climate change report (Watts et al., 2018), and the second was the UNICEF framework that highlights various drivers of nutrition outcomes at different levels (UNICEF, 1990, subsequently adapted for the Lancet Nutrition Series e.g., Black et al., 2013). The third framework builds on a recent article on climate and equity (Karlsson, Naess, Nightingale, & Thompson, 2018), which itself builds on a more substantial theoretical base on equity (see Harris & Nisbett, 2018), drawing particularly on work of the political philosopher John Rawls (Rawl, 1972) and feminist scholars such as Nancy Fraser (Fraser, 2011).
This meta-framework is one-directional as it was the cumulative effect of these pathways on the nutrition outcomes of the vulnerable that informed our research questions. However, there are many ways in which the layers of the framework can further perpetuate the threat of climate change via various multi-directional relationships within levels and through feedback loops. While they are out of the scope of this review, they are very important and deserve acknowledgement. For example, dietary intake, particularly the growing demand for meat and dairy in emerging economies, is fundamentally changing the scale of agricultural production and is a significant contributor to GHG emissions. Additionally, the way that vulnerable groups respond to the pressures of climate change may result in damaging coping behaviors. For example, deforestation and land degradation are short-term means of overcoming shocks, but with longterm impacts of increasing vulnerability and intensity of future shocks (IPCC, 2012). Another consideration is how mitigation and adaptation policy impact the most vulnerable, as there is potential for leaving these hard-to-reach groups behind, or even worse, doing more harm (Hasegawa et al., 2018;Nelson et al., 2009), leading to calls in the climate and equity literature for a more "just transition" to low carbon economies (Markkanen & Anger-Kraavi, 2019).

| Equity conceptualization
Equity is often mistaken with equality of outcomes, but a substantial literature has highlighted the fact that while studies of equity are predominantly concerned with the fact of such inequalities, they are as concerned or in some cases more concerned with the processes which led to those unequal outcomes in the first place (Harris & Nisbett, 2018). These may have deep roots, work in the field of health equity and the social determinants of health has advocated going beyond simple causes to the "causes of causes" (CSDH, 2008;Marmot, 2007). To draw on such nuanced conceptions of equity in our framework, we developed three concepts which extend and deepen a focus on equity, based on the forms of equity ("distributional," "recognition," and "procedural") highlighted by Karlsson et al., 2018. We adapt these concepts to make them accessible to a wider audience 1 labeling them: fairness, justice, and inclusion-as described below.

| Fairness
At their most basic, health inequities can be judged via differences in the distribution of health outcomes, or access to services or program coverage. That these differ according to socio-economic status-whether by wealth, gender, age, disability, sexuality and so on, highlights some of the "causes of causes" at play. In systematic reviews, the PROGRESS+ methodology has been developed to help search for and analyze health impacts of such difference (Box 1). Building on this, the concept of fairness (which has otherwise been referred to as distributional equity- Karlsson et al., 2018), focuses on the equitable distribution of outcomes, or access to services, or life chances, regardless of one's geography or identity. Fairness also references the need to ensure both equitable distribution of costs or impacts, as well as both positive and negative outcomes of climate-related policy. F I G U R E 1 Climate change and nutrition meta-framework. Meta-framework maps the impacts of climate change on malnutrition. The pathways pass through an essential equity dimension which mediates nutrition outcomes. Δ, variable; NCD, noncommunicable disease

| Justice
Fairness overlaps with the concept of justice, but justice (as in our legal systems) suggests a need to understand how these inequalities occurred in the first place and to provide some means of redressal. As the WHO definition suggests, people experience poor health outcomes not only because there is some natural or biomedical mechanism causing the outcome, but because of the policy, political, and social decisions that inadequately allocate scarce resources to people based on their identity in the first place. These resources might be education, or food, or wealth, or might be access to both preventative and curative health services and medicine. Furthermore, these inequities in resource allocation do not occur purely by chance but can reflect wider stigma and discrimination in society which can take place over generations. Justice implies that these inequities need to be recognized before they can be acted on (recognition equity ) and that different people's values and beliefs need to be taken into account in such action.

| Inclusion
The concept of inclusion (also referred to as process equity- Karlsson et al., 2018) focuses on who is involved in decision making, or processes of governance and accountability: who is accountable to whom. Decision-making and governance processes are equitable not just when they pay attention to inequitable processes, but when those most affected by decisions are adequately represented in them in the first place (or as disability and other counter-discrimination advocates have stated memorably "nothing about us, without us" 2 ).

| METHODS
A systematic mapping approach was undertaken as this proved flexible to accommodate the broad range of health, social, and environmental science literatures. The aim of a systematic mapping review is not to answer a specific question (as is the norm in a systematic review), rather, the aim is to collate, describe catalogue available evidence, while upholding rigorous and transparent processes of capturing relevant evidence. (James, Randall, & Haddaway, 2016). A systematic search was undertaken to identify relevant peer-reviewed research published between 2007 and 2019.

BOX 1. PROGRESS+ EQUITY THEMES
PROGRESS+ was developed as a tool within systematic reviews to give explicit consideration to equity in the design of health interventions. It is now widely applied beyond interventions and provides useful (evidencebased) categories to guide equity analysis and data extraction. It is now a deep-rooted component of systematic reviews, incorporated by Cochrane, and the Campbell group, among others. Representing the multidimensionality of the distribution of health outcomes, it is a flexible tool that allows further categories to be added, depending on the given topic. The year 2007 was chosen as the cut-off date as it was around this time that the IPCC were in the process of making conceptual linkages between climate change and nutrition that have been built on in all assessment reports since then (IPCC, 2007). The review team has diverse collective expertise in nutrition, social science, climate change and adaptation science, and risk and vulnerability analysis, which is very valuable in this interdisciplinary approach. A search syntax was created using the Population Exposure Comparison Outcome (PECO) model, with input from all reviewers into these terms (L. S., L. C., N. N., P. T., S. G.; Tables 2 and 3). Scoping searches were trialed in several databases to gauge the depth and relevance of published material. Search terms were further refined as a result, for example dropping terms that "exploded" results with irrelevant studies. A protocol was drafted following the scoping phase, however, due to the updated guidelines of the PROSPERO review registry, this review did not fit the criteria for review registration. Given the mapping nature of this review and type of evidence synthesis, this is unlikely to impact the reliability of our findings.
The final search was completed on January 15, 2019, in the PubMed, Web of Science and CABI abstracts databases. These databases were selected due to their interdisciplinary content and relevance across social science, health, and environmental fields. Search terms were adapted slightly for each database to reflect differing database search features. Search results were imported into the bibliographic software Mendeley and duplicates were excluded ( Figure. 2). The remaining studies (n = 5,304) were screened by title and abstract against predetermined eligibility criteria to determine relevance (Table 2). Two authors (L. S. and L. C.) completed the screening, each screened 2,562 studies, those deemed potentially includable were screened a second time by the alternate reviewer to ensure consistency in inclusion. Discrepancies were discussed and, where relevant, expert opinion was sought from the rest of the review team (N. N., P. T., and S. G.). A total of 320 studies were screened at a full-text level by LS and LC. Reasons for exclusion of studies were recorded, and 46 studies were determined to fully meet the inclusion criteria.
Only peer-reviewed primary research was included. 3 Beyond this, no further quality assessment of included studies was completed. Citations of included literature can be found in Supporting Infornation Item 2. The following information was extracted from the studies; country, region, study design, climate change components, nutrition outcomes, pathways of interaction, and sources of inequity (using the PROGRESS + themes). Two additional categories were added to PROGRESS+ that were relevant to the literature retrieved, that were not captured within original categories. These were enabling environment factors (weak governance, access to credit, provision of health insurance, country food import capacity, diminishing food aid) and climate change specific factors (unavailability of early warning systems, low adaptive capacity).

| Equity scoring of included studies
To enable a deeper understanding of how inequity is conceptualized in the literature, the studies were further assessed in relation to the three equity dimensions presented in Table 1. A scoring system was applied in order to differentiate studies with a strong focus on each of the equity dimensions, from those that just touched on them, or did not address these dimensions at all. The articles were read in full and the content assessed to see if/how dimensions of fairness, justice, and inclusion were incorporated. Each dimension received a score from 0 to 3, with 0 indicating no incorporation at all, 1 indicating some acknowledgement of the equity dimension but not in detail, 2 indicating that the article considered the dimension in a more thorough sense (albeit not as a core theme explored/measured), and 3 indicating that exploring the dimension was a fundamental objective of the research and/or measured in the methodology. Therefore, the maximum score is 9, indicating comprehensive integration of all three equity dimensions. This scoring system was not designed to reduce studies to a numerical value, or as a means of quality assessment. Nor was it designed as a critique of individual studies. Its purpose rather was to unpack the treatment of equity across a body of work that had already passed through a comprehensive search and eligibility screening process, satisfying the criterion of an equity focus within the context of climate change and malnutrition. While it is imperfect (given the overlap between dimensions and gray areas that cannot easily be captured within the three concepts), it does serve as a three-dimensional lens through which we can consider how equity is conceptualized and reported, helping to reveal trends and gaps in research.

| General characteristics
The included articles (n = 46) covered a range of regions, although the greatest coverage was of sub-Saharan Africa (n = 19; Table 4).
The majority of articles focused on undernutrition in their evaluation of how climate change affects nutrition outcomes (Table 4). When we look at nutrition outcomes investigated by the region of study, we see that the overwhelming focus in sub-Saharan Africa is on undernutrition, while diet-related noncommunicable diseases (NCDs) have only been included in articles focusing on North America, South Asia and East Asia and Pacific (Figure 3). The two articles that examined global trends focused on micronutrient deficiency (Myers, Wessells, Kloog, Zanobetti, & Schwartz, 2015;Nelson et al., 2018). Only one article included the potential for climate change to impact obesity (McIver et al., 2016).

Terms Syntax
Climate change terms "climat* change" OR" changing climate" OR "climate variability" OR "global warming" OR "climate shock*" OR "environmental change" OR "extreme climate" OR "extreme weather" OR "weather extreme" OR "climate extreme" OR "CO2 concentration*" OR "GHG emission*" AND Equity terms equity OR equitable OR inequit* OR equality OR inequality OR marginalized OR marginalization OR marginalize OR marginalise OR marginalised OR marginalization OR empower* OR disempower* OR vulnerable OR vulnerabilit* OR slum OR poverty OR impoverished OR race OR ethnicity OR ethnic OR culture OR indigenous OR refugee OR minorities OR gender OR religion OR poor OR "resource poor" OR "social capital" OR "social exclusion" OR "sexual orientation" OR morbidity OR disability OR disabled OR disparity OR deprivation OR deprived OR urban OR rural OR lowland* OR "market access" OR coastal OR highland* OR "land access" OR "rural growth" OR (infrastructure and access) OR caste OR tribe OR tribal OR wealth OR "pro-poor" OR "income gap" OR high-income OR low-income OR higher-income OR lower-income OR "social norm" OR "social change" OR "social behaviour" OR "socio-economic" OR elder OR elderly OR geriatric OR youth OR "young people" OR adolescen* OR power OR ((gender OR woman* OR women OR "man" OR "men" OR female OR male OR girl* OR boy*) and (equal* OR inequal* OR equit* OR inequit* OR dispar* OR *empower*)) AND Nutrition terms stunting OR wasting OR "low birth weight" OR anemia OR malnutrition OR undernutrition OR overweight OR obesity OR hunger OR emaciation OR marasmus OR "nutritional status" OR undernourished OR nutrition OR anthropometry OR "noncommunicable disease" OR "cardiovascular disease" OR diabetes OR overnutrition OR overnourished The two most commonly addressed climate change components in the included articles were rainfall variability and temperature variability. In some cases, particularly those articles using scenario modeling to evaluate potential future impacts, the climate change components were not specified and were addressed in general terms as undefined effects of climate change.
Using the PROGRESS+ framework to identify factors that affect equity, we categorized the literature by the factors discussed in each article. Place of residence and socioeconomic status were the most frequently included categories, followed by occupation/livelihood and specific personal characteristics (children, women, single motherhood, and morbidity; Figure 4). Although religion is a category within the PROGRESS+ framework, none of the articles included it in their analyses. Despite gender becoming a more frequent consideration within studies on climate change, agriculture and nutrition, only a small number of included articles in this review addressed gender as an equity component. Further details of examples within each category can be found in Supporting Infornation Item 1.

| Equity dimensions and scoring
On reviewing the scores (0-3) within the three categories of fairness, justice, and inclusion, we find a high level of variation of scores across studies and their characteristics ( Figure 5). Nine articles scored highly across all three dimensions F I G U R E 3 Nutrition outcomes reported across regions. Some studies report more than one nutrition outcome or region therefore total is greater than 46 (number of included studies)

F I G U R E 4 PROGRESS+ equity themes in included studies
F I G U R E 5 Equity scores ranked from lowest to highest for each included article. Each article can score a maximum of 3 per equity dimension (fairness, justice, and inclusion), therefore a maximum score of 9 (total score of 8 or 9) and are all either qualitative or mixed methods studies. Of the nine highest scoring articles on equity, four are focused on North America, three on Africa, and three on East Asia/Pacific/South Asia. 4 The articles that scored lowest on equity are modeling, cross-sectional, and other quantitatively focused studies. Articles score highest in the fairness dimension (mean score 2.6 out of maximum 3), lowest on the justice dimension (mean score of 1.0) and low on the inclusion dimension (mean score of 1.3).

| Framework pathways
To relate the articles included in the systematic review to the meta-framework we created through our framework scoping exercise, we identified the pathways described in each article and mapped these to our framework ( Figure. 6). Figure 6 illustrates the pathways identified in the included literature, from changes in climate to nutrition outcomes. Each vertical black bar represents a component that mirrors those in our meta-framework. The length of each black bar represents the volume of literature reporting on that component. Not all studies pass through each layer, therefore when there is missing information, this is represented with an asterisk.
F I G U R E 6 Pathways of interaction from climate change to nutrition outcomes. CC, climate change; NCD, noncommunicable disease; WASH, water, sanitation, and hygiene, *, some studies do not pass through each category of the pathway. When there is missing information this is represented with an asterisk. Each vertical black bar represents a component that mirrors those in the meta-framework. The length of each black bar represents the volume of literature reporting on that component Food security is the most widely reported pathway component, with over half of studies identifying this route from climate change to nutrition outcomes. We see that agricultural productivity and quality, physical capacity, and biodiversity all feed into food security. From the element of food security, dietary intake is studied most of all, ultimately leading to undernutrition as the main nutrition outcome reported (this includes stunting, wasting, underweight, thinness, low birth weight, and hunger). It is perhaps unsurprising to find this dynamic within the literature given the highly researched links between food security in the face of climate change and the ultimate impacts on food availability and quality. What is more nuanced about this pathway is the equity score across each dimension. Of the articles reporting on this pathway, the average score for fairness is 2.4, justice is 1.0, and inclusion is 1.3, giving an overall average equity score of 4.7 (out of 9). We see a trend of scoring highly in the identification of the distribution of fairness, with lower scores for recognizing how and why this exists and/or is perpetuated.
While the food pathway is well represented, we found a dearth of articles that met our criteria and that consider the ways climate change is affecting the underlying nutrition drivers of care capacity and health services, water, and sanitation. Dietary intake and undernutrition are the most common driver and outcome within the pathways examined, but this leaves a research gap around health status as an immediate driver and other nutrition outcomes (micronutrient deficiency, dietary diversity, overweight/obesity, and diet-related NCDs) in the context of how climate change is affecting nutrition.
No included articles examined the effects that climate change is having on fish or aquaculture productivity and which might affect nutrition outcomes despite this sector being very important to nutrition, especially among marginalized populations in many countries (Harper, Zeller, Hauzer, Pauly, & Sumaila, 2013;Lynch et al., 2016).

| DISCUSSION
Our analysis highlights four key lessons. First, we have identified a body of literature that explores the connections between climate change and malnutrition, mediated by some form of inequity, whether that be due to place of residence, occupation, gender, and so on. Through the application of equity scores across the articles included in our review, we find strong evidence of the inequitable distribution (fairness) of nutrition outcomes in the context of climate change. This is to be expected, as the search strategy and eligibility criteria were developed to identify such equity issues. However, when interrogated further to investigate the reasons (or "causes of the causes") for the perpetuation of discrimination (justice) overtime or the processes of exclusion or inclusion of vulnerable groups, the equity score of literature drops considerably. Thus, while we have evidence of that distribution of inequity in the context of climate change, we find much less evidence exploring how and why this inequity persists. This is an important gap: without this evidence, it becomes much harder to identify (let alone implement) equitable adaptation and mitigation solutions that are appropriate to the needs and struggles of populations vulnerable to climate change. A paucity of evidence exploring the inclusion of vulnerable groups as a process in decision making highlights another key gap in the research, particularly in the context of inclusive policy solutions.
Marginalized, or at-risk groups have often sustained generations of entrenched disadvantage due to systemic discriminatory systems (whether they be racial, patriarchal, colonial, and capitalist). Securing of land tenure rights for indigenous peoples is one such example of how inclusion can improve their legitimacy as stakeholders and decisionmakers, as well as play a vital role in climate change mitigation. "Secure tenure rights for indigenous groups provides a foundation for stewardship of fragile natural resources and development of innovative and equitable partnerships for the protection and restoration of tropical forests and other natural carbon sinks and the biodiversity they support." (Quan & Dyer, 2008). Furthermore, allocating productive assets such as land tenure rights in the name of women can enhance their tenure security and allow more benefits to flow to their dependants (Mitchell & McEvoy, 2019), improve economic independence of women, investment in child education, and smaller family size (Atinc et al., 2005), which are well understood underlying drivers of malnutrition (Black et al., 2013).
Second, the studies that score highly across the three dimensions we considered indicate the value of using qualitative or mixed methods in their analysis. In the nine studies that scored highly with an 8 or 9, the equity categories of gender and race, ethnicity, culture, and language are well represented, with an above-average representation compared with all the studies. The high-scoring gender studies (n = 4) detail factors such as women skipping meals to feed other family members in times of hunger, traveling further for water and fuel, and lacking access to productive resources and assets as sources of heightened vulnerability to poor nutrition outcomes (Adebo & Sekumade, 2013;Alston & Akhter, 2016;Ford & Beaumier, 2011;Githinji & Crane, 2014). The high-scoring studies exploring race, ethnicity, culture, and language largely focus on Inuit communities across North America, the way their lifestyles are changing due to melting ice patterns, and how this affects their mobility, traditional ways of hunting and fishing, and ultimately their diets (Ford & Beaumier, 2011;Wesche & Chan, 2010).
Such observations suggest the need for a holistic perspective, coupled with appropriate research methods that are capable of delving into the different dimensions that communities face. The articles that score the lowest on equity were modeling, cross-sectional, or other quantitatively-focused studies. This is not a criticism of these individual studies, as there is a time and place for research that models and estimates the impacts of climate change on health outcomes. However, if/when we are looking for evidence on the inequities facing vulnerable groups, and if the majority of evidence that is returned focuses on quantitative modeling studies that do not delve into deeper considerations of equity, then we have a problem as the solutions we choose will likely respond to this distributional understanding on equity, rather than investigating how and why it emerged in the first place.
There is increased recognition of the need to reshape the type of data we gather, and how it is collected. This is prioritized in the 2020 Global Nutrition Report which calls for the disaggregation of health and nutrition data for equity determinants such age, sex, ethnicity, education, wealth, disability, migration status, and geographic location, as well as qualitative data collection at the community level to increase understanding of their root causes (GNR, 2020). Participatory methods have been a development research tool that has the potential to shift the power dynamics when telling the stories of marginalized groups and co-create solutions. There is evidence of these methods in practice bringing success in climate adaptation and mitigation projects (Figueiredo & Perkins, 2013). However, caution is needed to ensure that longstanding power struggles and hierarchies are not reinforced by having every stakeholder around the same table, Sprain (2017) argues that there are specific paradoxes to participation in climate governance that need careful consideration.
Third, our analysis highlights that the existing literature is highly skewed toward a consideration of only one of many possible pathways from climate change impacts to nutrition outcomes: from direct effects on agricultural productivity through to food insecurity and thence to reduced dietary intake and undernutrition. This is undoubtedly a critically important pathway, but others may be equally important. We find few articles that meet our inclusion criteria addressing the ways in which climate change affects nutrition through underlying drivers other than food security such as care capacity and health services, water, and sanitation. This constitutes another major gap concerning health status as a driver of nutrition outcomes, particularly in the context of impeded physical capacity and stress. We found very few studies reporting in on obesity and diet-related NCDs.
It is important to note that additional articles were captured in our original search that linked increases in obesity to contributions to GHG emissions and climate change (e.g., Goodman et al., 2012;Squalli, 2014), but these were excluded from our review because, as noted in the conceptual framework section, we did not focus on that feedback loop in this research. The one included article reporting on obesity (McIver et al., 2016) discusses the likelihood of climate change affecting the local production of nutritious foods among Pacific Island nations and the potential increased reliance on imported, processed foods contributing to an increased prevalence of overweight and obesity. Such changes are not the result of individual choices but reflect rapidly changing food environments-a nutrition transition-which is influenced by a range of commercial drivers such as advertising, marketing, product formulation and lobbying of government policy . Such "commercial determinants" have been documented as influencing the available policy space both for food and healthy eating, as well as the climate crisis (Swinburn et al., 2019). In LMICs, such trends reflect a longer history of imbalanced trade stretching back to the colonial era, with fresh foods such as fruit in particular flowing one way toward wealthier countries and poorer countries a growing market for more unhealthy, processed products (Baker, Friel, Schram, & Labonte, 2016;Hawkes, 2010;Huang, 2010;Lopez, Loopstra, McKee, & Stuckler, 2017). Trade and food policy therefore can contribute to inequity in nutrition outcomes, as well as perpetuate climate damage. When food is considered a commodity, rather than essential for health, there is a proliferation of unhealthy and unsustainable foods (Thow & Nisbett, 2019).
Fourth, our review highlights the complexity and heterogeneity of the linkages between climate change and nutrition outcomes. The disciplinary mix of our author team is a good illustration: our analytical framework was developed as a result of collective expertise in nutrition, social science, climate change and adaptation science, and risk and vulnerability analysis. The value of inter-and multi-disciplinarity in this kind of research, when allied with the need for qualitative or mixed analytical methods, may partially explain the dearth of literature that explores the whole gamut of pathways between climate change and nutrition outcomes and the concentration on what the problem is rather than why the problem exists: such research can be time-consuming, challenging to manage, and potentially costly.

| CONCLUSIONS
Addressing the root causes of inequity can help mediate the negative impacts of climate change on nutrition outcomes. Our research indicates that inequity is most often considered in terms of the unfair distribution of vulnerability. But, we are missing a trick: in this era of sustainable development, lack of knowledge on fundamental justice, and the inclusion dimensions of equity, may severely limit the extent to which sustainable solutions to the challenges posed by climate change can be identified and implemented effectively. This is an important gap to fill if we are to broaden our understanding of deep-rooted inequities and how to reorient research and resources for equitable adaptation and mitigation action. One concrete step toward filling this gap is to develop a more sophisticated analytical framework than the one we developed here. The relationships between GHG emissions, climate, the physical environment, equity, different nutrition drivers, and ultimate nutrition outcomes are not linear. They include multiple feedback loops, as well as temporal and spatial dependencies. Improved understanding of these relationships could greatly enhance the frameworks and tools that could be brought to bear to identify equitable climate action. A second step is to embrace the need for inter-and multi-disciplinary teams in planning and implementing research in this arena, despite the undoubted challenges. Third, it is critical to share these findings and their implications with policymakers who are in a place to take action and enact plans and strategies that can help mitigate the inequitable effects of climate change on nutrition among populations vulnerable to these impacts.

ACKNOWLEDGMENTS
The authors wish to thank Ms. Ciara Gillespie for her assistance in extracting data from the full text of included articles, and Mr. Jason Chow for his design expertise in creating our Climate Change and Nutrition Meta Framework. This work was implemented as part of the CGIAR Research Program on Climate Change, Agriculture and Food Security (CCAFS), which is carried out with support from CGIAR Fund Donors and through bilateral funding agreements. For details please visit https://ccafs.cgiar.org/donors. The views expressed in this document cannot be taken to reflect the official opinions of these organizations. Production of this article was also undertaken as part of the CGIAR Research Program on Agriculture for Nutrition and Health (A4NH).

CONFLICT OF INTEREST
The authors declare no potential conflict of interest.
3 As a team, we discussed including gray literature and the potential value this would add to the review. We decided to exclude nonpeerreviewed materials, along with review and commentary pieces, as it was felt that these materials often reiterate the findings of peerreviewed empirical work that is already published. 4 One article looks at both North America and East Asia/Pacific by studying immigrants from the Marshall Islands to the US.