Climate change adaptation and mitigation in agriculture: a review of the evidence for synergies and tradeoffs

Agricultural practices that both support climate change mitigation and facilitate adaptation to a changing climate are critical for reducing greenhouse gas emissions while ensuring food security. This need has led to many claims regarding the potential for a variety of agricultural practices to achieve synergies between mitigation and adaptation in agriculture. However, the evidence for climate change mitigation and adaptation synergies in agriculture remains mixed. To evaluate such claims, we examined the evidence for these synergies by conducting a systematic review of peer-reviewed literature that make claims about outcomes for both climate change adaptation and mitigation in agriculture. Based on 87 articles identified, we show that synergistic outcomes are claimed more frequently than tradeoffs for all practices, yet the evidence was stronger for mixed and conflicting outcomes than for synergies. Indeed, claims of synergistic outcomes may be overstated, because these publications more often relied on secondary data rather than empirically evaluating adaptation and mitigation outcomes. We also show important gaps in the consideration and assessment of climate change adaptation and mitigation objectives and outcomes. This review highlights the critical need for more robust research, evidence, and evaluation of the adaptation and mitigation outcomes of agricultural practices, and the need to clarify the contexts of such results, in order to effectively support policies and practices that aim to promote synergistic outcomes and avoid conflicting outcomes.


Introduction
Agroecosystems cover over a third of earth's land surface [1].These lands are vital to sustaining roughly 7.8 billion people and are highly vulnerable to climate change.Indeed, a third of the variability in agricultural productivity is driven by climate factors such as temperature, precipitation [2], and extreme weather events, and in specific locations, these factors may account for up to 70% of the variability in production [3].Many food-producing regions around the world are already experiencing changes in climate that are projected to continue, with potentially dramatic negative impacts on agriculture [4].At the same time, while not all agricultural practices similarly contribute to climate change, agricultural activities are responsible for 10%-12% of global anthropogenic greenhouse gas (GHG) emissions [5].And further, including GHG emissions from pre-and postproduction of food and agriculture-driven land use change increases the contribution of agriculture to global anthropogenic emissions to a third [6,7].Yet, agricultural practices also have the potential to reduce these negative climate impacts [8].Thus, agriculture is affected by, contributes to, and can help mitigate climate change.Given the significant impact of climate change on agriculture and the need to meet mitigation targets, such as those set by the Paris Agreement, climate change mitigation and adaptation are increasingly simultaneous objectives of agricultural research and action as well as climate and development policy [9][10][11].
The growing need to respond to climate change via adaptation and mitigation has led to many claims of practices that achieve both adaptation and mitigation objectives.However, relationships between climate change mitigation and adaptation in agriculture are not always clear.Nevertheless, 'Climate-smart agriculture,' farming that enhances climate change mitigation, adaptation, and productivity, has been widely promoted as a technical solution to climate change [12,13].As farmers, researchers, and policymakers seek to demonstrate and promote climaterelevant agricultural practices, claims are often made that specific practices can deliver adaptation and mitigation synergies.While there have been several reviews addressing linkages between climate change adaptation and mitigation in agriculture, studies that comprehensively assess adaptation and mitigation objectives and provide evidence for the ability of agricultural practices to support both adaptation and mitigation outcomes remain limited [14][15][16].Indeed, the specific effects of agricultural practices on a broader suite of adaptation and mitigation objectives are often not quantified [17], and the nature of the evidence and significance of the effects often go unreported, such that for many agricultural practices there is insufficient data for global estimates of effect sizes [18].
In much of the literature, the evidence provided for assessing adaptation and mitigation outcomes is often not well substantiated, and the evidence underlying claims of synergies remains simultaneously understudied and open to necessary critique [19][20][21].For example, agroforestry and management practices that aim to increase soil carbon sequestration are often exemplified as practices that have high climate adaptation and mitigation benefits, yet many knowledge gaps remain.The lack of quantitative data detailing agroforestry practices and measuring outcomes challenges confidence in the degree to which adopting these practices will impact climate objectives [22].Even when outcomes are measured, many studies do not examine the longterm impacts of climate mitigation and adaptation practices-either for soil carbon sequestration [23] or agroforestry [24]-but long-term sequestration is a critical component to understanding the effectiveness of practices.
Finally, adaption and mitigation outcomes can also depend on interactions among environmental and management contexts, and even on interactions among multiple implemented agricultural practices.However, studies often focus on a singular practice or specific climate objective, without contextualizing the reported effects as a product of a suite of practices inherent to most agricultural activities [25].Increasingly, scholars are calling for more holistic consideration of agricultural practices [22,25,26] as well as more detail in quantifying and contextualizing climate change adaptation and mitigation outcomes [27].Further, these outcomes of agricultural practices often vary with regional and site-specific variables (e.g.biochar; [28]).While studies which examine agricultural practices in a specific location are important for contributing climate-relevant knowledge, the outcomes of these studies are often inappropriately used to claim effectiveness at broader regional or global scales [29].Thus, to understand the potential for climate-relevant agricultural practices and existing knowledge gaps, we must understand the gaps and uncertainties of studies claiming climate change adaptation and mitigation synergies in agriculture and examine the nature of evidence provided.
Here, we review claims of agricultural practices achieving both climate change mitigation and adaptation outcomes and provide an assessment of the evidence supporting those claims.We examine the peer-reviewed literature to address: (1) What agricultural practices and climate change objectives are being assessed?(2) To what extent are synergistic adaptation and mitigation outcomes claimed and how is evidence provided?(3) How robust is the evidence for these claims and what does this evidence show?(4) What are some of the gaps and uncertainties in the stated objectives, outcomes, and evidence that need to be considered?

Methods
This review examines the scientific literature reporting claims of both climate change mitigation and adaptation achieved through agricultural practices.We assess the evidence for these claims and characterize the themes to provide a base for future work.

Literature compilation
We compiled 87 peer-reviewed publications that provided claims of both climate change mitigation and adaptation outcomes in agriculture from 2000-2018 (supplementary table 1).In January 2019, we used the ISI Web of Knowledge search engine and the CGIAR Research Program on Climate Change, Agriculture and Food Security (CCAFS) program publication databases, as CCAFS is the primary international research program looking at both adaptation and mitigation in developing countries and produces one of the largest bodies of emerging agricultural climate adaptation and mitigation research.
We compiled peer-reviewed publications from CCAFS' Low-Emissions Development program and Climate-Smart Agriculture (CSA) program publication databases using the keywords 'adaptation' and 'mitigation' and using the ISI Web of Knowledge (ISI WoK) search engine, selected using the keyword search string 'agriculture * AND adapt * AND mitigate * AND climate.'This yielded 1147 publications, which were then refined to 276 publications with additional keyword search strings reflecting interactions among adaptation and mitigation outcomes: co-benefit * OR cobenefit * OR conflict * OR synergy * OR tradeoff * OR trade-off * OR 'adverse effect * ' OR 'negative effect * ' OR 'positive effect * ' .We further included articles based on salient references and citations in the selected literature, as well as publications recommended for inclusion by expert researchers working in this area and publications that had been previously coded for adaptation and mitigation outcomes produced by CCAFS (for full methods see [30]).These additional sources added 20 unique publications, for a total of 296 publication that were reviewed for study eligibility.
Publications with no direct relevance for both climate change adaptation and mitigation objectives in agriculture were removed based on title screening (n = 49) and abstract screening (n = 86).The remaining 161 articles were read in full and subsequently screened to ensure that the publications made claims regarding the achievement of both climate change adaptation and mitigation objectives in agriculture.Eighty-seven publications met these criteria and were comprehensively assessed for climate change adaptation and mitigation claims and evidence.

Conceptual framework
Our goals were to characterize trends in the emergent body of agricultural climate adaptation and mitigation literature and provide an in-depth review of the claims of and evidence for synergistic or conflicting outcomes.To consistently assess how publications considered and provided evidence for agricultural practices, adaptation and mitigation objectives, and claims of outcomes, we developed a methodology that used consistent terminology.We adopted the Food and Agriculture Organization (FAO) definition of adaptation as 'changes in processes, practices and structures to moderate potential damages from climate change, or to benefit from opportunities associated with such changes' [31].We adopted the Intergovernmental Panel on Climate Change definition of mitigation as 'actions that reduce the rate of climate change […] by limiting or preventing GHG emissions and by enhancing activities that remove these gasses from the atmosphere.' [32].
We examined how authors of a publication provided evidence for the effects of an agricultural practice or set of practices for contributions to climate change adaptation and mitigation objective(s).The authors' conclusions are claims about adaptation and mitigation outcome(s) and their relationship.Claims can reflect synergistic outcomes, which fulfil both adaptation and mitigation objectives, conflicting outcomes, which do not fulfil both adaptation and mitigation objectives, but show tradeoffs, negative or adverse consequences between mitigation and adaptation outcomes, or mixed outcomes, which contain both synergistic and conflicting outcomes, either by considering several different adaptation and mitigation objectives potentially fulfilled by a given agricultural practice, or by considering several different agricultural practices for a given adaptation or mitigation objective.
For each publication, we assumed that claims are substantiated by evidence of the effects of an agricultural practice or set of practices for contributions to both climate change adaptation and mitigation objective(s).This evidence could be produced in three ways: (1) use of empirical primary data (e.g.measured field data), (2) use of modelled data, (3) use of secondary data or literature (e.g.synthesis of publications or case studies, citations from other literature).In some instances, claims were not substantiated, with no evidence provided.

Coding and claims
All 87 publications selected were coded for adaptation and mitigation objective(s), agricultural practice(s), claim(s) of synergistic, mixed, or conflicting outcomes and study characteristics (e.g.geographical region, scale of focus, main methodological approach).We documented the evidence provided for claims and assessed the robustness of that evidence.These results were compiled in a database that allowed us to characterize trends and analyse the content of the publications.
For mitigation, we grouped objectives into two primary categories: those relating to GHG emissions and those relating to carbon sequestration, with secondary coding of GHG emissions avoidance or reduction and carbon sequestration in either biomass or soil.For adaptation, we grouped objectives into four primary categories based on FAOs' Adaptation Framework [31]: 1. agricultural production, 2. socioeconomic 3. institutions and policy and 4. natural resources and ecosystems.For agricultural practices, we recorded the specific practices considered (n = 317), and then grouped these practices into four main categories with subcategories: soil and water management (subcategories: tillage, fertilizer, other amendments, water and irrigation), crop production (subcategories: post-harvest activities, crop diversification, seed/crop alterations, timing alterations, intercropping and cover cropping), agroforestry (subcategories: in livestock systems, in cropping systems, other) and livestock management (subcategories: herd/animal management, feed management, grazing and pasture, and manure) to characterize trends more broadly.
We used number of publications (n = 87), and the number of publications assessing each main category agricultural practices (n = 178) and subcategory (n = 317) as the primary units of analysis, as some publications considered multiple agricultural practices with different claims.In these cases, we disaggregated the practices from a single publication to compare their claims both at the more specific practice level and across the broader practice groupings.For each agricultural practice, we calculated a synergy-to-conflict claims ratio by amassing all publications that examined that agricultural practice, summing all the synergistic claims (numerator in equation ( 1)), summing all the conflicting claims made for any given agricultural practice (denominator in equation ( 1)), and dividing the sum or number of synergistic claims by the sum or number of conflicting claims (equation ( 1)).Synergy-to-conflict claims ratios of less than 1 indicate more claims of conflicts than synergies, whereas greater than 1 indicate more claims of synergies than conflicts.
where S:C is the synergy-to-conflict claims ratio, and S is the total number of synergistic claims and C is the total number of conflicting claims.

Evidence assessment
We did not independently assess the adaptation or mitigation effects claimed within the publications (e.g. with a further literature review of agricultural practices), but instead recorded the publication's claim(s) and assessed the robustness of the evidence provided within the publication.Evidence robustness scores were assigned to each publication.These were calculated based on the relative robustness of the evidence provided for adaptation and mitigation outcomes as follows.First, we assessed the evidence a publication provided to assess climate adaptation and mitigation outcomes.We scored an adaptation score (AS) and a mitigation score (MS) separately and then summed the scores.Scores were as follows: no substantiation of claims = 0; claims based on secondary data external to the publication = 1 with an extra 0.5 point if the secondary data was analysed by the authors using a systematic literature review or meta-analysis; claims based on model results, where an established model with documented methodology was used = 2; claims based on empirical measurement (i.e.primary quantitative and/or qualitative data collection) = 3.Second, we scored the dominant methodological approach for the publication's overall assessment of agricultural practices objectives and outcomes using the same scoring system.Finally, publications focused on a single practice tended to be more thorough, while publications assessing a wide range of agricultural practices were often more surficial in their reviews.Therefore, if a publication had a dedicated focus on one specific agricultural practice or closely related practices, it was given an additional point of 1 towards the methodological assessment score (MAS).The publication score was combined with the AS and MS to give each publication an evidence robustness score, with the highest possible score of 10.
where ERS is evidence robustness score, AS is the adaptation score, MS is mitigation score, and MAS is the methodological assessment score.The highest possible ERS score is 10 and the lowest possible score is 0. We conducted a sensitivity analysis with both a more simplified and a more complex scoring system and found no significant differences in results (supplementary figure 1).

General trends
The 87 publications we analysed were focused on both agricultural climate adaptation and mitigation, were published from 2000 to 2018 and spanned the globe (figure 1).Around a third had a global or multi-continent focus, while two-thirds reported results from specific locations within (listed in order from most to least publications): Africa, Asia, Europe, Central and South America, North America, or Australia (figure 1(C)).The publications described and assessed a total of 317 agricultural practices for contributions to climate adaptation and mitigation outcomes at scales ranging from field-level to multiple continents.Overall, agricultural practices were predominantly related to the management of soil and water, crop production, agroforestry, and livestock, and were relatively well represented across geographic regions.
The 87 publications examined different agricultural mitigation and adaptation objectives, and a single publication often included more than one objective.For mitigation, objectives were fairly straightforward.Publications assessed agricultural practices for their ability to either reduce or avoid GHG emissions (n = 32) or to sequester carbon in soils or biomass (n = 13) or both (n = 34), with a few publications considering mitigation broadly without specifying mechanisms (n = 8).For adaptation, most publications focused on maintaining or increasing agricultural production as their main climate change adaptation objective (n = 57).Socioeconomic objectives predominated in 20% of the publications (n = 18).Fewer publications focused on institutions and policy (n = 8) or natural resources and ecosystem (n = 4) adaptation objectives.

Claims for synergies and tradeoffs
Of the 87 publications that addressed adaptation and mitigation outcomes, nearly all (89%) contained claims of agricultural practices contributing to synergies between climate mitigation and adaptation outcomes.About one third of the 87 publications made claims to only synergistic outcomes, half made claims of a mix of both synergistic and conflicting outcomes, and nine made claims of only conflicting outcomes.
The predominant methods for substantiating these claims in all publications were secondary data and literature methods (e.g.syntheses of previously  published literature, literature reviews) (49% of articles), empirical methods (e.g.field experiments, case studies) (44%), and model-based approaches (7%).Publications that claimed synergies relied more often on secondary data from other literature, whereas publications that claimed mixed and conflicting outcomes more frequently used empirical or modelling methods (figure 2; p = 0.03, Fisher's exact test).Further, publications that focused on a specific agricultural practice more often substantiated evidence with empirical or modelling methods, whereas those that focused on more suites of varied practices more often relied on secondary data (p = 0.005, Fisher's exact test).Finally, claim substantiation method did not differ significantly by geographic region or agricultural practice.

Evidence assessment
Publications examined a variety of agricultural practices, predominantly relating to soil and water Table 1.Synergy-to-conflict claims ratio for adaptation and mitigation outcomes by agricultural practice.Green indicates higher synergy-to-conflict claims ratios, while yellow indicates lower synergy-to-conflict claims ratios.Higher ratios indicate more claims of synergies relative to conflicts.Values less than 1 indicate more claims of conflicts relative to synergies.Evidence robustness scores have a minimum possible score of 1 and maximum of 10 for each publication, and average evidence robustness scores are presented here for all the publications reporting on that agricultural practice from light blue (least robust) to dark blue (most robust).management, crop production, agroforestry, and livestock management.While many publications used labels for suites of agricultural practices (e.g.CSA, conservation agriculture, sustainable rice intensification), we analysed the specific practices that were described to contribute to adaptation or mitigation objective(s).

Agricultural practices # of publications
We calculated the synergy-to-conflict claims ratio and average evidence robustness scores by specific agricultural practices and grouped these practices by agroforestry, soil and water management, crop production, and livestock (table 1).We further examined whether the publication considering the practice was focused primarily on that practice or considered a range of practices, and broke down the average evidence robustness scores by claim (table 2).Overall, publications' claims of synergistic or conflicting outcomes for adaptation and mitigation varied by agricultural practice, as represented in the synergy-toconflict claims ratios across practices.Higher ratios indicate more claims of synergies relative to conflicts.Publications assessing agroforestry practices (n = 36) had the highest synergy-to-conflict claims ratio, with an even higher synergy-to-conflict claims ratio for publications that focused specifically on agroforestry practices (n = 13).Publications assessing livestock practices (n = 29) had the lowest synergy-to-conflict claims ratio, with an even lower synergy-to-conflict claims ratio for publications focusing specifically on livestock practices (n = 8).
Agricultural practices with higher synergy-toconflict claims ratio often had lower average evidence robustness scores, whereas agricultural practices with lower synergy-to-conflict claims ratio had higher average evidence robustness scores (figure 3).Across all publications, those that focused on a specific agricultural practice had substantially higher evidence robustness scores than publications that assessed a variety of practices (table 2).For all agricultural practices, publications that claimed mixed or conflicting outcomes had higher evidence robustness scores than those that only claimed synergistic outcomes (table 2).

Gaps in objectives, outcomes, and assessment
This review revealed several important gaps, both in the consideration of climate change adaptation and mitigation objectives and outcomes, and in the assessment of outcomes.

Objectives and outcomes
Mitigation objectives were relatively well-described and evenly represented (supplementary figure 2), with most publications (80%) considering specific outcomes for climate change mitigation by GHG emissions reductions or avoidance (n = 53) and/or carbon sequestration in above ground biomass (n = 12) or in soils (n = 37).However, we found a lack of careful examination of the relationship between GHG emissions reductions and yield increases.For 20% of publications considering GHG emissions reductions, outcomes assessed included reducing emissions intensity (n = 10).While reducing emissions intensities may reduce GHG emissions per unit of food produced, total emissions will very likely increase if yields increase (e.g.reducing GHG emissions per animal product, such as litre of milk, while increasing overall amount of animal product, such as increasing milk production).And indeed, all ten of these publications assessed yield increases within adaptation outcomes and nearly all (n = 9) made some claims of synergies, with half claiming only synergies.
For adaptation, agricultural productivity was the dominant objective in two-thirds of the publications, with socioeconomic, institution and policy, and natural resource and ecosystem as primary adaptation objectives for the remaining third of the publications.Beyond this broader imbalance, we found two further important and related gaps.First, even if publications described socioeconomic or natural resource and ecosystem adaptation objectives such as 'food security and nutrition' or 'improved ecosystem functioning' , most often the outcome that was assessed was still agricultural productivity, primarily measured as yield.Far fewer publications examined outcomes such as nutrition and nutritional diversity, medicinal benefits, land tenure, or improvements to agricultural labour.Second, of all the publications that considered agricultural productivity as the main adaptation objective, only those using secondary data approaches claimed synergies without conflicts.Publications with more robust empirical and model approaches only claimed mixed and conflicting outcomes for agricultural productivity, yet claimed synergies more frequently for the other objectives (figure 4).

Assessment
Across all agricultural practices, adaptation outcomes were assessed with more robust evidence than mitigation outcomes (figure 5).Publications that substantiated claims with empirical data did not always measure both adaptation and mitigation outcomes.Empirical evidence was provided more frequently for only adaptation outcomes (n = 14) than only mitigation outcomes (n = 3).Of the 38 publications with empirical approaches, only nine measured both adaptation and mitigation outcomes within the same study of agricultural practice(s).Of the publications that measured both adaptation and mitigation objectives, most were focused on soil and water and crop production practices and claimed predominantly mixed and conflicting results.

Discussion
Synergistic agricultural practices-those practices that achieve both mitigation and adaptation objectivesare often offered as a solution for addressing climate change [9,10].We found that in the scientific literature, claims of synergies were relatively common for a wide range of agricultural practices related to soil, crop, agroforestry, and livestock management (e.g.no or low tillage practices, alternate wetting and drying in rice paddies, and planting trees as hedgerows and windbreaks in agriculture fields; supplementary table 2).However, the evidence for synergistic outcomes was less robust than the evidence for mixed or conflicting outcomes.Publications claiming synergistic outcomes most often relied on secondary evidence from the literature rather than empirical data.Of the nine articles (out of 87) that empirically assessed both adaptation and mitigation outcomes, only two reported synergistic outcomes, whereas the other seven found mixed or conflicting outcomes.Publications with the most robust evidence more frequently claimed mixed or conflicting outcomes.This suggests that claims of synergies may be overstated.
While few publications empirically assessed both the adaptation and mitigation outcomes of a given agricultural practice, there were many more studies (17) which did use empirical evidence to assess either an adaptation or mitigation outcome empirically and then relied on modelled or secondary data for the other objective.This may reflect financial or labor constraints, or it may be due to the scientific focus of the study remaining on either adaptation or mitigation outcomes-even when considering both objectives together.To increase the robustness of evidence in the future, the need for empirical data for both adaptation and mitigation outcomes together needs to be made clear, and more attention to identifying the reason for the gaps, and overcoming them, is important.

Adaptation outcomes and measurement
We found that adaptation outcomes were more often substantiated by empirical evidence than mitigation outcomes, and that adaptation was most often measured as agricultural productivity, specifically yield.Agricultural yields were often described as important metrics for measuring increases in food security, and increasing yields were considered a priority for adaptation outcome [33,34].However, publications we examined also noted the importance of considering whether the outcomes being measured contributed meaningfully to achieving stated objectives [16,35].
Relying on productivity measures alone does not necessarily provide insight about food security or the adaptation of agriculture to climate change and may leave other systematic challenges of food and nutritional access unexamined [35][36][37].Further, using agricultural productivity as the dominant measurement for climate change adaptation means that other dimensions of adaptation (e.g.improving biodiversity, nutrition, wellbeing) may remain poorly understood.Despite the dominance of agricultural productivity as a measure for climate change adaptation, many studies did highlight the importance of adaptation outcomes other than productivity, including focusing on self-sufficiency, rights to resources, and meaningful access to food, water, and health, instead of economic returns, strengthening subsistence livelihoods as opposed to market integration, and the importance of spiritual and aesthetic values [33,38,39].This suggests that successful adaptation is a balance of many different adaptation factors, and that tradeoffs may occur among these different adaptation outcomes [34].
Therefore, while using yield as the dominant metric for assessing climate change adaptation outcomes might streamline assessments and make for easier quantification, it comes at the cost of potentially misunderstanding, or misrepresenting, a wider range of beneficial adaptation practices.Given the variety of important outcomes, climate change adaptation should not be considered in a singular way.Meaningful assessment of climate adaptation requires that indicators properly correspond with the outcomes they are meant to address, and that multiple possible outcomes, not only yield, be considered.

Mitigation outcomes and measurement
While overall publications selected for in this study examined both GHG emissions and carbon sequestration, for those that empirically measured mitigation outcomes-almost all were measurements of GHG emissions.As with adaptation, climate mitigation requires considering multiple mitigation outcomes from any given practice, such as examining the effects of a practice on carbon sequestration in the soil in addition to GHG emissions.Many publications noted that this examining of multiple mitigation outcomes is critical, yet understudied [15,40,41].Further, these empirical studies often contained cautionary wording, noting that the measured mitigation outcomes may not adequately characterize the full picture of GHG emissions and carbon sequestration.For example, in a publication empirically assessing nitrous oxide (N 2 O) emissions reduction in rice paddies, authors state that 'the sampling regime utilized in the two studies might not have captured all possible N 2 O emission peaks because of the limited number of days on which field sampling was undertaken' [42].This difficulty in obtaining representative sampling is well-characterized for agricultural GHG emissions, where monitoring can be prohibitively expensive [43] and GHG emissions can pulse in hot spots and hot moments across the landscape [44,45].Inadequate sampling can also be a problem for measuring soil carbon, which often requires extensive (both laterally across a field or landscape, or with depth) and long-term sampling to determine the full impact of agricultural practices on soil carbon sequestration [23,46].
No matter the mitigation outcome being measured, limited sampling can result in high measurement variation, which can reduce or limit the level and appropriateness of statistical significance [47] and result in inaccurate estimates of climate mitigation [46,48].The challenges and limitations of measuring and estimating climate change mitigation outcomes is often noted in empirical studies, but we found these limitations less often referenced in publications that used secondary data or literature to claim mitigation outcomes-which may also contribute to the overstatement of synergistic claims.
For empirical measurement of mitigation outcomes, we also note that the appropriate unit for measuring successful outcomes must also be carefully considered.Measurement of GHG emissions may be characterized by absolute emissions, or emissions intensity (e.g.GHG emissions per unit product, such as a litre of milk).Empirical studies from the livestock sector note substantial tradeoffs between increasing production and increasing overall GHG emissions [49].While reducing emissions intensities may reduce GHG emissions per unit of food produced, total emissions will increase if yields increase.Thus, claims that practices are synergistic when they successfully mitigate emissions by reducing emissions intensity, while at the same time successfully support adaptation by increasing agricultural production may also be overstating synergies.Indeed, all publications reviewed in this study that assessed GHG reductions using emissions intensity paired that mitigation outcome with the adaptation outcome of yield increases.This leaves us with the question: should practices that only reduce emissions intensity be considered to have sufficient mitigation impacts for claiming synergistic effects, or should absolute reductions in GHG emissions be necessary?This is especially relevant in the livestock sector, which had the fewest synergies among the agricultural practices examined and yet as a sector contributes the most to global agricultural GHG emissions [50].

Adaptation and mitigation synergies and tradeoffs
We found that agricultural practices may be overstated in terms of their ability to deliver synergies between climate change adaptation and mitigation outcomes in two main ways: (1) practices had generally high claims of synergies across the publications that examined them, but often did not have robust evidence of adaptation and/or mitigation outcomes or (2) practices had a mix of synergies and conflicting outcomes claimed across the publications, but there was more robust empirical evidence for mixed or conflicting results.
When synergies were claimed without robust evidence (the first case), it was often when no empirical evidence was examined or provided and secondary sources, such as literature citations, were instead used to substantiate claims of adaptation and mitigation synergies.For example, some publications provided very few relevant citations to substantiate claims of synergies resulting from agricultural practice(s).However, other publications did provide thorough literature reviews to examine practices or suites of practices that might provide synergistic adaptation and mitigation outcomes.Claims based on robust literature reviews occurred more frequently in publications that focused on specific practices-such as using agroforestry practices to improve fallow fields and supplement livestock feed [51] and in conservation soil management practices to improve crop productivity and resource use efficiency [52]-though occasionally publications assessing a broader variety of agricultural practices also provided thorough literature reviews [53,54].
In other instances, publications did include empirical evidence, but only for either adaptation or mitigation outcomes-not for both, as discussed in section 4.1.For example [8,55,56] provide excellent evidence for adaptation outcomes for various suits of tillage practices, agroforestry, and crop diversification, but for mitigation, no empirical evidence was provided.While it may be enough to rely on literature citations for carbon sequestration occurring with the implementation of some practices (such as planting trees in cropland), this measurement gap remains problematic.For many practices, the outcomes, especially carbon sequestration or emissions mitigation, are dependent on geography, environmental conditions, and specific agricultural system and practice implementation [57, 58].Thus, it may not be enough to assume, for example, that carbon sequestration is happening regardless of where and how the practice is implemented and only based on literature evidence from a different environmental and management context.Instead, providing either robust empirical evidence of outcomes or relying on literature that matches environmental and management contexts would provide more robust claims of synergistic outcomes.
Finally, although claims of synergies substantiated by literature reviews may be overstated due to a lack of empirical evidence, they do draw attention to potential synergies and can have important and unique vantage points for understanding synergistic practices across multiple studies.For example, Aggarwal et al summarized a comprehensive review of climate-smart agriculture literature and showed that in studies across all regions, there was considerable yield advantage when a suite of practices was used, instead of a single practice implemented in isolation [8].
In the second case, we found that publications that used robust evidence to examine agricultural practices more often found mixed or conflicting outcomes.Practices with the most evidencebased tradeoffs included fertilizer and tillage practices, intermittent drainage in rice, and livestock management practices.In some cases, the practices that mitigated GHGs or sequestered carbon, were potentially bad for adaptation (e.g. the practices were more labour-intensive, or yields were lowered).We found this for both fertilizer and water management in rice [59] and in fodder improvements for livestock potentially encroaching on cropland [10].We also found the opposite tradeoff, where the agricultural practices achieved adaptation benefits, but GHG emissions were higher.This was demonstrated for no-tillage practices in rice and wheat systems in India [40] and in Kenya, where improving livestock feed increased milk production but also increased GHG emissions [60].Additionally, some publications found that practices with mixed or synergistic outcomes in some environmental and management contexts, had negative outcomes for both adaptation and mitigation in others.This was the case for intermittent drainage in rice systems, a practice with mixed results in, for example, in India [40], but implementation in Japan, resulted in lower rice yields and higher N2O emissions compared to conventional continual drainage [61].
For all publications, those that described tradeoffs in detail, were often able to outline many different and useful facets of the potential tradeoffs.For example, Duguma et al describes additional potential tradeoffs for the livestock fodder improvements such as 'the increasing expansion of enclosure-based fodder management system competes with the land available for crop production; there is a poor market access for the products though currently most products seem to be consumed locally; most recent discussions emphasize the expansion of Ngitili [trees on cropping and grazing lands] with limited look at the longterm implications of the expansion, for example, possibilities of woodland invasion which in the long run can enhance carbon sequestration but may limit the livestock feed production' [10].Encouraging more publications to provide these specifics for tradeoff considerations would lead to deeper and more holistic understanding of climate change adaptation and mitigation in agroecosystems.Altogether, a more critical, evidence-based, and comprehensive look at the synergies and tradeoffs resulting from agricultural practices will provide better information to understand and support climate responsive agriculture.

Environmental and management context
The empirical studies we reviewed described the potentially large role of the environmental and management context in affecting adaptation and mitigation outcomes-and thus the capacity of a practice to provide synergistic outcomes.For example, studies that empirically assessed adaptation and mitigation outcomes found that GHG emissions could be higher if some 'climate-smart' agricultural practices, like alley cropping [62] or intermittent drainage [61], were not implemented appropriately e.g. were applied in less favourable environmental or management contexts.Other practices, such as those focusing on agroforestry as a part of reforestation practices, may lead to 'leakages' , shifting deforestation of agricultural land to regions with weaker regulations [16].
Our review also revealed the frequent need of farmers and communities to achieve other critical, but often non-climate change related management goals.Publications that empirically examined both adaptation and mitigation often noted that, while there may be adaptation-mitigation tradeoffs, particularly with higher GHG emissions, other management goals may be more, or just as important, as achieving mitigation outcomes.For example, Dick et al note that 'while the addition of manure did increase N 2 O emissions, our results support Seneviratne (2001) who, after reviewing a range of mitigation strategies for tropical agriculture, concluded that the recycling of organic materials […] was the most feasible, realistic, and immediately applicable option' [63].Thus, while manure application may increase emissions in the tropics more than other practices, it may still be a better choice due to other more holistic management considerations.Similarly, Romasanta et al found that 'the complete removal of stubbles and rice straw residues, which had the highest reduction in GHG emissions, may not be beneficial for soil health in the long term, as it may potentially lead to depletion of other critical nutrients in rice paddies' [64].Climate-relevant agricultural practices need to consider and ultimately be responsive to environmental and management contexts, and may be balancing more immediate, local needs with longer-term global climate objectives [16].

Effect sizes
This review also provides evidence for the need to consider the significance of the effect size of adaptation versus mitigation outcomes.For example, practices with large, positive mitigation outcomes, but small positive adaptation effects may not contribute synergistic effects that are meaningful enough to be worthwhile for farmers and agricultural communities to implement.Practices with large, positive adaptation outcomes, but weak positive mitigation effects may not contribute synergistic effects that are meaningful to national or global climate policy targets.Griscom et al [18] suggests a maximum mitigation potential for natural pathways including agricultural practices, but ended up excluding many of the practices we examined (e.g.no-till agriculture, improved manure management, adaptive grazing) due to lack of data for sufficiently robust global extrapolation [18].
Including effect sizes is an understudied yet important component of both defining a synergy and evaluating the significance of an agricultural practice for climate change adaptation and mitigation objectives.While we did not quantitatively compare effect sizes in our review, we did include a qualitative assessment of adaptation and mitigation outcome effects for six studies across a range of agricultural practices and evidence robustness scores (supplementary table 3).This example shows that publications often report adaptation and mitigation outcomes in different ways.For example, studies that did not empirically assess the outcomes might have simple qualitative assessments of 'significant amounts' of carbon sequestered [16] or they may cite specific estimates, such as with carbon sequestration in Mg of C per hectare per year, and then scale those to global estimations [38].Those studies which empirically measured both outcomes also had a range of reporting, for example from product-related and arearelated GHG emissions changes over a previous nonadaptive baseline given in percents [65] to estimates of the global warming potential per unit of agricultural product yield-including estimates of emissions that could be mitigated due to decreased diesel use for zero tillage practices [52].This makes meaningful comparisons a huge challenge, particularly across agricultural practices and methods of evidence substantiation.We hope this review can lend voice to the need for future studies to improve on the reporting of effect sizes in adaptation and mitigation outcomes for better comparison and synthesis.

Three needs for more robust research
Both the lack of robust evidence, and what the existing evidence reveals, suggests that more caution is needed in claiming that synergistic climate change adaptation and mitigation outcomes are common.
To improve the evaluation of these outcomes, more empirical evidence is needed to substantiate claims of synergistic or 'win-win' outcomes, and to provide robust evidence for all the potential benefits and risks of agricultural practices.While identifying practices with synergistic adaptation and mitigation outcomes is important for meeting agricultural and socioeconomic goals, it is critical to evaluate practices with robust evidence, rather than focusing on the need to identify a synergistic practice.Focusing only on trying to substantiate synergies will ultimately obfuscate a deeper understanding of the relationships among climate change adaptation and mitigation objectives in agriculture.However, simply a call for more robust research for climate-relevant agriculture is insufficient.We conclude with three important needs for future research.
First, our review supports the need for clarity and precision in describing the agricultural practices implemented for delivering adaptation and mitigation outcomes.There is also a need to better consider how to aggregate specific, nuanced practices into broader categories for synthesis.By necessity, we aggregated relatively diverse practices into broad categories (e.g.livestock feed alterations and livestock herd management into livestock practices; supplementary table 2) as disaggregating these categories into individual practices did not allow us to provide a meaningful comparison of outcomes whereas these broader categories did.As suggested by Amadu et al, grouping agricultural practices into relevant aggregated groupings or typologies given specific management and environmental contexts could likely facilitate conceptual clarity and may help increase adoption [66], and we see a need for more explicit consideration on what aggregations are most useful for synthesizing adaptation and mitigation outcomes.We hope this analysis will serve as a call to generate more empirical research on a broader suite of adaptation and mitigation outcomes of agricultural practices and help to guide future, deeper reviews and analyses of adaptation and mitigation outcomes of agricultural practices at a finer scale-with more intentional consideration for conceptual clarity on groupings of practices for improved synthesis of outcomes and adoption.
Second, our review highlights the need to contextualize adaptation and mitigation results.This includes consideration (and clear definition) of the environmental and management context, as practices that may be synergistic in one environment or set of management practices, may not be synergistic in another.This also includes consideration of the effect size and spatial and temporal extent of an agricultural practice's impact on adaptation or mitigation outcomes.In particular, both an absolute effect size (e.g. total emissions reduction/increase) and intensity (e.g.emissions per unit of product) should be reported when relevant, to better inform management decisions.Additionally, as synthesis and comparison of adaptation and mitigation outcomes remains a challenge, future studies should take into consideration how to improve the understand ability and comparability of agricultural practices, the environmental and management contexts, and the adaptation and mitigation outcomes across broad categories of agricultural practices.
Third, our review suggests that truly successful adaptation and mitigation practices will need to meet multiple adaptation and mitigation outcomes, and potentially balance equally important agricultural goals.This likely requires supporting more collaborative, interdisciplinary, and participatory research efforts that consider a broader suite of adaptation and mitigation objectives and outcomes [16].In particular, there is a need to work more closely with farmers and communities directly to better understand their desired objectives and outcomes for climate adaptation beyond agricultural productivity and yield (e.g. if increased food security and nutritional access is a desirable objective, is increasing yield of a singular commodity crop without any other consideration a positive adaptation outcome?).The process of collaboratively identifying such outcomes may reveal previously unconsidered risks and rewards.For example, despite more robust evaluations of agricultural adaptation objectives, the evidence does not necessarily correspond to tangible adaptation outcomes for all within an agricultural community (e.g. may not help more rural communities [37]), and may actually undermine adaptive capacity in the longterm [67].And further, the pursuit of unexamined adaptation objectives may inadvertently perpetuate patterns of injustice [68][69][70][71].It is also possible that for communities that face imminent threats from climate change, the only available options for adaptation may involve mitigation trade-offs, which may still be preferable.Additionally, these practices which could release more GHG emissions (e.g. using manure for fertilizer) might also provide more opportunity for holistic agricultural relationships (e.g.closing the loop of nutrient production and distribution) which may precipitate agricultural system shifts that provide more meaningful mitigation opportunities at scale.
We believe it is critical to reflect carefully on these three needs for producing the meaningful and robust evidence needed for better evaluation of the adaptation and mitigation outcomes of agricultural practices.And in doing so, identifying and supporting interdisciplinary spaces of work and providing participatory engagement is important for the scientific community to help bring forth the research needed for making climate-relevant decisions in agricultural systems.

Conclusion
An increasing number of studies consider climate change adaptation and mitigation in agriculture.While the publications reviewed here claimed adaptation and mitigation synergies more often than tradeoffs, publications with conflicting outcomes were substantiated with more robust evidence.Further, certain climate change adaptation objectives, such as agricultural production, were considered much more frequently than other objectives, which demonstrates a significant and important lack of more comprehensive adaptation objectives from socioeconomic to broader ecosystem objectives.
Our review indicates that more empirical evidence of adaptation and mitigation outcomes is needed, and we further identify three considerations for improving the robustness of this evidence for evaluating adaptation and mitigation synergies: (1) clarity and precision of descriptions of the agricultural practices being examined, the adaptation and mitigation objectives and outcomes being assessed, and the evidence used to assess those outcomes (2) careful contextualization of outcomes with a consideration for synthesis and comparison; and (3) recognizing that successful adaptation or mitigation practices may need to meet-and balance between-multiple goals or outcomes.
This review demonstrates the importance of continuing to develop clear conceptual understandings and robust evidence for synergistic and conflicting adaptation and mitigation outcomes in agricultural systems.In providing a comprehensive evaluation of the evidence for adaptation and mitigation outcomes, we hope to help direct much needed research and institutional support toward work that provides deeper and more meaningful assessments of adaptation and mitigation outcomes of agricultural practices.This work is needed to better support policy and management decisions aimed at meeting climate change adaptation and mitigation goals.

Figure 1 .
Figure 1.(A) The number of publications over the last 30 yr in a Web of Science search that contain the key words 'agriculture' and 'climate' in combination with either 'adaptation' or 'mitigation' or both terms together.Green (A) diamonds and (B) circles represent publications we reviewed in-depth for our study, (B) characterized by geographic region where dot size and shade represent the number of publications focused on that region (larger and darker dots are more publications) and (C) agricultural practices considered.Numbers in parentheses are the total number of publications considering the agricultural practices and the geographic regions.

Figure 2 .
Figure 2. Number of publications claiming synergistic, mixed, and conflicting climate change adaptation and mitigation outcomes broken down by main methodological approach of the publication.Numbers in parentheses are total number of publications represented in each claim (x axis) and methodological approach (y axis).

Figure 3 .
Figure 3.The relationship between synergy-to-conflict ratio and average evidence robustness score, dot size corresponds with the total number of publications considering those practices, and colours represent the main agricultural practices, with subcategories labelled next to each dot (linear regression, p = 0.034, r-squared = 0.23).

Figure 4 .
Figure 4.The frequency of publications considering agricultural productivity or any other objective as the dominant climate change adaptation objective, broken down by main methodological approach (empirical, secondary data, model).Publication claims are in colour: green synergies, blue mixed, and red conflicting.

Figure 5 .
Figure 5. Average evidence robustness scores for adaptation outcomes and mitigation outcomes by agricultural practice.For each adaptation and mitigation outcome, evidence robustness scores were a minimum score of 0 (unsubstantiated) to maximum score of 3 (empirically measured).Across all publications, adaptation outcomes were substantiated with more robust evidence than mitigation outcomes (p = 0.005, t-test).

Table 2 .
Synergy -to-conflict claims ratio for adaptation and mitigation objectives by agricultural practice as broken down by all publications that assessed that agricultural practice and publications with that agricultural practice as their main focus.Green indicates higher synergy-to-conflict claims ratio, while yellow indicates lower synergy-to-conflict claims ratio.Higher ratios indicate more claims of synergies relative to conflicts.Evidence robustness scores have a minimum possible score of 1 and maximum of 10 for each publication, and average evidence robustness scores are presented here from light blue (least robust) to dark blue (most robust) for all publications that assessed that agricultural practice and publications with that agricultural practice as their main focus, broken out by claim.
Insights from Kenya Clim.Change 118 151-65 [61] Kudo Y, Noborio K, Shimoozono N and Kurihara R 2014 The effective water management practice for mitigating greenhouse gas emissions and maintaining rice yield in central Japan Agric.Ecosyst.Environ.186 77-85 [62] Kongsager R 2017 Barriers to the adoption of alley cropping as a climate-smart agriculture practice: lessons from maize cultivation among the Maya in Southern Belize Forests 8 260 [63] Dick J, Kaya B, Soutoura M, Skiba U, Smith R, Niang A and Tabo R 2008 The contribution of agricultural practices to nitrous oxide emissions in semi-arid Mali Soil Use Manage 24 292-301 [64] Romasanta R R et al 2017 How does burning of rice straw affect CH4 and N2O emissions?A comparative experiment of different on-field straw management practices Agric.Ecosyst.Environ.239 143-53 [65] Martin G and Willaume M 2016 A diachronic study of greenhouse gas emissions of French dairy farms according to adaptation pathways Agric.Ecosyst.Environ.221 50-59 [66] Amadu F O, McNamara P E and Miller D C 2020 Understanding the adoption of climate-smart agriculture: a farm-level typology with empirical evidence from southern Malawi World Dev.126 104692 [67] Noble I R et al 2014 Adaptation needs and options Climate Change 2014: Impacts, Adaptation, and Vulnerability.Part A: Global and Sectoral Aspects.Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (Cambridge University Press) pp 833-68 [68] Scoville-Simonds M, Jamali H and Hufty M 2020 The hazards of mainstreaming: climate change adaptation politics in three dimensions World Dev.125 104683 [69] Mbah M, Ajaps S and Molthan-Hill P 2021 A systematic review of the deployment of indigenous knowledge systems towards climate change adaptation in developing world contexts: implications for climate change education Sustainability 13 4811 [70] Eriksen S et al 2021 Adaptation interventions and their effect on vulnerability in developing countries: help, hindrance or irrelevance?World Dev.141 105383 [71] Morgan C, Brevik K, Barbieri L and Ament J 2023 Humans in/of/are nature: re-embedding reality in sustainability sciences Elementa 11 00083 [72] Berry P M, Brown S, Chen M, Kontogianni A, Rowlands O, Simpson G and Skourtos M 2015 Cross-sectoral interactions of adaptation and mitigation measures Clim.Change 128 381-93 [73] Branca G, Lipper L and Sorrentino A 2015 Cost-effectiveness of climate-related agricultural investments in developing countries: a case study New Medit.14 2 (available at: https://link.gale.com/apps/doc/A432894912/AONE?u=anon~883c1daf&sid=google Scholar&xid=1c82bc52) [74] Debaeke P, Pellerin S and Scopel E 2017 Climate-smart cropping systems for temperate and tropical agriculture: mitigation, adaptation and trade-offs Cah.Agric.26 34002 [75] Descheemaeker K, Oosting S J, Homann-Kee Tui S, Masikati P, Falconnier G N and Giller K E 2016 Climate change adaptation and mitigation in smallholder crop-livestock systems in sub-Saharan Africa: a call for integrated impact assessments Reg.Environ.Change 16 2331-43 [76] Di Gregorio M, Nurrochmat D R, Paavola J, Sari I M, Fatorelli L, Pramova E, Locatelli B, Brockhaus M and Kusumadewi S D 2017 Climate policy integration in the land use sector: mitigation, adaptation and sustainable development linkages Environ.Sci.Policy 67 35-43 [77] Jarvis A, Lau C, Cook S, Wollenberg E, Hansen J, Bonilla O and Challinor A 2011 An integrated adaptation and mitigation framework for developing agricultural research: synergies and trade-offs Exp.Agric.47 185-203 [78] La Rovere E L, Avzaradel A C and Monteiro J M G 2009 Potential synergy between adaptation and mitigation strategies: production of vegetable oils and biodiesel in northeastern Brazil Clim.Res.40 233-9 [79] Lobell D B, Baldos U L C and Hertel T W 2013 Climate adaptation as mitigation: the case of agricultural investments Environ.Res.Lett.8 015012 [80] Locatelli B, Fedele G, Fayolle V and Alastair B 2015 Synergies between adaptation and mitigation in climate change finance Int.J. Clim.Chang.Strateg.Manag.8 112-28 [81] McHenry M P 2009 Agricultural bio-char production, renewable energy generation and farm carbon sequestration in Western Australia Agric.Ecosyst.Environ.129 1-7 [82] Campbell B M, Thornton P, Zougmoré R, van Asten P and Lipper L 2014 Sustainable intensification: what is its role in climate smart agriculture?Curr.Opin.Environ.Sustain.8 39-43 [83] Deng A, Chen C, Feng J, Chen J and Zhang W 2017 Cropping system innovation for coping with climatic warming in China Crop J. 5 136-50 [84] Harvey C A et al 2014 Climate-smart landscapes: opportunities and challenges for integrating adaptation and mitigation in tropical agriculture Conserv.Lett.7 77-90 [85] Notenbaert A, Pfeifer C, Silvestri S and Herrero M 2017 Targeting, out-scaling and prioritising climate-smart interventions in agricultural systems: lessons from applying a generic framework to the livestock sector in sub-Saharan Africa Agric.Syst.151 153-62 [86] Sharma R, Chauhan S K and Tripathi A M 2016 Carbon sequestration potential in agroforestry system in India: an analysis for carbon project Agrofor.Syst.90 631-44