Publications

In applications across agriculture, ecology, and human development, machine learning with satellite imagery (SatML) is limited by the sparsity of labeled training data. While satellite data cover the globe, labeled training datasets for SatML are often small, spatially clustered, and collected for other purposes (e.g., administrative surveys or field measurements). Despite the pervasiveness of this issue in practice, past SatML research has largely focused on new model architectures and training algorithms to handle scarce training data, rather than modeling data conditions directly. This leaves scientists and policymakers who wish to use SatML for large-scale monitoring uncertain about whether and how to collect additional data to maximize performance. Here, we present the first problem formulation for the optimization of spatial training data in the presence of heterogeneous data collection costs and realistic budget constraints, as well as novel methods for addressing this problem. In experiments simulating different problem settings across three continents and four tasks, our strategies reveal substantial gains from sample optimization. Further experiments delineate settings for which optimized sampling is particularly effective. The problem formulation and methods we introduce are designed to generalize across application domains for SatML; we put special emphasis on a specific problem setting where our coauthors can immediately use our findings to augment clustered agricultural surveys for SatML monitoring in Togo.

Understanding food demand patterns is critical for developing countries in Sub-Saharan Africa (SSA), which are facing increasing urbanization, intensifying climate change, and rapid transitions in food production systems. This study models food demand in Zambia and Tanzania using data from the Living Conditions Monitoring Survey (LCMS) (2015 and 2022) and the Living Standards Measurement Survey (LSMS) (2015 and 2021). This study applies the Quadratic Almost Ideal Demand System (QUAIDS) model to analyze demand behavior and calculate expenditure and price elasticities of both food and nutrition. Urbanization is linked to significant shifts in dietary preferences and greater sensitivity to the prices of non-staple, nutrient-dense foods, reflecting the diversifying food demand in urbanized regions. Households allocate more resources to calories with increases in income, but the rate of increase slows as basic caloric needs are met. Conversely, expenditure nutrition elasticity for Vitamin B12 is particularly high, as rising incomes enable households to spend more on animal-based foods such as meat, fish, eggs, and dairy. Once basic caloric needs are satisfied, additional income is more likely to be directed toward diversifying the diet with higher-value, nutrient-dense foods rather than increasing carbohydrate consumption. The positive expenditure elasticities for both food consumption and nutrient intake in both countries emphasize that increasing household income remains a crucial strategy for improving food security.

Large-scale deployment of wind and solar power is essential for climate change mitigation but requires substantial land and offshore area due to their relatively low installed capacity densities. However, wide variation in installed capacity density estimates creates major uncertainty in projecting future spatial requirements. Here, we apply multiple spatial methods to satellite imagery to estimate installed capacity densities for 60 onshore wind, 24 offshore wind, and 54 utility-scale solar PV plants (including 36 fixed-tilt and 18 single-axis tracking) across six countries: Australia, China, Germany, India, United Kingdom, and the United States. For wind power plants, mean estimates of installed capacity density using the 8D buffer method (eight times the rotor diameter), the convex hull method with a 5D buffer, and the median Voronoi polygon (bounded by an 8D buffer) method are similar (4.0, 4.4, and 4.9 MW km⁻² for onshore wind and 4.6, 5.2, and 5.4 MW km⁻² for offshore wind, respectively) but the 5D buffer method provides a relatively larger estimate because of unaccounted areas within power plant boundaries. Single-axis tracking solar PV power plants have an approximately 33% lower mean installed capacity density than fixed-tilt power plants, with estimates of 36 and 29 MW km⁻² (without and with a 50 m buffer, respectively), compared to 54 and 43 MW km⁻² for fixed-tilt systems. Estimates vary more within countries and across methods than between countries, underscoring the importance of methods and assumptions in estimating spatial requirements for large-scale wind and solar deployment.

Rigorous evidence detailing the myriad interconnections between humans and ecosystems will be critical to slow the loss of biodiversity. Effective conservation interventions will depend upon a detailed understanding of the benefits that biodiversity provides to people, the ways that human activities drive biodiversity decline, and the potential for conservation policies to stem this decline. Although hundreds of papers explore these relationships, a careful review of this literature shows that the overwhelming majority of studies fall short of documenting causal relationships with biodiversity per se. However, a combination of data and methodological advances has led to rapid growth of quasi-experimental analyses that are advancing our understanding of human interactions with biodiversity in three domains. First, economists have provided valuable insights into the causes of biodiversity loss, which complement traditional ecological experiments by better reflecting real-world conditions. Second, quasi-experimental studies have begun to identify which policy interventions, in what contexts, have slowed biodiversity loss. Finally, recent quasi-experimental studies have shown that the loss of species can impose extremely large costs on humanity but that these costs vary widely depending upon the species and opportunities for human adaptation.

Climate change and anthropogenic activities threaten biodiversity and ecosystem services. Climate-smart conservation plans address these challenges by ensuring protection of some climate-resilient areas. However, integrating climate change in the design of conservation plans is often deemed too expensive, as it may require larger networks or protecting more costly sites from a conservation perspective. Using mangroves as a case study, we evaluate the efficiency of protecting mangroves in climate-smart versus climate-naïve reserve networks. We find that climate-smart conservation plans could provide sizable benefits (13.3%) for relatively moderate increases in protected area (+7.3%). Moreover, transboundary plans, involving cooperation among countries, require less area and protect more climate-resilient mangroves than nation-by-nation plans. Implementing these strategies would improve the current protected area network for mangroves, which currently has poor climate resilience. Our methodology could potentially be tested on other ecosystems, assuming sufficient information exists regarding their distribution, biodiversity, and resilience to climate change.

Forests are central to climate, biodiversity, and development goals, but effective monitoring and evaluation of their contributions depends on reliable data. Satellite-derived global forest cover and change datasets (GFDs) are widely used to address this need. However, differences in resolution, forest definition, and methodology challenge their use in research and policy, yet how GFD differences affect key forest-related estimates remains poorly understood. Here, we quantify global area-based spatial congruence among 10 GFDs and test their influence on three national policy-relevant estimates: carbon accounting, forest-poverty mapping, and biodiversity habitat. We find only 26% spatial congruence among GFDs at native resolution. This low congruence translates to an order-of-magnitude difference in national case study indicator estimates. We demonstrate that GFD selection fundamentally shapes monitoring and evaluation outcomes, particularly in biomes with fragmented or sparse tree cover. We provide a decision-support framework to guide GFD selection according to different forest-related science and policy applications.

Heat extremes pose substantial health risks during pregnancy and early childhood. High humidity exacerbates heat strain, but its long-term effects on health remain poorly understood. We compare the effect of prenatal exposure to extreme humid heat versus heat alone on child growth in South Asia, where high rates of child stunting meet rapidly accelerating hot-humid extremes. After adjusting for sociodemographic, seasonal, and spatial confounders, we use within-community variation in children’s ages to isolate the impact of prenatal exposures. We find that hot-humid exposures are much more detrimental to health than hot temperatures alone, with the potential to increase stunting in South Asia by over 3 million children by 2050. These findings underscore the importance of accounting for humidity when estimating and localizing climate change impacts.

The world relies on analyses by the United Nations-facilitated Integrated Food Security Phase Classification (IPC) to identify where populations are food insecure and to quantify the severity of these crises. IPC sub-national analyses are designed to be comparable over space and time in the more than 30 countries in which the IPC operates. Humanitarian agencies appear to regard these findings as authoritative and comparable, and as of 2024, used IPC analyses to guide more than six billion dollars of annual aid allocations. We study the consistency and comparability of IPC food insecurity analyses across time and space. Drawing on 1,849 IPC subnational analyses covering 742 million people from fifteen countries between 2019 and 2023, we show that IPC analyses face significant challenges related to data availability and food security measurement, resulting from underlying food security indicators that are often discordant. We find that the vast majority of IPC subnational analyses are consistent with IPC technical guidance, but that this guidance permits a wide range of classifications for a given set of food security indicators. We also find evidence that IPC subnational analyses vary in the way they use food security data, often weighing food security indicators differently across locations. While variation in how analyses use food security indicators can plausibly reflect varying contextual factors across countries, we find evidence that analyses weight indicators differently across time for the same location. Finally, we show that analyses do not treat closely correlated food security indicators as substitutes, suggesting inconsistency in the treatment of food security indicators across analyses. We discuss implications of these findings for policy and for the interpretation and use of IPC analyses by researchers and policymakers.

The Integrated Food Security Phase Classification (IPC) system is the official global method for classifying food insecurity. As of 2023, international agencies and governments use IPC analyses to allocate more than US$6 billion of humanitarian assistance annually. Here we evaluate data from approximately 1 billion people in more than 10,000 IPC subnational analyses conducted between 2017 and 2023. We find that IPC estimates understate the extent and severity of crises. Our primary estimates indicate that IPC subnational analyses underestimate the number of acutely hungry people in the world, missing approximately one in five. We find evidence of under-classification around the IPC threshold that determines whether an area is classified as ‘stressed’ or ‘in crisis’—a threshold meant to trigger deployment of humanitarian resources. Contrary to widely held assumptions, our findings suggest that IPC analyses are conservative; the prevalence and severity of acute hunger is probably considerably higher than global estimates indicate.

Aquatic food security is closely interconnected with multiple sustainable development goals (SDGs). Although assessing aquatic food security relies on understanding global trends in per capita production and consumption, there has been no comprehensive index to evaluate these trends in a country or regional context. Here, we develop a novel framework based on a comprehensive scoring system to assess changes in contemporary per capita aquatic food production and consumption trends (tendency, magnitude, and stability) across 177 countries in two time periods (1961–1990 and 1991–2019). Globally, 58.2 % of countries scored positive in production trends, and 57.6 % in consumption trends from 1961 to 1990. However, between 1991 and 2019, 57.1 % of the countries achieved negative production trend scores, while 68.4 % of countries maintained positive consumption trend scores, accompanied by greater stability in the trends. This significantly widened the positive gap between consumption and production trend scores, highlighting a growing mismatch between global consumption and production patterns. Meanwhile, aquaculture exhibited significantly higher trend scores than capture fisheries, accompanied by rapid global trade growth. Our findings indicate that the synergy between aquaculture and trade plays a crucial role in sustaining growth and enhancing the stability of aquatic food consumption worldwide.

Green hydrogen has the potential to address two critical challenges in a zero-carbon energy system: balancing seasonal variability of solar and wind in the electricity sector, and replacing fossil fuels in hard-to-electrify sectors. In this study, focusing on China, we deploy a provincial-scale energy system planning and operation model to examine the technical and cost-optimal potential of green hydrogen to fully remove carbon-based fuels in the electricity and hard-to-electrify sectors by 2050. Our results show that green hydrogen infrastructure can enable more cost-effective decarbonization of both the electricity and hard-to-electrify sectors. First, in the zero-carbon electricity sector alone, utilizing green hydrogen as long-duration storage enables a 17% reduction in the electricity-only cost (ZE scenario) relative to one without hydrogen. However, cost savings hinge on the availability of underground hydrogen storage. Second, coupling the electricity and hard-to-electrify sectors by sharing green hydrogen infrastructure reduces the combined energy system cost by 6% compared to a decoupled energy system. Third, the coupled energy system also makes green hydrogen comparable to fossil fuel-based gray and blue hydrogen costs in China. Allocating the entire savings realized in the coupled energy system to just the hydrogen used as fuel/feedstock in hard-to-electrify sectors yields a 24% reduction in the hydrogen-only cost relative to the cost under the decoupled system. Last, coupling hydrogen infrastructure between electricity and hard-to-electrify sectors yields a substantially different spatial pattern of hydrogen production. In the decoupled energy system, 80% of hydrogen demand in electricity and hard-to-electrify sectors is produced locally within the same provinces, but the coupled energy system cuts local production to 30%, shifting production to high renewable energy generating provinces. Understanding the spatial patterns of optimal hydrogen infrastructure siting will help plan an integrated electricity and hydrogen system that can cost-effectively decarbonize multiple sectors and China’s broader economy.

Recent innovations in task-agnostic imagery featurization have lowered the computational costs of using machine learning to predict ground conditions from satellite imagery. These methods hold particular promise for the development of imagery-based monitoring systems in low-income regions, where data and computational resources can be limited. However, these relatively simple prediction pipelines have not been evaluated in developing-country contexts over time, limiting our understanding of their performance in practice. Here, we compute task-agnostic random convolutional features from satellite imagery and use linear ridge regression models to predict maize yields over space and time in Zambia, a country prone to severe droughts and crop failure. Leveraging Landsat and Sentinel 2 satellite constellations, in combination with district-level yield data, our model explains 83% of the out-of-sample maize yield variation from 2016 to 2021, slightly outperforming a model trained on Normalized Difference Vegetation Index (NDVI) features, a common remote sensing approach used by practitioners to monitor crop health. Our approach maintains an 𝑅2 score of 0.74 when predicting temporal variation alone, while the performance of the NDVI-based approach drops to an 𝑅2 of 0.39. Our findings imply that this task-agnostic featurization can be used to predict spatial and temporal variation in agricultural outcomes, even in contexts with limited ground truth data. More broadly, these results point to imagery-based monitoring as a promising tool for assisting agricultural planning and food security, even in contexts where computationally expensive methodologies remain out of reach.

Strategically planning the phase-out of coal power is critical to achieve climate targets, yet current approaches often fail to account for the context-specific barriers and vulnerabilities to retirement. Here we introduce a framework that combines graph theory and topological data analysis to classify the US coal fleet into eight distinct groups based on technical, economic, environmental and socio-political characteristics. We calculate each non-retiring coal plant’s ‘contextual retirement vulnerability’ score, a metric developed to quantify susceptibility to retirement drivers using the graph-based distance to a coal plant with an announced early retirement. Separately, we identify ‘retirement archetypes’ that explain the key factors driving announced retirements within each group, which are used to inform group-specific strategies for accelerating retirements. Our findings reveal the diverse strategies that are required to accelerate the phase-out of remaining coal plants, including regulatory compliance, public health campaigns and economic incentives.

As climate change makes agricultural production shocks more frequent and severe, it is vital to understand their effect on farmer welfare, land use, and deforestation. Theoretically, a change in agricultural productivity could increase or decrease deforestation by changing demand for agricultural land and/or through the consumption of forests as a coping strategy. This paper uses the introduction of a crop pest to sub-Saharan Africa to estimate the effect of a negative agricultural productivity shock on deforestation. Using primary household data, we first find that farmers who were exposed to higher levels of fall armyworm saw substantial decreases in yield and food security. Using estimates of fall armyworm suitability in conjunction with machine-learning models of maize yields and deforestation, we find that the introduction of the fall armyworm induced a doubling of the deforestation rate in Zambia in the 3 y following the outbreak. This increase was driven both by increased agricultural land expansion and increased charcoal production as a coping strategy. These responses vary substantially over space. More remote areas experienced 23% lower FAW-induced deforestation compared with the sample average, suggesting that farmers with access to maize and charcoal markets may have increased deforestation as a response. Wealthier areas were also less likely to deforest in response to FAW pressure. In sum, our results suggest that negative agricultural productivity shocks may lead to a negative climate feedback, with farmers engaging in emissions-increasing strategies in response.

Over the last decade, attribution science has shown that climate change is responsible for substantial death, disability and illness. However, health impact attribution studies have focused disproportionately on populations in high-income countries, and have mostly quantified the health outcomes of heat and extreme weather. A clearer picture of the global burden of climate change could encourage policymakers to treat the climate crisis like a public health emergency.

Understanding the spawning strategies of large pelagic fish could provide insights into their underlying evolutionary drivers, but large-scale information on spawning remains limited. Here we leverage a near-global larval dataset of 15 large pelagic fish taxa to develop habitat suitability models and use these as a proxy for spawning grounds. Our analysis reveals considerable consistency in spawning in time and space, with 10 taxa spawning in spring/summer and 9 taxa spawning off Northwest Australia. Considering the vast ocean expanse available for spawning, these results suggest that the evolutionary benefits of co-locating spawning in terms of advantageous larval conditions outweigh the benefits of segregated spawning in terms of reduced competition and lower larval predation. Further, tropical species spawn over broad areas throughout the year, whereas more subtropical and temperate species spawn in more restricted areas and seasons. These insights into the spawning strategies of large pelagic fish could inform marine management, including through fisheries measures to protect spawners and through the placement of marine protected areas.

Sustainability requirements imposed on agricultural producers by downstream supply chain actors risk excluding smallholder farmers from upgraded markets. Here we investigated smallholder participation in sustainably certified palm oil mill markets in Indonesia. We developed and applied a conceptual model to estimate the importance of structural market access, smallholder capacity, and buyer/seller behavior in shaping mill smallholder sourcing. Smallholders who hold exclusive contracts with specific mills were overrepresented at certified mills. Independent smallholders unaffiliated with mills contributed one-third of regional oil palm production but 7% of certified mill supply. We found no evidence that independent smallholders exited markets after mill certification (“active” exclusion). Instead, only 36% of certified mills ever purchased from independent smallholders, and independent smallholder lands were less common around certified (29–38% of palm area) versus noncertified (41–42%) mills. To address such “passive” exclusion, supply chain governance programs should encourage participation of actors well-positioned to source from small-scale producers.

Recent public and private policies seek to end deforestation by regulating the production and trade of forest-risk commodities. The design, implementation, and evaluation of these policies rely on metrics that are typically bounded in scope by either territories or supply chains, and therefore only provide a partial account of deforestation on the ground. We argue that metrics linking deforestation and forest degradation to commodity production need to consider two distinct questions: (1) How much of today’s commodity production is associated with past deforestation? and (2) to what extent is today’s deforestation driven by the prospects of producing a specific commodity in the future? This paper describes how metrics can respond to these questions by being classified according to their commodity or deforestation focus. We propose common terminology to facilitate the communication and use of these perspectives and metrics. We then make the case for combining perspectives through the monitoring and reporting of multiple metrics by governments, companies, and non-governmental organizations alike to both assess progress and drive more coordinated action to reduce deforestation.

As climate change increasingly affects biodiversity and ecosystem services, a key challenge in ecology is accurate attribution of these impacts. Though experimental studies have greatly advanced our understanding of climate change effects, experimental results are difficult to generalise to real-world scenarios. To better capture realised impacts, ecologists can use observational data. Disentangling cause and effect using observational data, however, requires careful research design. Here we describe advances in causal inference that can improve climate change attribution in observational settings. Our framework includes five steps: (1) describe the theoretical foundation, (2) choose appropriate observational datasets, (3) estimate the causal relationships of interest, (4) simulate a counterfactual scenario and (5) evaluate results and assumptions using robustness checks. We demonstrate this framework using a pinyon pine case study in North America, and we conclude with a discussion of frontiers in climate change attribution. Our aim is to provide an accessible foundation for applying observational causal inference to estimate climate change effects on ecological systems.

Aquaculture interventions and policies are now fundamental in sustainability agendas, particularly in supporting small-scale fisheries and coastal communities. These policies often rely on the “blue transitions” theory of change, which posits that an expansion of aquaculture will aid in recovering declining fish stocks and enhancing livelihoods. However, the blue transitions theory is relatively new, leaving many aspects uncertain, especially regarding how transition stages unfold and impact communities as they are expected to transform livelihoods. Frequently, these policies adopt a top-down approach driven by political and corporate interests at global or national levels, emphasizing environmental and economic benefits while neglecting local social, cultural, and historical contexts. This study aims to identify gaps in current blue transition policies at the local level through two empirical case studies in Baja California Sur, Mexico. Additionally, it evaluates the suitability of existing frameworks for incorporating justice in food system transitions for seafood system transitions and provides insights for developing more equitable blue food policies. Using an exploratory mixed methods approach from 2021 to 2023, including ethnography, interviews, surveys, and focus groups, this research delves into the complexities of aquaculture policies for communities going through blue transitions. Findings indicate that these policies often prioritize economic development over social, cultural, and historical considerations, leading to injustices within communities. The case studies reveal impacts and challenges such as intra-community conflict, illegal fishing, and threats to food security and resilience, as well as benefits like momentary economic gains. Applying a framework for just food system transitions, we advocate for flexible, community-centric policies that recognize local heterogeneity and empower communities to shape their transitions, including deciding whether a transition is appropriate. This study underscores the limitations of viewing aquaculture as a panacea for small-scale fisheries’ challenges, emphasizing the need for holistic, multiscale management approaches. Contextualizing blue transitions within local realities and prioritizing food justice can promote just and equitable outcomes that address the nuanced needs of diverse coastal communities amidst global pressures.