Members of Duke University's Plastic Pollution Working Group examine the unequal distribution of benefits and burdens of plastics. They find the benefits of plastics to communities and stakeholders are principally economic, whereas their burdens fall largely on human health. The report stresses the need for policy design to include health burdens to all impacted stakeholders across all plastic life stages and urges the Intergovernmental Negotiating Committee to consider harmful effects across the entire plastic lifecycle and to apply the precautionary principle when drafting the upcoming international global plastic treaty.

There is evidence that transformers offer state-of-the-art recognition performance on tasks involving overhead imagery (e.g., satellite imagery). However, it is difficult to make unbiased empirical comparisons between competing deep learning models, making it unclear whether, and to what extent, transformer-based models are beneficial. In this paper we systematically compare the impact of adding transformer structures into state-of-the-art segmentation models for overhead imagery.

Aquatic foods are critical for food and nutrition security in Malawi, but it is unclear which populations benefit from different aquatic foods and what factors shape food access. Spatial analysis of food flows across value chains from Lake Malawi to domestic consumers shows that usipa (Engraulicypris sardella) reaches more consumers than chambo (Oreochromis karongae) across all Malawi districts. Spatial analysis of food flows can guide policy makers toward supporting fisheries that reach vulnerable populations and designing interventions that enhance physical and economic access to fish.

This analysis of coastal habitat policy in six US states—California, Florida, Massachusetts, North Carolina, Texas, and Washington—aims to identify promising policy approaches for improved protection and restoration of oyster reefs, mangroves, salt marshes, and seagrass.

Effective integration planning for small, distributed solar photovoltaic (PV) arrays into electric power grids requires access to high quality data: the location and power capacity of individual solar PV arrays. Unfortunately, national databases of small-scale solar PV do not exist; those that do are limited in their spatial resolution, typically aggregated up to state or national levels. While several promising approaches for solar PV detection have been published, strategies for evaluating the performance of these models are often highly heterogeneous from study to study.

Using a simulation tool, the authors investigate the effects created by a US state defecting from the wholesale electricity market in an organized electric grid on the states that remain in the coalition. The report finds that if a net-importing state defects, the remaining states’ producers are worse off and the remaining states’ consumers are better off. The opposite effect takes hold if the defecting state is a net-exporter. Furthermore, the authors find evidence that defection impacts the remaining states’ climate initiatives.

Though many challenges impede low- and middle-income countries’ access to clean cooking energy, cost barriers are perhaps most significant. This report discusses the role of subsidy and tax policies—levied on both the supply and demand side of this market—in affecting progress toward universal access to clean cooking. Moreover, we show that a “fear of spoiling the market” with such incentives finds little empirical support in the literature. This report offers recommendations to policy makers, in additional to a case study on clean cooking transitions in Nepal.

In a study published by the journal One Earth, Duke experts share findings from an examination of the types of commitments that corporations have made to address global plastic pollution. The authors find that, rather than tackle virgin plastics, most companies target packaging and general plastics and frequently emphasize recycling-related efforts. While many large and important companies are making commitments, significantly more efforts beyond plastic recycling are required to effectively address plastic pollution challenges.

What will it take to achieve a net-zero carbon emissions footprint for the US economy by 2050? This report from Energy Pathways USA helps strengthen the evidence base on what will be required for a robust US energy transition and elucidates key barriers and opportunities for reaching net-zero goals. The report also outlines future areas of focus for Energy Pathways USA, a Duke-based project that accelerates progress towards a net-zero carbon future by developing workable solutions with corporate partners across multiple key industries.

Transfer learning has been shown to be an effective method for achieving high-performance models when applying deep learning to remote sensing data. Recent research has demonstrated that representations learned through self-supervision transfer better than representations learned on supervised classification tasks. However, little research has focused explicitly on applying self-supervised encoders to remote sensing tasks.