Human and planetary health are inextricably linked, and both are jeopardized by the global food system. Dominant agricultural development policies, in alignment with UN Sustainable Development Goal 2 (Zero Hunger), favor strategies to increase production of foods with caloric quantity over nutritional quality. These Green Revolutionera approaches are not conducive to securing healthy diets for a growing global population. Currently, poor diets are the largest contributors to global morbidity, with one in three people worldwide experiencing a form of malnutrition. The resource requirements of this production-centric paradigm also contribute to environmental degradation, including deforestation, biodiversity loss, soil erosion, and freshwater contamination. Though the consequences of landscape and dietary simplification are global phenomena, the bulk of the burden is felt by communities in low-and middle-income countries (LMICs). These inequities have only been exacerbated by the recent COVID-19 pandemic. With the same regions disproportionately experiencing the impacts of climate change, there is an urgent need for food system transformations that support human and ecological health in tandem.

Forests and trees have a key role to play in this dual agenda. The environmental benefits from retaining trees on farms and in forests are well documented: agroforestry and reforestation have been cited as highly effective climate change adaptation and mitigation measures. Tree-based landscapes can also provide micronutrient-rich fruits and vegetables that are undersupplied and under-consumed, with recent research finding positive associations between tree cover and more diverse and nutritious diets. There is increasing recognition of how forests can contribute to dietary quality through provision of wild foods, sources of income, and ecosystem services which can boost agricultural production. However, few studies examining forest-diet linkages expand this scope to include trees outside of forests, despite the widespread prevalence of trees in agricultural landscapes. Additionally, the studies that successfully draw these links are often restricted by the use of broad indicators of dietary quality. As such, they do not indicate the degree to which trees can play a role in increasing the intake of key micronutrients vital to human health.

The aim of this thesis is to tease apart the pathways by which forests and tree-based landscapes can improve people’s diets, with an empirical emphasis on the contributions of on-farm trees. The featured research elucidates and evaluates the relative importance of mechanisms linking trees and diets through a mixed-methods, case study approach. The research is based on data collected over two rounds of extensive fieldwork in rural Malawi, a densely populated country with high rates of poverty and malnutrition. The primarily rural population has a strong dependence on subsistence agriculture and wood-sourced fuel, which has contributed to widespread deforestation in recent decades. This confluence of natural resource reliance, poverty, and malnutrition justifies Malawi as a case study country to examine the links between tree cover and dietary quality.

This thesis is highly interdisciplinary, integrating methods from botany, forestry, agronomy, and nutrition to investigate tree-diet linkages in a specific context. Through cross-sectoral collaboration, this research advances existing knowledge by a) synergistically addressing the role of forest and non-forest trees in tandem, and b) going beyond simple dietary quality metrics by assessing micronutrient adequacy levels for each respondent and the source of each food item consumed. The analysis primarily draws from a novel dataset for 460 households combining 1) dietary data from detailed 24-hr dietary recalls surveys (conducted twice within one week) in both the dry and wet seasons, 2) data on socioeconomic dimensions, forest use, and agricultural land use, and 3) tree cover estimates from high resolution 2019 PlanetScope imagery (3m resolution). This data is supplemented by insights from participatory observation, focus groups, and an agricultural field survey conducted with a subset of 60 households.

The research findings illustrate how trees from forests to farms can benefit Malawian diets across seasons. Quasiexperimental methods reveal that women living in areas with greater tree cover have higher vitamin A and iron adequacy in the dry season, and higher zinc, vitamin A, iron, and folate adequacy in the wet season. For example, an increase from zero tree cover to the maximum recorded value (62% - equating to approximately 195 ha in the 1km radius buffer) around a household can increase micronutrient adequacy by 10-16% in the dry season and 1026% in the wet season. Our results show that specifically, women from households that directly source food from on-farm trees have 8-15% higher levels of zinc, vitamin A, iron and folate adequacy in the dry season, and 6-12% higher levels of zinc, vitamin A and folate adequacy in the wet season compared to women from households that do not source food from on-farm trees. Critically, our analysis illustrates that having food trees on farm can help explain why tree cover (more broadly) supports dietary quality. In other words, food trees on farms mediate the relationship between tree cover in the landscape and micronutrient adequacy.

On-farm food trees influenced food consumption patterns as well. For example, we saw that in the dry season, households with on-farm food trees consumed on average 18% more from their own farms and purchased 19%

less food from the market, with seemingly no dependence on wild foods to supplement food consumption. This indicates that on-farm tree systems may influence how much a household consumes from their own agricultural production during certain times of the year. While households do not gain additional dietary benefits by using trees on farms for fuel and/or income, multi-purpose on-farm trees do not detract from dietary quality and have the potential to offer other livelihood benefits. By identifying associations between different types of on-farm trees and food consumption patterns, these results can help elucidate how different on-farm trees factor into subsistence or commercially-oriented production systems.

Most Malawians cannot afford to purchase the foods necessary to achieve a healthy diet on a daily basis. These findings, by highlighting the specific nutritional benefits of retaining trees in agricultural landscapes for direct provision, identify an avenue by which households can improve their supply of nutrient-rich foods regardless of income level. While a household can choose to source food from their own on-farm trees, their ability to source foods of similar micronutrient value from the market is dependent on larger, underlying structural aspects of the enabling economic, social, and political environment. Nutrition-oriented initiatives must accordingly address treediet linkages at the farm, regional, and national scales - with the objective of making nutritious foods available, accessible, and affordable to households through monetary and non-monetary means. Optimizing a portfolio of onfarm food trees with staggered harvest periods is recommended to ensure year-round availability of micronutrientrich foods on farms. Infrastructure investments are also necessary to improve trading networks for tree-based products, thereby increasing the availability of nutritious foods in rural markets. Policy reforms are needed to rebalance agricultural subsidies toward nutritious foods by integrating investments in tree-based farming into social protection systems. Through promoting trees in agrarian landscapes, these actions can synergistically address environmental and nutrition objectives for the well-being of humans and nature.