Quantifying Agrifood Production and Consumption at the U.S. Metropolitan and County Level to Inform Urban Food System Interventions

Abstract

There are persistent data gaps and a lack of harmonization among existing data on the U.S. Food System, especially at the subnational level, even around basic data relating to agrifood production, demand, and loss, and their associated environmental impacts. With rising policy focus on the food system (e.g., the Milan Food Policy Pact), particularly a focus on local agriculture as a lever to effect positive societal change, subnational data become important to model the effects of local food system related initiatives. With an aim of informing localization policies, the overarching goal of this thesis is to quantify agrifood production and consumption at the U.S. Metropolitan and County Level to inform urban food systems. The thesis contributes methods to synthesize relevant publicly available data on agrifood production, demand (consumptive use and consumption), and losses, and utilizes the aggregated, harmonized data to provide subnational metrics on the food system in a baseline case and in future food localization scenarios. The work is implemented in three steps with the following specific objectives: • Defining of concept of current local capacity, and demonstrating measurement for select food items across U.S. metro areas (Chapter 2) • Developing a county-level database of production, demand, and current local capacity for all 95 agrifoods in the American Diet (Chapter 3) • Modeling future scenarios of urban agricultural localization and their environmental impacts across 377 MSAs (Chapter 4) In total, the methods and data outlined in this thesis (Chapters 2 and 3) link agrifood production to dietary demand for all major agrifoods produced in the US (95 agrifoods), detailed at the metro- and county-level for all counties in the conterminous U.S. (3,109 counties), covering all food items reported to be consumed by Americans (12,070 food items), incorporating all relevant food losses (e.g. waste and weight losses along the supply change), in a manner that’s internally consistent and consistent with national data at both ends of the supply chain (i.e. harmonized), while delineating fresh agrifood demand from processed agrifood demand. These data are leveraged to compute agrifood localization metrics and associated environmental impact metrics. To measure agrifood localization, we define and quantify the current local capacity of U.S. counties and metro areas to satisfy their current local agrifood demand with current local production, or agrifood current local capacity (CLC) for short (Chapters 2 and 3). The CLC is a heuristic for the localization potential of the current local food system which may be of use to scholars and local policymakers, which can inform the analysis and discussions around food localization (i.e., increasing agricultural production in areas closer to the point of demand, often in the context of increasing urban agriculture), addressing both potential positive and negative environmental impacts, such as water withdrawal and fertilizer application (Chapter 4). Key Contributions. This dissertation elucidates and fills in key data gaps in the American food system and provides harmonized agrifood production, demand, loss, and impact data. Previous works have harmonized parts or all of the food system, such as the Loss-adjusted Food Dataset and the Complete Diet Model (Peters et al., 2016), but no previous dataset, to the best of my knowledge, lists the subnational production, demand, and associated loss of individual agrifoods with whole diet coverage. The methods outlined also are purposefully designed to leverage publicly available, well-stewarded datasets when available (e.g., U.S. Agricultural Census, thus providing methods that avoid costly inputs (e.g., high-quality satellite imagery, Input-Output economic data) and allowing for easy historical comparisons and future iterations on the data.