Visualizing politics: A feminist political ecology and participatory GIS approach to understanding smallholder farming, climate change vulnerability, and seed bank failures in Northern Ghana

Over the past three decades, seeds banks have emerged as a major strategy for building seed systems resilience to climate change. Often initiated and funded by non-governmental organizations, seeds banks have grown prolifically, but questions remain concerning their long-term sustainability. Despite their precipitous rise, and effectives during initial years, many seed banks cut back on activities or stop altogether once external NGO funding is withdrawn. This rise and fall of seed banks raise three questions worthy of examination: (1) What factors shape the sustainability of community seed banks? (2) Do community seed banks function as they are designed to be? (3) How well do seed banks target farmers based upon true underlying need? Drawing upon insights from feminist political ecology (FPE) and Participatory Geographic Information Systems (PGIS), this paper examines these questions with data collected from drought-prone northern Ghana. The empirical research includes participant-observations; 144 in-depth interviews; participatory geospatial data analysis; gender-disaggregated data validation workshops; and analyses of seed bank inventory, lending, and payment records. Contrary to previous work emphasizing the role of climate variability and crop failure, this paper highlights the centrality of rural politics in the sustainability of seed banks. Specifically, the findings show that the lack of respect for indigenous knowledge, pseudo seed borrowing, and local elite capture, all work together to undermine the sustainability of seed banks. When seed banks do not meet farmers' needs, the paper also demonstrates how farmers covertly resist such projects. Finally, the paper shows how through a repertoire of gender politics, village men undermine seed banks' vision of ensuring equitable and democratic access to seeds. Overall, the paper contributes to existing research by demonstrating how FPE and PGIS could be used in parallel to permit a more rigorous testing of claims of village and gender politics on the ground.