Conference Agenda

Agroforestry is a regenerative farming practice which can improve ecosystem health, soil productivity, and the socio-economic well-being of smallholder farmers facing growing pressures from climate change effects. Many studies explore factors associated with agroforestry adoption, but how to successfully increase adoption rates remains less understood. Encouraging and enabling farmers to plant more trees or more farmers to start planting is linked to both physical context and social networks. Using a rich dataset of over 50,000 members of the International Small Group and Tree Planting Program (TIST) in Uganda, we undertake a novel tripartite network analysis with farmers, planting locations and farmer groups as nodes to develop an understanding of what relational patterns influence agroforestry adoption and adoption rates.

The study compares two rice farmer social networks in Northeast Thailand (Isan province). The first rice farmer network is in Ban Phai District, with 39 interviewees, many of whom are active members of Learning Center. The second network, in Mueang Khon Kaen District, has 46 interviewees, all of whom are not active members of Learning Center. Ban Phai network has a greater network density than Mueang network does, due to two high centrality nodes that show high degree and betweenness density in both sustainability and innovation networks. Two high centrality nodes in Ban Phai network are connected through strong, lasting links. Mueang network, on the other hand, has less dense and fragmented sustainability and innovation networks. Learning Center provides its members group documentation sessions and greater connections with government officers to help organic rice program participation. Learning Center members tend to explain their decisions related to their rice farming activities through themes and narratives of extension officers, while non-Learning Center members are less likely to do so. Learning Center members also more tend to utilize Learning Center's innovative methods or machines without recognizing them as new or innovative. Social links generated by Learning Center help farmers to continue organic rice farming despite of the lack of irrigation, fiscal restraints, or a change of farmer's attitude toward the program. Extension officers of both network mention limited funding as a limiting factor of Learning Center expansion. This result suggests that funding Learning Centers could help Thailand’s agrarian transformation program to be successful.

Natural hazards and disasters pose significant threats to public health, particularly for vulnerable groups like the unhoused population. These individuals may not only require special resources (e.g., cooling stations in extreme heat situations) but are often not reachable by standard broadcast methods (e.g., TV or radio). Research suggests that during an emergency, multiple networks (satellite, cellphone, etc.) should be used to complement each other to facilitate robust and resilient communication. This study explores the potential of leveraging existing social ties within homeless communities, both physical and virtual, for effective emergency and protective action communication through SMS or other cellular methods. Given that not all individuals experiencing homelessness are equally connected to phones, we investigate the feasibility of reaching the entire social network within a specific geographic area (e.g., county) by activating the connected nodes. Consider, for example, a hypothetical scenario in which public health authorities in Seattle anticipate a heatwave. By utilizing text messaging, they could inform unsheltered individuals about the impending heatwave and the locations of nearby cooling stations more effectively, potentially mitigating the health risks posed by extreme weather events. To empirically test this solution, we used data from two large-scale representative surveys based on network data collection. The surveys employed respondent-driven sampling (RDS), which requires providing an incentive for referring someone in the target population. These RDS surveys used SMS messages to deliver these incentives. This realized data on the network of people experiencing homelessness provides key metrics for testing the reliability of diffusion over a network of unhoused people. Based on the literature, we fit Exponential Random Graph Modes (ERGM) to the resulting network data from the RDS surveys. We then simulate a complete network of relations between people experiencing homelessness in King County, WA. This is followed up with a simulation of the diffusion process over the network to examine the efficacy of emergency preparedness and response communication plans (e.g., SMS messaging outreach). By simulating similar scenarios within our constructed social network, we aim to quantify the extent of interconnectedness needed to spread information effectively and efficiently (we aim for 90% reach) and identify key factors that influence the reach within these networks along known equity issues such as gender or race/ethnicity. We will conclude with policy and future recommendations for improving the communication pathways for safety and responsiveness to natural hazards and disasters in the context of people experiencing homelessness.