My work on complex social-ecological systems includes work in Zimbabwe and California. As part of my work as an NSF Science, Engineering, and Education for Sustainability Fellow*, I attended the Santa Fe Institute’s Summer School on Complex Systems. We learned many methods and perspectives on the description, dynamics, and prediction of complex systems. I worked with two different groups of fellow summer school students on projects exploring the resilience of the social-ecological system studied by The Muonde Trust in Mazvihwa Communal Area, Zimbabwe, one using agent-based modeling, and the other using social networks analysis. More recent and ongoing projects include applying historical ecology and social networks analysis to questions of interest to the Karuk Tribe in Northern California, and, at the Center for Community and Citizen Science, supporting community-based dam removal research and analyzing human coastal use in marine protected areas.
Community-Based Dam Removal Research and Management
All throughout the western United States, municipal infrastructure is aging, and the needs of communities may have changed since many dams were built more than 50 years ago. There are therefore many projects underway to remove dams that provide little benefit and can pose disaster risks and cause environmental harm. The Center for Community and Citizen Science is supporting community-based science around dam removal projects in California, Montana, and Washington. My role is to help identify datasets that may be helpful for communities to collect and to advise them on how they might analyze these and other data they may have.
Human Coastal Use in Marine Protected Areas
The MPA Watch program collects data on human uses of coastal use (both inside and outside Marine Protected Areas) throughout California. Surveyors (who are largely volunteers) fill out a checklist of the different activities they see people engaging in, including walking their dogs, tidepooling, various kinds of fishing, on- and off-shore recreation, and a variety of recreational boating activities. As part of the State of California’s Decadal Management Review process, the Center for Community and Citizen Science is analyzing the MPA Watch archive to draw statewide conclusions about coastal use and the effectiveness of MPAs. My role is to help apply statistical modeling tools to the dataset, which like many participatory datasets has uneven sampling effort and requires special handling in order to draw statistically robust conclusions. We are wrapping up a first pass at fitting an occupancy model to the data, learning that some categories of human activity are more likely or less likely inside MPAs rather than outside.
Tribal Water Quality Social Networks
I am working with Sibyl Diver and the Karuk Tribe Department of Natural Resources to represent and analyze the diversity of their collaborations on a variety of water quality issues on the Klamath River in Northern California. With only five staff members, the Tribe works with almost 200 different organizations through direct contacts and coalition memberships, and a large number of these organizations are not governmental entities (despite the Tribe’s status as a sovereign nation). Our papers describing the ways in which the Tribe (and other tribes on the Klamath) are interacting with other organizations are in revision for Water Alternatives and Ecology & Society.
Historical and Contemporary Land-Use/Land-Cover Change on the Klamath River
I am also advising the Karuk Tribe on models of vegetation and fire dynamics. I have received approval for my involvement from the Karuk Resources Advisory Board through their Practicing Píkyav process. One particular project involves mapping, representation, and analysis of land-use/land-cover change via a historical ecology approach using historical aerial photos and more recent National Agricultural Imagery Program (NAIP) images. With Dan Sarna, I have reconstructed land-cover change at the Tribe’s research plot locations from the 1940s through the 2010s, and we are currently drafting papers describing our findings.
Agent-Based Modeling of a Feedback-Rich Complex Agro-Pastoral System*
The Muonde Trust has been engaged in a 35-year collaborative research project in Mazvihwa Communal Area, Zimbabwe. I engaged with the community research team to develop methods of modeling the resilience of their agro-pastoral system, synthesizing their long-term data to answer pressing concerns about sustainable environmental management. Together, we created an agent-based model of the farmers’ agro-pastoral system (see its GitHub repository, and its entry on the CoMSES Computational Model Library).
This model features the feedbacks between crops/arable land, cows/livestock, and woodland/grazing area/sacred forest (rambotemwa). For example, cows are needed to plough crops, but will also try to eat the crops, so woodland trees need to be cut down to make brushwood fences, but this isn’t desirable because the woodland has cultural, practical, and spiritual value as well as providing browse for livestock to eat. This is a classic situation of multiple-use management, so maximizing any one component will not result in a sustainable system, and in fact human management actions that support one component may unintentionally hurt a different component. Into this situation comes climate change and increasingly long droughts and erratic rainfall. Our model incorporates these feedbacks and management strategies as well as different ways to make rainfall more extreme at a year-to-year level. The focus of our High-Performance Computing (HPC) analysis of the model was on the ability of management techniques to keep the system sustainable for multiple uses (livestock, crops, and woodland) in the face of increasing climate variability.
In analyzing nearly 500,000 simulations, we observed lower sustainability for model runs with wider year-to-year variability in rainfall, and also for model runs that included management strategies that broke feedbacks between cows, woodland, and crops. Model runs that included management strategies that smoothed over year-to-year variation (for example, storing crop harvests for multiple years) improved the modeled system’s sustainability. We published a paper in PLoS ONE describing our analysis and community-based validation processes. We also examined how these results changed when different definitions of sustainability ‘success’ were used: annual harvest traded off with the long-term persistence of all three system components (cows, crops, and woodland); and different sets of management interventions supported different definitions. And different definitions of how much is ‘enough’ for persistence also gave different results. In particular, when balancing the three components, an intermediate proportion of land dedicated to crops gave the best model system sustainability. We published a paper in Ecology & Society detailing these results. Perhaps most important, however, is that the Muonde Trust has been able to use the model to successfully generate discussion among local leaders about changes in land-use planning policy that can increase the sustainability of their system; see the section on “Critical Participatory Data Science” for more on that aspect of this project.
Dynamics of Households and Kinship in Rural Zimbabwe*
In addition to creating the agro-pastoral model using Muonde’s data archive, we examined the community’s social resilience and compared the stability of households, genealogical kinship networks, and other structures through social network analysis. Household composition tended to change a great deal over the course of the 35-year dataset, with new households forming and moving to areas recently available for settlement. We also noted that the community made use of a variety of different strategies for sharing resources. See the SFI 2015 Complex Systems Summer School Proceedings page for our written report of these findings.
*Projects marked with the asterisk (*) were supported by the United States National Science Foundation under Grant No. 1415130. NSF requires the following statement to appear on any content generated through their funded research: “Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.”