To Plow or Not to Plow: An Investigation Into Land Use Changes and Consequences in the Northern Great Plains Using Systems Thinking and Dynamics

This dissertation aimed to investigate grassland conversion to row crop production in the Northern Great Plains (NGP), which has become a growing concern among many stakeholders. First, we explore the historical development of the land ethic concept in the United States and how those ethics have impacted land use policy and use of private lands (Chapter 4). Primary data gathered from semi-structured interviews of farmers, ranchers, and influential stakeholders were analyzed using stakeholder analysis methods to identify major factors considered in land use decisions, priorities of factors of each group, and to define relevant mental models describing each group’s view of the land ethic concept. Results show that these stakeholder groups prioritize land use factors qualitatively differently and possess strikingly different land ethics. Second, we develop a system dynamics (SD) model to integrate ecological, economic, and social components influencing land use decisions (Chapter 5). We conceptualized the model, identified key variables and reference modes, and established model boundaries through a group theory building process (Appendix 1). Components of the model included grassland cattle ranching, cropland farm production, rural community dynamics, land capability classes, public policies, and soil externalities. Individual producer decision-making factors were used to link the components and were found to drive land use change over time. The model was developed using the Vensim™ modelling environment, with years as the time unit. The model consisted of 375 state variables (i.e., stocks, flows, and converters), 32 look up tables, and 149 constants. The model’s key strengths were the incorporation of land quality and individual decision-making factors into standard physical SD structures. Model limitations included a lack of spatial input and output capabilities. Our evaluation followed rigorous SD procedures and indicated the model satisfactorily predicted historical data (Chapter 6). The model reproduced trends (i.e., reference mode) for key variables, including farmland area. The full analysis indicated the model was robust under extreme and sensitivity tests and proved to adequately predict land use as its intended purpose. Lastly, forecasts of land use change were predicted for 2012 to 2062 (Chapter 7). The base-case showed farmland area continued to increase through midcentury. Unmitigated, the soil environmental risk (SER) value of such changes would be comparable to conservative estimates of Dust Bowl-era externalities. Results showed that reducing livestock production costs, doubling conservation compliance requirements, and livestock-farming integration had the largest impact on grassland conservation and mitigating SER. Eliminating CRP and raising CIPS had the largest negative effects, with SER reaching Dust Bowl-era levels despite farming improvements. System archetypes were identified, including ‘fixes that backfire’, ‘success-to-the-successful’, and ‘eroding goals’. Our projections can be used aid policy-makers and interested stakeholders in enhancing the discussions currently taking place about land use transformation in the region. Complimentary work provided useful information related to watershed responses to conversion (Appendix 3) and land use legacies (Appendix 4) as well as methodological considerations within the SD process (Appendix 2)