Faculty of Science; School of Geography, Archaelogy and Environmental Studies; PhD Thesis / This thesis investigates agropastoral production and ecological conditions under which complex socio-political systems in the Shashe/Limpopo River Basin, southern Africa, periodically expanded and declined between ~AD 900 and 1700. Environmental reconstruction for this period, derived from multi-stable isotope analysis of modern and archaeological fauna from the area, demonstrate that agropastoral settlement and changes in their social, economic and political complexity were less driven by climate than previously had been assumed. Rather, at a relatively short-term climatic scale, these cultural events took place even as precipitation and temperature appeared to have fluctuated above and below the modern seasonal mean of ~350mm and ~22oC, conditions presently considered to be marginal for agropastoral production. Alternative to a climate driven model for settlement, ethnographies of traditional southern African agropastoral systems provide a comparative basis for understanding the range of environmental and social parameters that past agropastoralists in the Shashe/Limpopo River Basin may have employed to sustain population growth and intensify socio-political complexity in the face of short-and long-term climatic variability. Over a long-term climatic scale, the δ15N and δ18O values from Bos taurus and Ovis/Capra indicate that the initial settlement by Zhizo agropastoralists people, between AD 900 and 1010, took place under semi-arid conditions that were similar to, or only marginally wetter, than the present. This thesis suggests that the Zhizo settlement and their ‘capital’ site of Schroda were motivated by broader cultural factors, such as trade networks, and not solely by climate conducive for agriculture. As documented ethnographically, crops and livestock herds could have been sustained by taking advantage of various geographical features of the river basin, such as planting near outcrops where dammed water keep soils moist even in dry periods and using browse and crop fodder to offset diminished grazing lands. Results for sites dating between AD 1010 to 1415, support previous interpretations that the Leopard’s Kopje A and B cultural period ‘capitals’ of K2 and Mapungubwe, respectively, rose to prominence under a trend towards increased available moisture. The additional moisture would have facilitated the greater floodplain settlement recorded between AD 1010 and 129, which was most likely a response to increased population pressures of the capitals and the need to extend cultivated lands. This spatial shift was accompanied by an apparent greater management of livestock. The preliminary 87Sr/86Sr data, together with intra-annual δ18O and δ13C values, from B. taurus and Ovis/Capra indicates a geographical expansion of herd management took place with the transition from K2 to Mapungubwe. This thesis proposes that to sustain population and socio-political growth in the face of short-and long-term climatic variability, livestock management would need to be politically coordinated. Maintaining large-scale herds outside the river basin would have allowed for expansion of crop production onto previous river basin pasturelands, while extending territories or networks. Further, the δ15N and δ18O data indicates that the period of increased available moisture extended beyond the abandonment of Mapungubwe at AD 1290. Previous assumptions that link this event to the negative agricultural impact of a cool dry trend starting at ~AD 1300, as extrapolated from sub-continental scale climatic sequence, must be re-assessed. The isotopic data from Moloko/Khami cultural period sites suggest that drier conditions did not develop in the area until after ~AD 1450. Their initial settlement in the area during this drier period needs to be re-considered, as does the entire sequence from ~AD 900 onward, in terms of agropastoral production strategies within shifting natural, economic and political environments.

This thesis provides a theoretical framework for evaluating the causes of vacant land in shrinking cities. The focus of this thesis was New Orleans and St. Louis; these two cities were selected as the case studies because they are roughly of similar age, possess a common cultural and economic heritage, and have a geographic footprint which encompasses different environmental conditions. This thesis evaluated factors that contribute to patterns of land vacancy within these two cities. Factors included in this evaluation include employment and other economic and cultural opportunities, environmental and ecological conditions, social dynamics and conditions, governmental management decisions, and 'quality of life' stressors, such as proximity to major infrastructure and industrial development. The theoretical framework described in this thesis is intended to apply to other shrinking cities beyond the case studies. A geographic information system database using historical maps and population census data were created for each city and utilized to examine temporal patterns in the relationship between land vacancy and a variety of environmental, economic, and social factors. Maps from the time of the founding of each city were geo-referenced to create a depiction of the ecological conditions prior to European settlement at the sites of New Orleans in 1718 and St. Louis in 1764, respectively. Time-series data gathered from the United States population censuses were utilized to document spatial change of the two cities as they evolved. Homo sapiens like other species compete for habitat. Access to high quality habitat within the urban ecosystem is determined by contestation between individuals and social groups, through market mechanisms and through management decisions, both utilitarian and ideological. Corruption and violence may also be factors. Individual agency is a factor in this contestation but social and cultural structures can also work to limit individual choices, particularly for minorities and low income residents, and relegate many residents to suboptimum or marginal habitat. A data analysis of both New Orleans and St. Louis showed that the quantity and location of vacant land is primarily influenced by proximity to opportunities and by proximity to major risks which impact the quality of Homo sapiens habitat. The first of these is proximity to opportunities such as employment, education, and cultural resources. The second is the presence of natural hazards, such as flooding and geological hazards, as revealed by the analysis of the historical ecology of the city. The third is the impact of local government management decisions and social planning which has spatial implications, including racially-based zoning, racial covenants, redlining, and isolation from public services and facilities such as the segregation of public schools. These decisions are often the reflection of ideology and power relationships. A fourth driver of land vacancy is proximity to risks, notably industrial lands, but also the intrusion of major infrastructure projects such as the development of the railyards and rail corridor of St. Louis, the construction of the Industrial Canal in New Orleans, and the construction of Interstate highways through both cities. In some circumstances, such drivers that include the unintended consequences of utilitarian decisions. The fifth driver include socio-economic factors and the neighborhood effects of crime, and poor education. These five drivers act in different proportions in each city to influence land values which, in turn, drive levels of vacancy. This comparative investigation revealed that the impact of geophysical factors on land vacancy varies greatly between New Orleans and St. Louis. While much of New Orleans lies below sea level and is often subject to flooding and hurricanes, little of the vacant lands of St. Louis are impacted by geophysical factors. In contrast, management decisions and social planning have contributed significantly to the concentration of poverty and, in turn, land vacancy in both cities. While some of these management decisions are utilitarian in nature and intended to provide the greatest benefits for the most number of people, others are ideologically driven or reflect power relationships and in the case of both New Orleans and St. Louis, racism. Proximity to risks, such as active railroad tracks, major highways, and industrial development, also has a strong relationship to land vacancy in both cities. Land vacancy also has a strong spatial relationship with areas of low income, poor education, and crime and neighborhood effects. While an understanding of environmental history can provide a useful guide to vacant land policy, efforts to address the challenge of vacant lands must consider not only the symptoms but the underlying causes of vacancy, particularly economic and social factors. This thesis is addressed to planners, architects, urban designers, landscape architects, and elected and appointed government officials who work to address the challenges of shrinking cities. Though this thesis examined the causes of vacant land in two shrinking cities, future research should examine the application of the theoretical framework presented here to cities experiencing growth as well.