abstract / Geography / Doctoral / Doctor of Philosophy

As climate change impacts a growing number of aspects of economic activity, it is becoming ever more vital to understand how these effects will manifest. This work advances the study of the impacts of environmental change. First, I provide a panel analysis at the country level that identifies the effects of dust storms on economic activity in West Africa. I also find some evidence in the agricultural sector to support the finding of damaging effects. Second, I examine the extent to which dust storms and climate shocks affect migration in the same region. While temperatures and precipitation are found to play important roles, dust storms do not appear to have a significant influence. Third, I consider the role of adaptation to climate change in the United States. I build a model that predicts the locations most likely to be used in agriculture in the future, allowing for better forecasting of shifts in the areas used for agricultural production.

Female migration in South Africa has been on the increase over the years. This thesis attempts to look at the demographic and socio-economic factors that drive this increase using data from the Africa Centre Demographic Information System (ACDIS) during the period 2001 and 2008. Using data that provides for timing of events such as migration and births, the study analyses the time it took females to migrate. Migration was defined as having out-migrated the Demographic Surveillance Area (DSA) and never coming back. Migration levels were found to be high with 28 per cent of the females between 15-49 years of age out-migrating from the DSA. Models were created to explore the demographic and socio-economic factors controlling for other known determinant of migration. In the logistic regression, odds ratios showed that parity and childbearing status were important predicators of female migration. Females with four children were less likely to out-migrate the DSA (a 61 per cent less chance of migrating compared to females without children). Furthermore, pregnant females were not likely to migrate (a 45 per cent less chance of migrating compared to females who are not pregnant or breastfeeding). In a survival analysis, determinants of timing of migration showed that females with high parities had a higher survivorships to out-migration, compared to females who were pregnant. Hazard ratios also showed that females with four children are not likely to migrate compared to females with four children (a 7 per cent less hazard of migrating compared to females with no children). Age, marital status and educational attainment were also found to be predictors of female migration. Older females were less likely to migrate compared to younger females (females in the 44-49 age group had a 70 per cent less hazard of migrating compared to females in the 15-19 age group). Currently married and cohabiting females had a 29 per cent less hazard of migrating compared to never married females. Females with high educational attainment were more likely to migrate compared to females without education (an 18 per cent higher hazard of migrating compared to females without education). The timing of migration showed that pregnant females migrate after five years into the start of their pregnancy (date of conception). In conclusion, females with many children and females who are pregnant or breastfeeding are not likely to migrate. / Thesis (M.A.)-University of KwaZulu-Natal, Durban, 2011.