Climate change perception, crop diversification and land use change among small-scale farmers in the Midlands region of KwaZulu-Natal, South Africa : behavioural and microeconomic analyses.

Ph. D. University of KwaZulu-Natal, Pietermaritzburg 2015.

The future of small-scale agriculture in South Africa is facing the challenges of high population growth, land degradation, and agro-biodiversity loss. In the Midlands region of KwaZulu-Natal (KZN), the challenges are exacerbated by climate change (CC). Agricultural land use change (ALUC) is among farm-level options to exploit the synergy between local adaptation and global mitigation of CC, whilst restoring degraded lands and agro-biodiversity reservoirs. The literature contends that the pathways to climate-resilient sustainable land use start with the farming sector’s vulnerability to CC, and translate into behavioural change through farmers’ perceptions about the climate risk. Two major steps characterize behavioural response to changing distributions of weather patterns. Behavioural responses to CC begin with reducing the vulnerability to climatic variability (e.g. by using tactical adjustments such as crop diversification), followed by forward-looking integration of adaptation and mitigation into farm planning (e.g. through ALUC). The purpose of this study was to provide a systematic and detailed understanding of climate-driven ALUC in the setting of small-scale farming in South Africa. Taking the Midlands region of KZN as an illustrative case, this study aimed (i) to explore some meso-level aspects of climate-related agricultural vulnerability; (ii) to investigate the perceptions of small-scale farmers about CC and their socio-psychological, institutional and cultural determinants; (iii) to analyse the farmland use systems and assess the microeconomic determinants of seasonal crop diversification; and (iv) to assess the attitudes of small-scale farmers towards land use and the constraints governing ALUC decision-making. To assess the meso-level vulnerability of farming sector to CC in the Midlands region of KZN, an indicator approach was adopted. Various aspects of exposure, sensitivity and adaptive capacity of the farming sector were explored using a systematic review of available empirical evidence. The review suggested that the farming sector is exposed to a warming and wetting climate. The reviewed evidence also suggested that the farming sector is highly sensitive to CC due to high population densities, high rates of small-scale farming, low irrigation rates and susceptibility to land degradation. Nevertheless, diversified crop portfolios remain a major aspect of resilience among small-scale farmers. The reviewed evidence further indicated that farmers’ adaptive capacity is confounded by inadequate access to infrastructure, rural exodus, skills shortages, poor health status, and low level of social capital. The remaining objectives were empirically investigated based on primary data from a household survey of 152 small-scale farmers in the uMshwati local municipality. With regard to the second objective, a principle component analysis (PCA) of eight variables of perceived seasonal climatic abnormality yielded two dominant perceptual shapes. CCP1 score captured the extent to which an individual farmer perceives the winter season as cooling and the summer season as warming and drying. CCP2 score captured the extent to which an individual farmer perceives the winter season as warming and wetting, and summer season as drying. The results of a Double-Hurdle (DH) model showed that the probability of perceiving abnormal trends in the local seasonal climate increases with holistic affect, egalitarian worldviews, age, female-headedness and hilly and wetter agro-ecological regions, and decreases with education. The model results also showed that the CCP1 score increases with holistic affect and other factors related to personal experience such as age and distance to the river, signifying experience-based learning. The results further showed that CCP2 score increases with cognitive ability (knowledge) and related socio-demographic factors such as education, extension and trust, indicating analytic processing of climate information. With regard to the third objective, a two-stage cluster analysis of land use data unveiled a diversified farmland use matrix characterised by a maize-beans inter-cropping system coexisting with of mixed crop-livestock, sugarcane and timber mono-cropping systems. After accounting for potential endogeneity biases, the results of a logit transformation model showed positive effects of labour and landholding on the intensity of multiple-cropping among poorer and richer households, suggesting that responding to technological constraints is an important driver of seasonal crop diversification. Furthermore, the results showed that the intensity of multiple-cropping among richer households decreases with education due to faster shadow wage rise at higher farm assets position. Among poorer households, the intensity of crop diversification decreases with off-farm occupation and increases with distance to water sources, suggesting that mitigating income and production risks are key motivations for crop diversification. Regarding the fourth objective, the results of a Mixed-multinomial Logit model that accounts for potential endogeneity biases showed that the decisions about ALUC from seasonal crops towards forestry or sugarcane cultivation are rationally derived and driven by clear but heterogeneous preferences and trade-offs between crop productivity, food security, and labour requirements. These motivations override income generation and ecological sustainability incentives, the common policy foci in South Africa. The model results also indicated that the utility of planting sugarcane increases with size of landholding, suggesting economies of size. The preference for forest plantation decreases with household size, a Chayanov-like afforestation pattern. Furthermore, the results indicated that the preference for ALUC increases with the proportion of ALUC in the community/ward (suggesting peer group influence) and hilly and wetter agro-ecological conditions (due to higher land conversion cost and acute soil erosion). The empirical findings of this study pointed to the need for designing a region-specific CC communication strategy that involves affected farmers and locally trusted agricultural extensions agents, and aligns the information content to local beliefs, values and norms. The findings also inferred that reduced agro-biodiversity loss and enhanced resilience in the face of the increasing climatic variability through crop diversification could be major co-benefits of the ongoing land reform, and explicit strategies should focus on regions with widespread poverty and dryland farming. For supporting ALUC towards farmland afforestation, the findings advocated the promotion of timber-based agroforestry systems as an effective strategy to align public goals with private incentives. The findings further advocated incentive-based afforestation schemes (such as payment for ecosystem services) designed on a per-capita or equivalent-consumption basis and targeting farms located in regions with steeper slopes and higher climate variability.