Integration of agent-based and GIS-based modelling for geosimulation of human-elephant interactions in the Bunda district, Tanzania : A thesis submitted in partial fulfilment of the requirements for the Degree of Doctor of Philosophy at Lincoln University

Abstract

Human-elephant interactions (HEI) are one type of the human-wildlife interactions that cause several adverse impacts to communities near African wildlife reserves. However, HEI is considered to be the most disastrous because of hostility of elephants and sheer size. The adverse impacts of human-elephant interactions include human and elephant deaths and injuries, crop damage and hidden impacts. Crop damage is the most common HEI reported adverse impact. Existing spatial and socio-economic studies provide a clear insight into HEI occurrences but lack precise measurements for each factor to either minimise or eradicate HEI occurrences. The model was developed to recommend the best HEI scenario(s) for either reduction or eradication of HEI occurrences in the Bunda District. A field survey was conducted to collect local opinions on HEI. A total of 130 questionnaires were distributed and collected from 12 villages in the district. Closeness of the villages to protected areas was the main criterion used to select the study area villages. The results from the questionnaires and secondary data were used to develop rule sets for development, calibration and, validation of an agent-based model of HEI for the Bunda District, Tanzania. Spatial data on the location of hidden impacts and elephant crop damage were collected from 12 villages. Due to the complexity of hidden impacts and elephant crop damage, agricultural officers, wildlife officers, medical experts, and community development officers were consulted for clarification and consultation. The spatial data were used for kernel density estimation and hotspot analysis. Additionally, the spatial analysis was conducted to understand the relationship between HEI occurrences and environmental features. The spatial analysis showed the presence of more HEI incidents near the Grumeti Game Reserve than the Serengeti National Park. In addition, many incidents of HEI occurred within 2000 meters from rivers and protected areas. The results from a spatial analysis were used for model development, calibration and, verification and validation of the model. The model simulated, tested and evaluated 18 modelling scenarios. Model results from each scenario were analysed for comparative performance, where minimal recorded incidents of crop damage, human deaths, elephant deaths, and hidden impacts were the primary focus. In that case, the selection of the best performing scenario based on the magnitude of the reduction of adverse impact(s). The AGHEI recommends the best scenario that minimised human access to the river, conservation corridors as well as the reduction in elephant population size. However, for each selection of the best scenario(s), there were costs that a model user must incur, as there was no cost -ree scenario. Reduction of any of the adverse impacts may run counter to fiscal, conservation, land and socio-economic policies. Therefore, the model user may select the best scenario within the constraints of these policies. AGHEI allows conservationists to design, test and prescribe tested actions that can reduce HEI occurrences. This approach is possible in all countries with active elephant ranges worldwide, once modified. This AGHEI is specifically applicable to Bunda District, as it replicates the environment and behaviours of agents specific to Bunda. It serves a specific purpose and not a general-purpose tool. However, implementation of AGHEI to other areas is possible but it requires changes in the densities of elephants and humans in the models, as well as their attributes and the environmental characteristics of new areas. In addition, there must be changes in interaction rules to reflect the HEI dynamics of the areas of interest-specific to local populations. The thesis recommends further studies to investigate the distribution of HEI incidents in villages near Grumeti Game Reserve and Serengeti National Park, to evaluate, environmental legal and socio-economic implications for implementation of AGHEI modelling scenarios.