Development and Validation of a Nigeria-Specific Bovine Respiratory Disease Prediction Model Integrating Climate, Livestock Movement, and Pathogen Data

This dissertation develops and validates an original predictive model for bovine respiratory disease (BRD) incidence in Nigerian cattle, integrating climatic variables, livestock movement network data, and multi-pathogen surveillance results to fill a critical gap in evidence-based BRD management for West African cattle production systems. BRD, encompassing diseases caused by bovine respiratory syncytial virus, bovine herpesvirus-1, Mannheimia haemolytica, and Pasteurella multocida, is the leading cause of morbidity and mortality in feedlot and post-transit cattle globally, but no predictive model exists for the Nigerian context where distinctive climate patterns, breed characteristics, and movement systems operate. The dissertation employs a four-phase design: systematic characterisation of BRD pathogen distribution through multi-pathogen PCR surveillance on 600 cattle samples across 30 sites in five geopolitical zones; collection and analysis of climatic covariates including temperature, humidity, and harmattan intensity; construction of livestock movement networks from trader interview data; and model development using generalised additive models (GAMs) and machine learning ensemble methods validated by k-fold cross-validation. Primary data collection spans two years, encompassing wet and dry seasons. The original BRD Prediction Index (BRD-PI) is developed and validated, achieving area under the ROC curve of 0.87 in holdout samples. Seasonal harmattan onset and high-density market transit emerged as the strongest BRD risk amplifiers. The dissertation provides original contributions to veterinary epidemiological modelling methodology and practical BRD management tools for Nigerian feedlot and smallholder cattle systems. Keywords: bovine respiratory disease, predictive model, climate, livestock movement, Nigeria.