Multiscale drivers of hippopotamus distribution in the St Lucia Estuary, South Africa

Abstract

This study provides preliminary data on predictors of Hippopotamus amphibius (Linnaeus, 1758; hippo) distribution in St Lucia Estuary, the largest estuarine system in Africa and a key habitat for one of South Africa’s largest hippo populations. We use binary logistic regression models to evaluate selected habitat features as predictors of hippo occurrence at two spatial scales and a negative binomial model with log-link function to evaluate  predictors of frequency of use at a fine spatial scale. At the scale of the whole estuary, models indicate that hippos preferentially select diurnal refuge sites that are closer to river inlets and farther from human settlements. At a fine scale (within the Narrows in which more than 50% of the population resides), occurrence and frequency of use models suggested that hippos preferentially settle in sites with water depths between 0.5 and 1.49 m, that are farther from human settlements, closest to natural wetland vegetation and near neighbouring groups. Preliminary data on habitat variables influencing hippo distribution highlights the necessity to manage water levels, restore wetland floodplains, protect wetland vegetation and halt human settlement encroachment in order to ensure the viability of this UNESCO site and its hippo population.

Keywords: abstraction, drought, estuarine lake, megaherbivore, spatial analyses