A framework for predicting and visualizing the East African wildebeest migration‐route patterns in variable climatic conditions using geographic information system and remote sensing

• Published in: Ecological research Vol. 21; no. 4; pp. 530 - 543
• Main Authors: Musiega, Douglas E., Sanga‐Ngoie, Kazadi, Fukuyama, Kaoru
• Format: Journal Article
• Language: English
• Published: Tokyo Springer‐Verlag 01.07.2006; Blackwell Publishing Ltd
• Subjects: Animal migration; Climate change; Climatic conditions; Connochaetes taurinus; Dry season; Food resources; Geographic information systems; GIS; Hydrologic data; Migration; Rainfall; Rainy season; Regression analysis; Relative abundance; Remote sensing; Seasons; Serengeti–Mara; Simulation; Vegetation; Wildebeest; Wildlife; Africa, East
• ISSN: 0912-3814; 1440-1703
• DOI: 10.1007/s11284-006-0175-9

In the Serengeti–Mara ecosystem of East Africa, the migrating wildebeests’ (Connochaetes taurinus) response to food resource distribution and terrain complexity impacts their movement characteristics. This manifests in varying ways such as movement speed, direction, turning frequency, and moving distance. To investigate these characteristics, indices derived from vegetation quantity (normalized difference vegetation index, NDVI) and relief (slope) were used in our previous work to quantify the relationships between migration route location versus vegetation, relief complexity, and their combination. Least cost pathways determined using these indices were representative of approximate migration routes. The simulated routes were shown to be strongly influenced by vegetation during the dry season. However the impact of climatic change (rainfall) on route location was not investigated though known to influence vegetation recovery patterns. This paper specifically addresses the impact of climatic change on route location. The mean monthly rainfall data were used to classify the rainy and dry seasons in the Serengeti, the Western, and the Mara areas as normal, drier, or wetter than normal, over the 1986–1997 period. The classification is based on the mean monthly rainfall variability about the 11‐year seasonal mean. Regression analysis showed strong linear relationships between rainfall and mean NDVI for each one of the three areas. The subsequent seasonal classification based on the corresponding habitat vegetation characteristics (NDVI) revealed the relative variation of vegetation with rainfall. Using the derived general categories, migration routes are then predicted for the various categories using a “route attractivity index.” The seasonal migration routes were shown to change depending on the relative abundance of the rainfall during the dry season. Dry season migration routes tended toward areas with better vegetation activity, i.e., those characterized by higher NDVI gradients. Our results showed that during the western trek, wetter dry seasons have the effect of delaying the herds movement northwestward. During the northern trek, wetter dry seasons have the effect of delaying the tendency to move westward. However the variation in rainfall conditions during the rainy and dry season had no significant influence on the southern trek route location. We assume that predicted routes based on average general category conditions for different years are representative of main migration route patterns for similar seasons, therefore they are well suited for approximate route prediction, if the climatic characteristics of the year are known.