Spatiotemporal habitat dynamics of ungulates in unpredictable environments: The khulan (Equus hemionus) in the Mongolian Gobi desert as a case study

Abstract

In resource-poor environments many large herbivores do not perform seasonal migrations but show unpredictable, long-range movements within a given season. The few studies that have examined drivers for within season long-range movements suggest that these movements are a response to spatiotemporal dynamics of foraging resources. We tested this hypothesis and were especially interested in detecting dynamics of foraging habitat which may influence high mobility of khulan, Equus hemionus, during summers in the Dzungarian Gobi of Mongolia. We used six years of ground census data combined with remotely sensed imagery of vegetation productivity (NDVI) to build a dynamic habitat model. We subsequently predicted khulan habitat suitability for each of the seven 16-day NDVI intervals every summer between 2004 and 2009 and examined variability of the resulting 42 prediction maps to characterize spatiotemporal dynamics in khulan foraging habitat. Our analyses showed khulan summer foraging habitat was highly predictable with little spatiotemporal variability making it unlikely that broad scale variability of foraging habitats can explain the high mobility of khulan. The few and small areas that did show khulan habitat variability were related to locations around water sources. In addition, we found that khulan avoided habitats beyond 21 km from water sources. Together these findings suggest that water availability and switching among the sparsely located water bodies rather than broad scale dynamics of foraging habitats may be the key driver for the high mobility of khulan in the Dzungarian Gobi. Our findings highlight the importance of securing access to and connectivity among water bodies for wild ungulates and the need for further studies on possible drivers of nomadic movements in drylands.