COS 11-9
Linking statistics of movement to resource dynamics

Monday, August 5, 2013: 4:20 PM
L100B, Minneapolis Convention Center
Chris H. Fleming, Conservation Ecology Center, Smithsonian Conservation Biology Institute, Front Royal, VA
Justin M. Calabrese, Conservation Ecology Center, Smithsonian Conservation Biology Institute at the National Zoological Park, Front Royal, VA
Thomas Mueller, Department of Biology, University of Maryland, Frankfurt (Main), MD
Kirk A. Olson, Conservation Ecology Center, Smithsonian Conservation Biology Institute, Front Royal, VA
Peter Leimgruber, Conservation Ecology Center, Smithsonian Conservation Biology Institute, Front Royal, VA
William F. Fagan, Department of Biology, University of Maryland, College Park, MD
Background/Question/Methods

Determining universal laws of animal movement, which would necessarily be driven by the distribution and dynamics of resources, remains an open challenge in movement ecology. One step towards fulfilling this challenge is to compare the relevant statistics of the movement process and resource field, and examine what qualitative and quantitative similarities might be present. In many cases, the movement process and the resource field can be largely summarized by their mean and autocorrelation statistics. These statistics are not mere numbers, but functions of time and distance that can reveal characteristic time and length scales in the movement behavior and resource dynamics, as well as any governing behaviors. If there is any law of animal movement driven by resource dynamics, then time and length scales of the resource field statistics would necessarily be translated into time and length scales of the movement process statistics. Making such a comparison then allows one to test hypothetical mechanisms that drive movement behavior without specifying any particular model of resource driven movement. Finally one is confronted with the much easier challenge of constructing phenomenological models consistent with the data before positing any universal laws.

Results/Conclusions

Using Mongolian gazelle relocation data and NDVI data, we find considerable correspondence between the statistics of the gazelle and their resources.  Using variogram and likelihood analysis, the gazelle are discovered exhibit a two scale movement process. They randomly confine themselves to small areas of approximately 7 km in length for approximately 1/2 days time, while intermittently performing a random search within a larger area approximately 80 km in length with a crossing time of approximately 80 days. A spatial analysis of the NDVI data reveals an average patch size of 6 km, with an average patch cluster size of 70 km. Therefore the movement of the gazelle is consistent with a random search for patches, within a single cluster of patches, along with a brief foraging time for each patch. On the other hand, a temporal analysis of the NDVI data does not reveal any timescale consistent with the 1/2 day foraging time of the gazelle, which detracts from any proposed models wherein gazelle movement is driven by the rapid growth and decay of individual resource patches.