Improved topographic ruggedness indices more accurately model fine-scale ecological patterns

Authors: Thomas E. Dilts, Marcus E. Blum, Kevin T. Shoemaker, Peter J. Weisberg, Kelley M. Steward

Context

Topographic ruggedness has been examined in thousands of ecological studies and is a popular variable for characterizing habitat selection. Despite widespread adoption, ruggedness metrics are often applied uncritically and require systematic and thorough testing using both artificial landscapes and real-world applications.

Objectives

In this paper we introduce a technique that removes the correlation of topographic ruggedness with curvature in order to more accurately represent fine-scale surface ruggedness.

Methods

We test our modified version of several ruggedness metrics against traditional ruggedness metrics using three ideal ruggedness criteria in artificial landscapes. We further tested our modified ruggedness measures using 449 real mountain ranges in Nevada, USA. Using desert bighorn sheep as a case study, we tested both modified and uncorrected ruggedness measures and slope in a multiscale context in order to examine habitat selection by female bighorn sheep.

Results

The modified versions of the metrics passed all three criteria of the ideal ruggedness test and was able to accurately capture surface ruggedness. Modified versions of ruggedness differed from uncorrected versions by containing fewer highly rugged cells along ridgelines and drainages. Habitat relationships of desert bighorn sheep with ruggedness were scale dependent, such that female sheep selected for steep slopes at fine spatial scales and ruggedness at moderate spatial scales.

Conclusions

We demonstrate that there are three components to ruggedness: elevation variation, aspect diversity, and surface ruggedness representing first, second, and third generation ruggedness indices. Our technique for removing underlying topographic variation provides an improved mapping of surface ruggedness and augments the other two generations of ruggedness metrics.

Suggested Citation