•  
  •  
 

Abstract

Establishing sampling frameworks to monitor the occurrence of ecological indicators, and to identify the covariates that influence occurrence, are high priority needs for natural-resource restoration and management efforts. We utilized occupancy modeling to identify patterns of beaver occurrence and factors influencing these patterns (i.e., type and amount of vegetation cover) in the Grand Canyon of the Colorado River ecosystem. We used rafts and kayaks to access a stratified random sample of sites (i.e., 100 m long sections of river bank) and used repeated sampling procedures to sample for beaver sign (i.e., lodges, cuttings, tracks, and beaver observed). We quantified the type and amount of vegetation cover at each sampled section using a remote sensing GIS database of the riparian vegetation in the Grand Canyon. We first modeled occurrence of beaver sign as a function of the total amount of vegetation cover (summed across classes) and then determined the relative importance score for each of the 7 vegetation classes. Detection probability (p) was two times higher in kayaks (0.61) than in rafts (0.29). Occurrence of beaver sign (ψ) in sampled transects was widespread throughout the Grand Canyon (ψ = 0.74, SE = 0.06) and positively associated with total vegetation. The relative importance scores for Tamarix and Pluchea vegetation classes were 1.5 – 2.5 times larger than all other vegetation classes, indicating that occurrence of beaver sign was most strongly associated to the cover of these 2 vegetation classes. Our results imply that quantifying the amount of riparian vegetation in close proximity to a river helps determine the occurrence of an important ecological indicator in riparian systems, and underscores a useful and cost-effective method for monitoring riverine species use by explicitly accounting for detectability.

COinS