9781934874141-ch12

Community Ecology of Stream Fishes: Concepts, Approaches, and Techniques

A Modeling Framework for Assessing Long-Distance Dispersal and Loss of Connectivity in Stream Fish

Marco A. Rodríguez

doi: https://doi.org/10.47886/9781934874141.ch12

Abstract.—Long-distance dispersal (LDD; relatively infrequent displacements occupying the tails of the dispersal kernel) and habitat connectivity (the degree to which the landscape facilitates or impedes movement among resource patches) influence many important ecological processes. These processes include population spread and redistribution, regulation of local and regional population dynamics, colonization of newly available habitats, maintenance of diversity in variable environments, and transfer of energy and nutrients. Field studies have shown that both LDD and instream barriers can have marked effects on the distribution patterns and demographic isolation of stream fishes at various spatial scales. Traditional summary measures of spatial use at the individual level, such as home ranges, have limited utility for examining the effects of connectivity in the presence of LDD or instream barriers; however, simple models can be tailored to extract and synthesize this information efficiently. This study presents a modeling framework for quantifying LDD of marked fish as well as their movements in the presence of barriers of differing porosity or permeability. Simulations are used to illustrate the feasibility of the modeling approach and explore sample size and spatial scale requirements. Comparison of model parameters across systems, species, and time periods should provide insights into the contribution of movement to structuring fish communities in riverine landscapes. The proposed framework can help improve on methods currently used (e.g., to quantify characteristic scales of habitat use by using median displacements or other appropriate percentile measures instead of home ranges and to relate fish movements to environmental or individual predictors by robust analyses based on heavy-tailed rather than simple normal distributions).