Multi-model analysis reveals habitat and conservation needs of two Rio Grande fishes

Authors: Ryan Friebertshauser, Kevin Bestgen

Year: 2024

Abstract

As human demands for water outpace supply, and invasive biota expand, persistence of freshwater fishes native to dryland systems increasingly depends on ecologically informed conservation practices. Endemic to the Rio Grande basin, declining Rio Grande Chub Gila pandora, and Rio Grande Sucker Pantosteus plebeius, currently face these challenges, and in portions of their range, rely on active management such as habitat restoration and stocking to persist. Using multiple data types and modeling techniques, this work informs these practices through a holistic understanding of the ecological requirements of these species. We integrated remotely sensed data, reach-level habitat assessments, and fish community surveys, collected across these species’ shared range in Colorado and New Mexico, to develop two ecological models designed to: 1) predict regional suitability of stream reaches across a large spatial extent and 2) identify physiochemical and biotic characteristics that influence local presence. Driven largely by warmer water temperature, predictive spatial models identified a geographic trend of increasing suitability, for both species, towards the southern extent of our study area, with the limitation that quantity of non-ephemeral reaches there is limited. Inferential models using local measures identified the importance of deep pools for Rio Grande Chub and the absence of riffles for Rio Grande Sucker, as well as a positive relationship with water temperatures for both species, supporting trends from predictive models. Inferential models also highlighted a strong, negative effect of White Sucker Catostomus commersonii, (for Rio Grande Sucker) and Brown Trout Salmo trutta, (both species) relative abundances on presence. Spatially explicit predictions of suitability, useful to guide broad-scale efforts like exploratory sampling, can be integrated with conclusions from inferential modeling to inform more complex management actions such as habitat restoration and repatriation. This approach to predict broader spatial suitability followed by reach-level inference regarding habitat and invasive species offers a detailed view of these species ecological requirements, a technique applicable to conservation of other imperiled organisms native to dryland streams.