Using natural landscape and instream habitat to identify stream reference groups for bioassessment

TL;DR

This study simplifies stream classification in Oklahoma by grouping streams based on natural habitat and climate factors to improve ecological assessments.

Contribution

A simplified, hierarchical classification of Oklahoma streams using natural drivers instead of ecoregions for bioassessment.

Findings

Six stream groups were identified that capture regional and local habitat variability as effectively as 13 ecoregion-based groups.

Three stream archetypes (Plains, Valley, Rocky) were defined based on shared habitat characteristics across Oklahoma.

The new classification is not influenced by human activities and better reflects natural ecological patterns.

Abstract

Grouping streams into reference groups based on their similarities is a critical component of developing multi-metric indices to evaluate biotic integrity. The use of level III ecoregions is a common approach that has been successful for many geopolitical regions. However, the diversity in ecoregions across the state of Oklahoma results in excessive complexity in the application of reference groups. In this study, we sought to simplify Oklahoma’s reference groups by considering the natural drivers of species distribution in wadeable streams.

We used two K-means clustering algorithms to create hierarchical stream groups. In the first K-means analysis, we grouped wadeable streams using several climatic and geologic features expected to influence fish distributions across the state (Tier I). Next, we subdivided the stream groups identified in the Tier I analysis based on water chemistry and instream habitat (Tier II). We used classification trees and between class multivariate analyses to validate and understand the resulting groups and define them as ‘stream archetypes’ based on distinguishing habitat characteristics.

The two K-means analyses resulted in six groups that represented both regional differences and reach-scale geomorphological differences. We determined that there were two regions, West and East that were primarily distinguished by mean annual precipitation. Tier II clustering identified Plains and Valley stream groups in both the East and West regions. These groups represented shallow, run-dominated, sand bed streams, and relatively deep, pool dominated, silty streams, respectively. The East region possessed two additional stream groups (i.e., Hills and Highlands groups), each with rocky substrate and riffle channel units that broke up the runs and pools in the stream. Of these, the Hills group was more prone to having pools than runs and had low nutrient concentrations compared to the Highlands group. We found that the six groups explained variability in habitat and water chemistry as well as the previous 13 ecoregion-based reference groups. Further, we confirmed that the groups were not the result of common anthropogenic influences on stream ecosystems. The six groups were subsequently organized into three stream archetypes, Plains (East and West), Valley (East and West) and Rocky (Hills and Highlands). We expect these six hierarchical stream groups and three archetypes to be useful in evaluating biotic communities in Oklahoma moving forward.

Figures