Predictive mapping of the biotic condition of conterminous U.S. rivers and streams

Abstract

Understanding and mapping the spatial variation in stream biological condition could provide an important tool for conservation, assessment, and restoration of stream ecosystems. The USEPA's 2008-2009 National Rivers and Streams Assessment (NRSA) summarizes the percentage of stream lengths within the conterminous United States that are in good, fair, or poor biological condition based on a multimetric index of benthic invertebrate assemblages. However, condition is usually summarized at regional or national scales, and these assessments do not provide substantial insight into the spatial distribution of conditions at unsampled locations. We used random forests to model and predict the probable condition of several million kilometers of streams across the conterminous United States based on nearby and upstream landscape features, including human-related alterations to watersheds. To do so, we linked NRSA sample sites to the USEPA's StreamCat Dataset; a database of several hundred landscape metrics for all 1:100,000-scale streams and their associated watersheds within the conterminous United States. The StreamCat data provided geospatial indicators of nearby and upstream land use, land cover, climate, and other landscape features for modeling. Nationally, the model correctly predicted the biological condition class of 75% of NRSA sites. Although model evaluations suggested good discrimination among condition classes, we present maps as predicted probabilities of good condition, given upstream and nearby landscape settings. Inversely, the maps can be interpreted as the probability of a stream being in poor condition, given human-related watershed alterations. These predictions are available for download from the USEPA's StreamCat website. Finally, we illustrate how these predictions could be used to prioritize streams for conservation or restoration.

A review of macroinvertebrate‐ and fish‐based stream health modelling techniques

“…When used for classification, RF determines the most frequent class across all trees for each observation within the out-of-bag portion. Hill et al, 2017). Therefore, because a large number of trees provides limited generalization errors, RF method prevents overfitting (Prasad et al, 2006).…”

“…Therefore, because a large number of trees provides limited generalization errors, RF method prevents overfitting (Prasad et al, 2006). Hill et al, 2017). Many of these studies have been focused on predicting macroinvertebrate taxa richness and composition (Álvarez-Cabria, González-Ferreras, Peñas, & Barquín, 2017;Booker et al, 2015;Chinnayakanahalli, Hawkins, Tarboton, & Hill, 2011;Patrick & Yuan, 2017;Vander Laan, Hawkins, Olson, & Hill, 2013;Waite et al, 2014).…”

“…Random forests Booker et al, 2015;Chinnayakanahalli et al, 2011;Clapcott et al, 2017;Death et al, 2015;Fox et al, 2017;Hill et al, 2017;Patrick & Yuan, 2017;Vander Laan et al, 2013;Waite et al, 2012Álvarez-Cabria et al, 2017Fukuda & De Baets, 2016;Fukuda et al, 2013;Grenouillet et al, 2011;He et al, 2010;Kwon et al, 2015 Chon, 2011;Gazendam et al, 2016;Goethals et al, 2007;Mathon et al, 2013; Ans M. Mouton et al, 2010;Sauer et al, 2011Chon, 2011Fukuda et al, 2013;…”

“…Therefore, health assessments of these ecosystems are critical to promoting their protection and restoration (Beechie et al, 2010). During the last few decades, many environmental legislations have been increasingly supportive of the introduction of biological assessments in local, regional, and national monitoring programs (Hering et al, 2010;Hill et al, 2017). Typically, different metrics and indices are derived from biological assessments to represent the stream health condition.…”

Understanding Landscape Influences on Aquatic Fauna across the Central and Southern Appalachians

“…As data availability and computing power both increase, the ability to perform spatial analysis across large geographic extents and at small grain size is increasing. Examples of large scale spatial analyses across the United States include; Esselman et al (2013), who used fish as indicators, the US Environmental Protection Agency's Wadeabe Stream Assessment (2006) which included benthic macroinvertebrates and fish as indicators, and Hill et al (2017) who modeled biological stream condition across the conterminous U.S. based on anthropogenic and natural watershed features. Multimetric indices (MMIs) are commonly used in studies that attempt to understand anthropogenic influences on the condition of biological aquatic communities over vast geographies (Harris and Silveira, 1999;Hering et al, 2004;Pont et al, 2006;Pont et al, 2009).…”