Quantifying relationships between watershed characteristics and hydroecological indices of Missouri streams

Sean J Zeiger; Jason A Hubbart

doi:10.1016/j.scitotenv.2018.11.205

Quantifying relationships between watershed characteristics and hydroecological indices of Missouri streams

Sci Total Environ. 2019 Mar 1:654:1305-1315. doi: 10.1016/j.scitotenv.2018.11.205. Epub 2018 Nov 14.

Authors

Sean J Zeiger¹, Jason A Hubbart²

Affiliations

¹ University of Missouri, School of Natural Resources, 203-T ABNR Building, Columbia, MO 65211, USA. Electronic address: ZeigerS@Missouri.edu.
² Institute of Water Security and Science, West Virginia University, 4121 Agricultural Sciences Building, Morgantown, WV 26506, USA; Davis College, Schools of Agriculture and Food, and Natural Resources, West Virginia University, 4121 Agricultural Sciences Building, Morgantown, WV 26506, USA. Electronic address: Jason.Hubbart@mail.wvu.edu.

PMID: 30841403
DOI: 10.1016/j.scitotenv.2018.11.205

Abstract

There is an ongoing need for multidisciplinary investigations that will lead to policy changes that target and reduce natural and anthropic alterations to hydroecological indices important for regional environmental flows management. The hydroecological indices assessed in this study were all deemed ecologically relevant due to causal linkages with hydrogeomorphology, physical habitat, water quality, and/or ecological processes. Watershed characteristics (i.e. topography, land use and land cover (LULC), soils, and geomorphic variables) and hydroecological data were assessed for general trends between ecoregions at gauged watersheds (n = 115) in Missouri, USA. Univariate ordinary least squares (OLS) and multivariate least absolute shrinkage and selection operator (LASSO) regression models were fit to selected hydroecological indices, and models were validated using a split-site approach. Key results included: 1) significant differences (p ≤ 0.05) were observed between hydroecological indices of different ecoregions, particularly low flows statistics; 2) urban land use was associated with moderate (0.25 < R² adj. > 0.75) to strong (R² adj. ≥ 0.75) influence on more hydroecological indices (31 of 171 indices) compared to other LULC indices and watershed characteristics assessed, especially urban land use - high flow frequency relationships (5 of 11 indices; 0.77 ≤ R² ≥ 0.85); and 3) univariate ordinary least squares (OLS) regression models performed better overall relative to least absolute shrinkage and selection operator (LASSO) regression models at validation sites. Given the ecological relevance of the hydroecological indices assessed in this study, results indicated management efforts should focus on mitigating urban land use influence on environmental flows in Missouri, USA. These results hold important implications for other regions globally, where urban land use has altered environmental flows.

Keywords: Ecohydrology; EflowStats; Environmental flows; Hydroecology.