Contents lists available at ScienceDirect Ecological Indicators journal homepage: www.elsevier.com/locate/ecolind High spatial resolution landscape indicators show promise in explaining water quality in urban streams Xiaofeng Ruan a, ⁎ , Jieying Huang a , Dave A.R. Williams a,b , Karly J. Harker a , Sarah E. Gergel a a Department of Forest and Conservation Sciences, University of British Columbia, 2424 Main Mall, Vancouver, BC V6T 1Z4, Canada b Department of Forest and Conservation Sciences, Department of Forest Resources Management, University of British Columbia, 2424 Main Mall, Vancouver, BC V6T 1Z4, Canada ARTICLE INFO Keywords: RapidEye LiDAR Urban planning Aquatic health Land cover monitoring Impervious surfaces Riparian vegetation Ecosystem services Watershed ABSTRACT Urban streams are subject to a myriad of complex stressors associated with impervious surfaces, which impact delivery of pollutants. Landscape indicators (e.g. the amount and arrangement of land cover) are widely used to help monitor water quality, yet their application in highly urbanized settings remains poorly understood. Here, we explored the utility of landscape indicators for explaining water chemistry and bacterial concentrations in streams throughout the Greater Vancouver Regional District, British Columbia, Canada. First, we compared high and moderate spatial resolution imagery to evaluate their relative efficacy in characterizing impervious surfaces. Second, we linked a suite of landscape indicators to major chemical properties and pollutants (e.g. dissolved oxygen, nitrate, turbidity, fecal coliform, E. Coli.) using a series of comparative statistical models. Land cover at 30-m resolution often over-estimated the proportional area of impervious surface by 0.10–0.30, sometimes doubling the amount detected using finer resolution imagery. Finer 5- and 2-m resolution imagery demonstrated little difference in estimating impervious cover. Therefore, 5-m resolution imagery was adopted for spatial analysis in the study region. Watershed-level landscape indicators explained nearly 50% of the variability in wet season concentrations of dissolved oxygen and fecal coliform. When considering the spatial configuration of impervious areas, edge density within the riparian zone explained more than 30% of the variability in fecal coliform and E. Coli. Our approach is potentially transferable to many cities worldwide facing similar challenges in monitoring urban streams. 1. Introduction Degradation of rivers and streams is linked to massive landscape changes occurring in recent decades. Conversion of natural vegetation to anthropogenic land uses (e.g. urban growth and agriculture) is often accompanied by increases in impervious surfaces as well as fragmen- tation of natural vegetation (Faulkner, 2004; Riebsame et al., 1994). These changes can amplify the transport of nutrients and sediment to receiving waters resulting in deteriorating conditions in many urban streams (Allan, 2004; Arnold and Gibbons, 1996; Paul and Meyer, 2008). A large body of research has demonstrated that land cover and its spatial arrangement can serve as a proxy indicator of stream status (Amiri and Nakane, 2009; Carey et al., 2011; Collier and Clements, 2010; Gergel et al., 2002). Termed “landscape indicators”, they have been widely used to identify the potential drivers of change in aquatic systems (Johnson et al., 1997; Gergel et al., 2002). Landscape indicators complement time-consuming field sampling, provide information over broader spatial scales, and enable comparison among numerous streams. One of the most commonly used landscape indicators of water quality in urban streams is the percentage of impervious surface in the surrounding watershed (e.g. Brabec et al., 2002; Sunde et al., 2016; Torbick and Corbiere, 2015). In urban landscapes, increases in im- pervious surfaces often come at the expense of forest and vegetation cover. As a result, infiltration is reduced (Paul and Meyer, 2008) which contributes to greater runoff volume, higher peak discharges, and in- creased loading of pollutants and sediments (Amiri et al., 2012; Hogan et al., 2014; Zhou et al., 2014). Increased impervious surfaces can also lead to increased concentrations of pathogens (e.g. fecal coliform) in streams and alter water temperatures and biotic assemblages (Pullanikkatil et al., 2015; Schoonover and Lockaby, 2006; Steuer et al., 2010). The explanatory power of landscape indicators can be quite https://doi.org/10.1016/j.ecolind.2019.03.013 Received 31 July 2018; Received in revised form 4 March 2019; Accepted 9 March 2019 ⁎ Corresponding author at: Department of Forest and Conservation Sciences, University of British Columbia, 2424 Main Mall, Vancouver, BC V6T 1Z4, Canada. E-mail addresses: 2000rxf@gmail.com (X. Ruan), huangjieying1991@gmail.com (J. Huang), Dave.Williams@forestry.ubc.ca (D.A.R. Williams), harkerkj@gmail.com (K.J. Harker), sarah.gergel@ubc.ca (S.E. Gergel). Ecological Indicators 103 (2019) 321–330 1470-160X/ © 2019 Elsevier Ltd. All rights reserved. T