RETRIEVING RIVER ATTRIBUTES FROM REMOTELY SENSED DATA: AN EXPERIMENTAL EVALUATION BASED ON FIELD SPECTROSCOPY AT THE OUTDOOR STREAM LAB C. J. LEGLEITER * AND B. T. OVERSTREET Department of Geography, University of Wyoming, Laramie, Wyoming, USA ABSTRACT This paper summarizes experiments conducted at the Outdoor Stream Lab to gain insight on the relationships between a channel’s spectral characteristics and the river attributes of interest to scientists and managers. Improved understanding of these connections would advance the application of remote sensing to fluvial systems. Motivated by this objective, we examined the following: (i) interactions among local hydraulics, water surface roughness and surface reflectance; and (ii) the influence of periphyton development and streambed disturbance on bottom reflectance. These experiments could thus lead to novel methods for inferring flow velocity and substrate type, respectively, from passive optical image data. Our initial results indicate the following: (i) water surface reflectance can compromise spectrally based depth retrieval by reducing the range of useful wavelengths and weakening correlations between band ratios and depth, implying that removing the surface-reflected component of the total radiance could facilitate bathymetric mapping; (ii) water surface roughness was influenced by local hydraulics and was positively cor- related with water surface reflectance, suggesting that observations of surface reflectance could be used to infer flow velocities; (iii) substrate spec- tral characteristics were influenced by the degree of periphyton development, implying that algal biomass might be estimated from remotely sensed data; and (iv) similarly, bottom reflectance differed significantly for intact versus disturbed substrates, suggesting that areas of streambed disturbance might be identified via remote sensing. Future work will focus on further elucidating connections between surface reflectance, water surface texture and flow velocity, and developing quantitative relationships between bottom reflectance, periphyton abundance and the degree of streambed disturbance. Copyright © 2013 John Wiley & Sons, Ltd. key words: remote sensing; field spectra; depth retrieval; water surface reflectance; velocity estimation; bottom reflectance; substrate mapping; in-stream habitat Received 20 November 2012; Accepted 30 April 2013 INTRODUCTION AND OBJECTIVES Remote sensing has emerged as a powerful tool for mapping river morphology and in-stream habitat at a higher spatial resolution and over a greater extent than would be feasible via conventional ground-based methods (e.g. Marcus and Fonstad, 2008; Carbonneau et al., 2011). Although this approach clearly has outstanding potential to advance scien- tific understanding of fluvial systems and facilitate river management (Marcus and Fonstad, 2010), achieving these goals will require prospective users of remotely sensed data to develop a certain degree of confidence in these methods—that confidence must be justified (Legleiter and Roberts, 2009). Our previous research has examined the physical processes that both enable and limit the application of remote sensing to riverine environments, primarily in the context of retrieving information on water depth from passive optical image data (e.g. Legleiter et al., 2004; Legleiter et al., 2009). In keeping with this cautious, incremental approach, this study sought to establish connections between a channel’s spectral characteristics and other river attributes, in addition to depth, of interest to scientists and managers. To assess the feasibility of deriving river information from in situ and/or remote observations of reflectance, we conducted a series of controlled experiments at the National Center for Earth Surface Dynamics’ Outdoor Stream Lab (OSL). The unique facilities available at the OSL allowed us to manipulate and carefully measure various factors that influence the upwelling spectral radiance from a river and potentially could be inferred from optical image data. Our experiments thus were intended to not only refine existing algorithms for estimating flow depth but also to explore the possibility of mapping other channel characteristics, such as velocity and bottom composition, via remote sensing. The investigation thus involved a combination of field spectroscopy and measurements of bed topography, water surface elevation (WSE), flow characteristics and bottom composition. More specifically, this study pursued the following objectives: (1) Quantify the influence of the air–water interface on the upwelling radiance signal by recording reflectance spec- tra above and immediately below the water surface. *Correspondence to: C. J. Legleiter, Department of Geography, University of Wyoming, 1000 E. University Ave., Laramie, Wyoming 82071, USA. E-mail: Carl.Legleiter@uwyo.edu RIVER RESEARCH AND APPLICATIONS River Res. Applic. (2013) Published online in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/rra.2674 Copyright © 2013 John Wiley & Sons, Ltd.