Variability in fish and water hydrogen and oxygen stable isotope values in the nearshore region of a large water body

Abstract

Incorporating organismal hydrogen (δ²H or δD) and oxygen (δ¹⁸O) stable isotope ratios into aquatic food web research can help elucidate relative reliance on shoreline inputs and/or feeding at different trophic levels; however, their interpretation is complicated by the fact that aquatic organisms derive hydrogen and oxygen from both their diet and the ambient water in which they reside. We examined spatial and seasonal patterns in δ²H and δ¹⁸O of field-collected water and small-bodied fishes. Samples were gathered from nearshore Lake Michigan, a dynamic region of a large, increasingly oligotrophic freshwater lake. We examined seasonal and spatial δ²H and δ¹⁸O values of surface and bottom water, and compared spatial δ²H and δ¹⁸O values of young-of-year yellow perch (Perca flavescens) and seasonal δ²H and δ¹⁸O values of round goby (Neogobius melanostomus). Nearshore δ²H and δ¹⁸O water values were more variable than previously-described offshore values. Variation in nearshore water δ²H and δ¹⁸O values was likely related to differential precipitation and/or discharges from tributaries, and there may be a surface dilution effect which led to lower δ¹⁸O values in nearshore surface waters than bottom waters. In our study, fish tissue δ²H and δ¹⁸O also varies more spatially than seasonally, and some findings suggest that fish tissue δ²H may reflect feeding at higher trophic levels. Though characteristics of the study system affect their interpretation, we suggest that δ²H and δ¹⁸O can be another tool to assess food web structure, with δ²H in particular having potential to resolve questions when δ¹³C or δ¹⁵N are inconclusive.

Introduction

Hydrogen (δ²H or δD) and oxygen (δ¹⁸O) stable isotope ratios have been used to examine water movement and evaporation processes in a variety of aquatic systems (e.g., Genereux and Hooper, 1998, Pham et al., 2009). Though they have potential to increase understanding of aquatic food webs, particularly when assessing relative reliance on allochthonous or autochthonous inputs to a system (e.g., Finlay et al., 2010), their use in food web studies has thus far been relatively limited (reviewed by Vander Zanden et al., 2016). Laboratory studies (e.g., Soto et al., 2013) have demonstrated potential for δ²H to reflect feeding at different trophic levels, and δ²H may provide a complementary tracer to the more commonly used δ¹⁵N. Given that δ²H and δ¹⁸O help elucidate migration patterns in humans and terrestrial wildlife (e.g., Bowen et al., 2005, Ehleringer et al., 2008), they could potentially also help assess movement and site fidelity of aquatic organisms. However, a major issue complicating interpretation is that aquatic organisms derive hydrogen and oxygen from both their diet and the ambient water in which they reside (Soto et al., 2013, Vander Zanden et al., 2016; Coulter et al., 2017).

To help refine understanding of how δ²H and δ¹⁸O reflect diet and habitat of freshwater fishes, we examined seasonal and spatial patterns in δ²H and δ¹⁸O values of water and small-bodied fishes collected from nearshore Lake Michigan, USA (i.e., <15 m depth). Lake Michigan has become increasingly oligotrophic in the past decade in response to sweeping food web changes (Barbiero et al., 2012), and the nearshore region is highly heterogeneous and dynamic. Substrates in the northern and western regions of nearshore Lake Michigan are primarily rocky while the south and eastern regions are primarily sand (GLAHF, 2020); and locally, there may be patches of hard substrate embedded in softer substrates or vice versa (e.g., Creque et al., 2010, Great Lakes Aquatic Habitat Framework (GLAHF), 2020). Moreover, the magnitude and characteristics of allocthonous loadings are spatially variable; there are more and larger tributaries on the eastern side of the lake as opposed to the western side, and surrounding land cover is primarily urban or agricultural around the southern basin and forested in the north (Robertson and Saad, 2011). Tributary discharge rates and nutrient loadings vary not only spatially, but also temporally (e.g., Robertson and Saad, 2011), with generally greatest discharges in the spring. In addition, upwelling and downwelling events cause the nearshore regions to experience quick temperature changes along with in-or-outfluxes of water (Plattner et al., 2006). Upwelling events in particular are more common along the western side of the lake than the eastern (e.g., Höök et al., 2004, Plattner et al., 2006).

Past studies of δ²H and δ¹⁸O patterns in Lake Michigan water and fishes have contrasted values in the main lake against connected tributary environments (e.g., Dufour et al., 2008, Senegal et al., 2020) or examined the mass balance of the whole lake (e.g., Jasechko et al., 2014). While Jasechko et al. (2014) found limited spatial or temporal variation of δ²H and δ¹⁸O of water collected from offshore Lake Michigan, studies of Lake Michigan tributaries documented relatively great variation of water isotope values (e.g., Dufour et al., 2005, Senegal et al., 2020). In the current study, we explored a) spatial and seasonal trends in δ²H and δ¹⁸O of nearshore Lake Michigan water, including nearshore depth effects; b) correlations between water isotope ratios and δ²H and δ¹⁸O of fish collected in the same location; and c) patterns of sampling-event corrected fish δ²H and δ¹⁸O collected over different sites and seasons. For a subset of fish, we examined correlations between δ²H and δ¹⁸O and two measures more commonly used in food web studies: δ¹³C and δ¹⁵N. We expected that nearshore water δ²H and δ¹⁸O values would be more variable than offshore values documented by Jasechko et al. (2014); that, at a given location, runoff dilution effects would lead to surface water being depleted in the heavier isotope as compared to bottom water (e.g., Jameel et al., 2018); and that water and fish collected from sites close to larger tributaries would exhibit stronger precipitation and runoff signals than those collected near smaller tributaries (Doucett et al., 2007). We expected that both δ²H and δ¹⁸O of fish tissue would vary spatially and seasonally; however, with the lowest δ¹⁸O values observed during times of greatest tributary inflow and near northern tributaries with more ¹⁸O-depleted runoff, and competing effects of water and allochthonous food input producing less consistent patterns for δ²H.