Next Article in Journal
Risk Assessment of Coastal Flooding under Different Inundation Situations in Southwest of Taiwan (Tainan City)
Next Article in Special Issue
Thermal Characteristics of a Beaver Dam Analogues Equipped Spring-Fed Creek in the Canadian Rockies
Previous Article in Journal
Developing Pseudo Continuous Pedotransfer Functions for International Soils Measured with the Evaporation Method and the HYPROP System: II. The Soil Hydraulic Conductivity Curve
Previous Article in Special Issue
Riparian Vegetation Structure Influences Terrestrial Invertebrate Communities in an Agricultural Landscape
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Forested Riparian Zones Provide Important Habitat for Fish in Urban Streams

1
Norwegian Institute for Water Research (NIVA), 0349 Oslo, Norway
2
Institute of Landscape Ecology, University of Münster, 48149 Münster, Germany
3
Department of Aquatic Sciences and Assessment, Swedish University of Agricultural Sciences, 75007 Uppsala, Sweden
4
Aquatic Ecology Research Unit, Department of Animal Sciences and Aquatic Ecology, Ghent University, 9000 Ghent, Belgium
5
Helmholtz Centre for Environmental Research-UFZ, Department of Computational Landscape Ecology, 04318 Leipzig, Germany
6
Department of Systems Ecology and Sustainability, University of Bucharest, 050095 Bucharest, Romania
7
Research Institute of the University of Bucharest, 050663 Bucharest, Romania
8
Freshwater Biological Section, Department of Biology, University of Copenhagen, 2100 Copenhagen, Denmark
9
Water@leeds and School of Geography, University of Leeds, Leeds LS2 9JT, UK
*
Author to whom correspondence should be addressed.
Water 2021, 13(6), 877; https://doi.org/10.3390/w13060877
Submission received: 4 February 2021 / Revised: 5 March 2021 / Accepted: 13 March 2021 / Published: 23 March 2021
(This article belongs to the Special Issue Ecosystem Functioning in Rivers and Riparian Zones)

Abstract

:
Riparian zones form a boundary between aquatic and terrestrial ecosystems, with disproportionate influences on food web dynamics and ecosystem functioning in both habitats. However, riparian boundaries are frequently degraded by human activities, including urbanization, leading to direct impacts on terrestrial communities and indirect changes that are mediated through altered connectivity with adjacent aquatic ecosystems. We investigated how riparian habitat influences fish communities in an urban context. We electrofished nine urban site pairs with and without forested riparian buffers, alongside an additional 12 sites that were located throughout the river networks in the Oslo Fjord basin, Norway. Brown trout (Salmo trutta) were the dominant fish species. Riparian buffers had weak positive effects on fish densities at low to moderate levels of catchment urbanization, whereas fish were absent from highly polluted streams. Subtle shifts in fish size distributions suggested that riparian buffers play an important role in metapopulation dynamics. Stable isotopes in fish from buffered reaches indicated dietary shifts, pointing to the potential for a greater reliance on terrestrial-sourced carbon. Combining these results, we postulate that spatially-mediated ontogenetic diet shifts may be important for the persistence of brown trout in urban streams. Our results show that using a food web perspective is essential in understanding how riparian buffers can offset impacts in urban catchments.

1. Introduction

Streams and their adjacent riparian zones form tightly coupled networks in landscapes, which are characterized by intimate ecosystem and food web linkages across the land-water interface [1,2,3]. For instance, stream and terrestrial ecosystems are bound by reciprocal flows of organic matter and prey [3,4,5]. Reciprocal linkages include inputs of terrestrial detritus that help to fuel aquatic food webs, and emerging winged adults of aquatic insects as an important prey source for riparian consumers, such as spiders, ground beetles, and birds [6,7,8,9]. Riparian zones are biodiversity hotspots, providing essential habitat refugia for terrestrial and aquatic biota in the face of global change drivers, such as landscape disturbances and extreme weather events [2,10]. Streams and riparian habitats are important societally through the provision of key ecosystem services to humans (e.g., water purification, carbon storage, and fishing) [11,12]. However, human pressures from land use, such as urbanization, frequently degrade stream–riparian networks, with adverse impacts on cross-habitat linkages and ecosystem services [7,11,13].
Urbanization is a driver of global environmental change that disproportionately impacts freshwater ecosystems [14,15]. Streams draining urban catchments are often highly degraded, characterized by biodiversity losses and altered ecosystem functioning [16,17,18]. The ubiquity of these responses has led to the “urban stream syndrome” being a key focus of cross-disciplinary research [19]. Urban areas are often highly modified, including hugely transformed riparian zones that are typified by vegetation removal, imperviousness and soil degradation, channel modification (e.g., culverting, bank fixation, and channelization), and in-stream habitat deterioration, including greatly altered chemical, hydrological, and thermal dynamics [13,16]. These changes have led to the decoupling of the stream from terrestrial food webs [7] and declines in key organismal groups, including stream fishes.
Fish community responses to urbanization are less understood than for other biota [20], although there is evidence that urban land uses are associated with losses of sensitive species and increases in tolerant species leading to community homogenization [21,22,23]. Altered physical habitat, declines in water quality, and the loss of riparian cover have all been suggested as key proximate drivers of urban impacts on fish communities [20]. Many fish are also top predators in streams, and their loss may have profound effects on urban stream food webs by altering energy flows and destabilizing trophic interactions [24]. A decrease of fish in urban streams can reduce the amenity values for human populations and may lead to declines in recreational fishing, an important ecosystem service contributing to physical activity, stress-alleviation, and social interaction [25]. Consequently, there is an urgent need for management approaches for improving fish communities and environmental quality in urban stream–riparian networks, with increasing focus on the utility of forested riparian buffers to help conserve and rehabilitate degraded habitats [26].
Despite best management practices advocating a catchment wide approach for improving water quality and longitudinal connectivity in streams and rivers [27], there is a growing consensus that local remediation efforts, such as forested riparian buffers, offer multiple co-benefits in urban environments [28,29]. Riparian vegetation is particularly important for fish, as it affects light regimes, thermal dynamics, water quality, as well as habitat and food availability [30]. However, urban land use and riparian degradation frequently covary, which means that few studies have disentangled the relative importance of catchment urbanization from riparian land use [19,31,32]. Environmental contingencies, such as those arising from catchment-wide pressures, generate uncertainties that may explain why few regulations exist for forested buffers in stream–riparian networks. This challenge requires targeted case studies that focus on key food web compartments to develop the general framework that is needed for implementing forested buffers in human-impacted landscapes. Here, we wanted to investigate how top consumers (fish) in urban stream food webs respond to forested riparian buffers.
In our study, we investigated the effects of forested riparian buffers on fish communities in the urbanized Oslo Fjord basin (Norway). We sampled 30 sites over gradients of riparian land use and catchment urbanization. The stream sites ranged from having natural woody riparian vegetation to being completely devoid of a vegetated riparian zone. Sites were distributed within the river network, so the effect of catchment urbanization (e.g., urban impacts and altered longitudinal connectivity) could be assessed. Nested within the 30 were 18 site pairs, each with an upstream unbuffered reach and a nearby downstream site with a forested riparian buffer. We wanted to know whether forested riparian buffers confer benefits to stream fishes and if positive influences are environmental contingent (i.e., dependent on upstream catchment degradation).
Thus, we specifically tested three hypotheses:
  • Size distributions of fish differ with the presence of a riparian buffer. Our a priori expectation was that larger-sized fish would be more abundant in buffered sites compared to unbuffered, as there is more available habitat in terms of water depth and cover against predators.
  • Fish densities are positively related to improved riparian condition after accounting for upstream human impacts. To test this hypothesis, we applied the conceptual framework that was introduced by Burdon et al. [33], suggesting that responses to riparian buffers are determined by the extent of a local transition (i.e., improved riparian condition) and its interaction with the environmental context (i.e., the level of upstream catchment degradation).
  • The presence of a forested riparian buffer alters the energy flow paths to stream fishes. We predicted that the “brown” food web channel (i.e., driven by allochthonous carbon sources with cascading effects throughout the food web) would be more dominant in buffered sites. To test this, we analyzed carbon stable isotopes (δ13C) from stream fishes. Stable isotopes (e.g., nitrogen and carbon) are widely used to elucidate food web properties due to predictable changes in isotopic signatures between consumers and resources, where consumer tissues are enriched in the heavy isotope relative to the food resources in their diet [34].

2. Materials and Methods

2.1. Study Area and Study Design

The Norwegian capitol Oslo is located at the Inner Oslo Fjord (59′55 N, 10′45E). It is the most populous city in Norway with a current population of almost 700,000 inhabitants, and it is the fastest growing major city in Europe [35]. The population has increased by 20% between 2007 and 2018 [36] and large parts of the city are undergoing redevelopment to provide more residential areas by replacing previously industrialized areas. A major city planning strategy from the 1930s was to establish blue-green corridors throughout the city, connecting the inner Oslo Fjord to the surrounding upland forested areas along the stream channels. Oslo has a history of catchment modifications [37,38,39], thus representing a suitable urban environment for testing our hypotheses.
We investigated fish communities in 30 stream sites across Oslo and specifically addressed the effects of forested riparian buffers as a part of the BiodivErSA-funded CROSSLINK project (Figure 1) [27]. Eighteen of the sites were paired with and without forested riparian buffers at the scale of the sampling reach (i.e., nine pairs, each consisting of an upstream “unbuffered” and a downstream “buffered” site). The site pairs were always relatively close together (mean distance: 350 m) to ensure a similar stream size in terms of discharge with the upstream reach not having a riparian buffer. Paired sites were located in urban impacted landscapes. The buffered sites were selected to fulfill some key requirements: minimum buffer length exceeded 50 m upstream from the downstream end of the sampling reach, and the buffer width was 2–3 x the wetted stream width dominated by larger trees on both sides of the stream. Common tree species (Diameter at Breast Height, DBH > 5 cm) forming riparian buffers included the deciduous Alnus incana, Fraxinus excelsior, and Acer platanoides, whereas unbuffered sites were mostly open and dominated by grass, shrubs, and small trees [29]. Unbuffered sites usually only had a few trees (DBH > 5 cm) along the stream banks. The buffer widths ranged from 10–150 m and buffer lengths from 80–750 m. The additional sites were situated along the river continuum in the same systems, from headwaters to they entered the Oslo Fjord. “Reference” sites were typically located in the forested upstream catchment and constituting pristine, least impacted sites, while “Longitudinal” sites were located lower down in the stream-network, being subjected to accumulated levels of urbanization and with different configurations of the riparian buffer (Figure S1, Supplementary Materials).
To ensure consistency, the sampling reaches were wadeable, 1st–3rd order (i.e., approximately 2–5 m wide), and with a stable streambed (i.e., not frequently hydrodynamically disturbed) that is dominated by gravels and cobbles. The standardized length of our study sites was 30 m. The starting point of our surveys was the downstream end of each site, which, in the case of buffered sites, was located as far as downstream within the woody riparian buffer simultaneously having the largest lateral extent. For unbuffered sites, this meant that the location was chosen as far downstream in an unshaded/slightly shaded network position as possible.

2.2. Assessment of Riparian Habitat Conditions, Stream Width and Catchment Land Uses

The riparian habitat characteristics were assessed during field surveys along each study site using standardized field protocols. We used the Riparian Condition Index (RCI) that was introduced by Burdon et al. [28], which contains 13 variables to assess the quality of the riparian zone. The variables include channel shading, buffer width, buffer intactness, vegetation composition of buffer and adjacent land, groundcover of buffer and adjacent land, as well as bank stability, soil quality (denitrification potential and drainage), livestock access, land slope, and rills/channels. Briefly, observers rank aspects of the riparian zone from poor (1) to excellent (5) for each of the bank to indicate the quality and integrity of the riparian zone (Table S1, Supplementary Materials). The scores were summed to provide an index of riparian habitat quality (RCI). For the analysis of total riparian condition, the scores of the two banks (left and right) of the same site were averaged to provide a single value for riparian condition at each site.
We recorded cross-sectional measurements of wetted channel width at 5-6 transects that were distributed in a stratified random approach throughout each individual 30 m long study site and calculated stream area.
Data on upstream land use cover for stream catchments of each study site were obtained from the CORINE Land Cover Inventory [40].

2.3. Water Quality Data

We collected water samples in bottles during two different seasons (autumn 2017, summer 2018). The water samples were collected from below the water surface (i.e., 10 cm) in the channel thalweg at the downstream end of each site. Site pairs were sampled on the same day. Water samples were stored cold and refrigerated upon return to the laboratory. The water samples were analyzed in NIVAs accredited laboratory following international standards (https://www.niva.no/en/services/laboratory-services, accessed on 26 November 2020). Water samples were analyzed for total nitrogen (NS 4743:1993), ammonium (NH4N; [41]), nitrite- and nitrate-nitrogen (NO2N + NO3N; NS 4745:1991), total phosphorus (NS 4725:1984), dissolved reactive phosphorus (PO4P; NS 4724), specific conductivity, and pH (NS-EN ISO 10523). Specific conductivity was collected at the time of sampling while using a handheld instrument (WTW pH/Cond 3410i, Weilheim, Germany). For further details on water quality sampling, see [28].

2.4. Fish Survey and Stable Isotope Sampling

Thirty sites were sampled at base flow conditions in summer 2018. We electrofished sites at 230 V using a Terik Technologies FA4 generator (Levanger, Norway) to estimate the species composition and fish population density in each reach. All reaches were sampled using two independent runs with successive removals. The catch efficiencies were always sufficiently high to calculate population densities using the formulae from Serber & Le Cren [42]. The densities were both calculated per m stream and per m2 (Table S2, Supplementary Materials). Fish were measured (length in cm and weight in g) and identified before up to ten individuals of each species were fin-clipped for stable isotope analysis and returned to the stream upon completion of sampling. Samples for stable isotope analysis were stored on ice upon return to the laboratory. Fish were collected in accordance with ethical and legal standards. Oslo Municipality provided approval.

2.5. Stable Isotope Analysis (SIA)

The stable isotope samples were frozen and then freeze-dried (LyoDry compact, Mechatech systems LTD, Bristol, UK) for a minimum of 48 h at −45 °C. Freeze-dried samples were ground to a homogenous powder and individually encapsulated into tin capsules. The encapsulated samples were analyzed for their carbon and nitrogen ratios using a PDZ Europa ANCA-GSL elemental analyzer interfaced to a PDZ Europa 20-20 isotope ratio mass spectrometer (Sercon Ltd., Cheshire, UK) at the Stable Isotope Facility (University of California, Davis, CA, USA). For our data analyses, we focus on δ13C stable isotopes to track resource use of stream fish related to the presence/absence of a forest buffer (see below [43,44]).

2.6. Data Analysis

Our study investigated the effects of forested riparian buffers on fish communities in an urban context and tested whether positive impacts of woody vegetation are environmental contingent. To do this, we analyzed data from the CROSSLINK project on riparian integrity, fish communities, food web responses (carbon stable isotopes, δ13C), and catchment-wide human impacts (upstream land use and water quality) from the Oslo Fjord catchment.
1. Effects of vegetated riparian buffers on size distribution of stream fish were inspected visually using histogram plots and summary statistics (i.e., mean ± SD, median, skewness) for each site type (i.e., reference, buffered, unbuffered, and longitudinal sites).
2. To study a context dependent influence of forest buffers on fish, we first tested the relationship between the Riparian Condition Index (RCI) and fish abundance (i.e., number of fish per site).
Any variability that is caused by unknown factors has the potential to reduce the statistical power and sensitivity of biomonitoring indicators when comparing responses across human impact gradients [45]. Thus, the inherent weaknesses of observational data combined with inappropriate statistical analyses have limited the utility of field monitoring data for detecting and managing environmental stressors [46,47]. For riparian habitat, whilst poor conditions are expected to limit fish numbers, the same does not apply for good conditions, since other human impacts (e.g., water quality, hydrological connectivity) and natural phenomena (e.g., hydrodynamic disturbance, biotic interactions) introduce variability in the abundances of fish observed at different sites with similar riparian conditions. We used quantile regression to test whether fish abundances were limited by riparian conditions and if other factors introduce variability in the relationship between riparian habitat and fish abundances. Quantile regression is a statistical method used to estimate the conditional quantiles of a response variable distribution in a linear model to provide a more complete view of possible causal relationships between variables that are involved in ecological processes [48]. Quantile regression can be used to estimate the rate of change for any quantile (or percentile) of a response variable to a limiting factor [45,49]. Regression quantiles are ascending sequences of planes that are above an increasing proportion of sample observations as the values of the quantiles (t) increase [48]. This property of regression quantiles enables the estimation of the rate of change (slope) for any quantile of the data, and not just along the central tendency, as in ordinary least squares (OLS) regression [45]. We predicted that the upper percentiles of fish abundances would be limited by the riparian condition, whereas the lower percentiles could be limited by a suite of measured and unmeasured factors. In combination, these predictions would form a wedge-shaped relationship between the riparian condition and fish abundances. We estimated quantile regression coefficients and 90% confidence intervals for the 0.05, 0.25, 0.5, 0.75, and 0.95 percentiles using the quantreg R package. Hypothesis tests and CI were bootstrapped using the xy method with 200 replications. We also tested the relationship between riparian condition and fish abundances using a generalized linear mixed effects model (GLMM) assuming a Poisson distribution for count data. The GLMM differed from quantile regression by assessing the central tendency, as in ordinary least squares (OLS) regression, but accounts for the non-independence of Site as a random factor. The GLMM was fitted using the glmer function in the lme4 R package.
Our second approach adopted the conceptual framework that was introduced in Burdon et al. [28] and applied to our buffered/unbuffered site pairs. We used General Linear Models (GLM) to test the influence of riparian condition (RCI), upstream human activities (i.e., upstream catchment degradation “urbanization”), and their interaction on fish densities. This approach enabled us to test responses of fish communities to changes in riparian conditions, whilst accounting for upstream catchment degradation. We calculated the log response ratio, a common effect size metric that was calculated as log proportional change in the means of treatment and control group (e.g., buffered vs unbuffered), for fish density and RCI, with the former being the response variable (dFISH) and the latter a predictor (i.e., the magnitude of change from an “unbuffered” to a “buffered” riparian state, dRCI). In our model, we included fish density calculated per m stream, because our estimates of fish density calculated per area (m2) revealed that stream reach area was not independent of buffer/no buffer and it was likely to under-estimate population effects (Figure S2, Supplementary Materials). We used Principal Components Analysis (PCA) to describe upstream catchment degradation “urbanization” (i.e., environmental context). The PCA decomposed log-transformed water quality variables (i.e., total inorganic nitrogen, ammonium, nitrite- and nitrate-nitrogen, total phosphorus, dissolved reactive phosphorus, specific conductivity, and pH) and logit-transformed upstream land-use cover variables (i.e., % of the catchment area covered by urban, arable cropping, forest, natural features, water, wetlands, and other) into site scores (Axis 1, hereafter referred to PC1), explaining 56% of total variation (Table S3, Supplementary Materials). We scaled the PC1 scores by a constant (the min value) to help aid the interpretation of interactions.
3. Effects of forest buffers on food web properties were elucidated based on carbon stable isotopes (δ13C). We run linear mixed-effect models (LMM) to test whether buffers influence differences in fish-δ13C signatures whilst accounting for body size, and if buffers influence the relationship of body size with fish-δ13C signatures. One site pair (Mærradalsbekken; cf. Table S2, Supplementary Material) was stocked with hatchery fish that were fed during their rearing and thus might confound our expectations of isotopic signatures in natural populations of brown trout. Thus, we excluded this outlier site pair from our analyses. The LMMs included a varIdent function to enable different variances per stratum for site type [50], and we used site “pairs” as a random effect to account for non-independence of site pairs. The LMM was fitted with the lme function using the nlme package. We used the function testInteractions in the phia R package to test whether the size-δ13C relationship was significant in each site type. All tests were conducted in R [51].

3. Results

3.1. Fish Communities and Forested Buffers in Urban Streams

We collected a total of 1085 fish and a total biomass of 16,942 g from 24 sites across Oslo (no fish were present at six sites): 136 individuals in forested reference sites (a total biomass of 1321 g, n = 4), 239 individuals in unbuffered sites (total biomass of 4041 g, n = 6), 388 individuals in buffered sites (total biomass of 6665 g, n = 6), and 322 individuals in longitudinal sites (total biomass of 4915 g, n = 8). A total of 4 fish species were caught with brown trout (Salmo trutta) being the dominant species (N = 1026; Table S4, Supplementary Materials). The trout population in the streams was mainly composed of 0+, 1+ and 2+ year classes, based on the length measurements and the maximum size of the trout that were caught by electrofishing ranged from 13.3 to 28.1 cm in the individual streams. There were no differences in the mean size between site types, including the comparison between site pairs (F3,9 = 0.752, p = 0.5476), nor was there a general difference in the abundance of young of the year (YoY). However, we found a greater number of fish of the intermediate size class (~10 cm) in the buffered sites (Figure 2). This was evidenced by the greater median size in buffered sites, whereas the unbuffered sites had a more right-skewed size distribution that was driven by the larger sized individuals (Figure 2).

3.2. Context Dependent Influences of Forested Buffers on Stream Fish

We assessed if improved riparian conditions (i.e., forested buffers) can influence fish abundances and densities and whether a positive effect was contingent on the environmental context (i.e., upstream catchment degradation described as “urbanization”). An initial test showed that the riparian condition is important for fish numbers, with a significant positive influence of the Riparian Condition Index (RCI) on fish abundances (F1,15 = 23.4, p < 0.001, Table S5, Supplementary Materials). However, plotting the data indicated a “wedge-shaped” relationship (Figure 3), which we tested using quantile regression. The upper percentiles (0.95, 0.75) showed a significant positive influence of riparian condition (p < 0.05), but the non-significant relationship for lower percentiles indicated the potential for other factors to limit fish abundances (p > 0.05). Several sites with a high RCI scores had low fish abundances, or, in some cases, no fish present (Figure 3, Table S2, Supplementary Materials). Subsequently, we tested the interaction between the change in fish densities (log response ratio between paired buffered/unbuffered sites, LRR), the improvement of riparian condition (LRR), and upstream human activities (PC1). We found evidence for an interaction between catchment urbanization and riparian integrity, where the positive influence of forested buffers on fish densities at low and moderate pollution (PC1) becomes neutral at high pollution (Figure 4, Table 1).

3.3. Effects of Forested Buffers on Urban Stream Food Webs

To test our hypothesis (3) that the “brown” food web channel (i.e., driven by allochthonous carbon sources) would be more dominant in buffered sites, we analyzed δ13C from a total of 244 biological samples (i.e., fin clips) that were collected from stream fishes during the field surveys (Salmo trutta: n = 219, Salvelinus fontinalis: n = 24, Phoxinus phoxinus: n = 1). Overall, fish from buffered sites generally encompassed a narrower δ13C range (from −27.5 ‰ to −23.3 ‰) than fish from unbuffered sites (from −30.4 ‰ to −24.5 ‰) and shifted towards more enriched δ13C—i.e., the detrital-driven “brown” food web channel (median buffered: −25.5 ‰ vs. median unbuffered: −26.8 ‰; Figure 5 and Figure 6). Interestingly, δ13C from our forested reference sites showed a bimodal distribution with a smaller peak at −29.0 ‰ and the larger peak at −25.5 ‰. The isotopic composition of fish from longitudinal sites showed the largest variability and ranged from −30.8 ‰ to −22.9 ‰.
We saw evidence for two food web pathways leading to the largest brown trout in our streams. These pathways differed between the buffered and unbuffered sites after accounting for fish size in our statistical models (Figure 6, Table 2). One pathway characterized by depleted δ13C values typical of algal-derived carbon was predominant in the unbuffered sites, while the enriched δ13C values, characteristic of terrestrial-derived carbon, was more representative of the buffered sites. The relationship between fish length and δ13C was statistically significant for both buffered and unbuffered sites (p < 0.001), but differed in their slopes (Figure 6, Table 2). This shift in isotopic signatures between the buffered vs. unbuffered sites was accompanied by the greater number of intermediate-sized fish (~10 cm) in the buffered sites making forest buffers a source of these fish (Figure 2, Figure 6). The larger bodied fish (> 17 cm) showed less fidelity to buffers and a convergence in isotopic signatures (Figure 6).

4. Discussion

Urbanization is a driver of global environmental change leading to multiple impacts on watersheds, including the decoupling of the stream from semi-terrestrial and terrestrial food webs [7] and declines in key organismal groups, including stream fishes [21,22,23]. Although best management practices have advocated a catchment-wide approach for improving habitat quality in streams and rivers, there is a growing recognition that local restoration efforts, such as forested riparian buffers, offer multiple co-benefits in urban environments [28]. Forested riparian buffers provide crucial terrestrial habitat, help to reverse the “urban stream syndrome”, and offer a nature-based solution to offset the impacts of climate change [28,29]. We sought to explore the effects of forested riparian buffers on stream fish in the heavily urbanized Oslo Fjord basin, Norway.

4.1. Effects of Forest Buffers on Stream Fish in an Urban Context

We found, in concordance with other studies, weak positive effects of forested riparian vegetation on fish populations [32,52]. A positive effect on fish abundances seemed to be limited by other factors, as a number of sites with a high Riparian Condition Index (RCI) had low fish densities, or, in some cases, no fish present (Figure 3). The “wedge-shaped” relationship between RCI and fish abundances highlights this assumption [53]. We expected that other human impacts from urbanization, such as source pollution interferes with, and, in some cases, completely disrupts, stream-riparian buffer interactions, thereby limiting their potentially positive effects on fish populations. Urban land uses have been shown in several studies to be detrimental to fish populations [21,22,23,54,55] as well as other aquatic biota [16,17,56,57,58]. Given the presence of point-source and diffuse pollution in our urban streams with measured impacts on indicators of stream health [29], it is highly likely that water quality is a major limiting factor for populations of fish species, like brown trout. A test of the “environmental context” hypothesis [28] supported our assumption, as higher fish densities in the presence of a forested riparian buffer declined when the upstream catchment became too heavily degraded (both in terms of water quality and upstream land use, PC1). Given that some streams were fishless in our study (at the upper end of our urbanization gradient, PC1), we ended up with a relative low number of sites for our analyses and, hence, a limited statistical power. However, we believe that our ability to detect a positive effect using this modest dataset provides a useful indicator of the potential value of riparian buffers in an urban context.
More sites across different urban contexts are needed to verify our findings, given highly populated watersheds frequently experience multiple pressures and stressors in addition to biogeographical differences in fish assemblages. However, the benefits of additional replication with independent sites are constrained due to the finite spatial extent of urban areas and the differing intensities of urban land-uses. Increasing the sample size would require careful attention to context-driven differences in responses. For instance, in the Oslo Fjord catchment, we would not have been able to add further sampling sites without potentially confounding the study design by choosing low-impact peri-urban sites or alternatively, risking pseudo replication by increasing the number of non-independent study sites.
Another complementary approach that could help to disentangle context-driven (e.g., different combinations of stressors) effects on aquatic biota are experiments using mesocosms [59]. Mesocosms can reduce environmental complexity while maintaining essential background conditions of natural systems and providing the inferential power of replicated experiments. Most stream-side mesocosm experiments have focused on disentangling stressor interactions in agricultural landscapes [60,61,62], but this approach can also be used to investigate urban stressors [63]. Future studies in urban streams can further utilize the advantages of mesocosm experiments to quantify stressor impacts on aquatic biota and the mechanisms underpinning potential solutions that are related to riparian vegetation (e.g., thresholds for riparian vegetation types after which the positive effect of riparian buffers are lost).
Our data also suggest that measures of population densities per meter stream length is a better indicator than a measure per sampling area of the positive effects of riparian zones. We attribute this to the fact that riparian zones have a number of modifying effects on stream morphology [28], one being an increase in stream width as compared with non-buffered stream reaches [64,65,66] (Figure S2, Supplementary Materials). Hence, using an area-based measure is likely to under-estimate population effects.

4.2. Forest Buffers foster Tangled Stream-Riparian Food Webs in Urban Areas

Our findings add to the oft-quoted principle that “you are what you eat” by suggesting “you are where you eat”. We posit that the absence of a forest buffer resulted in a dietary shift towards the putative algal-driven “green” food web channel, as indicated by more 13C depleted stable isotopes in fish. We speculate that trout were more dependent on smaller grazing invertebrates, including chironomids in unbuffered stream reaches (Figure 7). However, more C13 enriched stable isotopes in fish from buffered sites pointed to the potential for greater reliance on terrestrial-sourced carbon (i.e., the “brown” food web channel). We speculate that the generally larger sized detritivores (shredders such as limnephilid caddisflies utilizing allochthonous CPOM) makes them a more favorable food for trout [67]. Trout are known to be size-specific predators and the higher overall energetic benefits of preying on large food items, within (in the upper end) the prey handling size range of an individual, have been shown in a number of studies [68].
Forested riparian buffers may have increased the proportions of terrestrial invertebrates in the diet of trout. Fluxes of terrestrial invertebrates falling to the water surface constitute a secondary allochthonous source of energetically beneficial prey for stream fish and their availability is directly linked to the presence of overhanging riparian vegetation [7,69]. Allochthony has been shown to help stabilize food webs in the face of environmental stress [70] and support food webs after restoration [71]. We believe that our results further suggest that forest buffers may help to stabilize urban stream food webs by providing a detrital “slow” energy channel in addition to the “fast” energy channel based on algae [72,73].
We also found evidence for subtle shifts in trout size distributions which suggesting that riparian buffers play an important role in metapopulation dynamics: intermediate-sized fish (~10 cm) were more abundant in buffered reaches as compared to the upstream unbuffered sites. This size class might be more sensitive to the presence of overhead cover [74], as their intermediate size makes them more susceptible to visual predators, which include bird species (Phalacrocorax spp.) [75]. The presence of riparian buffers may have encouraged the intermediate-sized fish to have greater site fidelity, whilst the larger sized fish were more mobile with larger home ranges [76,77,78].
Our statistical analyses showed that the δ13C of trout was influenced by the interaction between site type and fish body size. The larger bodied fish (>17 cm) in our study showed a convergence in isotopic signatures indicative of a similar diet in this size class (Figure 6 and Figure 7). We posit that these larger fish had more similar diets due to their larger home ranges and less fidelity to a single site type. Alternatively, their diets may have become more similar due to the size structure of prey assemblages, with a preference for larger-bodied caddisflies and terrestrial invertebrates available in the drift at both site types. Either way, it seems likely that prey from riparian buffers are important for these larger fish due to their more 13C enriched isotopic signatures.
Combining our abundance results with the stable isotopes signatures of brown trout, we postulate that spatially-mediated ontogenetic diet shifts may be important for the persistence of trout in urban streams [24]. The detrital “slow” energy channel that is provided by the presence of woody riparian vegetation appears to be an important feature of trout metapopulation dynamics, because these habitat types supported the highest abundances of intermediate-sized fish, and stable isotopes of larger fish from both site types converged on a carbon isotopic signature consistent with detrital and terrestrial food sources. The peak in abundances of intermediate-sized fish in reaches with forest buffers suggest that buffered sites are a source of fish at this life stage. However, future research should test the relative importance of ecosystem size (e.g., wetted channel width), habitat refugia (i.e., overhead cover and reduced predation risk), and resource availability (i.e., aquatic detritivores and terrestrial inputs of prey) in driving the abundances of trout in urban streams. Our findings emphasize that using a food web perspective is essential for better understanding how riparian buffers can offset impacts in urban catchments. This is consistent with studies highlighting the need to incorporate measures of ecosystem functioning into environmental assessments to better understand pressure–response relationships [79,80,81].
We observed an interesting pattern in fish carbon isotopic signatures at our forest reference sites. The forest sites showed a bimodal distribution with a smaller cluster at −29‰ indicating a strong reliance on putative autochthonous food sources (Figure 5). This cluster was dominated by fish that were sampled from the same study site. This site was influenced by an upstream lake, meaning that phyto- and zooplankton drifting into the study reach likely enabled a more depleted 13C isotopic signature to be expressed in the receiving stream food web.

5. Conclusions

Urbanization impacts stream ecosystems globally, causing biodiversity losses, altered system functioning, and reduced service provisioning. Consequently, there are growing efforts to rehabilitate and restore degraded urban stream-riparian networks [82]. Addressing the “urban stream syndrome” requires multiple mitigation tools. The enhancement and conservation of forested riparian buffers with their multiple benefits offers a cost-efficient solution in already degraded urban freshwater ecosystems as well as a key habitat feature to preserve when new areas are urbanized. Our study demonstrates that riparian buffers have the potential to mitigate the effects of urbanization on fish to a certain degree of environmental impact. However, with increasing urban impacts, the positive influence is de-coupled when habitat conditions become too heavily degraded in terms of upstream catchment deterioration (water quality and land use).
We also advance our knowledge of urban stream ecosystems using a food web perspective to reveal the importance of detrital and terrestrial carbon sources that are provided by riparian buffers. Our results suggest that spatially-mediated ontogenetic diet shifts may be important for the persistence of brown trout in urbanized streams of the Oslo Fjord basin.
The modest improvements that are suggested here may grow in magnitude when added to the broad portfolio of benefits provided by forested riparian buffers [28,29,83,84,85]. However, maximizing positive outcomes for fish in urban streams requires coordinating the placement of buffers with other management strategies, including controlling and mitigating point-source pollution. To conceptualizing and planning the optimal configuration of forested riparian zones at the catchment scale (i.e., width, length, position in the stream network), however, requires context-specific considerations and the forecasting of different scenarios. Future urban planning should implement modelling frameworks, such as CoMOLA [86,87], to optimize the spatial allocation and extent of riparian zones along urban stream networks to promote biodiversity, ecosystem functioning, and services to enhance human well-being, thus ultimately helping to create sustainable and livable cities of the future.

Supplementary Materials

The following are available online at https://www.mdpi.com/2073-4441/13/6/877/s1; Figure S1: Experimental design in the Oslo Fjord Basin with pristine or least-impacted ‘Reference’ sites (blue), site pairs with an ‘Unbuffered’ upstream site (orange) and a downstream ‘Buffered’ site (green) with a woody riparian buffer on both banks, and ‘Longitudinal’ sites (red) that were typically located further downstream from the other site types to capture cumulative land use impacts; Figure S2: Boxplots of channel widths from buffered and unbuffered sites in the Oslo Fjord Basin; Table S1: Description of scores for riparian attributes used to calculate the Riparian Condition Index; Table S2: Information about buffered/unbuffered sites including RCI, total number of fish per site, density per m reach, and density per m2; Table S3: Variables and loadings for Principal Components Analysis (PCA) decomposing catchment land use and water quality variables into a single index of urbanization for all 30 sites in the Oslo basin. PC1 explained 56.3% of the variation among sites, whereas PC2 explained 16.5%; Table S4: Fish species and their abundances from study sites across the Oslo Fjord basin; Table S5: Results from the Generalised Linear Mixed model testing the relationship between the Riparian Condition Index (RCI) and fish numbers.

Author Contributions

Conceptualization, B.K., F.J.B., B.G.M. and N.F.; methodology, all authors; software, B.K. and F.J.B.; validation, B.K., F.J.B., B.G.M. and N.F.; formal analysis, B.K. and F.J.B.; investigation, B.K., P.T.M., J.T., J.H. and N.F.; resources, F.J.B., B.G.M. and N.F.; data curation, B.K., F.J.B., B.G.M. and N.F.; writing—original draft preparation, B.K.; writing—review and editing, all authors; visualization, B.K. and F.J.B.; supervision, F.W., P.G., G.R., B.G.M. and N.F.; project administration, B.K., F.J.B., M.A.E.F., P.G., G.R., B.G.M. and N.F.; funding acquisition, P.G., G.R., B.G.M. and N.F.; All authors have read and agreed to the published version of the manuscript.

Funding

This research was conducted as part of the CROSSLINK project, titled “Understanding cross-habitat linkages between blue and green infrastructure to optimize management of biodiversity, ecosystem services and multiple human uses”. The CROSSLINK project was funded through the 2015-2016 BiodivERsA COFUND call for research proposals, with national funders: the Swedish Research Council for Sustainable Development (FORMAS, project 2016-01945) and the Swedish Environmental Protection Agency; The Research Council of Norway (NFR, project 264499); The Research Foundation of Flanders (FWO, project G0H6516N), Belgium; the Romanian National Authority for Scientific Research and Innovation, CCCDI—UEFISCDI, project BiodivERsA3-2015-49-CROSSLINK, within PNCDI III; and the Federal Ministry of Education and Research (BMBF, Germany, project FKZ: 01LC1621A).

Institutional Review Board Statement

The study was conducted in accordance with ethical and legal standards, and approved by Oslo Municipality (23 May 2018).

Informed Consent Statement

Not Applicable.

Data Availability Statement

Data supporting the results are available in the Supplementary Material.

Acknowledgments

We would like to thank James Edward Sample for providing the land use data. We are grateful to many colleagues who assisted in the field: Eivind Ekholt Andersen, Teis Friberg, Birk Fogde Ørnskov, Benoit Demars, and Markus Lindholm assisted in the field. Ellinor Ramberg helped process samples for stable isotope analysis. We would also like to thank the anonymous reviewers for their comments that improved the manuscript.

Conflicts of Interest

The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

References

  1. Gregory, S.V.; Swanson, F.J.; McKee, W.A.; Cummins, K.W. An ecosystem perspective of riparian zones: Focus on links between land and water. Bioscience 1991, 41, 540–551. [Google Scholar] [CrossRef]
  2. Naiman, R.J.; Decamps, H. The ecology of interfaces: Riparian zones. Annu. Rev. Ecol. Syst. 1997, 28, 621–658. [Google Scholar] [CrossRef] [Green Version]
  3. Polis, G.A.; Anderson, W.B.; Holt, R.D. Toward an intergration of landscape and food web ecology: The dynamics of spatially subsidized food webs. Annu. Rev. Ecol. Syst. 1997, 28, 289–316. [Google Scholar] [CrossRef] [Green Version]
  4. Power, M.E.; Rainey, W.E. Food webs and resource sheds: Towards spatially delimiting trophic interactions. In The Ecological Consequences of Environmental Heterogeneity; Hutchings, M.J., John, E.A., Stewart, A.J., Eds.; Blackwell Science: London, UK, 2000; pp. 291–314. [Google Scholar]
  5. Nakano, S.; Murakami, M. Reciprocal subsidies: Dynamic interdependence between terrestrial and aquatic food webs. Proc. Natl. Acad. Sci. USA 2001, 98, 166–170. [Google Scholar] [CrossRef] [Green Version]
  6. Vannote, R.L.; Minshall, G.W.; Cummins, K.W.; Sedell, J.R.; Cushing, C.E. The river continuum concept. Can. J. Fish. Aquat. Sci. 1980, 37, 130–137. [Google Scholar] [CrossRef]
  7. Baxter, C.V.; Fausch, K.D.; Saunders, C.W. Tangled webs: Reciprocal flows of invertebrate prey link streams and riparian zones: Prey subsidies link stream and riparian food webs. Freshw. Biol. 2005, 50, 201–220. [Google Scholar] [CrossRef]
  8. Hering, D.; Plachter, H. Riparian ground beetles (Coeloptera, Carabidae) preying on aquatic invertebrates: A feeding strategy in alpine floodplains. Oecologia 1997, 111, 261–270. [Google Scholar] [CrossRef]
  9. Paetzold, A.; Schubert, C.J.; Tockner, K. Aquatic Terrestrial Linkages Along a Braided-River: Riparian Arthropods Feeding on Aquatic Insects. Ecosystems 2005, 8, 748–759. [Google Scholar] [CrossRef]
  10. Naiman, R.J.; Decamps, H.; Pollock, M. The role of riparian corridors in maintaining regional biodiversity. Ecol. Appl. 1993, 3, 209–212. [Google Scholar] [CrossRef] [PubMed]
  11. Hanna, D.E.L.; Raudsepp-Hearne, C.; Bennett, E.M. Effects of land use, cover, and protection on stream and riparian ecosystem services and biodiversity. Conserv. Biol. 2020, 34, 244–255. [Google Scholar] [CrossRef] [PubMed]
  12. Riis, T.; Kelly-Quinn, M.; Aguiar, F.C.; Manolaki, P.; Bruno, D.; Bejarano, M.D.; Clerici, N.; Fernandes, M.R.; Franco, J.C.; Pettit, N.; et al. Global Overview of Ecosystem Services Provided by Riparian Vegetation. Bioscience 2020, 70, 501–514. [Google Scholar] [CrossRef]
  13. Allan, J.D. Landscapes and riverscapes: The influence of land use on stream ecosystems. Annu. Rev. Ecol. Evol. Syst. 2004, 35, 257–284. [Google Scholar] [CrossRef] [Green Version]
  14. Grimm, N.; Faeth, S.H.; Golubiewski, N.E.; Redman, C.L.; Wu, J.; Bai, X.; Briggs, J.M. Global Change and the Ecology of Cities. Science 2008, 319, 756–760. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  15. Vörösmarty, C.J.; McIntyre, P.B.; Gessner, M.O.; Dudgeon, D.; Prusevich, A.; Green, P.; Glidden, S.; Bunn, S.E.; Sullivan, C.; Liermann, C.R.; et al. Global threats to human water security and river biodiversity. Nature 2010, 467, 555–561. [Google Scholar] [CrossRef] [PubMed]
  16. Paul, M.J.; Meyer, J.L. Streams in the urban landscape. Annu. Rev. Ecol. Syst. 2001, 32, 333–365. [Google Scholar] [CrossRef]
  17. King, R.S.; Baker, M.E.; Kazyak, P.F.; Weller, D.E. How novel is too novel? Stream community thresholds at exceptionally low levels of catchment urbanization. Ecol. Appl. 2011, 21, 1659–1678. [Google Scholar] [CrossRef]
  18. Wiederkehr, F.; Wilkinson, C.L.; Zeng, Y.; Yeo, D.C.; Ewers, R.M.; O’Gorman, E.J. Urbanisation affects ecosystem functioning more than structure in tropical streams. Biol. Conserv. 2020, 249, 108634. [Google Scholar] [CrossRef]
  19. Walsh, C.J.; Roy, A.H.; Feminella, J.W.; Cottingham, P.D.; Groffman, P.M.; Morgan, R.P. The urban stream syndrome: Current knowledge and the search for a cure. J. North. Am. Benthol. Soc. 2005, 24, 706–723. [Google Scholar] [CrossRef]
  20. Meador, M.R. Historical changes in fish communities in urban streams of the south-eastern United States and the relative importance of water-quality stressors. Ecol. Freshw. Fish. 2019, 29, 156–169. [Google Scholar] [CrossRef]
  21. Wang, L.; Lyons, J.; Kanehl, P.; Bannerman, R.; Emmons, E. Watershed urbanization and changes in fish communities in south- eastern Wisconsin streams. J. Am. Water Resour. Assoc. 2000, 36, 1173–1189. [Google Scholar] [CrossRef]
  22. Scott, M.C.; Helfman, G.S. Native invasions, homogenization, and the mismeasure of integrity of fish assemblages. Fisheries 2001, 26, 6–15. [Google Scholar] [CrossRef]
  23. Walters, D.M.; Leigh, D.S.; Bearden, A.B. Urbanization, sedimentation, and the homogenization of fish assemblages in the Etowah River Basin, USA. Hydrobiologia 2003, 494, 5–10. [Google Scholar] [CrossRef]
  24. Rooney, N.; McCann, K.; Gellner, G.; Moore, J.C. Structural asymmetry and the stability of diverse food webs. Nature 2006, 442, 265–269. [Google Scholar] [CrossRef] [PubMed]
  25. Melstrom, R.T.; Lupi, F.; Esselman, P.C.; Stevenson, R.J. Valuing recreational fishing quality at rivers and streams. Water Resour. Res. 2015, 51, 140–150. [Google Scholar] [CrossRef]
  26. Coelho, D.; Hughes, S.J.; Varandas, S.; Cortes, R.M.V. Conservation benefits of riparian buffers in urban areas: The case of the Rio Corgo (north Portugal). Fundam. Appl. Limnol. Archiv Hydrobiol. 2014, 185, 55–70. [Google Scholar] [CrossRef]
  27. Bernhardt, E.S.; Palmer, M.A. River restoration: The fuzzy logic of repairing reaches to reverse catchment scale degradation. Ecol. Appl. 2011, 21, 1926–1931. [Google Scholar] [CrossRef] [PubMed]
  28. Burdon, F.J.; Ramberg, E.; Sargac, J.; Forio, M.A.E.; De Saeyer, N.; Mutinová, P.T.; Moe, T.F.; Pavelescu, M.O.; Dinu, V.; Cazacu, C.; et al. Assessing the Benefits of Forested Riparian Zones: A Qualitative Index of Riparian Integrity Is Positively Associated with Ecological Status in European Streams. Water 2020, 12, 1178. [Google Scholar] [CrossRef] [Green Version]
  29. Mutinová, P.T.; Kahlert, M.; Kupilas, B.; McKie, B.G.; Friberg, F.; Burdon, F.J. Benthic Diatom Communities in Urban Streams and the Role of Riparian Buffer. Water 2020, 12, 2799. [Google Scholar] [CrossRef]
  30. Zalewski, M.; Thorpe, J.E.; Naiman, R.J. Fish and riparian ecotones—A hypothesis. Ecohydrol. Hydrobiol. 2001, 1, 11–24. [Google Scholar]
  31. Walsh, C.J.; Waller, K.A.; Gehling, J.; Mac Nally, R. Riverine invertebrate assemblages are degraded more by catchment urbanisation than by riparian deforestation. Freshw. Biol. 2007, 52, 574–587. [Google Scholar] [CrossRef]
  32. Roy, A.H.; Freeman, B.J.; Freeman, M.C. Riparian influences on stream fish assemblage structure in urbanizing streams. Landsc. Ecol. 2007, 22, 385–402. [Google Scholar] [CrossRef]
  33. Burdon, F.J.; Reyes, M.; Alder, A.C.; Joss, A.; Ort, C.; Räsänen, K.; Jokela, J.; Eggen, R.I.L.; Stamm, C. Environmental context and magnitude of disturbance influence trait-mediated community responses to wastewater in streams. Ecol. Evol. 2016, 6, 3923–3939. [Google Scholar] [CrossRef] [Green Version]
  34. Gannes, L.Z.; Obrien, D.M.; del Rio, C.M. Stable isotopes in animal ecology: Assumptions, caveats, and a call for more laboratory experiments. Ecology 1997, 78, 1271–1276. [Google Scholar] [CrossRef]
  35. Statistikkbanken. Population Statistics: Annually, Estimated Figures. Available online: https://www.ssb.no/en/befolkning/statistikker/folkemengde/aar-berekna (accessed on 26 November 2020).
  36. Oslo Kommune. Statistikkbanken. Available online: http://Statistikkbanken.oslo.kommune.no (accessed on 26 November 2020).
  37. Allan, I.; Jenssen, M.T.S.; Braaten, H.F.V. Priority Substances and Emerging Contaminants in Selected Norwegian Rivers—The River Monitoring Programme 2018; Norwegian Institute for Water Research: Oslo, Norway, 2019. [Google Scholar]
  38. European Commission. Oslo—European Green Capital 2019; Publications Office of the European Union: Luxembourg, 2018; Available online: https://ec.europa.eu/environment/europeangreencapital/wp-content/uploads/2019/Oslo%20Brochure_EGCA%202019.pdf (accessed on 26 November 2020).
  39. Oslo Kommune. Faktaark—Oslos Byvasdrag. Available online: https://usercontent.one/wp/www.osloelveforum.org/wp-content/uploads/2015/10/Faktaark-Oslos-byvassdrag-2017.pdf (accessed on 26 November 2020).
  40. CLC. Corine Land Cover (CLC) Inventory; European Environment Agency, EEA: Copenhagen, Denmark, 2018; Available online: https://land.copernicus.eu/ (accessed on 26 November 2020).
  41. Kerouel, R.; Aminot, A. Fluorometric determination of ammonia in sea and estuarine waters by direct segmented flow analysis. Mar. Chem. 1997, 57, 265–275. [Google Scholar] [CrossRef]
  42. Serber, G.A.F.; Le Cren, E.D. Estimating population parameters from catches large relative to the population. J. Anim. Ecol. 1967, 36, 631–643. [Google Scholar] [CrossRef]
  43. Post, D.M. Using stable isotopes to estimate trophic position: Models, methods and assumptions. Ecology 2002, 83, 703–718. [Google Scholar] [CrossRef]
  44. Brauns, M.; Boëchat, I.G.; de Carvalho, A.P.; Graeber, D.; Gücker, B.; Mehner, T.; von Schiller, D. Consumer-resource stoichiometry as predictor of trophic dis- crimination (Δ13C, Δ15N) in aquatic invertebrates. Freshw. Biol. 2018, 63, 1240–1249. [Google Scholar] [CrossRef]
  45. Schmidt, T.S.; Clements, W.H.; Cade, B.S. Estimating risks to aquatic life using quantile regression. Freshw. Sci. 2012, 31, 709–723. [Google Scholar] [CrossRef]
  46. Pacheco, M.A.W.; McIntyre, D.O.; Linton, T.K. Integrating chemical and biological criteria. Environ. Toxicol. Chem. 2005, 24, 2983–2991. [Google Scholar]
  47. Linton, T.K.; Pacheco, M.A.W.; McIntyre, D.O.; Clement, W.H.; Goodrich-Mahoney, J. Development of bioassessment-based benchmarks for iron. Environ. Toxicol. Chem. 2007, 26, 1291–1298. [Google Scholar] [CrossRef] [PubMed]
  48. Cade, B.S.; Noon, B.R. A Gentle Introduction to Quantile Regression for Ecologists. Front. Ecol. Environ. 2003, 1, 412–420. [Google Scholar] [CrossRef]
  49. Burdon, F.J. Agriculture and Mining Contamination Contribute to a Productivity Gradient Driving Cross-Ecosystem Associations Between Stream Insects and Riparian Arachnids; Contaminants and Ecological Subsidies: The Land-Water Interface; Kraus, J.M., Walters, D.M., Mills, M.A., Eds.; Springer: Cham, Switzerland, 2020. [Google Scholar]
  50. Zuur, A.F.; Ieno, E.N.; Walker, N.J.; Saveliev, A.A.; Smith, G.M. Mixed Effects Models and Extensions in Ecology with R; Springer: New York, NY, USA, 2009. [Google Scholar]
  51. R Core Team. A Language and Environment for Statistical Computing; R Foundation for Statistical Computing: Vienna, Austria, 2019; Available online: http://www.R-project.org/ (accessed on 26 November 2020).
  52. Roy, A.H.; Faust, C.L.; Freeman, M.C.; Meyer, J.L. Reach-scale effects of riparian forest cover on urban stream ecosystems. Can. J. Fish. Aquat. Sci. 2005, 62, 2312–2329. [Google Scholar] [CrossRef]
  53. Friberg, N. Pressure-response relationships in stream ecology: Introduction and synthesis. Freshw. Biol. 2010, 55, 1367–1381. [Google Scholar] [CrossRef]
  54. Marchetti, M.P.; Lockwood, J.L.; Light, T. Effects of urbanization on California’s fish diversity: Differenti- ation, homogenization and the influence of spatial scale. Biol. Conserv. 2006, 127, 310–318. [Google Scholar] [CrossRef]
  55. Morgan, R.P.; Cushman, S.F. Urbanization effects on stream fish assemblages in Maryland, USA. J. N. Am. Benthol. Soc. 2005, 24, 643–655. [Google Scholar] [CrossRef]
  56. Cuffney, T.F.; Brightbill, R.A.; May, J.T.; Waite, I.R. Responses of benthic macroinvertebrates to environmental changes associated with urbanization in nine metropolitan areas. Ecol. Appl. 2010, 20, 1384–1401. [Google Scholar] [CrossRef]
  57. Lundquist, M.J.; Weixing, Z. Aquatic insect diversity in streams across a rural–urban land-use discontinuum. Hydrobiologia 2019, 837, 15–30. [Google Scholar] [CrossRef]
  58. Mackintosh, T.J.; Davis, J.A.; Thompson, R.M. The influence of urbanisation on macroinvertebrate biodiversity in constructed stormwater wetlands. Sci. Total Environ. 2015, 536, 527–537. [Google Scholar] [CrossRef]
  59. Birk, S.; Chapman, D.; Carvalho, L.; Spears, B.M.; Andersen, H.E.; Argillier, C.; Auer, S.; Baattrup-Pedersen, A.; Banin, L.; Beklioğlu, M. Impacts of multiple stressors on freshwater biota across spatial scales and ecosystems. Nat. Ecol. Evol. 2020, 4, 1060–1068. [Google Scholar] [CrossRef]
  60. Piggott, J.J.; Salis, R.K.; Lear, G.; Townsend, C.R.; Matthaei, C.D. Climate warming and agricultural stressors interact to determine stream periphyton community composition. Glob. Chang. Biol. 2015, 21, 206–222. [Google Scholar] [CrossRef]
  61. Piggott, J.J.; Niyogi, D.K.; Townsend, C.R.; Matthaei, C.D. Multiple stressors and stream ecosystem functioning: Climate warming and agricultural stressors interact to affect processing of organic matter. J. Appl. Ecol. 2015, 52, 1126–1134. [Google Scholar] [CrossRef]
  62. Davis, S.J.; Oh Uallacháin, D.; Mellander, P.-E.; Matthaei, C.D.; Piggott, J.J.; Kelly-Quinn, M. Chronic nutrient inputs affect stream macroinvertebrate communities more than acute inputs: An experiment manipulating phosphorus, nitrogen and sediment. Sci. Total Environ. 2019, 683, 9–20. [Google Scholar] [CrossRef]
  63. Burdon, F.J.; Yaohui, B.; Reyes, M.; Tamminen, M.; Staudacher, P.; Mangold, S.; Singer, H.; Räsänen, K.; Joss, A.; Tiegs, S.D.; et al. Stream microbial communities and ecosystem functioning show complex responses to multiple stressors in wastewater. Glob. Chang. Biol. 2020, 26, 6363–6382. [Google Scholar] [CrossRef]
  64. Sweeney, B.W.; Bott, T.L.; Jackson, J.K.; Kaplan, L.A.; Newbold, J.D.; Standley, L.J.; Hession, W.C.; Horwitz, R.J. Riparian deforestation, stream narrowing, and loss of stream ecosystem services. Proc. Natl. Acad. Sci. USA 2004, 101, 14132–14137. [Google Scholar] [CrossRef] [Green Version]
  65. Quinn, J.M.; Cooper, A.B.; Davies-Colley, R.J.; Rutherford, J.C.; Williamson, R.B. Land use effects on habitat, water quality, periphyton, and benthic invertebrates in Waikato, New Zealand, hill-country streams. N. Z. J. Mar. Freshw. Res. 1997, 31, 579–597. [Google Scholar] [CrossRef]
  66. Turunen, J.; Elbrecht, V.; Steinke, D.; Aroviita, J. Riparian forests can mitigate warming and ecological degradation of agricultural headwater streams. Freshw. Biol. 2021, 66, 1–14. [Google Scholar] [CrossRef]
  67. Jensen, H.; Kahilainen, K.K.; Amundsen, P.-A.; Gjelland, K.O.; Tuomaala, A.; Malinen, T.; Bøhn, T. Predation by brown trout (Salmo trutta) along a diversifying prey community gradient. Can. J. Fish. Aquat. Sci. 2008, 65, 1831–1841. [Google Scholar] [CrossRef]
  68. Sih, A.; Christensen, B. Optimal diet theory: When does it work and when does it fail? Anim. Behav. 2001, 61, 379–390. [Google Scholar] [CrossRef]
  69. Mason, C.F.; MacDonald, S.M. The input of terrestrial invertebrates from tree canopies to a stream. Freshw. Biol. 1982, 12, 305–311. [Google Scholar] [CrossRef]
  70. Kraus, J.M.; Pomeranz, J.F.; Todd, A.S.; Walters, D.M.; Schmidt, T.S.; Wanty, R.B. Aquatic pollution increases use of terrestrial prey subsidies by stream fish. J. Appl. Ecol. 2015, 53, 44–53. [Google Scholar] [CrossRef] [Green Version]
  71. Kupilas, B.; Friberg, N.; McKie, B.G.; Jochmann, M.A.; Lorenz, A.W.; Hering, D. River restoration and the trophic structure of benthic invertebrate communities across 16 European restoration projects. Hydrobiologia 2016, 769, 105–120. [Google Scholar] [CrossRef]
  72. Moore, J.C.; Berlow, E.L.; Coleman, D.C.; de Ruiter, P.C.; Dong, Q.; Hastings, A.; Johnson, N.C.; McCann, K.S.; Melville, K.; Morin, P.J.; et al. Detritus, trophic dynamics and biodiversity. Ecol. Lett. 2004, 7, 584–600. [Google Scholar] [CrossRef]
  73. Rooney, N.; McCann, K.S.; Moore, J.C. A landscape theory for food web architecture. Ecol. Lett. 2008, 11, 867–881. [Google Scholar] [CrossRef]
  74. Penaluna, B.E.; Dunham, J.B.; Noakes, D.L.G. Instream cover and shade mediate avian predation on trout in semi-natural streams. Ecol. Freshw. Fish. 2015, 25, 405–411. [Google Scholar] [CrossRef]
  75. Barrett, R.T.; Røv, N.; Loen, J.; Montevecchi, W.A. Diets of shags Phalacrocorax aristotelis and cormorants P. carbo in Norway and possible implications for gadoid stock recruitment. Mar. Ecol. Prog. Ser. 1990, 66, 205–218. [Google Scholar] [CrossRef]
  76. Clapp, D.E.; Clark, R.D.; Diana, J.S. Range, activity, and habitat of large, free-ranging brown trout in a Michigan stream. Trans. Am. Fish. Soc. 1990, 119, 1022–1103. [Google Scholar] [CrossRef]
  77. Young, M.K. Brown trout mobility in south-central Wyoming streams. Can. J. Zool. 1994, 72, 2078–2083. [Google Scholar] [CrossRef]
  78. Young, M.K. Summer movements and habitat use by Colorado River cutthroat trout (Oncorhynchus clarki pleuriticus) in a small, montane stream. Can. J. Fish. Aquat. Sci. 1996, 53, 1403–1408. [Google Scholar] [CrossRef]
  79. Palmer, M.A.; Febria, C.M. The heartbeat of ecosystems. Science 2012, 336, 1393–1394. [Google Scholar] [CrossRef] [PubMed]
  80. Friberg, N. Impacts and indicators of change in lotic ecosystems. WIREs Water 2014, 1, 513–531. [Google Scholar] [CrossRef]
  81. Jackson, M.; Weyl, O.L.F.; Altermatt, F.; Durance, I.; Friberg, N.; Dumbrell, A.J.; Piggott, J.J.; Tiegs, S.D.; Tockner, K.; Krug, C.B.; et al. Recommendations for the next generation of global freshwater biological monitoring tools. Adv. Ecol. Res. 2016, 55, 615–636. [Google Scholar] [CrossRef] [Green Version]
  82. European Commission. Towards an EU Research and Innovation Policy Agenda for Nature-Based Solutions & Re-Naturing Cities; European Commission: Brussels, Belgium, 2015. [Google Scholar]
  83. Ramberg, E.; Burdon, F.J.; Sargac, J.; Kupilas, B.; Rîsnoveanu, G.; Lau, D.C.P.; Johnson, R.K.; McKie, B.G. The Structure of Riparian Vegetation in Agricultural Aquatic-Terrestrial Linkages. Water 2020, 12, 2855. [Google Scholar] [CrossRef]
  84. Forio, M.A.E.; De Troyer, N.; Lock, K.; Witing, F.; Baert, L.; De Saeyer, N.; Rîsnoveanu, G.; Popescu, C.; Burdon, F.J.; Kupilas, B.; et al. Small Patches of Riparian Woody Vegetation Enhance Biodiversity of Invertebrates. Water 2020, 12, 3070. [Google Scholar] [CrossRef]
  85. Popescu, C.; Oprina-Pavelescu, M.; Dinu, V.; Cazacu, C.; Burdon, F.J.; Forio, M.A.E.; Kupilas, B.; Friberg, N.; Goethals, P.; McKie, B.G.; et al. Riparian Vegetation Structure Influences Terrestrial Invertebrate Communities in an Agricultural Landscape. Water 2021, 13, 188. [Google Scholar] [CrossRef]
  86. Strauch, M.; Cord, A.; Pätzold, C.; Sven, L.; Kaim, A.; Christian, S.; Seppelt, R.; Volk, M. Constraints in multi-objective optimization of land use allocation—repair or penalize? Environ. Model. Softw. 2019, 118, 241–251. [Google Scholar] [CrossRef]
  87. Schwarz, N.; Hoffmann, F.; Knapp, S.; Strauch, M. Synergies or Trade-Offs? Optimizing a Virtual Urban Region to Foster Plant Species Richness, Climate Regulation, and Compactness Under Varying Landscape Composition. Front. Environ. Sci. 2020, 8, 16. [Google Scholar] [CrossRef] [Green Version]
Figure 1. Location of study sites along the stream-riparian networks in the Oslo Fjord basin, Norway.
Figure 1. Location of study sites along the stream-riparian networks in the Oslo Fjord basin, Norway.
Water 13 00877 g001
Figure 2. Histogram plots showing the distribution of brown trout (Salmo trutta) body lengths (cm) among the four site types: Pristine or least-impacted ‘Reference’ sites (=Forest) that were typically located further upstream due to urbanization in lower reaches, site pairs with an ‘Unbuffered’ upstream site and a downstream ‘Buffered’ site with a woody riparian buffer on both banks, and ‘Longitudinal’ sites that were typically located further downstream from the other site types to capture cumulative impacts of urban land uses. Mean and SD, median, and skewness of each site type are given in each panel.
Figure 2. Histogram plots showing the distribution of brown trout (Salmo trutta) body lengths (cm) among the four site types: Pristine or least-impacted ‘Reference’ sites (=Forest) that were typically located further upstream due to urbanization in lower reaches, site pairs with an ‘Unbuffered’ upstream site and a downstream ‘Buffered’ site with a woody riparian buffer on both banks, and ‘Longitudinal’ sites that were typically located further downstream from the other site types to capture cumulative impacts of urban land uses. Mean and SD, median, and skewness of each site type are given in each panel.
Water 13 00877 g002
Figure 3. Slopes from estimated quantile regression coefficients for the 0.05, 0.25, 0.50, 0.75, and 0.95 percentiles. Solid lines indicate significant relationships (p < 0.05), whereas dashed lines indicate non-significant relationships (p < 0.05). The results show a “wedge-shaped” pattern of variation in fish abundances relative to the Riparian condition Index (RCI), where good riparian condition places an upper constraint on fish numbers, but other factors limit fish densities, even with good riparian habitat.
Figure 3. Slopes from estimated quantile regression coefficients for the 0.05, 0.25, 0.50, 0.75, and 0.95 percentiles. Solid lines indicate significant relationships (p < 0.05), whereas dashed lines indicate non-significant relationships (p < 0.05). The results show a “wedge-shaped” pattern of variation in fish abundances relative to the Riparian condition Index (RCI), where good riparian condition places an upper constraint on fish numbers, but other factors limit fish densities, even with good riparian habitat.
Water 13 00877 g003
Figure 4. Context dependent fish responses to forest buffers in the Oslo Fjord basin. Increasing urbanization (PC1) weakened the potential for riparian buffer (dRCI) to have a positive effect on fish densities (dFISH). PC1 explains 56% variation in catchment-wide human impacts. Each point represents a site pair (unbuffered and buffered sites) in the Oslo Fjord basin.
Figure 4. Context dependent fish responses to forest buffers in the Oslo Fjord basin. Increasing urbanization (PC1) weakened the potential for riparian buffer (dRCI) to have a positive effect on fish densities (dFISH). PC1 explains 56% variation in catchment-wide human impacts. Each point represents a site pair (unbuffered and buffered sites) in the Oslo Fjord basin.
Water 13 00877 g004
Figure 5. Distribution of δ13C signals across the four site types in Oslo, Norway: forested reference sites (n = 4; isotope samples: 40), unbuffered (n = 6; isotope samples: 53), buffered (n = 6; isotope samples: 60) and longitudinal (n=8; isotope samples: 91) sites across Oslo. In each box plot, the center line indicates the median, the box limits indicate the upper and lower quartiles, the whiskers indicate the 1.5× interquartile range.
Figure 5. Distribution of δ13C signals across the four site types in Oslo, Norway: forested reference sites (n = 4; isotope samples: 40), unbuffered (n = 6; isotope samples: 53), buffered (n = 6; isotope samples: 60) and longitudinal (n=8; isotope samples: 91) sites across Oslo. In each box plot, the center line indicates the median, the box limits indicate the upper and lower quartiles, the whiskers indicate the 1.5× interquartile range.
Water 13 00877 g005
Figure 6. Relationship between isotope signatures (δ13C) and fish body size (cm) whilst accounting for site type (unbuffered and buffered, N = 93 biological samples). Separate regression lines for buffered (green) and unbuffered (orange) sites. The density plots along margins show differences in the distribution of values (top, δ13C; right, fish length) between the buffered and unbuffered sites.
Figure 6. Relationship between isotope signatures (δ13C) and fish body size (cm) whilst accounting for site type (unbuffered and buffered, N = 93 biological samples). Separate regression lines for buffered (green) and unbuffered (orange) sites. The density plots along margins show differences in the distribution of values (top, δ13C; right, fish length) between the buffered and unbuffered sites.
Water 13 00877 g006
Figure 7. Conceptual model showing the two food web channels to the top: the algal-driven “green” food web channel in unbuffered sites (left) and the “brown” food web channel relying on terrestrial-sourced carbon (right). The peak in abundances of intermediate-sized fish in reaches with forest buffers suggest that buffered sites are a source of fish at this life stage. Larger fish from both site types converged on a diet consistent with detrital and terrestrial food sources.
Figure 7. Conceptual model showing the two food web channels to the top: the algal-driven “green” food web channel in unbuffered sites (left) and the “brown” food web channel relying on terrestrial-sourced carbon (right). The peak in abundances of intermediate-sized fish in reaches with forest buffers suggest that buffered sites are a source of fish at this life stage. Larger fish from both site types converged on a diet consistent with detrital and terrestrial food sources.
Water 13 00877 g007
Table 1. Results from the General Linear Models (GLM) testing the influence of riparian condition (RCI), upstream human activities (i.e., upstream catchment degradation “urbanization”), and their interaction on fish densities. The model use log response ratios to describe the magnitude of change in the response (dFISH) and predictor (dRCI) variables between site-pairs (i.e., unbuffered upstream sites and downstream, buffered sites) whilst controlling for the existing level of ecological impairment (i.e., upstream PC1). PC1 is the Axis 1 sites scores from a Principal Components Analysis (PCA) explaining 56% variation in catchment-wide human impacts. CI, 95% confidence interval.
Table 1. Results from the General Linear Models (GLM) testing the influence of riparian condition (RCI), upstream human activities (i.e., upstream catchment degradation “urbanization”), and their interaction on fish densities. The model use log response ratios to describe the magnitude of change in the response (dFISH) and predictor (dRCI) variables between site-pairs (i.e., unbuffered upstream sites and downstream, buffered sites) whilst controlling for the existing level of ecological impairment (i.e., upstream PC1). PC1 is the Axis 1 sites scores from a Principal Components Analysis (PCA) explaining 56% variation in catchment-wide human impacts. CI, 95% confidence interval.
Response PredictorsEstimatesCIp
dFISH(Intercept)2.430.14–4.710.04
dRCI−9.54−20.32–1.240.07
PC1−1.16−2.74–0.410.12
dRCI * PC15.07−1.93–12.060.12
Observations9
Nagelkerke R20.62
Table 2. Results from the mixed models testing the influence of riparian buffer on carbon stable isotopes (δ13C) in fish, whilst accounting for fish body size, and testing if buffers influence the relationship of body size with fish δ13C signatures. Rm2, marginal R2; Rc2, conditional R2.
Table 2. Results from the mixed models testing the influence of riparian buffer on carbon stable isotopes (δ13C) in fish, whilst accounting for fish body size, and testing if buffers influence the relationship of body size with fish δ13C signatures. Rm2, marginal R2; Rc2, conditional R2.
ResponsePredictorsnumDFdenDFF-Valuep-ValueRm2Rc2
δ13C(Intercept)18110,809<0.0010.6110.907
Site type146.5640.062
Body size (mm)181118.0<0.001
Site type: Body size1818.75<0.001
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Kupilas, B.; Burdon, F.J.; Thaulow, J.; Håll, J.; Mutinova, P.T.; Forio, M.A.E.; Witing, F.; Rîșnoveanu, G.; Goethals, P.; McKie, B.G.; et al. Forested Riparian Zones Provide Important Habitat for Fish in Urban Streams. Water 2021, 13, 877. https://doi.org/10.3390/w13060877

AMA Style

Kupilas B, Burdon FJ, Thaulow J, Håll J, Mutinova PT, Forio MAE, Witing F, Rîșnoveanu G, Goethals P, McKie BG, et al. Forested Riparian Zones Provide Important Habitat for Fish in Urban Streams. Water. 2021; 13(6):877. https://doi.org/10.3390/w13060877

Chicago/Turabian Style

Kupilas, Benjamin, Francis J. Burdon, Jens Thaulow, Johnny Håll, Petra Thea Mutinova, Marie Anne Eurie Forio, Felix Witing, Geta Rîșnoveanu, Peter Goethals, Brendan G. McKie, and et al. 2021. "Forested Riparian Zones Provide Important Habitat for Fish in Urban Streams" Water 13, no. 6: 877. https://doi.org/10.3390/w13060877

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop