River ecosystems are sustained by various water sources, such as springs and runoffs, which form heterogeneous physicochemical environments (Lusardi et al., 2021). The source-associated heterogeneity of river networks is expected to enhance not only the beta diversity (Reiss et al., 2016; Sakai et al., 2021a), but also the gamma diversity of animal communities by driving interspecific interactions among the distinctive water sources (e.g., Katahira et al., 2017; Sakai et al., 2021b, in press). Therefore, understanding the spatiotemporal distribution pattern of animals between spring-fed and runoff streams is important for maintaining the health of river networks.

Spring-fed streams formed by permanent groundwater discharge are widespread worldwide (Cantonati et al., 2012). While spring-fed streams draining from karst and volcanic terrains often show flashy discharge (Stevens et al., 2021), those in other terrains with lower water permeability generally have highly stable temperature and water discharge rates (Sear et al., 1999; Sakai et al., 2021a). While the great water permeability in karst and volcanic terrains often forms a number of spring-fed streams in a single river system, other terrains generally have sparse spring-fed streams. Thus, studies on spring-fed streams have primarily advanced in karst and volcanic terrains for replicable sampling design (e.g., Mattson et al., 1995; Lusardi et al., 2016).

Meanwhile, highly stable temperature and flow regimes of the sparse spring-fed streams may particularly be important to provide heterogeneous river environments. For example, the stable temperature and flow regime lead to the formation of unique animal communities (Reiss et al., 2016; Sun et al., 2020; Sakai et al., 2021a) and may also provide temporal habitats, such as refuges for mobile animals from environmental extremes (Inoue and Ishigaki, 1968; Kollaus and Bonner, 2012; Sakai et al., in press). In particular, the stable flow regime in spring-fed streams induces fine sediment settlements and thereby increases the populations of burrowing detritivores (Sakai et al., 2021a), resulting in higher abundance of macroinvertebrates compared to runoff streams (Füreder et al., 2001; Barquín and Death, 2006; Lusardi et al., 2016). This higher abundance of macroinvertebrates may imply the potential importance of spring-fed streams for foraging habitats for fishes. Thus, this study focused on such sparse, but highly stable spring-fed streams as important ecosystems in river networks.

While roles of spring-fed streams in providing thermal refuges were well recognized (Morita et al., 2011; Kemp et al., 2017; Koizumi et al., 2017; Wang et al., 2020; Togaki et al., 2023), how the unique animal communities in spring-fed streams provide food resources for fishes is largely unexplored. The present study aimed to examine the potential functions of sparse spring-fed streams in a clastic terrain (primarily composed of sandstone and mudstone), one of the dominant geological conditions in Japan, by investigating the seasonal changes in population density and stomach contents of juvenile Oncorhynchus masou masou in adjacent spring-fed and runoff tributaries. Although this study lacked replications for spring-fed streams, the results can contribute to understanding how the sparse, but highly stable spring-fed streams function in providing both thermal refuges and foraging habitats for fishes.

MATERIALS AND METHODS

Study site.—This study was conducted in the Shubuto River System in Kuromatsunai Town, Hokkaido, northern Japan (42.64°N, 140.34°E). The basin area encompasses 367 km² of montane and lowland regions. The geological composition of the area is sandstone and mudstone, including fossil seashells and tuff from the Cainozoic. The mean annual precipitation and air temperature from 2009 to 2018, measured at the nearby Kuromatsunai Automated Meteorological Data Acquisition System weather station (4 km northwest of the study site), were 1,615.8 mm and 7.5°C, respectively. The dominant tree species in riparian zones of the study area are Salix spp. and Quercus crispula, and the dominant understory plants are Sasa kurilensis and Reynoutria sachalinensis. The Shubuto River System includes a perennial spring-fed tributary with significant groundwater discharge (ca. 0.1 m³/s) in the lowland region (Fig. 1).

We selected two adjacent tributary reaches (spring-fed and runoff) in the lowland area as our study sites (Fig. 1). Each study reach was 40 m in length. No dams or weirs prevent the migration of fish among the sites. The tributaries had few heterogeneous structures such as riffle-pools, step-pools, or cascades.

The Shubuto River System harbors three species of Oncorhynchus (O. keta, O. masou masou, and O. mykiss; Miyazaki, 2017). Oncorhynchus mykiss is an alien species absent from the study streams, while the other two native species are present in the study streams. Oncorhynchus masou masou is one of the most predominant fish in the study streams, and its population includes both anadromous and resident individuals; most females and some males become anadromous after approximately two years of growth in the river (Miyazaki, 2017). Oncorhynchus keta is anadromous and generally runs up the Shubuto River from mid-September, where it spawns until mid-November, after 2–8 years of growth in the North Pacific Ocean (Miyazaki, 2017). While O. keta rarely migrates to the runoff tributary during their spawning season, this species migrates annually to the spring-fed tributary.

Environmental condition and hydrological monitoring.—Wet-ted width, water depth, and current velocity in each study reach were measured under base-flow conditions. Current velocity and water depth were measured 12 times (3 m interval) using an electromagnetic current meter (VE10; Kenek Co., Tokyo, Japan) and folding scale within the study reaches between 24–29 August, 8–10 October, and 17–20 November 2017, and 29–30 May 2018. Wetted width was measured five times (10 m intervals) using a measuring tape at the same time as the current velocity and water depth measurements. Water level and temperature loggers (HOBO CO-U20L-04; Onset, Bourne, MA, USA) were fixed to the streambeds of each study reach using metal tubes and wire, and hydraulic pressure and temperature were recorded hourly. We preliminarily recorded water levels at the monitoring points using a folding scale and constructed linear regression models to analyze the relationship between water level (cm) and hydraulic pressure (kPa). Based on these models, hydraulic pressure data recorded by the loggers were transformed into water level data. For the study reaches, the water level at 1800 hrs on 25 July 2018 was set as 0 cm, and hydrographs of subsequent water level and temperature changes were constructed up to 1800 hrs on 25 July 2019.

Benthic macroinvertebrates.—We sampled benthic macroinvertebrates, drifted macroinvertebrates, and fishes between 24–29 August, 8–10 October, and 17–20 November 2017, and 29–30 May 2018, to assess the seasonal changes in prey biomass and population density of juvenile O. masou masou. We used Surber nets (25 ×ばつ 25 cm, 0.5 mm mesh sieve) to sample benthic macroinvertebrates in four randomly placed quadrats per reach in each season (4 quadrats ×ばつ 2 reaches ×ばつ 4 seasons). Prior to sampling the macroinvertebrates, fine sediment cover within each quadrat of the Surber nets was estimated using the method of Sakai et al. (2013, 2021a). Sampled benthic materials were placed into a plastic tray, and macroinvertebrates with particulate matter were thoroughly rinsed in non-woven bags. The samples were preserved immediately in 70% ethanol. Macroinvertebrates were separated and identified to the order level with a stereomicroscope (SZ61; Olympus, Tokyo, Japan). Samples from each order were dried in an oven (FC-610; Advantec, Tokyo, Japan) at 60°C for 24 h and weighed to estimate benthic macroinvertebrate biomass (mg/m²).

Drifted macroinvertebrates.—Four driftnets (25 ×ばつ 25 cm, 0.5 mm mesh sieve) were positioned slightly upstream of the upper limit of each study reach in the early morning and dusk during each season (4 driftnets ×ばつ 2 reaches ×ばつ 2 times of day ×ばつ 4 seasons). The duration of sampling was approximately 1 hour, and the volume of water flowing through the driftnets was estimated from the duration of sampling, water depth, and current velocity, measured using an electromagnetic current meter (VE10; Kenek Co., Tokyo, Japan) at the opening of the driftnet. The drifted macroinvertebrate samples were preserved immediately in 70% ethanol, identified to the order level, and then dried and weighed using the same procedures applied for benthic macroinvertebrates. Drifted macroinvertebrate biomass was expressed as density (mg/m³), calculated from the macroinvertebrate dry weight and water volume flowing through the driftnets.

Fish.—We conducted electrofishing using a backpack electrofisher (LR-20; Smith-Root Inc., Vancouver, WA, USA) and hand nets in the study reaches. Collection was performed in both the early morning and at dusk during each season. During previous collection efforts conducted either in the early morning or at dusk, only the stomach contents of juvenile O. masou masou were sampled, whereas subsequent collection events involved both stomach content sampling and estimation of the population density of juvenile O. masou masou. Whether the survey was conducted first in the early morning or at dusk depended on the schedule. Stomach contents were randomly extracted from 15 individuals selected from each sample (15 individuals ×ばつ 2 reaches ×ばつ 2 times of day ×ばつ 4 seasons) using stomach pumps. If a sample did not contain 15 or more individuals, stomach contents were extracted from all individuals in the sample. Standard body length was measured for each individual.

To estimate the population density of juvenile O. masou masou, the upper and lower limits of the reaches were gently partitioned using nets to prevent fishes from entering or exiting the reach during collection. Three-pass electrofishing was conducted from lower to upper positions to collect as many individuals of O. masou masou as possible. Mean wetted width multiplied by reach length (40 m) was used as survey area for accurate calculation of fish population density (individuals/m²; Sakai et al., 2021b). After the fish surveys, all individuals were released back into their original habitats.

Statistical analysis.—The population density of juvenile O. masou masou in the tributary reaches was estimated using the maximum likelihood method proposed by Zippin (1956), as the number of collected individuals decreased substantially between the first and third sampling events. Differences in the total biomass and biomasses of each order of benthic and drifted macroinvertebrates between the tributaries were tested using Welch's t-test in each sampling event. The effect of tributary type (spring-fed and runoff) on stomach contents of juvenile O. masou masou was tested using analyses of covariance (ANCOVAs), with tributary type as factor and body size as covariate. The ANCOVAs were performed for total biomass and biomasses of each order of prey in the stomach contents in each sampling period. The biomass data were log-transformed to normalize distributions and standardize variance structures prior to the statistical analyses. All statistical procedures were performed using R software (ver. 4.2.1; R Core Team, 2022), with the FSA (Ogle et al., 2020) and car (Fox et al., 2022) packages.

RESULTS

The runoff and spring-fed tributaries had similar wetted width and water depth, but the current velocity was slower in the spring-fed tributary (Table 1). The hydrological monitoring results indicated that the flow and temperature regimes in the spring-fed tributary were highly stable compared to the runoff reach (Fig. 2). Hydrographs of the runoff reach showed dramatic water level increases, especially during typhoon season (August to September) and snowmelt season (April to May), whereas the spring-fed reach did not show such peaks (Fig. 2A). The water temperature range was 4.5–11.6°C in the spring-fed tributary, and 0–21.8°C in the runoff tributary (Fig. 2B).

Table 1

Mean ± standard deviation values of environmental variables measured in each study reach.

The proportion of fine sediment cover on the streambed was markedly greater in the spring-fed tributary than in the runoff tributary (Table 1). In addition, total benthic macroinvertebrate biomass was one order of magnitude greater in the spring-fed tributary than in the runoff tributary, and the macroinvertebrate assemblage of the spring-fed tributary was characterized by large biomass of dipterans across all four seasonal sampling events (Table 2). The dipterans were mostly chironomid midges (Pagastia spp.), which burrow in fine sediments and consume detritus (collector-gatherers; Merritt et al., 2008). Amphipods (Sternomoera yezoensis), which were absent in the runoff tributary, were found in the spring-fed tributary in all the sampling events and their biomass was significantly greater in the spring-fed tributary than in the runoff tributary in May (Table 2). The runoff tributary showed significantly greater biomass of herbivorous ephemeropterans including Epeorus spp. and Baetis spp. (Merritt et al., 2008) across the sampling events (Table 2).

Table 2

Mean ± standard deviation values (mg/m²) of total biomass and biomasses of benthic macroinvertebrates. Bold characters indicate statistically significant differences (Welch's t-test, P < 0.05).

In some seasons, the biomasses of drifted dipterans and amphipods were significantly greater in the spring-fed tributary and that of drifted ephemeropterans were significantly greater in the runoff tributary, similarly to the results on benthic macroinvertebrates (Table 3). However, total biomass of drifted macroinvertebrates was not significantly different between the tributaries in all the sampling events (Table 3).

Table 3

Mean ± standard deviation values (mg/m³) of total biomass and biomasses of drifted macroinvertebrates. Bold characters indicate statistically significant differences (Welch's t-test, P < 0.05).

Seasonal changes in the population density of juvenile O. masou masou were similar between the tributaries, with higher values in August and November, but densities were one order of magnitude higher in the spring-fed than runoff tributary across the sampling events (Fig. 3). In particular, the population density in the spring-fed tributary in November was remarkably high (5.2 individuals/m²).

While the total biomass of stomach contents of juvenile O. masou masou was greater in the spring-fed tributary than runoff tributary in May, those did not differ between the tributaries in the other sampling events (Table 4, Fig. 4). Dipterans and amphipods were consumed more in the spring-fed tributary than runoff tributary in August and May, and eggs of O. keta were consumed significantly more in the spring-fed tributary than runoff tributary in October and November (Table 4). Meanwhile in the runoff tributary, some terrestrial prey (hemipterans and hymenopterans) were consumed significantly more than in the spring-fed tributary in October and November (Table 4).

Table 4

Mean ± standard deviation values (mg/m³) of total biomass and biomasses of each order in stomach contents of juvenile Oncorhynchus masou masou. Bold characters indicate statistically significant differences (analysis of covariance, P < 0.05).

DISCUSSION

The spring-fed tributary had a more stable flow regime compared to the runoff tributary. The bottom of the spring-fed tributary was primarily covered with fine sediments, suggesting that the stable flow regime caused settlement of fine particles on the streambed, in agreement with Lusardi et al. (2016) and Sakai et al. (2021a). Furthermore, across the seasonal sampling events, the remarkably higher benthic macroinvertebrate biomass observed in the spring-fed tributary was supported by the predominance of burrowing chironomid detritivores, similar to other studies (Füreder et al., 2001; Sakai et al., 2021a). Further, the substantial numbers of dipterans drifted in the spring-fed tributary and were consumed by juvenile O. masou masou in August and May. These results suggest that the stable flow regime in spring-fed tributaries can provide attractive habitats for burrowing macroinvertebrates and foraging habitats for fishes.

The spring-fed tributary also harbored an amphipod species (Sternomoera yezoensis) that is considered to be a crenobiont or crenophilous species (Kuribayashi et al., 1994; Kusano and Ito, 2004), and the species was also consumed by juvenile O. masou masou in August and May. Generally, because such amphipods may be stenothermal (Sun et al., 2020) and vulnerable to flush (Barquín and Death, 2004; Meyer et al., 2007), the stable temperature and flow regimes found in the spring-fed tributary may allow such crenobiont or crenophilous species to be available as food resources for fishes.

In October and November, juvenile O. masou masou changed their primary food resource to eggs of O. keta in the spring-fed tributary. It is well known that some O. keta preferentially spawn their eggs in upwelling water habitats, including spring-fed streams (Kobayashi, 1968; Milligan et al., 1984; Geist et al., 2002; Clawson et al., 2022). Thus, the preferential aggregation of the redds of O. keta in the spring-fed tributary can potentially enhance the food availability of juvenile O. masou masou (Sakai et al., 2021b). In the spring-fed tributary, a facility had previously been constructed to capture mature O. keta taking advantage of their migration to irreversibly trap them in an artificial pool. Trapped individuals are then transported to another tributary of the Shubuto River System for egg collection and fry rearing. Thus, due to the one-way transport of individuals, eggs found in the spring-fed tributary are presumed to have spawned within the tributary, and not to have drifted from the facility. However, because the reared fry are released back to the spring-fed tributary, the population of O. keta likely includes both wild and hatchery-reared individuals. It remains unknown if preferential migration and spawning of wild O. keta into spring-fed streams triggers the aggregation of mobile consumers, such as juvenile O. masou masou under natural condition lacking hatchery stocking programs.

In the runoff tributary, more ephemeropterans were sometimes consumed by juvenile O. masou masou than in the spring-fed tributary, but not across the sampling events as observed in benthic biomass. Meanwhile, more terrestrial invertebrates including hemipterans and hymenopterans were consumed compared to the spring-fed tributary in October and November. Because the runoff tributary did not possess abundant aquatic prey biomass, like dipterans and eggs of O. keta in the spring-fed tributary, the population of O. masou masou in the runoff tributary might depend more on resource subsidies from terrestrial environments.

Although biomass of benthic macroinvertebrates was one order of magnitude greater in the spring-fed tributary than runoff tributary across the seasonal sampling events, that of drifted macroinvertebrates was similar between the tributaries. Because population density of juvenile O. masou masou was also one order of magnitude greater in the spring-fed tributary, the contradiction observed between benthic and drifted macroinvertebrates is attributable to the difference of predation pressure by O. masou masou between the tributaries. Thus, the unique habitat characteristics of spring-fed tributaries possibly enhance the carrying capacity, relevant to foraging habitats for stream fishes.

In the summer season, the cooler spring-fed tributary with abundant aquatic prey was presumably suitable thermal refuge for juvenile O. masou masou. Mayama (1992) and Sato et al. (2001) reported that juvenile O. masou masou stagnates foraging activity at 18°C and 24°C, respectively. The maximum temperature in the runoff tributary was 22°C, which was much higher than the spring-fed tributary (12°C), suggesting that cooler spring-fed streams may be suitable for avoiding stagnation of the growth of juvenile O. masou masou in summer. In the winter season, the warmer spring-fed tributary could also function as a suitable thermal refuge. Moreover, the substantial mass of eggs spawned by both wild and hatchery-reared O. keta made the tributary a more attractive refuge. These factors could explain the remarkable aggregation of juvenile O. masou masou in the spring-fed tributary. Meanwhile, Inoue and Ishigaki (1968) reported that juvenile O. masou masou also aggregated in a spring-fed tributary lacking redds of O. keta in the winter. Therefore, the suitability of spring-fed tributaries as winter thermal refuges may remain high even though the tributaries do not possess the marine-derived resource subsidy.

The present study revealed two ecosystem functions of spring-fed tributaries important to the seasonal changes in population density of juvenile O. masou masou. First, the stable flow regime of spring-fed streams can form depositional habitats and harbor abundant detritivores that burrow into fine sediments; these macroinvertebrates can provide a significant food resource for juvenile O. masou masou. Second, the stable temperature regime of spring-fed streams provides thermal refuges for juvenile O. masou masou during both hot and cold seasons. Therefore, the population density of juvenile O. masou masou peaked during those seasons, perhaps due to both abundant food resources and thermal stress avoidance. Although replication for spring-fed tributaries was lacking in this study, the results highlighted the importance of the unique habitat characteristics of spring-fed tributaries as drivers of animal community dynamics. Further studies are warranted to better understand the interactions between animal communities and water source-associated heterogeneity of river networks.

DATA ACCESSIBILITY

Unless an alternative copyright or statement noting that a figure is reprinted from a previous source is noted in a figure caption, the published images and illustrations in this article are licensed by the American Society of Ichthyologists and Herpetologists for use if the use includes a citation to the original source (American Society of Ichthyologists and Herpetologists, the DOI of the Ichthyology and Herpetology article, and any individual image credits listed in the figure caption) in accordance with the Creative Commons Attribution CC BY License.

ACKNOWLEDGMENTS

A portion of this study was supported by JSPS KAKENHI Grant numbers 2629181 and 19K20491, and Kuromatsunai Biodiversity Conservation Research Grant (2017). Dr. Izumi Washitani provided invaluable comments on the study plan. We thank the field assistance by Dr. Kosei Takahashi, Mr. Hitoshi Saito, Nobuo Hatai, Katsuya Iwabuchi, and Kengo Ebihara. Invaluable comments that improved the contents of this manuscript were provided by Associate Editor Matthew Craig and the anonymous reviewers. The authors declare no conflict of interest.