Isotope evidence for preferential dispersal of fast ...

This article was downloaded by: [Alexander Cerwenka] On: 07 January 2015, At: 00:33 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Isotopes in Environmental and Health Studies Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/gieh20

Isotope evidence for preferential dispersal of fast-spreading invasive gobies along man-made river bank structures a

b

b

Joerg Brandner , Karl Auerswald , Rudi Schäufele , Alexander F. c

Cerwenka & Juergen Geist a

d

Wasserwirtschaftsamt Regensburg, Regensburg, Germany

b

Click for updates

Lehrstuhl für Grünlandlehre, Technische Universität München, Freising, Germany c

Department of Ichthyology, Bavarian State Collection of Zoology (ZSM), München, Germany d

Aquatic Systems Biology Unit, Technische Universität München, Freising, Germany Published online: 02 Jan 2015.

To cite this article: Joerg Brandner, Karl Auerswald, Rudi Schäufele, Alexander F. Cerwenka & Juergen Geist (2015): Isotope evidence for preferential dispersal of fast-spreading invasive gobies along man-made river bank structures, Isotopes in Environmental and Health Studies, DOI: 10.1080/10256016.2014.993978 To link to this article: http://dx.doi.org/10.1080/10256016.2014.993978

PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or

howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content.

Downloaded by [Alexander Cerwenka] at 00:33 07 January 2015

This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/termsand-conditions

Isotopes in Environmental and Health Studies, 2014 http://dx.doi.org/10.1080/10256016.2014.993978

Isotope evidence for preferential dispersal of fast-spreading invasive gobies along man-made river bank structures


Joerg Brandnera , Karl Auerswaldb∗ , Rudi Schäufeleb , Alexander F. Cerwenkac and Juergen Geistd Regensburg, Regensburg, Germany; b Lehrstuhl für Grünlandlehre, Technische Universität München, Freising, Germany; c Department of Ichthyology, Bavarian State Collection of Zoology (ZSM), München, Germany; d Aquatic Systems Biology Unit, Technische Universität München, Freising, Germany a Wasserwirtschaftsamt

(Received 26 June 2014; accepted 31 October 2014)

Dedicated to Professor Dr Hanns-Ludwig Schmidt on the occasion of his 85th birthday Invasive round goby Neogobius melanostomus and bighead goby Ponticola kessleri have successfully colonized freshwater and coastal habitats worldwide. The objective was to use stable isotope analyses to study the foraging and movement of both species at small spatial scales in the Upper Danube River, considering 861 samples from two different years, seasons and sides of the river in an area where limited mixing at a confluence occurs. A difference in δ 13 C of 1 ‰ between gobies from both river sides was observed in both species and reflected the isotope spacing in their dominant benthic prey Dikerogammarus villosus. These results suggest an absence of goby movement across the Danube River which was unexpected, given the fast spread of gobies at invasion fronts. It can be concluded that their dispersal is highly preferential with longitudinal movement likely being facilitated by artificial rip-rap structures along river banks, which provide shelter and food. Keywords: animals; anisotropy; biological invasion; carbon-13; fish; isotope ecology; migration; nitrogen-15; tagging; trophic level

1.

Introduction

Invasive species are important drivers of global biodiversity loss [1,2] and one of the major threats to global freshwater biodiversity [3–5]. In the last two decades, invasive Ponto-Caspian fish species, in particular bighead goby Ponticola kessleri (Günther, 1861) and round goby Neogobius melanostomus (Pallas, 1814) (both: Teleostei: Perciformes: Neogobiidae), have expanded their distribution ranges in European freshwater and coastal ecosystems (reviewed in Kornis et al. [6]), with N. melanostomus having even colonized habitats on both sides of the Atlantic Ocean [7]. Rapid spread and a high potential to cause ecological regime shifts have mobilized substantial scientific interest in these species worldwide (see [6]). Among other key attributes or qualities facilitating invasive success, potent invaders such as bighead and round *Corresponding author. Email: [email protected] © 2014 Taylor & Francis


2

J. Brandner et al.

goby are characterized by their rapid range expansion, super-dominant population growth [8– 10], rapid genomic adaptation [11] and plasticity in several life history traits [12–14]. Due to their benthic sedentary lifestyle, their morphology and suspected small home ranges, gobies are expected to have only poor natural dispersal ability, especially in upstream direction [15,16]. However, spread rate estimates range from 500 m to 27 km yr–1 in selected areas [6,13,17,18] of riverine ecosystems. Moreover, the current invasion of the upper Danube River by additional species of Ponto-Caspian neogobiids [19] appears to be a fast-running process [13]. In contrast to the broad knowledge on upstream-directed migration patterns, little is known about the mobility of invasive gobies on small spatial scales. Typically, investigations on movement and dispersal patterns of fishes are carried out using telemetry or mark–recapture experiments. In large river systems such as the Danube, such experiments are difficult to conduct since telemetry or passive integrated transponder tagging would require the application of comparatively large tags on small specimens, with the potential that behaviour may be affected by the methodology. Analyses of specimens without prior treatment would thus be advantageous to study fish movement on small spatial scales. In cases where different habitats provide food of different isotopic composition, stable isotope analysis (SIA) has been shown to be a powerful tool to identify life history and migration patterns in a wide variety of animal taxa including mammals, birds, reptiles and insects (for review see [20,21]) as well as fishes [22–26] and in particular marine goby [27]. The ratios of stable isotopes of carbon (δ 13 C) and nitrogen (δ 15 N) reflect long-term feeding patterns and were used for trophic niche assessment in fish (e.g. [28,29]), including bighead and round goby (e.g. [12,13,30]. δ 15 N values of a consumer are typically increased by 3–4 ‰ relative to its diet (e.g. [31,32]). In contrast, there are limited differences ( ≤ 1 ‰) between heterotrophic organisms and their diet in the case of δ 13 C, since carbon moves through food webs (e.g. [32]) and therefore δ 13 C typically can be used to determine the ultimate sources of carbon for an organism if δ 13 C values of the sources differ [29]. This is especially true for invasive gobies since Brandner et al. [12] have shown that there is no difference between their diet and somatic δ 13 C. Thus, differences in dietary signals along a spatial gradient can potentially be used to trace foraging behaviour and movement. In this study, a naturally occurring difference in δ 13 C at the confluence of the Regen River with the Danube River was used to test for potential differentiation of foraging areas in two invasive gobies. Given the small distance of only about 100 m between both shorelines and the longdistance spread of gobies, we hypothesize that the difference between both river sides should at least be attenuated in the δ 13 C of the fish tissue.

2. 2.1.

Material and methods Sampling

At nine sampling sites (A–J) along the Upper Danube River (Figure 1), sampling was carried out at both river sides at four occasions in 2010 and 2011. Most of the Upper Danube catchment is agricultural land except for one of its large tributaries, the river Regen, which drains a mainly forested mountain range with high rainfall and lower input of fertilizer. The sampling sites A and B are located upstream the confluence of the Regen River into the Danube River; sampling sites C and D are directly downstream the mouth of the river Regen (Figure 1); sampling sites C and D differ in their shoreline substrate (rip-rap vs. gravel; Table 1). Sites E–J are located downstream a hydroelectric dam, where complete mixing of water occurs. At all sampling sites, physicochemical conditions (electrical conductivity at 25 °C (EC), dissolved oxygen, water temperature, pH and turbidity) were measured using handheld multi 350i and Turb 355T (both WTW GmbH, Weilheim, Germany). Gobies were collected from both


Isotopes in Environmental and Health Studies

3

Figure 1. Map of the sampling sites (A–J) along the Upper Danube River with drainage area and location in Europe. The confluents discharging water with significantly lower electrical conductivity compared with the Danube River are denoted by numbers: (1) River Naab and (2) River Regen, filled circles mark important cities. The connection of the Rhine–Main–Danube canal at the city of Kelheim is displayed. The confluence of the rivers Danube and Regen within the city of Regensburg is presented in magnification. Here, the sampling sites C and D, hydroelectric dam (3) and lock Regensburg (4) as well as the harbours Westhafen, Ölhafen (5) and Osthafen (6) are indicated by numbers. Bridges and buildings of the city are not displayed to increase readability.

shorelines by electrofishing following Brandner et al. [33]. Both the early (March–June) and late (August–October) annual growth periods of fish were covered to consider seasonal feeding effects as suggested from previous studies (e.g. [12]). Since stable isotope composition can be influenced by ontogenetic diet shifts (shown in N. melanostomus [12,13]) and since many morphometric indices assume isometry of body proportions in fish of varying size (e.g. [34]), specimens were size-class selected, targeting a total length (LT ) of 8–12 cm and a 1:1 sex ratio to achieve comparability of samples. LT was measured to the nearest mm, total body mass (M T ) was weighed to the nearest 0.2 g, and sex was determined by the morphology of the urogenital papilla following Kornis et al. [6]. Specimens were sacrificed using a lethal dose of anaesthetic and immediately frozen on dry ice to avoid tissue degradation. In total, 861 specimens were analysed. To exclude that the prey of the gobies differed between both sides at sites C and D, which would confound the interpretation of their body δ 13 C, we estimated the volumetric proportions of different food items in the digestive tract following [12]. Additionally, 15 N abundance was used to determine the position of both species within the food web. We sampled the invasive mussel Dreissena polymorpha (Pallas, 1771), which is regarded a food web-specific baseline genus regarding 15 N because it mainly consumes primary producers that incorporate mineral

4

J. Brandner et al.

Table 1.

Characteristics of sampling sites and specimens.

Sampling site


Latitude Longitude MQ (m3 s–1 ) Distance to river mouth (km) Distance to upstream site (km) Distance to next dam upstream (km) Distance to next dam downstream (km) Distance to river Regen inflow (km) River width (m) Substrate (both sides) EC (µS cm–1 ) O2 (mg L–1 ) pH Temperature (°C) Turbidity (NTU) Number of P. kessleri LT (cm) of P. kessleri M T (g) of P. kessleri Number of N. melanostomus LT (cm) of N. melanostomus M T (g) of N. melanostomus

B

C

D

E

11°59 15.6 48°57 10.8 351 2395 – 7 14 –18 103 rip-rap 560 ± 54 544 ± 28 11 ± 2 11 ± 2 8.3 ± 0.1 8.3 ± 0.1 15 ± 4 15 ± 4 8±7 6±3 15 16 10 ± 2 10 ± 2 16 ± 10 20 ± 15 13 13 10 ± 1 10 ± 1 14 ± 6 13 ± 5

12°09 5.3 49°01 11.5 444 2375 20 6 21 2 118 rip-rap 459 ± 24 562 ± 23 11 ± 1 11 ± 2 8.3 ± 0.2 8.2 ± 0.2 15 ± 1 15 ± 2 6±2 5±2 16 17 9±3 9±2 11 ± 10 8±6 16 16 10 ± 1 10 ± 1 15 ± 6 12 ± 4

12°10 4.3 49°00 54.6 444 2374 1 7 21 3 114 gravel 469 ± 27 557 ± 27 11 ± 1 10 ± 1 8.2 ± 0.1 8.2 ± 0.1 15 ± 1 16 ± 2 6±5 4±1 14 16 10 ± 2 10 ± 1 12 ± 10 9±5 17 16 10 ± 1 9±1 13 ± 3 11 ± 4

12°21 28.6 48°58 32.1 459 2353 21 1 23 24 142 rip-rap 495 ± 21 492 ± 25 11 ± 3 12 ± 3 8.3 ± 0.3 8.4 ± 0.2 15 ± 5 14 ± 5 6±1 5±2 17 15 10 ± 2 10 ± 2 13 ± 8 13 ± 8 16 17 10 ± 1 9±1 16 ± 5 14 ± 6

Notes: MQ is the arithmetic mean annual discharge (long-term data at water-gauge ‘Regensburg-Schwabelweis’; available at www.gkd.bayern.de/ provided by the Bavarian Environment Agency); all distances (in km) are measured along the river; values represent arithmetic means ± SD (where repeated measurements were made); EC is electrical conductivity (at 25 °C); LT is total length, M T is fish wet biomass; NTU is Nephelometric Turbidity Unit; in cases where two values are recorded, the upper line always applies to the left side and the lower to the right side; significant differences between both sides at a sampling site are highlighted in bold.

nitrogen. D. polymorpha thus allows calculation of trophic positions [30]. However, due to its sessile lifestyle, it is unsuitable to integrate the large variation in δ 13 C between microhabitats (e.g. pools and riffles, flow-exposed vs. non-exposed sides of stones; see [35–37]). Furthermore, it is not a prey for P. kessleri and it is only rarely consumed by N. melanostomus in the Upper Danube River [12]. According to Brandner et al. [12], the most important prey item for invasive gobies in the Upper Danube River is the benthic amphipod Dikerogammarus villosus (Sovinskij, 1894). Based on D. villosus, the spacing in δ 13 C of prey between both river sides was assessed. In total, 12 individuals of D. villosus and 8 individuals of D. polymorpha were collected at each side of sampling site C in April 2014. Benthic invertebrates were held in tap water for 24 h to empty their guts. Entire bodies of D. villosus and muscle tissue of D. polymorpha were then snap-frozen using liquid nitrogen. All samples were stored at –18 °C until further preparation for SIA. 2.2.

Stable isotope analysis

After thawing and ultrasonic cleaning, pieces of fish flank muscle tissue (about 0.5–1.0 cm3 ) were defatted with a chloroform–methanol (2:1) solution since lipids are depleted in 13 C



5

compared with whole organisms and the lipid content of animal tissue samples is variable [29]. All fish muscle, amphipod and mussel samples were oven-dried (40 °C for 48 h) and ground to homogenous powder using a mixer mill. Fish and mussel samples of 0.3–0.4 mg were weighed into tin cups. Amphipod samples were weighed into silver cups and fumigated with HCl [38] to remove their exoskeleton carbonates. Only the δ 13 C values of the amphipods are reported and compared with that of fish tissue because δ 15 N is altered by HCl fumigation [38]. In contrast, only the δ 15 N values of the mussel samples were used and compared with δ 15 N in fish tissue for side-specific calculation of trophic positions according to Brush et al. [30]. In this case, δ 13 C was not used because D. polymorpha is not a prey item that would contribute significant amounts of carbon to both invasive gobies. Samples were combusted in an isotope ratio mass spectrometer (Delta plus, Finnigan MAT, Bremen, Germany) interfaced (via ConFlo III, Finnigan MAT, Bremen, Germany) with an elemental analyser (NA 1108, Carlo Erba, Milan, Italy). Each sample was measured against a reference gas (CO2 , N2 ) which was previously calibrated against an IAEA secondary standard (IAEA-CH6 for CO2 , accuracy of calibration < 0.15 ‰ SD; IAEA-N1 and IAEA-N2 for N2 ; and accuracy of calibration < 0.20 ‰ SD). After every tenth sample, a solid internal lab standard with similar C/N ratio as the sample material (bovine horn) was run as a blind control that was also previously calibrated against the international standards. The long-term precision for the internal lab standards was better than 0.20 ‰ (SD) for δ 13 C and δ 15 N. 2.3.

Data analysis

The relative isotope ratios δ 13 C and δ 15 N were calculated as (Rsample /Rstandard ) − 1, where R is the ratio of the heavy and the light isotopes, and the standards are Vienna-PeeDee Belemnite (V-PDB) for carbon and AIR for nitrogen, respectively. The trophic position is given by [39] (δ 15 Nconsumer − δ 15 Nbaseline ) + 2, 3.4 where ‘consumer’ indicates the respective goby species and ‘baseline’ refers to baseline organism (D. polymorpha). The constant 3.4 is the diet-tissue enrichment per trophic level [29], and the constant 2 is the trophic level of D. polymorpha [30]. For comparisons of mean values between species and river sides, one-way ANOVA was used if the criteria for parametric testing were fulfilled. Alternatively, nonparametric Mann–Whitney U tests or Kruskal–Wallis tests (Bonferroni corrected) were applied. Significance was accepted at p ≤ 0.05. Statistical analyses and plots were computed using Statistica (version 6.1, StatSoft Inc., Tulsa, OK, USA), PAST [40], CoPlot (CoPlot Software, Monterey, CA, USA) and Excel 2010 (Microsoft™). The tests described above examine whether significant differences between species or sides exist, but they do not answer the question whether these differences are identical or not, i.e. whether the difference between sides is identical for predator and prey. A movement of some gobies from one side to the other would mix the isotopic information of both sides. This would reduce the difference between side-specific means until body turnover of the side-switching gobies has been completed and their bodies reflect the prey of the new side. Until then, the side-specific means in δ 13 C of P. kessleri and N. melanostomus would be smaller than the respective difference between sidespecific means of their dominant prey D. villosus. Thus, the confidence interval for the difference between both river sides in δ 13 C of the analysed species was calculated according to Sachs [41] by assuming equal variances. The standard deviation of the difference is then given by (s2l (nl − 1) + s2r (nr − 1)) sD = (nl + nr − 2)

6

J. Brandner et al.


and the confidence interval of the difference results from 1 1 + , CD = t sD nl nr where s, n, C and t represent the standard deviation, the number of samples, the confidence interval and the t-value, respectively. The indices indicate the left (l) and the right (r) side and the difference (D). The t-value was taken for p < 0.05 and (nl + nr − 2) degrees of freedom. A movement of single predatory individuals between two contrasting sides should also cause a larger scatter (SD) among the specimens collected at one side compared with the scatter that evolves in situations where sides do not differ and where the scatter thus only reflects the individual differences in prey preference, health status and other reasons of isotope variability. Hence, as a second test for the movement between river sides, the residuals between the measured values and the respective means for each side, sampling site and species were calculated. These residuals were then grouped for sampling sites C and D, where movement should lead to large residuals due to the difference between sides, and sampling sites B and E, where even with movement small residuals should result due the identical δ 13 C of both sides. The homogeneity of variances was then tested.

3.

Results

Upstream of the confluence of the river Regen with the Danube River and downstream of the dam, EC was almost identical on both sides (sampling sites A, B and E to J in Figure 2), while EC differed between both sides at sampling sites C and D, reflecting the difference in rainfall, soils and fertilizer input between the Danube and the Regen catchments. The mean EC on the right side (dominated by Danube water) at sampling site C significantly (n = 8; p < .001; Kruskal–Wallis test) exceeded that on the left river side (dominated by Regen water) by about 103 µS cm–1 ; at sampling site D, 1.3 km downstream of sampling site C, EC at the right side was still higher by 88 µS cm–1 (n = 8; p < .01; Kruskal–Wallis test) than at the left side (Table 1). This indicated two parallel flowing water bodies until they were eventually mixed at a dam (sampling site E), below of which EC did not differ anymore between sides. The differences in EC between river sides for the following sampling locations (F–J) remained small because no other large tributary with contrasting rainfall and land use enters the Danube River. The difference of δ 13 C in fish between both sides was generally small except for sampling sites C and D. The difference varied between 0.03 ‰ at sampling site H (in total of both sides: 119 specimen) and 0.35 ‰ at sampling site J (in total of both sides: 123 specimen). In contrast, fish differed significantly in δ 13 C values between both river sides at sampling sites C and D (Figure 3, lower panel) where the differences in EC were also most pronounced. Here, δ 13 C values were lower on the right side (Danube water). Similar to EC, no difference between both sides could be detected before the inflow of the river Regen (sampling site B) and after the dam (downstream of sampling site E). Hence, the large contrast between both sides at C and D provided a situation where the movement of invasive gobies across the river could be studied, while B and E served as control. Except for EC, no differences in the physicochemical parameters between both sides and all four sampling sites were found that could confound the results. Also, no significant differences between sides and sampling sites in LT and M T of the analysed specimens were detected (Table 1). The observed contrast in δ 13 C between both sides at sampling sites C and D for both species was obviously not associated with a shift in feeding behaviour because the δ 15 N values did not differ. This was also confirmed considering gut content analyses (Table 2), which indicated


EC 25°C ( S cm-1)

600

A

B

C/D

F

G

H

J

550 500 450

direction of flow

400

inflow 2400

2350

2300

2250

Distance to river mouth (km) Figure 2. Average electrical conductivity on the left side (left half-moons) and on the right side (right half-moons); error bars show 95 % intervals of confidence of the mean; error bars are only given directly after the inflow of the river Regen to improve readability but are similar for all other sampling sites. Dashed box indicates sampling sites used in the following figures.

EC25 oC ( S cm-1)

600 550 500 450

C (0/00)

-27

-28

13


E

7

-29

-30 B

C D Sampling location

E

Figure 3. Electrical conductivity (upper panel) and δ 13 C of gobies (mean of N. melanostomus and P. kessleri and four sampling occasions during two years, lower panel) at the sampling sites upstream of the river Regen inflow (B) and downstream (C, D and E). Left half-moons indicate the left river side and right half-moons indicate the right side. Error bars show 95 % intervals of confidence of the mean; error bars are only given directly after the inflow of the river Regen to improve readability but are similar for all sampling sites.

highly similar feeding patterns at both river sides for both species. However, the δ 15 N values in N. melanostomus significantly (p < .001; Kruskal–Wallis test) exceeded those in P. kessleri by about 1.5 ‰, indicating a niche separation of half a trophic level between both species. δ 13 C of the most important prey D. villosus significantly (p < .05; Mann–Whitney U) differed by 0.75 ‰ between both sides (Figure 4). Even more important, the difference between both sides was identical for prey (D. villosus) and predators. This was equally true for P. kessleri and N. melanostomus even though P. kessleri usually had higher δ 13 C values (mean of 407 specimen from all sites: − 28.8 ‰) than N. melanostomus (mean of 454 specimen from all sites: − 29.1 ‰) due to the different trophic niches. Hence, no indication could be found that any

8

J. Brandner et al.

Table 2. Mean δ 15 N values and trophic positions of D. polymorpha, P. kessleri and N. melanostomus and relative volume contribution of major food items to the total gut content at sites C and D depending on river side.


D. polymorpha

P. kessleri

N. melanostomus

Property

Left

Right

Left

Right

Left

Right

δ 15 N (‰) 95 % CI (‰) TP Crustacea (%) Chironomidae (%) Mollusca (%) Other items (%) n

10.7 1.0 2.0

10.6 0.8 2.0

13.2 0.3 2.7 74 12 3 11 28

12.9 0.2 2.7 74 11 0 16 30

14.6 0.2 3.2 55 28 10 7 32

14.6 0.2 3.2 58 26 7 9 31

8

8

Note: CI is the confidence interval of the mean δ N; TP is the trophic position and n is number of samples. 15

D. villosus N. melanostomus

C

P. kessleri N. melanost.

D

P. kessleri 0

0.5 Difference

1

1.5

(0/00)

δ 13 C

Figure 4. Differences between the values of both river sides (left minus right) for both goby species at sampling sites C and D and for the feed source D. villosus; error bars indicate 95 % intervals of confidence.

movement of the predators between the river sides had attenuated the isotope spacing resulting from their prey. This is in line with the results of the test of homogeneity of variance for the residuals between the measurement and the respective mean of a cohort (defined by sampling site, the river side and the species) which also did not indicate any movement or exchange of specimens across the river. At sites B and E, where such a movement would not have increased the scatter within a cohort, the SD was 0.544 ‰ (n = 227). A longitudinal movement from site C or D was also unlikely to bias this SD, given that the distance to these sites was 20 km and that dams separated the sampling sites (Table 1). This SD thus quantifies the natural variability within a cohort. At sites C and D, where a movement across the river should have caused a larger scatter of the residuals, the SD was 0.573 ‰ (n = 128) and the test of homogeneity of variance clearly rejected heterogeneity (p = 0.502). The slightly larger SD was caused by an outlier (residual: 2.42 ‰) but this outlier specimen could also not be assigned to an animal that had switched the sides because the deviation occurred in the other direction.

4.

Discussion

This study used stable isotope analyses and the naturally occurring contrast between two water bodies downstream the confluence of two rivers to explore movement patterns in two invasive goby species. Analysing a large number of specimens from P. kessleri and N. melanostomus, no indication for movement of specimens across the river was found. Neither the side contrast in



9

prey was attenuated in the predator tissue, nor was the variability among specimens increased in this situation where a large contrast in δ 13 C existed within a distance of less than 100 m between the two banks. The lack of movement of gobies between river sides was unexpected since both studied species are considered fast invaders for which a linear propagation velocity at the invasion front from 500 m to 27 km yr–1 was previously observed [6,13,17,18]. This means that the daily movement distance is on average about 40 m, and this estimate is already conservative since seasonal variation due to spawning is likely and since the movement of individual specimens is probably greater than the detectable shift of the invasion front. The analysis of dispersal patterns of both invasive species using SIA became feasible because different habitat patches in close vicinity sufficiently differed in δ 13 C. Our results require the discussion of three issues: (i) Is the sensitivity of our analysis based on the isotope contrast of both water bodies sufficient for the detection of movement across the river? (ii) Which mechanisms may lead to the observed mobility pattern? (iii) What is the reason for the differentiation in δ 13 C of both water bodies?

4.1.

Sensitivity

A mismatch between the isotope composition of prey and predator after an individual predator has moved to the other side of the Danube will gradually disappear due to the steady turnover of the predator tissue and its replacement by new tissue incorporating isotopes derived from the new diet [42–44]. Guelinckx et al. [27] quantified that the half-lives of carbon and nitrogen in muscle tissues of juvenile sand gobies (Pomatoschistus minutus; Pallas, 1770) are approximately 25 and 28 d. The value of carbon (and nitrogen) cannot be directly applied to our situation because the apparent turnover rate is influenced by metabolic activity (strongly depending on weight and temperature) and growth. The influence of weight resulting from the relation of feed intake to body size follows metabolic body weight and thus leads to a decrease in turnover rate with the ( − ¼) power of body weight [43,45,46]. The juveniles analysed in [27] had an initial body mass of 1 g while our specimen had a mean mass of 13 g. The ( − ¼) power rule thus predicts that the half-life for muscle tissue for the size class used in our study should be 47 d. A different approach, predicting turnover directly from body mass [47,48], leads to a similar value of 45 d. Guelinckx et al. [27] carried out their experiments at 17 °C while the average water temperature during our samplings was 15 °C, and the annual mean temperature of the Upper Danube River at sampling site C (Figure 1) was only 11.5 °C (raw data from 2005 to 2013 at water-gauge ‘RegensburgSchwabelweis’; available at www.gkd.bybn.de; data from the Bavarian Environmental Agency). The temperature effect on turnover follows e0.08T [48]. At a water temperature of 15 °C, half-life thus should be 20 % longer than at 17 °C. At 11.5 °C, half-life would be 55 % longer, leading to a predicted half-life of 73 d. Finally, Guelinckx et al. [27] determined their apparent half-lives from fast growing juvenile gobies which doubled their weight within 90 d. Growth adds material that is entirely derived from the new diet while the old body tissue is gradually turned over [42,49]. Thus, the apparent turnover including growth is much faster than the true turnover of old body tissue. Relative growth is much smaller for older animals than for juveniles as analysed in [27]; and in natural environments, food supply is limited and energy expenditure is higher than in laboratory experiments. Lynch and Mensinger [50] reported that N. melanostomus increases in length by only 30 % per year. This is in line with Grul’a et al. [51] who showed that N. melanostomus need more than three to four years to grow from 40 mm (the size of the juveniles in [27]) to 100 mm (the size of our specimens), which also corresponds to a growth rate of about 30 % per year, while doubling weight within 90 d corresponds to a growth rate of 400 % per year. We can thus safely assume that half-lives in our specimens should at least be 100 d or longer, which means that half of the isotope contrast between both sides should still exist 100 d after an individual had switched sides.

10


4.2.

J. Brandner et al.

Mechanisms of preferential dispersal

Structurally degraded sites close to navigable water bodies are exposed to an increased invasion risk by alien species [52]. Ships have been proposed as important vectors for the invasion of gobies, and their leap from their Ponto-Caspian origin to the Laurentian lakes can only be explained by such mechanisms (reviewed in [6]). Transport by ships is thus likely to extend from stepping-stone hotspots like harbours [53], which is in line with the observed genetic population structure [11]. It, however, fails to explain the rather constant dispersal rate and the time shift of about five years between the ongoing sympatric invasion of N. melanostomus (first recorded 2004 [54]), and the earlier arrival of P. kessleri (first recorded 1999 [55]) in the Upper Danube River. The spread of invasive gobies mainly proceeds along the navigable Danube River, but does not affect its hydromorphologically less modified tributaries like the river Regen. Rip-rap structures are typically used along navigable rivers for bank reinforcement and to prevent shoreline damage induced by ship wash. It is well known that rip-rap structures are a highly preferred habitat of gobies [56] since they provide shelter and optimum prey conditions [12]. Our study shows that the movement and dispersal of gobies also seems to be facilitated by these bank structures whereas the gobies hardly seem to take the risk to swim away from them to cross the river. While plasticity in life history traits is often regarded as a key property to the success of invasive species [11,12,14,17,57], such plasticity does not seem to exist regarding the movement of N. melanostomus and P. kessleri, resulting in preferential movement along man-made bank structures. This insight offers options to slow down the invasion process considering the serious and lasting effects on native communities by both species [58–61]. The restoration of structural deficits in rivers and the consideration of invasive species in conservation management are two great remaining challenges in the management of aquatic ecosystems in the context of the policy goals formulated in the European Water Framework Directive [62]. However, both aspects are difficult to address in large river ecosystems such as the Danube River. 4.3.

Potential origin of the isotope differentiation

Even though our study was primarily designed to use the differentiation in δ 13 C between the water bodies from the rivers Danube and Regen to analyse the movement patterns of two goby species and their prey, there is also some indication on which factors have caused the observed differences in δ 13 C of food webs. Principally, the difference could be allochthonous, mirroring differences in carbon input (endogenous vs. exogenous) from the respective catchments. The catchment of the Upper Danube River upstream of site B (Figure 1) is dominated by calcareous sediments and intense agricultural use. In contrast, the river Regen catchment mostly comprises felsic rock, thus allowing less intense agriculture and more forest. These differences can explain why the EC of the Regen River was much lower than EC of the Danube River. In addition, allochthonous carbon of the forested Regen catchment is likely to be older carrying a smaller Suess effect [63] than the carbon of the Upper Danube catchment. However, the visible algal growth in both water bodies during the vegetation period indicates that autochthonous carbon is probably by far more important than allochthonous sources, which is a general observation [64]. An isotope difference in autochthonous carbon would also result from a difference in the exchange between atmospheric CO2 and respiratory CO2 originating from the saprobic decomposition of organic matter within the water column and the river bed [65]. A restricted exchange between atmospheric and respiratory CO2 will increase the proportion of respiration-derived CO2 taken up by photosynthesis, resulting in lower δ 13 C values. A restricted exchange is likely to be found in the Danube River, which is, on average, deeper and slower flowing due to the shorter distance between dams than the river Regen which drains a mountain area, thus being more shallow and more turbulent. This interpretation was corroborated by the very low δ 13 C in


11

fish at sampling site E. While EC at sampling site E clearly resulted from the mixing of Danube and Regen waters, δ 13 C in fish at sampling site E was even lower than in the Danube water at sampling sites C and D and thus could not result from the mixing of Regen and Danube waters. Sampling site E was located just 1.8 km downstream of a large dam (dam height 7.3 m and maximum width 336 m) that likely reduced the exchange of respiratory and atmospheric CO2 and can thus lead to the lowest δ 13 C within all sampling sites. The reasons for the difference in δ 13 C and for the difference in EC hence are different, but the coincidence of the deviation of the parameter under focus (δ 13 C) with the deviation in EC allowed an easy identification of both water bodies.


5.

Conclusions

This study used the dietary signals along a fluvial gradient to trace movement of aquatic animals. The dispersal of invasive gobies appears to be highly anisotropic, following artificial rip-rap structures along the river banks. The fast longitudinal movement of gobies reported in other studies is potentially supported by those structures which reduce flow current, provide shelter and prey. Knowledge on the non-random spread of invasive gobies revealed in this study may be used to develop measures against their further spread. In addition to classical methods such as telemetry, the use of stable isotope analyses was shown to provide an elegant tool to analyse movement, migration and dispersal in aquatic animals if there is sufficient heterogeneity between different habitats. Acknowledgments We thank all owners of the local fishing rights and the ‘Fischereifachberatungen’ for their permission to carry out electrofishings.

Disclosure statement No potential conflict of interest was reported by the authors.

Funding Part of the sampling for this study was financially supported by the German Research Council DFG [grant numbers GE2169/1-1 (AOBJ: 569812) and SCHL567/5-1]. The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.

References [1] Sala OE, Chapin III FS, Armesto JJ, Berlow R, Bloomfield J, Dirzo R, Huber-Sannwald E, Huenneke LF, Jackson RB, Kinzig A, Leemans R, Lodge D, Mooney HA, Oesterheld M, Poff NL, Sykes MT, Walker BH, Walker M, Wall DH. Global biodiversity scenarios for the year 2100. Science. 2000;287:1770–1774. [2] Mooney HA, Cleland EE. The evolutionary impact of invasive species. Proc Natl Acad Sci USA. 2001;98:5446– 5451. [3] Dudgeon D, Arthington AH, Gessner MO, Kawabata Z, Knowler DJ, Lévêque C, Naiman RJ, Prieur-Richard A, Soto D, Stiassny MLJ. Freshwater biodiversity: importance, threats, status and conservation challenges. Biol Rev. 2006;8:163–182. [4] Keller RP, Geist J, Jeschke JM, Kühn I. Invasive species in Europe: ecology, status, and policy. Environ Sci Europe. 2011;23:1–17. doi:10.1186/2190-4715-23-23. [5] Geist J. Integrative freshwater ecology and biodiversity conservation. Ecol Indic. 2011;11:1507–1516. [6] Kornis MS, Mercado-Silva N, Vander Zanden MJ. Twenty years of invasion: a review of round goby Neogobius melanostomus biology, spread and ecological implications. J Fish Biol. 2012;80:235–285. [7] Corkum LD, Sapota MR, Skóra KE. The round goby, Neogobius melanostomus, a fish invader on both sides of the Atlantic Ocean. Biol Invasions. 2004;6:173–181.


12

J. Brandner et al.

[8] Sakai AK, Allendorf FW, Holt JS, Lodge DM, Molofsky J. The population biology of invasive species. Annu Rev Ecol Evol Syst. 2001;32:305–332. [9] Gutowsky LFG, Fox MG. Occupation, body size and sex ratio of round goby (Neogobius melanostomus) in established and newly invaded areas of an Ontario river. Hydrobiologia. 2011;671:27–37. [10] Brandner J. Ecology of the invasive neogobiids Neogobius melanostomus and Ponticola kessleri in the upper Danube River [PhD thesis]. München: Technische Universität München; 2014. [11] Cerwenka AF, Brandner J, Geist J, Schliewen UK. Strong versus weak population genetic differentiation after a recent invasion of gobiid fishes (Neogobius melanostomus and Ponticola kessleri) in the upper Danube River. Aquat Invasions. 2014;9:71–86. [12] Brandner J, Auerswald K, Cerwenka AF, Schliewen UK, Geist J. Comparative feeding ecology of invasive PontoCaspian gobies. Hydrobiologia. 2013;703:113–131. [13] Brandner J, Cerwenka AF, Schliewen UK, Geist J. Bigger is better: characteristics of round gobies forming an invasion front in the Danube River. PLoS ONE. 2013;8(9):e73036. [14] Cerwenka AF, Alibert P, Brandner J, Geist J, Schliewen UK. Phenotypic differentiation of Ponto-Caspian gobies during a contemporary invasion of the upper Danube River. Hydrobiologia. 2014;721:269–284. [15] Wolfe RK, Marsden JE. Tagging methods for the round goby (Neogobius melanostomus). J Great Lakes Res. 1998;24:731–735. [16] Ray WJ, Corkum LD. Habitat and site affinity of the round goby. J Great Lakes Res. 2001;27:329–334. [17] Brownscombe JW, Fox MG. Range expansion dynamics of the invasive round goby (Neogobius melanostomus) in a river system. Aquat Ecol. 2012;46:175–189. [18] Brownscombe JW, Masson L, Beresford DV, Fox MG. Modeling round goby Neogobius melanostomus range expansion in a Canadian river system. Aquat Invasions. 2012;7:537–545. [19] Haertl M, Cerwenka AF, Brandner J, Borcherding J, Geist J, Schliewen U. First record of Babka gymnotrachelus (Kessler, 1857) from Germany (Teleostei, Gobiidae, Benthophilinae). Spixiana. 2012;35:155–159. [20] Hobson KA. Tracing origins and migration of wildlife using stable isotopes: a review. Oecologia. 1999;120:314– 326. [21] Rubenstein DR, Hobson KA. From birds to butterflies: animal movement patterns and stable isotopes. Trends Ecol Evol Syst. 2004;19:256–263. [22] Hansson S, Hobbie JE, Elmgren R, Larsson U, Fry B, Johansson S. The stable nitrogen isotope ratio as a marker of food-web interactions and fish migration. Ecology. 1997;78:2249–2257. [23] Herzka SZ. Assessing connectivity of estuarine fishes based on stable isotope ratio analysis. Estuar Coast Shelf Sci. 2005;64:58–69. [24] Hesslein RH, Capel MJ, Fox DE, Hallard KA. Stable isotopes of sulfur, carbon, and nitrogen as indicators of trophic level and fish migration in the lower Mackenzie River basin, Canada. Can J Fish Aquat Sci. 1991;48:2258–2265. [25] McCarthy ID, Waldron S. Identifying migratory Salmo trutta using carbon and nitrogen stable isotope ratios. Rapid Commun Mass Spectrom. 2000;14:1325–1331. [26] Kennedy BP, Folt CL, Blum JD, Chamberlain CP. Natural isotope markers in salmon. Nature. 1997;387:766–767. [27] Guelinckx J, Maes P, Van Den Driessche B, Geysen F, Dehairs F, Ollevier F. Changes in delta 13C and delta 15N in different tissues of juvenile sand goby Pomatoschistus minutus: a laboratory diet-switch experiment. Mar Ecol–Prog Ser. 2007;341:205–215. [28] Vander Zanden MJ, Fetzer WW. Global patterns of aquatic food chain length. Oikos. 2007;116:1378–1388. [29] Post DM. Using stable isotopes to estimate trophic position: models, methods, and assumptions. Ecology. 2002;83:703–718. [30] Brush JM, Fisk AT, Hussey NE, Johnson TB. Spatial and seasonal variability in the diet of round goby (Neogobius melanostomus): stable isotopes indicate that stomach contents overestimate the importance of dreissenids. Can J Fish Aquat Sci. 2012;69:573–586. [31] DeNiro MJ, Epstein S. Influence of diet on the distribution of nitrogen isotopes in animals. Geochim Cosmochim Acta. 1981;45:341–351. [32] Peterson BJ, Fry B. Stable isotopes in ecosystems studies. Annu Rev Ecol Evol Syst. 1987;18:293–320. [33] Brandner J, Pander J, Mueller M, Cerwenka AF, Geist J. Effects of sampling techniques on population assessment of invasive round goby. J Fish Biol. 2013;82:2063–2079. [34] Anderson RO, Neumann RM. Length, weight, and associated structural indices. In: Murphy BR, Willis DW, editors. Fisheries techniques. 2nd ed. Bethesda, MD: American Fisheries Society; 1996. p. 447–482. [35] France RL. Differentiation between littoral and pelagic food webs in lakes using stable carbon isotopes. Limnol Oceanogr. 1995;40:1310–1313. [36] Fry B. Stable isotope ecology. New York: Springer; 2006. [37] Finlay JC, Khandwala S, Power, ME. Spatial scales of carbon flow in a river food web. Ecology. 2002;83:1845– 1859. [38] Harris D, Horwath WR, van Kessel C. Acid fumigation of soils to remove carbonates prior to total organic carbon or 13 C isotopic analysis. Soil Sci Soc Am J. 2001;65:1853–1856. [39] Cabana G, Rasmussen JB. Comparison of aquatic food chains using nitrogen isotopes. Proc Natl Acad Sci USA. 1996;93:10844–10847. [40] Hammer Ø, Harper DAT, Ryan PD. PAST: paleontological statistics software package for education and data analysis. Palaeontol Electron. 2001;4:1–9. [41] Sachs L. Angewandte Statistik [Applied statistics]. Berlin: Springer; 1978.



13

[42] Fry B, Arnold C. Rapid 13 C/12 C turnover during growth of brown shrimp (Penaeus aztecus). Oecologia. 1982;54:200–204. [43] Hobson KA, Clark RG. Assessing avian diets using stable isotopes. 1. Turnover of 13C in tissues. Condor. 1992;94:181–188. [44] Auerswald K, Wittmer MHOM, Zazzo A, Schäufele R, Schnyder H. Biases in the analysis of stable isotope discrimination in food webs. J Appl Ecol. 2010;47:936–941. [45] Martínez del Rio C, Wolf N, Carleton SA, Gannes LZ. Isotopic ecology ten years after a call for more laboratory experiments. Biol Rev. 2009;84:91–111. [46] Braun A, Auerswald K, Vikari A, Schnyder H. Dietary protein content affects isotopic carbon and nitrogen turnover. Rapid Commun Mass Spectrom. 2013;27:2676–2684. [47] Weidel BC, Carpenter SR, Kitchell JF, Vander Zanden MJ. Rates and components of carbon turnover in fish muscle: insights from bioenergetics models and a whole-lake 13 C addition. Can J Fish Aquat Sci. 2011;68:387–399. [48] Guelinckx J, Maes J, Geysen B, Ollevier F. Estuarine recruitment of a marine goby reconstructed with an isotopic clock. Oecologia. 2008;157:41–52. [49] MacAvoy SE, Macko SA, Arneson LS. Growth versus metabolic tissue replacement in mouse tissues determined by stable carbon and nitrogen isotope analysis. Can J Zool. 2005;83:631–641. [50] Lynch MP, Mensinger AF. Temporal patterns in growth and survival of the round goby Neogobius melanostomus. J Fish Biol. 2013;82:111–124. [51] Grul’a D, Balážová M, Copp GH, Kováˇc V. Age and growth of invasive round goby Neogobius melanostomus from middle Danube. Cent Eur J Biol. 2012;7:448–459. [52] Früh D, Stoll S, Haase P. Physicochemical and morphological degradation of stream and river habitats increases invasion risk. Biol Invasions. 2012;14:2243–2253. [53] Wiesner C. New records of non-indigenous gobies (Neogobius spp.) in the Austrian Danube. J Appl Ichthyol. 2005;21:324–327. [54] Paintner S, Seifert K. First record of the round goby. Neogobius melanostomus (Gobiidae), in the German Danube. Lauterbornia. 2006;58:101–107. [55] Seifert K, Hartmann F. Die Kesslergrundel (Neogobius kessleri Günther 1861), eine neue Fischart in der deutschen Donau [The bighead goby (Neogobius kessleri Günther 1861), a new fish species in the German Danube]. Lauterbornia. 2000;38:105–108. German. [56] Sindilariu PD, Freyhof J, Wolter C. Habitat use of juvenile fish in the lower Danube and the Danube Delta: implications for ecotone connectivity. Hydrobiologia. 2006;571:51–61. [57] Bøhn T, Sandlund OT, Amundsen PA, Primicerio R. Rapidly changing life history during invasion. Oikos. 2004;106:138–150. [58] Ricciardi A. Facilitative interactions among aquatic invaders: is an “invasional meltdown” occurring in the Great Lakes? Can J Fish Aquat Sci. 2001;58:2513–2525. [59] Crossman EJ, Helm E, Cholmondeley R, Tuininga K. First record for Canada of the Rudd, Scardinius erythrophthalmus, and notes on the introduced Round Goby, Neogobius melanostomus. Can Field Nat. 1992;106:206–209. [60] Freyhof J. Immigration and potential impacts of invasive freshwater fishes in Germany. Annual Report 2002. Berlin: Leibniz-Institute of Freshwater Ecology and Inland Fisheries; 2003. p. 51–58. [61] Karlson AML, Almqvist G, Skóra KE, Appelberg M. Indications of competition between non-indigenous round goby and native flounder in the Baltic Sea. ICES J Mar Sci. 2007;64:479–486. [62] Geist J. Trends and directions in water quality and habitat management in the context of the European Water Framework Directive. Fisheries. 2014;39:219–220. [63] Verburg P. The need to correct for the Suess effect in the application of δ 13 C in sediment of autotrophic Lake Tanganyika, as a productivity proxy in the Anthropocene. J Paleolimnol. 2007;37:591–602. [64] Finlay JC. Stable carbon isotope ratios of river biota: implications for energy flow in lotic food webs. Ecology. 2001;82:1052–1064. [65] Finlay JC. Controls of streamwater dissolved inorganic carbon dynamics in a forested watershed. Biogeochemistry. 2003;62:231–252.