Moving from multiple pass depletion to single pass timed electrofishing for fish community assessment in wadeable streams

Fisheries Research xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Fisheries Research journal homepage: www.elsevier.com/locate/fishres

Research Paper

Moving from multiple pass depletion to single pass timed electrofishing for fish community assessment in wadeable streams ⁎

Ronan Matson , Karen Delanty, Samuel Shephard, Brian Coghlan, Fiona Kelly Inland Fisheries Ireland, 3044 Lake Drive, CityWest Business Campus, Dublin D24 Y265, Ireland

A R T I C L E I N F O

A B S T R A C T

Handled by George A. Rose

Area-delineated multiple-pass depletion electrofishing (ADEF) can be resource-intensive. It may not capture fish community state when resource limitations mean that the number of sites sampled in a system is insufficient to account for ecological heterogeneity. Rapid assessment techniques such as single-pass timed electrofishing could be more efficient and support increased sample size, but it is important to understand how resulting catches will relate to existing ADEF data/time series. Paired ADEF (3-pass) and single-pass timed (10-min) electrofishing (TEF) samples were collected for sites in the River Barrow catchment (Ireland) in 2015. Paired samples were used to derive species-specific size-based conversion factors (CF) for extrapolating TEF catch up to a predicted ADEF Pass 1 catch. Applying these CF to a set of independent ‘validation’ TEF samples (2008–2013) produced fish catch estimates similar to observed ADEF Pass 1 catches (2008–2015). Species size-distributions in extrapolated TEF data were also similar to those in paired ADEF Pass 1 samples. Pass 2 and Pass 3 catches by species were then predicted from extrapolated validation TEF catches using a regression model. Cumulative catch curves fitted to these predicted catches were similar to those fitted to observed ADEF catch data for most species. TEF samples provide estimates of fish species catch and size-distribution that can be extrapolated following a simple protocol and compared with ADEF data for small streams measuring up to 10 m in width.

Keywords: Water Framework Directive Biological quality elements Electrofishing Monitoring Rapid bioassessment

1. Introduction Fish communities are widely used for assessing the ecological state of rivers and streams (Fausch et al., 1984; Angermeier and Karr, 1986; FAME CONSORTIUM, 2004). Biomonitoring of European aquatic ecosystems using fish and other biological quality elements has developed substantially since the implementation of the Water Framework Directive (WFD) in 2000 (2000/60/EEC) (Birk et al., 2012). The WFD requires European Union (EU) member states to conduct on-going assessment of the ecological status of surface waters. Assessing change in state typically requires standardised sampling methods that can provide consistent and measurable metrics of fish community (e.g. species diversity and abundance). Standardized survey protocols attempt to ensure that sampling is repeatable, that estimated metrics are comparable among years and systems, and that measurement uncertainty can be expressed. Metrics estimated with standardized sampling can be invaluable for identifying vulnerable populations, monitoring changes in state related to the management of specified pressures, communicating with stakeholders and providing direction for more in-depth assessments. Large-scale standardized surveys, such as those used for the fish community element of the WFD, can be resource-intensive (Foley et al.,

⁎

2015; Kelly et al., 2007; Meador et al., 2003). Economic and staff resources may limit the spatial coverage of sampling, including survey stretch length (Foley et al., 2015; Meador et al., 2003), the number of sites that can be sampled and the extent of complementary data collection, including abiotic measurements. When designing a sampling and assessment programme, a balance is required between objectives, available resources and the importance of results (Kelly et al., 2007). Achieving such a balance can involve a trade-off between sampling effort and adequately recording species diversity and a measure of species abundance or density. Ideally, any sampling method should maximize the return of information with the minimum effort. Electrofishing is a well-established technique used by fishery biologists for sampling river fishes and it is frequently the most non-destructive, effective and cost-efficient approach. The effectiveness of electrofishing surveys depend on many different factors including power output and frequency, gear configuration (number of sets and operators), the shape and size of the electrodes, conductivity and the size of the channel being surveyed (Beaumont et al., 2002). The optimum sample size for estimating species richness may be that after which species richness reaches an asymptote; a number of methods have been deemed effective for this purpose (Angermeier and

Corresponding author. E-mail address: Ronan.Matson@fisheriesireland.ie (R. Matson).

http://dx.doi.org/10.1016/j.fishres.2017.10.009 Received 1 February 2017; Received in revised form 9 October 2017; Accepted 13 October 2017 0165-7836/ © 2017 Published by Elsevier B.V.

Please cite this article as: Matson, R., Fisheries Research (2017), http://dx.doi.org/10.1016/j.fishres.2017.10.009

Fisheries Research xxx (xxxx) xxx–xxx

R. Matson et al.

TEF technique so that it can be compared directly with existing catch time series from the established ADEF method. The proposed TEF method may provide another useful tool for fisheries researchers and managers looking to carry out large sampling programmes in a more cost-effective manner.

Smogor, 1995; Lyons, 1992; Paller, 1995). Attaining sufficient samples to reach the asymptote can be achieved by adding additional sites (Bateman et al., 2005; Paller, 1995; Wyatt, 2002) or by increasing the length of the surveyed stretch (Hughes et al., 2002; Temple and Pearsons, 2003). The number of fished sites that must be sampled to reach a richness asymptote is low in Irish rivers due to limited species diversity and this makes the possibility of reaching an asymptote much greater than it would be in much more diverse systems elsewhere (Chao et al., 2009) or in systems where there are many rare species (Gotelli and Colwell, 2011). Therefore, for species diversity alone in these systems, less resource intensive sampling efforts can suffice. Assessment of fish abundance or density at a given site is usually conducted by removing fish in a series of successive passes. The added sampling effort increases the chances of recording species missed during previous passes within the system and also increases the likelihood of achieving a species richness asymptote. The quantitative area-delineated multiple pass depletion method of electrofishing (ADEF) (Zippin, 1956; Bohlin et al., 1989) has been used for over 60 years to support such assessment of river fish communities (Schmutz et al., 2007; Teixeira-de Mello et al., 2014). When applying the ADEF, sampling sites are isolated using stop nets up- and down-stream. The method is highly effective but resource intensive, particularly on wider streams. A simpler electrofishing sampling approach is the semi-quantitative single pass electrofishing survey, which uses the fish caught in one pass to calculate a minimum estimate of the fish population. For more specific fish assessments, such as the assessment of juvenile salmonids at the catchment scale, a rapid semi-quantitative timed electrofishing method (TEF) may be used, such as that developed by Crozier and Kennedy (1994). Such a technique has been used to undertake catchment wide surveys of juvenile Salmo salar and Salmo trutta (0+) fry (e.g. Crozier and Kennedy, 1995; Gargan and Roche, 2011). The use of a standard unit of time generates an abundance index that can be compared among sites. Semi-quantitative timed (10 min pass) electrofishing was adapted by Inland Fisheries Ireland (IFI, 2009) to monitor salmonid and other fish species and to cover all morphological features (riffle/glide pool) present at a given site. Multiple pass electrofishing like ADEF involves more staff equipment and time than TEF. ADEF may also have greater environmental impact; electrofishing can affect fish both physiologically and behaviourally depending on the level of exposure and may take up to 24 h to recover (Mesa and Schreck, 1989). Electrofishing mortality rates increase with sampling intensity (Habera et al., 1996; Panek and Densmore, 2013). Reducing the sampling effort by using TEF could help offset some of these negative impacts (Kocovsky et al., 1997; Panek and Densmore, 2013). Both resourcing and environmental impact issues create an incentive to shift towards TEF, and this shift would be acceptable under guidelines within the European standards (CEN 2003). However, there are various issues to be considered before adjusting a sampling programme that contributes to state assessment time series (e.g. WFD). A key potential disadvantage is that TEF may be less efficient than ADEF at catching species or size groups that are rare or that have low catchability (Bertrand et al., 2006; Paller, 1995; Vehanen et al., 2012), which may bias community level assessment metrics. However, research suggests that possible information loss due to reduced precision and accuracy in single-pass electrofishing can be compensated for or even improved upon by the potential to use resource-savings to increase sampling coverage (Bateman et al., 2005; Jones and Stockwell, 1995; Paller, 1995) and/or to sample a more representative range of sites in heterogeneous river systems (Edwards et al., 2003; Bateman et al., 2005; Reid et al., 2008). When introducing new sampling methods, it is important that the new data can be comparable with existing time series and associated ecological state assessments. Implementation of more efficient TEF methods will benefit from a robust protocol for adjusting catches to support comparison with current ADEF catches. The current study aimed to (1) compare catches between TEF and ADEF methods, and (2) develop a protocol for extrapolating catch from the rapid assessment

2. Material and methods Paired samples were collected from the River Barrow catchment (Ireland) in 2015, and comprised fish catches at the same site using both ADEF and TEF with sampling events spaced closely in time. These samples were used to support development of a catch extrapolation protocol for expanding observed TEF catch to a predicted ADEF catch. Independent archived data sets from other Irish rivers collected using each of the two electrofishing methods were then used to validate the proposed catch conversion protocol, i.e. to compare observed and predicted ADEF catch. 2.1. Study area and sites The River Barrow catchment in south-eastern Ireland was chosen as a suitable study area due to its relatively diverse network of tributaries (Fig. 1). The Barrow is Ireland’s second longest river measuring over 180 km in length, with a catchment of approximately 3000 km2. The tributaries of the Barrow catchment include lowland streams draining peatland, streams traversing agricultural and pastureland, and small to high gradient channels. Independent alkalinity data (EPA 2013–2014) from sites on tributaries within the River Barrow catchment range from between 40 and 345 mg/l CaCO3. The underlying geology of the catchment is varied, consisting predominantly of limestone to the west and granite to the east. Of the 36 paired sites (P-Sites), 22 could be classified as having calcareous underlying geology, with the remaining 14 siliceous. The length of each survey site was measured using a measuring tape, from downstream to upstream stop net for ADEF samples and start to finish points for TEF samples. Wetted width was taken as the mean of at least three transects, spaced evenly along the length sampled. Depth was recorded at five locations along each transect, with the mean of all values taken as the mean depth. The maximum depth was taken as the maximum of all depths recorded. 2.2. Fish sampling at paired sites (P-Sites) Thirty-six sites were surveyed using both ADEF and TEF electrofishing methods at low flow levels during July and August 2015 (Fig. 1). Sites were selected to cover a diverse range of habitats, including riffle, glide and pool. Estimates of recovery time after electrofishing range from 1 to 24 h (Peterson and Cederholm, 1984; Mesa and Schreck, 1989) and are dependent on factors including amount of exposure, number of shocks and handling stress. ADEF and TEF surveys were split between different teams of operators. Both methods were conducted on alternating river sites concurrently but with a minimum gap of two weeks and maximum gap of one month between samples at any given site to enable the fish and river habitat to recover to a normal state. Electrofishing was carried out in an upstream direction using bank-based units (Electracatch International (now SmithRoot) WFC4 units with Honda 10I Inverter generators) set with a typical output of 50 Hz and 100–200 V. All surveys began at the downstream end of a riffle where possible. Differences in physical dimension (depth, width and surface area) were tested using paired t-t-tests with alpha = 0.05 (PAST, 2015). All fish caught were held in buckets of fresh cold oxygenated water until processing, which was carried out promptly after each electrofishing pass. After processing, fish were returned to the river as quickly as possible to avoid further stress. All fish except for lamprey were identified to species level and counted; lampreys were identified to genus. Fish fork-lengths (cm) and weights (to the nearest 0.5 g) were recorded. 2

Fisheries Research xxx (xxxx) xxx–xxx

R. Matson et al.

Fig. 1. Map of Ireland showing the River Barrow catchment and all sites surveyed (A-DEF, P-Sites and A-TEF).

sampled is considered satisfactory when sampling adds no further species to the list. Barbour et al. (1999) suggest two possible approaches, fixed-distance (based on a standard length of channel) or proportional-distance (based on the width of the channel). The

2.2.1. Area delineated electrofishing (ADEF) method The appropriate length of a river survey stretch can depend on the purpose of a survey and whether qualitative or quantitative information is required. According to Beklemishev’s rule (Tarwid, 1956), the length 3

Fisheries Research xxx (xxxx) xxx–xxx

R. Matson et al.

approach used for ADEF sampling in the current study followed the same protocol used for collecting data for the WFD surveillance monitoring programme in Ireland: the CEN standards for sampling fish with electricity (CEN, 2003). These standards recommend surveying a minimum channel length of 20 m in streams less than 5 m in width and a minimum of 50 m in channels with a width between 5 m and 15 m. All P-Sites satisfied these criteria. Sites were isolated at both ends by stop-nets to prevent loss or recruitment of fish. The number of operators and equipment units required varied between one and three depending on the width of the river channel surveyed. This approximates to one operator per 5–6 m of channel width (one operator at 13 sites, two operators at 21 sites and 3 operators at 1 site).

•

Step 1: Determine form of observed cumulative catch data from ADEF electrofishing. Six candidate regression models were fitted to A-ADEF data pooled across all species caught, where the independent variable x is pass number (1–3) and the dependent variable y is cumulative fish catch. The Akaike Information Criterion (AIC) was used to select the bestfitting (lowest AIC) model, where models with ΔAIC ≤ 2 were considered competing. Plots were used to check for homogeneity of residuals. The selected best fitting model was y = a + b (log x). This model was then fit to the observed A-ADEF cumulative catches for each species separately. Step 2: Derive univariate conversion factors CF to extrapolate TEF catch by species to predicted ADEF Pass 1 catch. Paired TEF and ADEF Pass 1 catches (P-Sites) were used to derive preliminary univariate species-specific conversion factors CF that could be used to expand TEF catch to a predicted ADEF Pass 1 catch. A CF for each fish species f was derived from catch as:

2.2.2. Timed electrofishing (TEF) method The TEF sampling, adapted from IFI (2009), involved only two staff per site and one equipment unit with no stop-nets used to isolate the survey stretch (36 TEF sites, paired with the 36 ADEF sites). Electrofishing took place by wading upstream at a steady pace for exactly 10 min (continuous clock time). 2.3. Comparison between electrofishing methods 2.3.1. Compare species diversity between ADEF and TEF methods The efficacy of both sampling methods was compared in terms of the species richness (taxa) information they acquired. Species rarefaction curves were generated analytically using EstimateS (Colwell, 2013) for each of ADEF and TEF. The S(est) output (previously the Mao Tau estimator, Colwell et al., 2004) is a sample-based rarefaction method that calculates the total species richness with each additional sample collected (smoothed by averaging randomisations) based on the total species recorded across all samples. For this purpose, Salmo trutta and Salmo salar were each split into two separate ‘taxa’: fry (aged 0+) and parr (aged 1+ and older). For each sampling method, any taxa recorded at only a single site were removed a priori; this included Platichthys flesus (only present at one location very close to the sea) and Rutilus rutilus, Gobio gobio and Perca fluviatilis (typically deeper water species and rare in the shallow study system). For ADEF surveys (N = 36 and 11 taxa), Platichthys flesus, Perca fluviatilis and Rutilus rutilus were excluded, while for TEF surveys (N = 36 and 11 taxa) only Rutilus rutilus was excluded.

S

CFf =

• P-Sites: Paired ADEF and TEF data collected at 36 sites on the Table 1 Summary of three datasets used to develop catch conversion protocol.

A-ADEF P-Sites A-TEF a

(P-ADEF) (P-TEF)

Period of survey

No of sites

Taxon richness

2008–2014 2015 2015 2009–2014

266 36 36 309

15 14a 12a 15

∑s=1

(ADEFp1fs / TEFfs ) S

where ADEFp1fs is ADEF Pass 1 catch of species f in sample s, TEFfs is TEF catch of species f in sample s, and S is the total number of paired samples in which species f was caught. Step 3: Compare species size-distributions between P-Sites. Size-distributions for sampled species having sufficient records were compared statistically (Kolmogorov-Smirnov tests, alpha = 0.01) between paired ADEF Pass 1 (P-Sites) and extrapolated TEF (P-Sites) catches to test whether ADEF Pass 1 sample size-structure could be reproduced using extrapolated TEF data. This process showed that the univariate CF defined in (Step 2) tended to over-weight small fish, and so an adjusted multiplier (CF × 0.75) was applied to fish smaller than the 25th percentile of length by species. This generic multiplier for small fish was a pragmatic solution for multiple species, but a precise value for individual species could be analytically-derived. Step 4: Test species CF using archived validation datasets. The length-based bivariate CFf (Step 3) were used to expand catch by species in an independent ‘validation’ TEF dataset (A-TEF). A t-test (alpha = 0.01) for each species was then used to compare these extrapolated A-TEF catches (number per sampling site) with observed Pass 1 results from a large archived ADEF dataset (A-ADEF). In addition, bootstrapped means (boot function in R with 2000 re-sampling events) and 95% confidence intervals (CI) were estimated from the extrapolated A-TEF catch estimates by sample for each species, and these statistics were compared visually with the mean and 95% CI of Pass 1 catch for observed A-ADEF by species. These comparisons clarified whether bivariate CF from the 2015 paired hauls (P-Sites) could be applied retroactively to independent historical data. Step 5: Predict ADEF electrofishing catch curves from extrapolated TEF catch. The regression model (y = a + b × log x) selected in (Step 1) was used to predict cumulative ‘ADEF’ catches (Pass 2 and 3) and associated cumulative catch curves by species from extrapolated archived TEF (ATEF) samples. These predicted cumulative catches and curves were then compared with observed (archived) ADEF catches (A-ADEF) by testing for a significant ‘electrofishing method effect’ in the regression of

2.3.2. Protocol for predicting ADEF catch from TEF catch A Five-step protocol was developed to extrapolate single pass TEF fish catches to a predicted ADEF Pass 1 catch and then to a predicted ADEF three pass cumulative catch curve. An alternative approach might have been to model the ADEF cumulative catch directly from TEF catch, but the incremental process is more flexible for other users and allows a clearer picture of where error may emerge. The protocol compared predicted curves from extrapolated TEF catch with observed ADEF cumulative catch curves. Three data sets (Table 1 and Fig. 1) were used to develop and test the protocol; there were no differences in the equipment used to collect each dataset. This analysis used the R statistical software (R Core Team, 2015).

Dataset

Barrow catchment in 2015 (see above). ADEF (A-ADEF) validation dataset consisting of 410 wadeable surveys undertaken at 266 different sites, most of which were collected as part of Ireland’s national WFD surveillance monitoring programme (2008–2015). Sampling at each site comprised three passes and produced a cumulative catch curve. Archived TEF (A-TEF) validation dataset consisting of 309 surveys at unique sites (2008–2014). These data were collected as part of an Irish Environmental River Enhancement Programme (IFI, 2012).

• Archived

Salmo trutta and Salmo salar split into two taxa.

4

Fisheries Research xxx (xxxx) xxx–xxx

R. Matson et al.

Table 2 Physical characteristics of electrofishing sites for P-Sites (paired ADEF and TEF (n = 36)). P-Sites

Length (m)

Width (m)

Surface area (m2)

Mean depth (m)

Max depth (m)

P-ADEF Mean 39.71 St-Dev 2.08

5.70 1.60

228 68

0.21 0.09

0.44 0.14

P-TEF Mean St-Dev

5.38 1.66

151 65

0.22 0.09

0.42 0.16

27.86 7.22

Table 3 Univariate and size-based species conversion factors (CF) for extrapolating TEF catch to predicted ADEF pass 1 catch. 25%len is the 25th percentile of length (cm) by species: the size-based CF is applied to fish smaller than this threshold. For Leuciscus leuciscus and Anguilla anguilla (*) there were insufficient data to split the sample into two size components.

cumulative catch on pass, by applying the model: cumulative catch = a + b × log (pass) + method, where method is a categorical effect of electrofishing method (A-TEF vs. A-ADEF).

3.1. Study sites There was a marginal difference in stream width between ADEF (mean = 5.70) and TEF (mean = 5.38) sampling sites (P < 0.05, t = −2.6123) (Table 2). There was no significant difference in site depth between both ADEF (mean = 0.21 m) and TEF (mean = 0.22 m) samples. As expected, survey length (m) and total surface area sampled (m2) differed between sampling methods. There were no major weather events (e.g. rainfall) over the sampling period.

Fifteen fish taxa were recorded in A-ADEF and 13 in A-TEF: Abramis brama, Salmo trutta (brown trout and sea trout), Anguilla anguilla, Platichthys flesus, Gobio gobio, Lampetra spp., Phoxinus phoxinus, Pungitius pungitius, Perca fluviatilis, Esox lucius, Rutilus rutilus, Salmo salar, Barbatula barbatula and Gasterosteus aculeatus (Table 1). Thirteen fish taxa were recorded in the P-Sites (P-ADEF (12 taxa) and P-TEF (10taxa)), but five species (Platichthys flesus, Gobio gobio, Pungitius pungitius, Rutilus rutilus and Perca fluviatilis) had too few records (< 2 paired sites) to produce a CF.

100% of total fish species richness

9

75% of total fish species richness

6

20

Size-based CF

Salmo trutta Leuciscus leuciscus Anguilla anguilla Lampetra spp. Phoxinus phoxinus Salmo salar Barbatula barbatula Gasterosteus aculeatus Mean

2.11 3.81 3.25 1.74 1.55 2.01 2.14 3.12 2.47

7 * * 10 4 5 6 2 *

1.58 * * 1.31 1.16 1.51 1.61 2.34 *

Fig. 2. A comparison of species richness between tested electrofishing methods using rarefaction.

TEF

12

10

25%len (cm)

3.2.2. Protocol for extrapolating TEF to ADEF Step 1: Determine form of observed cumulative catch data from ADEF electrofishing. The cumulative catch in archived 3-pass ADEF electrofishing data (A-ADEF) was characterized by a regression model: cumulative catch = a + b × log (pass). Step 2: Derive univariate conversion factors (CF) to extrapolate TEF catch to ADEF Pass 1 catch. Preliminary univariate CF could be derived for the most abundant eight species in the paired samples (P-Sites). These CF are applied as a multiplier: TEF catch × Univariate CFf = ADEF Pass 1 catch (Table 3). A mean value can be applied for less abundant species. Step 3: Compare species size-distributions between ADEF Pass 1 and extrapolated TEF catch. Initial comparisons of size-frequency between sampling methods showed that the univariate CF typically over-weighted catches of the

3.2. Comparison between electrofishing methods

0

Univariate CF

3.2.1. Compare species diversity between ADEF and TEF methods Based on the rarefaction curves (Fig. 2), it took between two and three ADEF samples to reach 75% of the total taxa diversity present in the system. The ten-minute surveys reached this level of diversity between three and four samples. Both methods reached an asymptote (100% of all available taxa (N = 11), excluding rare taxa) after 20 samples.

3. Results

ADEF

Species

30

5

Fisheries Research xxx (xxxx) xxx–xxx

R. Matson et al.

Fig. 3. Compare species size-frequency distributions between ADEF Pass 1 and extrapolated TEF catch: Size-distributions for sufficiently abundant sampled species in the paired samples (PSites). Species size-distributions were statistically similar for all tested species (Two-sample Kolmogorov-Smirnov test, P > 0.01).

differ for Salmo trutta, Barbatula barbatula, Gasterosteus aculeatus and Pungitius pungitius.

smallest length classes. An adjusted CF (CF × 0.75) was then applied to TEF catch for individuals smaller than the 25th percentile of length by species, where 0.75 was an arbitrary value that down-weighted the influence of small fish. Extrapolating TEF data using these size-based (bivariate) CF (Table 3) produced size-frequency distributions that were statistically similar (ks test, P > 0.01) to those for observed ADEF Pass 1 data for all tested species in the 2015 paired samples (P-Sites). Only six fish species (Salmo trutta, Lampetra spp., Phoxinus phoxinus, Salmo salar, Barbatula barbatula and Gasterosteus aculeatus) had sufficient records to produce size data that could be compared statistically between paired samples (Fig. 3). As an example of applying CF, Salmo trutta > 7 cm should be extrapolated as: no. of ADEF Pass 1 Salmo trutta > 7 cm = no. of TEF Salmo trutta > 7 cm × 2.11, and Salmo trutta ≤7 cm extrapolated as: no. of ADEF Pass 1 Salmo trutta < 7 cm = no. of TEF Salmo trutta < 7 cm × 1.58. Step 4: Test species CF using a validation TEF dataset (A-TEF). Estimates of mean Pass 1 catch from extrapolated TEF data (A-TEF) were statistically similar (t test, P < 0.01) to observed ADEF Pass 1 values (A-ADEF) for all species except Anguilla anguilla and Pungitius pungitius. This similarity was evident in the visual comparison of mean and 95% confidence intervals by species (Fig. 4). Step 5: Predict ADEF electrofishing catch curves from extrapolated TEF catch. Catch and fitted cumulative catch curves by species, predicted from extrapolated A-TEF catches were typically similar to observed A-ADEF catches and fitted curves (Fig. 5). There was no significant difference (sampling method effect, log-linear regression, P > 0.01) between observed and extrapolated catch curves for Anguilla anguilla, Platichthys flesus, Gobio gobio, Lampetra spp., Phoxinus phoxinus, Perca fluviatilis, Esox Lucius, Rutilus rutilus or Salmo salar. Cumulative catch curves did

4. Discussion There are strong incentives to move from area delineated multiple pass electrofishing (ADEF) to more efficient single pass timed and single pass area delineated electrofishing techniques (e.g. TEF) due to resource savings. We found that a rapid assessment TEF approach can achieve similar species diversity estimates to ADEF with less sampling effort. A simple and robust protocol for extrapolating catch data from TEF is presented to support direct comparison with cumulative catch from area delineated depletion (ADEF) electrofishing methods in small streams (less than 10 m in wetted width). Catches and cumulative catch curves extrapolated from TEF catches using this protocol were similar to observed ADEF results for most tested fish species. Resources saved by using rapid assessment approaches such as TEF could be applied to extending the spatial or temporal coverage of sampling programmes, e.g. for the WFD. 4.1. Moving from ADEF to TEF sampling Rapid bioassessment techniques are commonly used to assess fish in rivers all over the world (Barbour et al., 1999; Kleynhans, 1999; Utrup and Fisher, 2006). Many authors demonstrate that a simple approach can be just as effective for assessing the fish community structure as a more resource-intensive approach, as long as the sample unit is increased accordingly (Bateman et al., 2005; Hughes et al., 2002). Careful consideration should be made of survey goals, e.g. whether species diversity or population estimates are required. TEF requires fewer resources and can sample a greater number of 6

Fisheries Research xxx (xxxx) xxx–xxx

R. Matson et al.

ADEF

TEF

ADEF

TEF

ADEF

TEF

60 40 20 0

Gasterosteus aculeatus Gasterosteus aculeatus

Gobio gobio

Gobio gobio

Lampetra spp.

Lampetra spp.

ADEF

TEF

ADEF

TEF

ADEF

TEF

Perca fluviatilis

Perca fluviatilis

Phoxinus phoxinus

Phoxinus phoxinus

Platichthys flesus

Platichthys flesus

ADEF

TEF

ADEF

TEF

ADEF

TEF

Pungitius pungitius

Pungitius pungitius

Rutilus rutilus

Rutilus rutilus

Salmo salar

Salmo salar

ADEF

TEF

ADEF

TEF

ADEF

TEF

Salmo trutta fario

Salmo trutta fario

ADEF

TEF

60 40 20 0

60 40 20 0

60 40 20 0

60 40 20

Fig. 4. Test species CF using a validation TEF dataset: Extrapolated mean TEF (A-TEF) catches by species with bootstrap 95% CI compared to observed mean Pass 1 catches with 95% CI in an archived ADEF electrofishing dataset (A-ADEF).

Barrow and validated using a wide range of site types across Ireland.

sites in the same time that it would take to intensively sample fewer sites using ADEF. Wider sample coverage can yield more complete catchment-level information. This additional coverage might be of particular benefit in systems that are ecologically heterogeneous, although the current analysis does not evaluate this effect in Irish streams. Rapid bioassessment approaches such as this also have important logistical benefits. In the TEF used in the present study, no stopnets were used to isolate the sampled river stretch. This technique enables more rapid surveys, but has the potential disadvantage of fish displacement, i.e. the recruitment and escape of fish from the sample stretch (Edwards et al., 2003; Lobón-Cerviá and Utrilla, 1993). Previous work suggests that the effects of such displacement do not render the method ineffective (Bateman et al., 2005; Edwards et al., 2003; Reid et al., 2008). By removing the need for stop-nets to isolate the survey stretch, the time required for equipment biosecurity protocols is also drastically reduced, as is the space required when transporting large heavy equipment on surveys. Furthermore, the spread of invasive species including plants, molluscs and small fish can be reduced when moving from site to site within the same or different catchments (Caffrey et al., 2015). The stop nets required for ADEF have a greater capacity to harbour unnoticed biological material and cannot be reliably disinfected in the field, unlike other equipment including waders, buckets and fishing nets. The TEF technique used in the current study can be applied to small wadeable streams less than 10 m in width. It is applicable across a wide range of channel types, as represented by the sites surveyed in the River

4.2. Catch extrapolation protocol TEF catches can differ from ADEF catches by under-representing species and size-classes of lower catchability (Odenkirk and Smith, 2005; Vehanen et al., 2012; Zalewski and Cowx, 1990). Our protocol attempts to address these issues by applying ‘bivariate’ species-specific corrections that extrapolate catch numbers while accounting for sizebased differences in catch within species. These corrections produce Pass 1 catch estimates and size-distributions that are similar to observed ADEF data (Figs. 3 and 4). A simple regression model then uses extrapolated TEF catch to predict 3-pass cumulative ‘catch’ that is also similar to observed ADEF catch (Fig. 5). An obvious potential concern is that our approach is too dependent on the input data and that it will not function in other systems or periods. A validation process was been undertaken by applying the protocol to two large independent datasets covering a wide range of catchments, channels and site types over an eight year period (2008–2015). This validation suggests some generality in the approach; Vehanen et al. (2012) found comparable differences between untimed single pass and three-pass electrofishing across different river systems. Other assessed systems with differing fish communities and underlying geology will require calculation of targeted species-specific extrapolation factors for timed electrofishing catch. It may also be necessary to modify the protocol to address particular issues with extrapolating catch for given species or systems. This

7

Fisheries Research xxx (xxxx) xxx–xxx

R. Matson et al.

80

500

200

400

75

60

150

300

50

40

100

200 100

50

0

0 1

2

3

25

20 0

0 1

Lampetra spp.

2

3

1

Phoxinus phoxinus

2

3

1000

150 100

500 200

250

0 2

3

50 0

0

0 1

1

Esox Lucius

2

3

1

Rutilus rutilus

10

2

3

1

2

3

1

Salmo salar

400

800

300

600

200

400

100

100

200

50

0

0 1

2

3

2

3

Barbatula barbatula

200 150

5 0

3

600 400

100

2 Perca fluviatilis

750

200

1

Pungitius pungitius

0 1

2

3

1

2

3

Gasterosteus aculeatus

1500 1000 500 0 1

2

3

Fig. 5. Predict probable ADEF electrofishing catch curves from extrapolated TEF catch: Observed ADEF electrofishing (A-ADEF) catches with fitted curves and modelled catch and fitted curves predicted from extrapolated archived TEF (A-TEF) catches by species. Cumulative catch curves differed for Salmo trutta, Barbatula barbatula, Gasterosteus aculeatus and Pungitius pungitius.

kind of national ‘interpretation’ of state indicators (e.g. Shephard et al., 2011) is accepted within EU environmental directives and should not be a problem as long as modifications are transparent and sufficiently simple that they have application outside a particular case. A strong similarity was found between modelled catch data from extrapolated TEF catch and observed ADEF catch curves for the Barrow catchment and for two larger Irish datasets. This result suggests that conducting a fish community assessment using extrapolated TEF data might produce a similar outcome to an assessment based on observed ADEF data, and that a shift between methods could be relatively simple. Any move towards rapid assessment based on timed electrofishing might demand a more comprehensive ‘benchmarking’ process, across a broader range of river typologies, a wider geographical range or in larger channels over 10 m in width.

difference was only slight and probably due to small differences in sampling location rather than a change in channel conditions. Therefore, the current findings, while applicable to a wide range of different river types, should only be transferred to channels of a similar width and depth. The freshwater ichthyofauna of Ireland is relatively limited in comparison to other ecoregions around Europe and further reduced when only rivers are considered. Perca fluviatilis and Rutilus rutilus are not rare species to Ireland or to the River Barrow’s deeper sections. However, their restricted distribution in the shallow wadeable streams characteristic of this river’s tributaries, limited the data that could be collected pertaining to them and used to develop conversion factors. This result suggests that cumulative catch curves based on bivariate CF estimated from the 2015 paired samples (P-Sites) can be applied to historical electrofishing data, but that further adjustment of curves may be necessary for some species. Platichthys flesus is a marine and estuarine species only found in the lower reaches of a river, a site type beyond the focus of this study. Baldwin and Aprahamian (2011) found that there was no difference in catch efficiency between multi-species surveys and eel-specific surveys, while the ICES Working Group on Eels (WGEEL) reported that eel density was substantially lower during multi-species surveys when compared to eel targeted surveys (ICES, 2007). Previous WFD electrofishing work conducted in Ireland (Delanty et al., 2017; Kelly et al., 2015), suggests that eels are often more prevalent in the second and third passes after disturbance during the first pass. Lampetra spp. present some difficulties when applying standard electrofishing methods and the described TEF method will probably suffer similar limitations. Although Lampetra spp. may emerge during the process, estimations of their abundance should be treated very cautiously. Regardless, standard

4.3. Methodology limitations The River Barrow’s large catchment area provided a useful study area for this work because it is relatively diverse for its size, with varied underlying geology characterised by calcareous limestone in the north and west and siliceous granite and slate to the east. While the main channel itself was too deep and wide for inclusion in this study, its numerous sub-catchments were ideal as they drain land used for a variety of different purposes including pasture, tillage farming, woodland and peatbog. The topography of the river varies from source to sea, with steep slopes in the upper reaches (4%) and gentle gradients (0.004%) in its lower watershed (Delanty et al., 2017). The majority of the paired sites used in this study (P-Sites) were less than 10 m in width, requiring only one or two operators to cover them effectively. Although there was a difference in width for P-Sites between both surveys, this

8

Fisheries Research xxx (xxxx) xxx–xxx

R. Matson et al.

Occasional Paper No. 50. Rome: Food and Agriculture Organization of the United Nations. Chao, A., Colwell, R.K., Lin, C.W., Gotelli, N.J., 2009. Sufficient sampling for asymptotic minimum species richness estimators. Ecology 90 (4), 1125–1133. Colwell, R.K., Mao, C.X., Chang, J., 2004. Interpolating, extrapolating, and comparing incidence-based species accumulation curves. Ecology 85, 2717–2727. Colwell, R.K., 2013. EstimateS: Statistical Estimation of Species Richness and Shared Species from Samples. Version 9. User's Guide and application published at: http:// purl.oclc.org/estimates. Crozier, W.W., Kennedy, G.J.A., 1994. Application of semi-quantitative electrofishing to juvenile salmonid stock surveys. J. Fish Biol. 45, 159–164. Crozier, W.W., Kennedy, G.J.A., 1995. The relationship between a summer fry (0+) abundance index, derived from semi-quantitative electrofishing, and egg deposition of Atlantic salmon, in the River Bush, Northern Ireland. J. Fish Biol. 47 (6), 1055–1062. Delanty, K., Kelly, F.L., McLoone, P., Matson, R., O’Briain, R., Gordon, P., Cierpal, D., Connor, L., Corcoran, W., Coyne, J., Feeney, R., Morrissey, E., 2017. Fish Stock Assessment of the River Barrow Catchment 2015. Inland Fisheries Ireland, 3044 Lake Drive, Citywest Business Campus, Dublin 24, Ireland. Edwards, M.R., Combs, D.L., Cook, S.B., Allen, M., 2003. Comparison of single-pass electrofishing to depletion sampling for surveying fish assemblages in small warm water streams. J. Freshw. Ecol. 18, 625–634. FAME CONSORTIUM, 2004. Manual for the Application of the European Fish Index – EFI. A Fish Based Method to Assess the Ecological Status of European Rivers in Support of the Water Framework Directive. Version 1.1. January 2005. Fausch, K.D., Karr, J.R., Yant, P.R., 1984. Regional application of an index of biotic integrity based on stream fish communities. Trans. Am. Fish. Soc. 113 (1), 39–55. Foley, K., Rosenberger, A., Mueter, F., 2015. Effectiveness of single-pass backpack electrofishing to estimate juvenile coho salmon abundance in Alaskan headwater streams. Fish. Sci. 81 (4), 601–610. Gargan, P.G., Roche, W.K., 2011. Catchment-wide electro-fishing as a means of salmon stock assessment in Ireland. Atlantic Salmon Trust J. 2011, 25–27. Gotelli, N.J., Colwell, R.K., 2011. Estimating species richness. Biol. Divers. 12, 39–54. Habera, J.W., Strange, R.J., Carter, B.D., Moore, S.E., 1996. Short-term mortality and injury of rainbow trout caused by three-pass AC electrofishing in a southern Appalachian stream. North Am. J. Fish. Manag. 16 (1), 192–200. Harvey, J., Cowx, I., 2003. Monitoring the River, Brook and Sea Lamprey, Lampetra fluviatilis, L. planeri and Petromyzon marinus. Conserving Natura 2000 Rivers Monitoring Series No. 5. English Nature, Peterborough. Hughes, R.M., Kaufmann, P.R., Herlihy, A.T., Intelmann, S.S., Corbett, S.C., Arbogast, M.C., Hjort, R.C., 2002. Electrofishing distance needed to estimate fish species richness in raftable Oregon rivers. North Am. J. Fish. Manag. 22 (4), 1229–1240. ICES, 2007. Report of the 2007 Session of the Joint EIFAC/ICES Working Group on Eels. CM2007/ACFM, 23: 142p and country reports. IFI, 2009. Environmental River Enhancement Programme Annual Report 2009. Inland Fisheries Ireland, Dublin, Ireland. IFI, 2012. Environmental River Enhancement Programme (EREP); Field Survey Methodologies. Inland Fisheries Ireland, Dublin, Ireland. Jones, M.L., Stockwell, J.D., 1995. A rapid assessment procedure for the enumeration of salmonine populations in streams. North Am. J. Fish. Manag. 15 (3), 551–562. Kelly, F., Champ, T., McDonnell, N., Kelly-Quinn, M., Harrison, S., Arbuthnott, A., Giller, P., Joy, M., McCarthy, K., Cullen, P., Harrod, C., 2007. Investigation of the Relationship between Fish Stocks, Ecological Quality Ratings (Q-Values), Environmental Factors and Degree of Eutrophication. Environmental Protection Agency, Wexford. Kelly, F.L., Matson, R., Delanty, K., Connor, L., O’Briain, R., Gordon, P., Corcoran, W., Coyne, J., Morrissey, E., Cierpal, D., Rocks, K., 2015. Sampling Fish for the Water Framework Directive, Rivers 2016. Inland Fisheries Ireland, 3044 Lake Drive, Citywest Business Campus, Dublin 24, Ireland. Kleynhans, C., 1999. The development of a fish index to assess the biological integrity of South African rivers. Water SA 25 (3), 265–278. Kocovsky, P.M., Gowan, C., Fausch, K.D., Riley, S.C., 1997. Spinal injury rates in three wild trout populations in Colorado after eight years of backpack electrofishing. North Am. J. Fish. Manag. 17 (2), 308–313. Lobón-Cerviá, J., Utrilla, C.G., 1993. A simple model to determine stream trout (Salmo trutta L.) densities based on one removal with electrofishing. Fish. Res. 15 (4), 369–378. Lyons, J., 1992. The length of stream to sample with a towed electrofishing unit when fish species richness is estimated. North Am. J. Fish. Manag. 12 (1), 198–203. Meador, M.R., McIntyre, J.P., Pollock, K.H., 2003. Assessing the efficacy of single-pass backpack electrofishing to characterize fish community structure. Trans. Am. Fish. Soc. 132 (1), 39–46. Mesa, M.G., Schreck, C.B., 1989. Electrofishing mark–recapture and depletion methodologies evoke behavioral and physiological changes in cutthroat trout. Trans. Am. Fish. Soc. 118 (6), 644–658. Odenkirk, J., Smith, S., 2005. Single-versus multiple-pass boat electrofishing for assessing smallmouth bass populations in Virginia rivers. North Am. J. Fish. Manag. 25 (2), 717–724. PAST: Paleontological Statistics software package for education and data analysis. Palaeontologia Electronica 4(1) 9 pp. Paller, M.H., 1995. Relationships among number of fish species sampled, reach length surveyed, and sampling effort in South Carolina coastal plain streams. North Am. J. Fish. Manag. 15 (1), 110–120. Panek, F.M., Densmore, C.L., 2013. Frequency and severity of trauma in fishes subjected to multiple-pass depletion electrofishing. North Am. J. Fish. Manag. 33 (1), 178–185. Peterson, N.P., Cederholm, C.J., 1984. A comparison of the removal and mark-recapture

electrofishing methods are generally considered ineffective for sampling Lampetra spp. quantitatively, and a targeted approach should be employed (e.g. Harvey and Cowx, 2003). 4.4. Resource and biosecurity implications Based on the resources used during this study, a team of four people driving two four-wheel-drive vehicles full of equipment, could survey two sites (approximately 45 m) per day using the ADEF method. By dividing these people into two teams it was possible to survey 10 sites (approximately 35 m in length) in a day using the TEF method. Such evaluations should inform decisions about changing sampling protocols and whether single pass electrofishing is appropriate for most fisheries applications (including WFD assessment), or whether its main role will be as a tool for screening rivers in order to identify sites that require more intensive study. Furthermore, additional biosecurity measures required for cleaning stop-nets also make heavy demands on both time and monetary expense that can be avoided through application of the TEF method. It is recommended that a detailed cost benefit analysis be carried out to estimate potential savings. Acknowledgements The authors would like to thank their colleagues in IFI. The A-ADEF dataset was compiled as part of Inland Fisheries Irelands national Water Framework Directive fish monitoring programme funded by the Department of Communications, Climate Action and Environment. The A-TEF data set was compiled as part of a series of catchment-wide studies on fish community in arterially drained catchments, funded by the Office of Public Works (Drainage Division). We would also likely to acknowledge the many angling clubs, fishery and land owners who granted permission and access during fish stock assessment surveys. We also acknowledge the Irish Environmental Protection Agency for providing water chemistry data relating to the River Barrow Catchment, collected as part of the EU WFD monitoring programme. The manuscript includes Ordnance Survey Ireland data reproduced under OSi Copyright Permit No. MP 007508. Unauthorised reproduction infringes Ordnance Survey Ireland and Government of Ireland copyright. © Ordnance Survey Ireland, 2017. References Angermeier, P.L., Karr, J.R., 1986. Applying an index of biotic integrity based on stream fish communities: considerations in sampling and interpretation. North Am. J. Fish. Manag. 6 (3), 418–429. Angermeier, Paul L., Smogor, Roy A., 1995. Estimating number of species and relative abundances in stream-fish communities: effects of sampling effort and discontinuous spatial distributions. Can. J. Fish. Aquat. Sci. 52 (5), 936–949. Baldwin, L., Aprahamian, M., 2012. An evaluation of electric fishing for assessment of resident eel in rivers. Fisheries Research 123, 4–8. Barbour, M.T., Gerristen, J., Snyder, B.D., Stribling, J.B., 1999. Rapid Bioassessment Protocols for Use in Wadeable Streams and Rivers: Periphyton, Benthic Macroinvertebrates and Fish, second edition. EPA. 841-B-99-002. U.S. Environmental Protection Agency, Office of Water, Washington, D.C. Bateman, D.S., Gresswell, R.E., Torgersen, C.E., 2005. Evaluating single-pass catch as a tool for identifying spatial pattern in fish distribution. J. Freshw. Ecol. 20, 335–345. Beaumont, W.R.C., Taylor, A.A.L., Lee, M.J., Welton, J.S., 2002. Guidelines for Electric Fishing Best Practice. Environment Agency R & D Technical Report W2-054/TR. Bertrand, K.N., Gido, K.B., Guy, C.S., 2006. An evaluation of single-pass versus multiplepass backpack electrofishing to estimate trends in species abundance and richness in prairie streams. Trans. Kansas Acad. Sci. 109, 131–138. Birk, S., Bonne, W., Borja, A., Brucet, S., Courrat, A., Poikane, S., van de Bund, W., Zampoukas, N., Hering, D., 2012. Three hundred ways to assess Europe’s surface waters: an almost complete overview of biological methods to implement the Water Framework Directive. Ecol. Indic. 18, 31–41. Bohlin, T., Hamrin, S., Heggberget, T.G., Rasmussen, G., Saltveit, S.J., 1989. Electrofishing—theory and practice with special emphasis on salmonids. Hydrobiologia 173 (1), 9–43. CEN EN 14011 8, 2003. Water Quality – Sampling of Fish with Electricity. European Committee for Standardization, Brussels 18 pp. Caffrey, J.M., Gallagher, C., Dick, J.T.A., Lucy, F., 2015. Aquatic invasive alien species: top issues for their management: outcomes. In: Freshwater Invasives: Networking for Strategy: IFI/EIFAAC Conference (FINS). Galway, Ireland, 9–11 April 2013. EIFAAC

9

Fisheries Research xxx (xxxx) xxx–xxx

R. Matson et al.

Temple, G.M., Pearsons, T.N., 2003. Comparison of Single vs. Multiple Pass Electrofishing Effort to Monitor Fish Populations in Wadeable Streams. Yakima River species interactions studies, Bonneville Power Administration, Annual Report, 2004. pp. 32–54. Utrup, N.J., Fisher, W.L., 2006. Development of a rapid bioassessment protocol for sampling fish in large prairie rivers. North Am. J. Fish. Manag. 26 (3), 714–726. Vehanen, T., Sutela, T., Jounela, P., Huusko, A., Mäki‐Petäys, A., 2012. Assessing electric fishing sampling effort to estimate stream fish assemblage attributes. Fish. Manag. Ecol. 20 (1), 10–20. Wyatt, R.J., 2002. Estimating riverine fish population size from single-and multiple-pass removal sampling using a hierarchical model. Can. J. Fish. Aquat. Sci. 59 (4), 695–706. Zalewski, M., Cowx, I.G., 1990. Factors affecting the efficiency of electric fishing. Fishing with Electricity: Applications in Freshwater Fisheries Management. Fishing News Books Oxford, UK, pp. 89–111. Zippin, C., 1956. An evaluation of the removal method of estimating animal populations. Biometrics 12, 163–189.

methods of population estimation for juvenile coho salmon in a small stream. North Am. J. Fish. Manag. 4 (1), 99–102. R Core Team, 2015. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/. Reid, S.M., Jones, N.E., Yunker, G., 2008. Evaluation of single-pass electrofishing and rapid habitat assessment for monitoring redside dace. North Am. J. Fish. Manag. 28 (1), 50–56. Schmutz, S., Cowx, I.G., Haidvogl, G., Pont, D., 2007. Fish-based methods for assessing European running waters: a synthesis. Fish. Manag. Ecol. 14 (6), 369–380. Shephard, S., Reid, D.G., Greenstreet, S.P., 2011. Interpreting the large fish indicator for the Celtic Sea. ICES J. Mar. Sci. 68 (9), 1963–1972. Tarwid, K., 1956. Beklemishev’s criterium of faunistic reliability of quantitative samples. Ekol. Pol. B 2, 27–31. Teixeira-de Mello, F., Kristensen, E.A., Meerhoff, M., González-Bergonzoni, I., BaattrupPedersen, A., Iglesias, C., Kristensen, P.B., Mazzeo, N., Jeppesen, E., 2014. Monitoring fish communities in wadeable lowland streams: comparing the efficiency of electrofishing methods at contrasting fish assemblages. Environ. Monit. Assess. 186 (3), 1665–1677.

10