Detailed Insight into Gillnet Catches: Fish Directivity and Micro Distribution

Abstract

Gillnets are widely used in research and commercial fishery activities. As passive gear, gillnets can be selective and dependent on the diel migration of fish. In areas with limited littoral extent, inshore–offshore migration may cause bias in the gillnet catch. Our hypothesis was that some factors, such as gillnet saturation, fish depletion, or chemical cues, could be the cause of the bias. We used a total of 66 CEN gillnets deployed at Římov Reservoir parallel to the shore at different positions of littoral-pelagic gradient. Individual fish direction was recorded from inshore, offshore, or unknown direction (i.e., entangled fish). A total of 5791 fishes from nine different species were caught. For most fish, it was possible to determine their directivity, and most fish were captured in littoral or first pelagic gillnets. Shallower and deeper benthic gillnets differed in their bleak (Alburnus alburnus) catch. No significant differences were found between fish directions. At the species level, only asp (Leuciscus aspius) and ruffe (Gymnocephalus cernua) showed differences between the captured directions in one case. The results support the assumption that gillnet capture is a random process that to a great extent is connected to random local movements. This is good news for fish monitoring projects. Sampling catch is likely to reflect true changes in the fish community, and not the effects of the deployment of the sampling gear. The experiment also showed that fish directivity statistics can be used for investigation of fish behavior and gear performance.

1. Introduction

Gillnets are among the most widely used gear in research and commercial applications. Gillnets are also basic tools for sampling fish communities to monitor the ecological quality of inland waters [1,2,3]. The use of gillnets for scientific purposes has been standardized in Europe [4], and basic advantages of multimesh gillnets include the ability to catch a wide range of fish sizes and species, the possibility of installation in various habitats, ease of operation, and low cost [5,6,7]. As passive gear, gillnets can show some selectivity in relation to their catch, and therefore, this may influence the representative abundance of some species or their size distribution [1,6,8,9]. Because of the massive use of CEN (Comité Européen de Normalisation [4]) multimesh gillnets for fish monitoring, it is important to understand whether the sampling results are influenced by fish swimming directivity or micro distribution.

Fish diel migration is defined as the shift of habitats or depths on a daily basis [10,11]. In Holarctic waters, fish older than young-of-year (YOY) migrate inshore in the evening, and offshore to deeper, open water habitats in the morning, while the YOY migrate offshore during the evening and inshore during the morning [12,13,14,15]. This migration is normally driven by light intensity, and balances the activities of searching for food and predation avoidance [11,13,16]. As the capture of fish in gillnets is correlated with the migration of the species and its intensity [8,17], it would be possible to assume that the local difference in the intensity of the migrations and density of migrants would be detectable in the gillnet catches.

In areas with a shorter distance between the littoral and pelagial zones, inshore–offshore movements may not represent the majority of fish movement trajectories. Most swimming distances are likely to be related to foraging and random movements within the habitat after relatively short inshore/offshore migration is completed. These random movements may account for most of the gillnet catches (Figure 1). If the gillnet is installed perpendicular to the shoreline or to the isobaths, the sample would lose depth sensitivity, as the benthic gillnet would span through several depth layers. The depth gradient is usually recorded as the most important gradient for fish distribution [7,17,18]. Therefore, in order to maintain vertical resolution of the samples, researchers are forced by steep bottom morphology to deploy the gillnets along the isobath lines or the shoreline (Figure 1).

Diel migration is probably not the only factor influencing directional catch. Gillnet saturation is also a proven phenomenon that influences catch [5,20]. The catch from earlier phases of migration can theoretically decrease the probability of new catches in later periods of migration due to gillnet occupancy and diminishing the number of fish available locally. This can lead to uneven gillnet catches [20]. In addition, fish can theoretically be alerted to danger by different chemical cues including predator odors and disturbance olfactory cues from conspecifics [21]. These effects can influence the catch, as the first fish caught would start to release cues and “scare” the fish from the gillnet area [22,23,24].

In a theoretical scenario, a gillnet set parallel to the shore or isobath could capture differences in the direction in which the fish were caught. When gillnets were set in the evening, the catch of fish migrating inshore (Figure 1, fish A) was more likely than the catch of fish moving offshore (Figure 1, fish B). No study has been found on the differences between fish capture and the direction of capture on gillnets. The aim of this study was to verify the common assumption that the direction of fish capture is a random process and has no directional influence on gillnet catch. Our hypothesis is that the shallowest gillnets would mostly be affected by the fish movement direction because gillnets may prevent some fish from migrating toward the shore (Figure 1). As a result, there may be a noticeable difference in catch directivity between shallow and deep benthic gillnets. We expect that benthic gillnets will show more variation in catch sizes as they are typically placed in areas with the highest fish abundance within the waterbody.

2. Materials and Methods

We used the fish sampling design reported by Moraes et al. (2021) [9], and the experiment was carried out at Římov Reservoir, South Bohemia, Elbe/Vltava Catchment, Czech Republic (48°50′55.0″ N 14°29′14.0″ E). Sampling dates were from 30 July to 2 August 2019. European Standard gillnets (ESGs), manufactured by Pokorny site, Brloh, Czech Republic, were deployed in six different locations in the Římov Reservoir. Benthic ESG gillnets 1.5 m in height × 30 m in length and 2.5 m mesh panels for each of the 12 mesh sizes was deployed in the littoral, while the pelagic gillnet 3 m in height × 30 m in length and 2.5 m mesh panels for each of the 12 mesh sizes was deployed in the open water. The mesh sizes of the ESG followed a geometric series with a ratio of approximately 1.25 (5, 6.25, 8, 10, 12.5, 15.5, 19.5, 24, 29, 35, 43, and 55 mm) in a random order.

For each location, we coded the gillnets; the first letter was M or S, with M representing “mild slope” and S for “steep slope” (see Figure 2). The second letter was P or B, with P representing “pelagic gillnets”, followed by a number (MP1–MP3, SP1–SP3), and B for “benthic gillnets”. The benthic gillnets had a third letter, where the last letter S represented “shallow gillnet” (MBS and SBS) at a depth of 0–1.7 m, and the last letter D for “deep gillnet” (MBD and SBD) at a depth of 1.5–3 m. The last code was the Center, which was a pelagic gillnet that separated the mild and steep slopes. During our analyses, the P2 and P3 of both slopes did not show a significant difference between their totals and species level as they were merged into mild true pelagic (MTP) and steep true pelagic (STP) (Figure 2).

We used seven pelagic gillnets and four benthic gillnets per each of six localities, so a total of 66 gillnets were deployed during this experiment. All gillnets were set parallel to the shore. The first benthic gillnet was deployed at a depth of 1.5–1.7 m, filling nearly the entire water column in the most inshore region. A second benthic gillnet was deployed at a depth of 2.8–3.0 m, filling the second half of the 0–3 m depth interval. The morphology of the flooded river valley caused one slope to be mild (sedimentary shore) and the opposite slope to be steep (eroded shore; Figure 2). Sampling was performed in accordance with CEN (Comité Européen de Normalisation) standard procedures, with the gillnets deployed 2 h before sunset and lifted 2 h after sunrise [4]. After being removed from the gillnet, the fish were identified to the species level, and the standard length (SL) was measured to the nearest millimeter and weighed in grams (g).

The catch per unit of effort (CPUE) and biomass per unit of effort (BPUE) were defined as the number of individuals per 1000 m² of gillnet area per night. We marked the gillnets with color-coded floaters to ensure that we were recording the direction in which the fish were approaching the gillnets before being caught. When the fish were removed from the gillnets, we recorded the direction in which they approached the gillnet (i.e., inshore or offshore). A smaller proportion of fish fell out of the gillnet during retrieval or were too entangled, preventing a proper definition of the direction they came from, and were thus classified as direction unknown. As the unknown category had few fishes per gillnet, it was not used in the models. The first pelagic gillnets (MP1 and SP1) were deployed at depths of above 3.5 m so that the gillnets did not touch the bottom. Further away pelagic gillnets P2, P3, and Center were deployed equidistantly along the transversal gradient of the reservoir (between right and left bank).

CPUE and BPUE were calculated as the mean of the total number of individuals divided by the total sampling effort (gillnet surface area). CPUE was calculated for individual species as well as for the entire fish assemblage.

Quasi-Poisson generalized linear models (GLMs) were applied to describe the differences in fish CPUE values with the direction of capture, as our data did not follow a normal distribution, but a positive skewed distribution with high deviation. The stats package was used to compute all GLMs [25]. All data analyses were performed using R software version 4.1.2 [25].

The global model for the gillnet analyses and the species level analyses follows the formula:

$$\mathit{G}\mathit{L}\mathit{M}=\mathit{V}\mathit{a}\mathit{l}\mathit{u}\mathit{e}~\mathit{D}\mathit{i}\mathit{r}\mathit{e}\mathit{c}\mathit{t}\mathit{i}\mathit{o}\mathit{n}\mathbf{\ast }(\mathit{N}\mathit{e}\mathit{t}\mathit{c}\mathit{o}\mathit{d}\mathit{e}+\mathit{L}\mathit{o}\mathit{c}\mathit{a}\mathit{l}\mathit{i}\mathit{t}\mathit{y})$$

where Value represents the mean of the gillnet in relation to the CPUE, BPUE, or SL. The direction refers to the inshore/offshore direction of capture of the gillnet, with the exception of the Center gillnet, where they were classified by the direction of the slope they were captured. Netcode represents the code for the gillnet position (for example, SP1) and the localities sampled in the reservoir. For the representation of the p-value, we followed the following standard: *** = p < 0.001, ** = p < 0.01, * = p < 0.05, and ns = p ≥ 0.05.

The coefficient of variation (CV) was also calculated for the inshore and offshore direction per gillnet:

$$\mathit{C}\mathit{V}=\mathit{\sigma }/\mathit{\mu }$$

where σ is the standard deviation of the direction-oriented CPUE and μ is the mean of the direction. The CV compares the degree of variation between the directions and also between gillnets.

3. Results

A total of 5791 fish of nine different species were caught: 76.19% bleak (Alburnus alburnus); 13.78% roach (Rutilus rutilus); 5.13% perch (Perca fluviatilis); 2.19% ruffe (Gymnocephalus cernua); 0.98% asp (Leuciscus aspius); 0.98% bream (Abramis brama); 0.50% pikeperch (Sander lucioperca); 0.14% rudd (Scardinius erythrophthalmus); and 0.10% wels catfish (Silurus glanis).

The model of CPUE showed no significance between the gillnets and the directions of capture (p-value = ns). Most fish were captured at the littoral or at P1 gillnets (Figure 3). The very highest fish density was recorded at the shallowest gillnets at the mild slope (Figure 4,

Table 1. Mean catch per unit of effort (inds. 1000 m² of gillnets) ± standard deviation by the direction of capture of each gillnet. p-value denotes the significance of the inshore/offshore direction fish CPUE. p = p-value, ** = p < 0.01, and ns = p ≥ 0.05.

	MBS			MBD			MP1
Species	Inshore	Offshore	p	Inshore	Offshore	p	Inshore	Offshore	p
Abramis brama	11.11 ± 18.59	29.63 ± 53.82	ns	3.7 ± 9.07	25.93 ± 29.54	ns	1.85 ± 4.54	5.56 ± 6.09	ns
Alburnus alburnus	611.11 ± 381.49	777.78 ± 688.24	ns	307.41 ± 380.62	437.04 ± 494.25	ns	437.04 ± 101.51	477.78 ± 170.4	ns
Gymnocephalus cernua	25.93 ± 21.85	51.85 ± 43.7	ns	66.67 ± 81.95	40.74 ± 40.77	**	0 ± 0	0 ± 0	ns
Leuciscus aspius	18.52 ± 25.98	22.22 ± 37.18	ns	11.11 ± 27.22	11.11 ± 12.17	ns	1.85 ± 4.54	0 ± 0	ns
Perca fluviatilis	155.56 ± 149.4	125.93 ± 62.33	ns	96.3 ± 66.91	192.59 ± 143.1	ns	7.41 ± 9.07	3.7 ± 5.74	ns
Rutilus rutilus	303.7 ± 165.5	314.81 ± 92.61	ns	162.96 ± 87.39	229.63 ± 97.03	ns	38.89 ± 32.77	14.81 ± 15.18	ns
Sander lucioperca	11.11 ± 18.59	7.41 ± 11.48	ns	7.41 ± 11.48	7.41 ± 18.14	ns	0 ± 0	0 ± 0	ns
Scardinius erythrophthalmus	0 ± 0	0 ± 0	ns	0 ± 0	0 ± 0	ns	0 ± 0	0 ± 0	ns
Silurus glanis	0 ± 0	0 ± 0	ns	3.7 ± 9.07	0 ± 0	ns	1.85 ± 4.54	0 ± 0	ns
Gillnet	1137.04 ± 465.59	1329.63 ± 674.09	ns	659.26 ± 455.35	944.44 ± 539	ns	488.89 ± 104.94	501.85 ± 151.6	ns
	MTP			Center			STP
Species	Inshore	Offshore	p	Mild	Steep	p	Inshore	Offshore	p
Abramis brama	1.85 ± 4.32	0 ± 0	ns	0 ± 0	3.7 ± 9.07	ns	4.63 ± 10	1.85 ± 4.32	ns
Alburnus alburnus	392.59 ± 119.51	414.81 ± 96.67	ns	348.15 ± 138.9	362.96 ± 124.66	ns	361.11 ± 127.61	342.59 ± 68.96	ns
Gymnocephalus cernua	0 ± 0	0 ± 0	ns	0 ± 0	0 ± 0	ns	0 ± 0	0 ± 0	ns
Leuciscus aspius	0.93 ± 3.21	1.85 ± 4.32	ns	0 ± 0	1.85 ± 4.54	ns	1.85 ± 6.42	2.78 ± 5.03	ns
Perca fluviatilis	0.93 ± 3.21	0.93 ± 3.21	ns	0 ± 0	1.85 ± 4.54	ns	0.93 ± 3.21	2.78 ± 5.03	ns
Rutilus rutilus	31.48 ± 22.14	19.44 ± 12.65	ns	20.37 ± 20.39	14.81 ± 16.73	ns	21.3 ± 16.04	18.52 ± 13.68	ns
Sander lucioperca	2.78 ± 5.03	0.93 ± 3.21	ns	1.85 ± 4.54	1.85 ± 4.54	ns	1.85 ± 4.32	0.93 ± 3.21	ns
Scardinius erythrophthalmus	0.93 ± 3.21	0 ± 0	ns	0 ± 0	0 ± 0	ns	0.93 ± 3.21	0 ± 0	ns
Silurus glanis	0 ± 0	0 ± 0	ns	0 ± 0	0 ± 0	ns	0 ± 0	0 ± 0	ns
Gillnet	431.48 ± 124.08	437.96 ± 101.22	ns	370.37 ± 146.51	387.04 ± 146.47	ns	392.59 ± 126.72	369.44 ± 70.61	ns
	SP1			SBD			SBS
Species	Inshore	Offshore	p	Inshore	Offshore	p	Inshore	Offshore	p
Abramis brama	1.85 ± 4.54	0 ± 0	ns	14.81 ± 26.91	11.11 ± 12.17	ns	11.11 ± 27.22	0 ± 0	ns
Alburnus alburnus	455.56 ± 168.36	625.93 ± 269.31	ns	44.44 ± 68.85	70.37 ± 172.37	ns	555.56 ± 800.49	237.04 ± 244.92	ns
Gymnocephalus cernua	0 ± 0	0 ± 0	ns	29.63 ± 33.46	62.96 ± 55.18	ns	62.96 ± 80.02	59.26 ± 53.82	ns
Leuciscus aspius	12.96 ± 8.36	1.85 ± 4.54	**	0 ± 0	0 ± 0	ns	14.81 ± 18.14	18.52 ± 16.73	ns
Perca fluviatilis	5.56 ± 9.3	5.56 ± 9.3	ns	55.56 ± 41.57	77.78 ± 43.89	ns	85.19 ± 32.71	70.37 ± 68.01	ns
Rutilus rutilus	35.19 ± 25.74	61.11 ± 48.05	ns	159.26 ± 87.11	144.44 ± 94.02	ns	344.44 ± 154.44	329.63 ± 235.77	ns
Sander lucioperca	1.85 ± 4.54	0 ± 0	ns	0 ± 0	0 ± 0	ns	3.7 ± 9.07	7.41 ± 18.14	ns
Scardinius erythrophthalmus	0 ± 0	7.41 ± 5.74	ns	0 ± 0	0 ± 0	ns	3.7 ± 9.07	0 ± 0	ns
Silurus glanis	0 ± 0	0 ± 0	ns	3.7 ± 9.07	7.41 ± 11.48	ns	3.7 ± 9.07	0 ± 0	ns
Gillnet	512.96 ± 186.91	701.85 ± 284.27	ns	307.41 ± 152.94	374.07 ± 258.93	ns	1085.19 ± 907.74	722.22 ± 324.09	ns

). All benthic gillnets contained a reasonable proportion of bleak, roach, perch, and ruffe. Bleak abruptly declined toward deeper and steeper sites, while in the pelagic gillnets, the bleak heavily dominated (Table 1).

In relation to the fish BPUE, no significant difference was found between the nets (p-value = ns) (Figure 5). The highest amount of biomass was in the MBS (

Table 2. Mean biomass per unit of effort BPUE (Kg) ± standard deviation by the direction of capture of each gillnet. For the full species scientific names, see Table 1. p = p-value, and ns = p ≥ 0.05.

	MBS			MBD			MP1
Species	Inshore	Offshore	p	Inshore	Offshore	p	Inshore	Offshore	p
A. brama	0.37 ± 0.29	3.35 ± 2.28	ns	0.13 ± 0.13	5.35 ± 4.66	ns	0.03 ± 0.03	0.93 ± 0.8	ns
A. alburnus	13.28 ± 3.37	14.67 ± 4.22	ns	6.04 ± 2.94	9.23 ± 3.96	ns	9.91 ± 0.97	11.23 ± 1.67	ns
G. cernua	0.19 ± 0.07	0.34 ± 0.13	ns	0.33 ± 0.16	0.16 ± 0.06	ns	0 ± 0	0 ± 0	ns
L. aspius	2.68 ± 1.66	1.19 ± 0.81	ns	4.17 ± 4.17	1.49 ± 1.08	ns	0.09 ± 0.09	0 ± 0	ns
P. fluviatilis	9.18 ± 2.61	7.22 ± 2.57	ns	3.88 ± 2.01	5.25 ± 2.4	ns	1.28 ± 0.78	0.81 ± 0.53	ns
R. rutilus	10.46 ± 3.3	9.04 ± 1.33	ns	4.33 ± 1.01	7.5 ± 2.13	ns	3.98 ± 1.71	1.94 ± 1.28	ns
S. lucioperca	3.88 ± 2.6	0.9 ± 0.6	ns	1.3 ± 0.82	0.58 ± 0.58	ns	0 ± 0	0 ± 0	ns
S. erythrophthalmus	0 ± 0	0 ± 0	ns	0 ± 0	0 ± 0	ns	0 ± 0	0 ± 0	ns
S. glanis	0 ± 0	0 ± 0	ns	0.92 ± 0.92	0 ± 0	ns	1.44 ± 1.44	0 ± 0	ns
Gillnet	40.03 ± 26.42	36.72 ± 9.24	ns	21.11 ± 25.84	29.56 ± 18.46	ns	16.73 ± 5	14.92 ± 2.5	ns
	MTP			Center			STP
Species	Inshore	Offshore	p	Mild	Steep	p	Inshore	Offshore	p
A. brama	0.04 ± 0.04	0 ± 0	ns	0 ± 0	0.13 ± 0.13	ns	0.14 ± 0.08	0.06 ± 0.05	ns
A. alburnus	8.86 ± 0.98	9.52 ± 0.69	ns	7.53 ± 1.33	8.22 ± 1.06	ns	8.4 ± 0.92	7.96 ± 0.68	ns
G. cernua	0 ± 0	0 ± 0	ns	0 ± 0	0 ± 0	ns	0 ± 0	0 ± 0	ns
L. aspius	0.24 ± 0.24	0.16 ± 0.11	ns	0 ± 0	1.07 ± 1.07	ns	0.35 ± 0.35	0.31 ± 0.22	ns
P. fluviatilis	0.21 ± 0.21	0.1 ± 0.1	ns	0 ± 0	0.25 ± 0.25	ns	0.3 ± 0.3	0.49 ± 0.29	ns
R. rutilus	1.97 ± 0.96	1.88 ± 0.63	ns	0.89 ± 0.77	0.12 ± 0.06	ns	0.7 ± 0.36	0.59 ± 0.34	ns
S. lucioperca	0.41 ± 0.3	1.05 ± 1.05	ns	1.09 ± 1.09	0.66 ± 0.66	ns	0.69 ± 0.49	0.82 ± 0.82	ns
S. erythrophthalmus	0.49 ± 0.49	0 ± 0	ns	0 ± 0	0 ± 0	ns	0.46 ± 0.46	0 ± 0	ns
S. glanis	0 ± 0	0 ± 0	ns	0 ± 0	0 ± 0	ns	0 ± 0	0 ± 0	ns
Gillnet	12.21 ± 4.26	12.71 ± 5.6	ns	9.51 ± 3.96	10.45 ± 4.78	ns	11.04 ± 3.88	10.23 ± 4.25	ns
	SP1			SBD			SBS
Species	Inshore	Offshore	p	Inshore	Offshore	p	Inshore	Offshore	p
A. brama	1.65 ± 1.65	0 ± 0	ns	0.17 ± 0.13	0.2 ± 0.1	ns	0.23 ± 0.23	0 ± 0	ns
A. alburnus	9.91 ± 1.78	12.29 ± 1.5	ns	0.68 ± 0.44	1.44 ± 1.44	ns	10.02 ± 5.51	4.6 ± 1.93	ns
G. cernua	0 ± 0	0 ± 0	ns	0.19 ± 0.08	0.46 ± 0.17	ns	0.35 ± 0.14	0.38 ± 0.14	ns
L. aspius	4 ± 2.41	0.08 ± 0.08	ns	0 ± 0	0 ± 0	ns	4.68 ± 3.88	1.02 ± 0.37	ns
P. fluviatilis	1.87 ± 1.41	0.42 ± 0.38	ns	2.77 ± 1.08	4.04 ± 2.35	ns	4.71 ± 3.17	2.46 ± 1.11	ns
R. rutilus	0.71 ± 0.22	1.22 ± 0.69	ns	4.46 ± 2.42	4.98 ± 2.39	ns	5.26 ± 1.06	5.16 ± 2.05	ns
S. lucioperca	0.29 ± 0.29	0 ± 0	ns	0 ± 0	0 ± 0	ns	4.49 ± 4.49	0.45 ± 0.45	ns
S. erythrophthalmus	0 ± 0	1.73 ± 0.94	ns	0 ± 0	0 ± 0	ns	1.06 ± 1.06	0 ± 0	ns
S. glanis	0 ± 0	0 ± 0	ns	0.54 ± 0.54	0.42 ± 0.33	ns	0.03 ± 0.03	0 ± 0	ns
Gillnet	18.43 ± 12.19	15.73 ± 5.94	ns	8.81 ± 7.9	11.53 ± 9.24	ns	30.83 ± 17.47	14.07 ± 6.91	ns

).

The variances of the CPUE values of the gillnets were the highest at the benthic gillnets (Figure 6). The pelagic gillnets showed lower levels of variance.

At the species level, the difference in CPUE between the capture directions also did not show significant differences in almost all gillnets and species, with the exception of ruffe in MDB and asp in SP1 (Table 1).

For both BPUE and SL, the results at the species level showed no significant difference between the directions in which the fish were captured in the different gillnets (p = ns) (Table 2 and

Table 3. Mean standard length (mm) ± standard deviation by the direction of capture of each gillnet. For the full species scientific names, see Table 1. p = p-value, ns = p ≥ 0.05.

	MBS			MBD			MP1
Species	Inshore	Offshore	p	Inshore	Offshore	p	Inshore	Offshore	p
A. brama	110.67 ± 25.42	150 ± 58.55	ns	118	163.57 ± 96.39	ns	92	163 ± 93.5	ns
A. alburnus	114.13 ± 16.97	109.48 ± 15.24	ns	110.8 ± 15.89	112.69 ± 18.3	ns	116.2 ± 15.69	117.83 ± 14.74	ns
G. cernua	67.57 ± 9.25	62 ± 18.49	ns	56.94 ± 15.1	53.55 ± 12.47	ns	-	-	ns
L. aspius	201 ± 57.6	152.67 ± 8.62	ns	243.33 ± 144.68	190 ± 73.82	ns	148	-	ns
P. fluviatilis	97 ± 74.3	104.79 ± 70.45	ns	88.78 ± 64.89	77.98 ± 53.31	ns	203.75 ± 28.69	222.5 ± 24.75	ns
R. rutilus	103.01 ± 37.84	99.39 ± 32.59	ns	103.84 ± 22.76	102.5 ± 35.69	ns	128.86 ± 72.27	142.5 ± 84.48	ns
S. lucioperca	290 ± 76.97	210 ± 28.28	ns	240 ± 0	167 ± 74.95	ns	-	-	ns
S. erythrophthalmus	-	-	ns	-	-	ns	-	-	ns
S. glanis	-	-	ns	340	-	ns	480	-	ns
	MTP			Center			STP
Species	Inshore	Offshore	p	Mild	Steep	p	Inshore	Offshore	p
A. brama	80 ± 56.57	-	ns	-	116 ± 19.8	ns	106.6 ± 26.6	109.5 ± 36.06	ns
A. alburnus	115.88 ± 16.11	116.5 ± 16.13	ns	113.98 ± 17.06	116.08 ± 15.97	ns	116.37 ± 17.86	116.65 ± 17.17	ns
G. cernua	-	-	ns	-	-	ns	-	-	ns
L. aspius	260	180 ± 7.07	ns	-	340	ns	217.5 ± 88.39	185 ± 56.79	ns
P. fluviatilis	225	175	ns	-	190	ns	255	201.67 ± 42.52	ns
R. rutilus	107.15 ± 61.32	126.05 ± 73.02	ns	90.27 ± 56.46	74.25 ± 4.13	ns	91.39 ± 45.12	91.05 ± 43.5	ns
S. lucioperca	184.67 ± 125.37	450	ns	360	305	ns	305 ± 49.5	413	ns
S. erythrophthalmus	265	-	ns	-	-	ns	260	-	ns
S. glanis	-	-	ns	-	-	ns	-	-	ns
	SP1			SBD			SBS
Species	Inshore	Offshore	p	Inshore	Offshore	p	Inshore	Offshore	p
A. brama	340	-	ns	80 ± 7.07	92 ± 12.53	ns	97.33 ± 11.02	-	ns
A. alburnus	114.12 ± 17.26	110.25 ± 17.23	ns	104 ± 8.61	113.42 ± 11.96	ns	107.51 ± 16.37	110.73 ± 14.65	ns
G. cernua	-	-	ns	63.75 ± 12.07	65.35 ± 16.62	ns	61.76 ± 7.08	64.88 ± 7.6	ns
L. aspius	245.86 ± 94.84	140	ns	-	-	ns	225.5 ± 129.71	153.8 ± 7.69	ns
P. fluviatilis	256.67 ± 23.09	123 ± 84.01	ns	105.53 ± 61.03	102.24 ± 63.86	ns	103.26 ± 67.86	96.32 ± 54.27	ns
R. rutilus	91 ± 27.48	83.09 ± 31.45	ns	96.77 ± 34.98	96.49 ± 43.62	ns	84.57 ± 22.1	83.36 ± 24.64	ns
S. lucioperca	230	-	ns	-	-	ns	460	167.5 ± 3.54	ns
S. erythrophthalmus	-	188.75 ± 66.88	ns	-	-	ns	220	-	ns
S. glanis	-	-	ns	290	210 ± 56.57	ns	115	-	ns

).

4. Discussion

The results of our study suggest that the direction of fish capture is a random process with no observable influence of direction on the gillnet catch rates. This lack of directionality was consistent across species, size categories, and gillnet locations. In addition, our findings highlighted a significant species-specificity in the gillnet catches, which varied according to the horizontal and vertical positioning of the gillnet. The study supports the assumption that the catchability of gillnets is largely random, further supporting the reliability of gillnet-based data.

Except for one gillnet deployment setup for asp and ruffe, our results showed no significant directional preference for fish entering the gillnets. This indicates that the probability of fish entering the gillnets from the inshore direction was similar to that from the offshore/deep direction. Various species and size categories exhibited intensive movements between the offshore and inshore areas as well as deeper benthic habitats, particularly during twilight periods when habitat switching occurs [12,15,26]. This habitat shift is driven by the need to reduce predation risk and increase foraging opportunities [12]. In accordance with CEN standards, our study deployed gillnets during both dusk and dawn, capturing fish movement in both directions (inshore-offshore/deep benthic or offshore/deep benthic-inshore) during these key transition periods. This comprehensive deployment strategy ensured that the gillnets accounted for fish movements associated with twilight habitat switching. Furthermore, during foraging, fish are expected to engage in area-restricted search (ARS) behavior, characterized by less direct movement, higher turning rates, and reduced speeds [19]. This ARS behavior involves low directionality and a high degree of stochasticity in movement. As a result, the fish’s less-predictable movements likely contributed to the similar probability of gillnet captures from either direction, further supporting the random nature of gillnet catchability in such conditions.

Shallow-installed benthic gillnets usually contain the highest fish catch [16,27]. Our assumed interferences (gillnet saturation, repellent effect of already captured fish, local decrease in fish abundance) would be expected to have the highest influence in littoral gillnet catches. With the saturation of the gillnet [20], which comes with a higher abundance, the capture limit of the gillnet would be reached sooner in benthic gillnets than in pelagic gillnets. Only the shallowest gillnet on the steep slope showed a prevalence of inshore direction offshore, which could mean that the gillnet may represent a kind of barrier, and less fish are likely to enter the gillnet from the inshore. However, this difference was not statistically significant, and the other gillnet deployments in Figure 4 exhibited the opposite pattern. These results suggest that this phenomenon is not important. At the densities of up to over one fish per square meter of gillnet (1000 inds. per 1000 m² of gillnets, Figure 4), we found no effect of gillnet saturation, repellent effect of already captured fish, or local decrease in fish abundance by previous catch.

The shallowest benthic habitats usually harbor the most diverse community [9,28,29]. As standard sampling of European water uses a depth resolution of three meters [4,7] the shallowest layers can be easily overlooked when researchers sample the horizon 0–3 m, which in practice means setting the net at the bottom depth 2–3 m. The very littoral can harbor shallow water specialists [28], which could be neglected. Our investigation of fish micro distribution showed no indication of such shallow water specialists. The catches at the shallowly installed gillnets differed mainly by the abundance of ubiquitous bleak.

Our study was conducted over a short period of four days at the end of July, following CEN recommendations for the selection of sampling time and to minimize the effects of environmental variability. The ideal period for sampling fish communities to assess their ecological state is when the fish are actively utilizing the full productivity of the water body. This optimal sampling window is outside the spawning or overwintering season, when fish undertake extensive migrations and aggregations [30,31,32,33]. In the conditions of Central Europe, this safe sampling period coincides with the end of July, August, and September, when the distribution of fish seems to be proportional to the production potential of the waterbody [7]. Outside this window, the distribution of fish is usually more variable and less predictable, which can affect the representativeness of the sampling. To minimize possible interference from weather changes or other environmental factors, experiments to evaluate the performance of sampling gear performance should be conducted within a short, well-defined time frame. For this reason, our study was intentionally short and intensive, spanning four warm and sunny days at the transition from July to August. Under these conditions, we can reasonably assume that our findings are representative of the optimal sampling conditions during the peak period from July to September.

5. Conclusions

Our study introduced a simple but feasible approach for testing fish diel movement directivity and the validation of the premise that gillnets catch fish independently of the direction they approach the gillnet. The results support the assumption that gillnet capture is a random process that is, to a great extent, connected to stochastic local movements. This result is encouraging for fish monitoring projects as it suggests that sampling results are likely to reflect actual changes in fish communities and are not biased by the orientation of the sampling gear. We also found that the differences in catches between the 0–1.5 m and 1.5–3 m depth layers were primarily due to variations in the abundance of bleak, while other species in these depth ranges had similar catch rates. This suggests that there are no species that are specifically restricted to extremely shallow habitats that would be overlooked or underestimated by standard gillnet sampling at a 0–3 m depth. This is another positive result for monitoring programs as it confirms the general representativeness of the commonly used gillnet methods. Future studies could further verify our results in deeper water layers. The net alignment statistics presented in this study can be applied in a variety of future studies, especially when investigating specific migration patterns and gillnet performance. Furthermore, the investigation of potential species interactions within the gillnet catches represents an interesting challenge for future research. It is possible that different species influence each other’s catchability, and this interaction could be an important factor to consider when interpreting gillnet data.