Submitted 16 March 2018Accepted 11 September 2018Published 12 October 2018

Corresponding authorEnric Aparicioenricapariciogmailcom

Academic editorBenjamin Letcher

Additional Information andDeclarations can be found onpage 16

DOI 107717peerj5730

Copyright2018 Aparicio et al

Distributed underCreative Commons CC-BY 40

OPEN ACCESS

Movements and dispersal of browntrout (Salmo trutta Linnaeus 1758) inMediterranean streams influence ofhabitat and biotic factorsEnric Aparicio1 Rafel Rocaspana2 Adolfo de Sostoa3 Antoni Palau-Ibars4 andCarles Alcaraz5

1GRECO Institute of Aquatic Ecology University of Girona Girona Catalonia Spain2Gesna Estudis Ambientals Linyola Catalonia Spain3Department of Evolutionary Biology Ecology and Environmental Sciences University of BarcelonaBarcelona Catalonia Spain

4Department of Environment and Soil Sciences University of Lleida Lleida Catalonia Spain5 IRTA Marine and Continental Waters Sant Carles de la Ragravepita Catalonia SpainThese authors contributed equally to this work

ABSTRACTDispersal is a critical determinant of animal distribution and population dynamicsand is essential information for management planning We studied the movementpatterns and the influence of habitat and biotic factors on Mediterranean brown trout(Salmo trutta) by mark-recapture methods in three headwater streams of the EbroBasin (NE Iberian Peninsula) Fish were sampled by electrofishing on five occasionsover 18ndash24 months and movements of over 3000 individually tagged trout (age 1+onwards) were recorded Most of the tagged fish exhibited limited movement and wererecaptured within 100 m from the initial capture section Small seasonal differencesin the movement pattern were observed but in two of the streams displacementdistances increased prior the spawning period in autumn The frequency distributionsof dispersal distances were highly leptokurtic and skewed to the right and fitted well toa two-group exponential model thus trout populations were composed of mobile andstationary individuals the latter being the predominant component in the populations(711ndash875 of individuals) The mean dispersal distances for fish captured at least inthree sampling events ranged 207ndash454 m for the stationary group and 2294ndash5405 mfor the mobile group Moving brown trout were larger than non-moving individualsand exhibited higher growth rates in two of the streams Habitat features were notconsistently linked to movement rates but there were some interaction effects betweenstream and habitat characteristics such as depth cover and water velocity

Subjects Aquaculture Fisheries and Fish Science Ecology Freshwater BiologyKeywords River connectivity Headwaters Mark-recapture Salmonidae Management

INTRODUCTIONDispersal defined as the movement of individuals between locations is a key process thatallows fish to occupy themost suitable habitat for survival growth and breeding by adaptingto both temporal and spatial changes in environmental or biotic conditions (Railsback et

How to cite this article Aparicio et al (2018) Movements and dispersal of brown trout (Salmo trutta Linnaeus 1758) in Mediterraneanstreams influence of habitat and biotic factors PeerJ 6e5730 DOI 107717peerj5730

al 1999 Lucas amp Baras 2001) Fish movement can be highly variable between individualsspecies and streams (Rodriacuteguez 2002) Most research on fish dispersal in streams showsheterogeneous populations comprising both sedentary and mobile individuals (Rodriacuteguez2002 Rasmussen amp Belk 2017) The movement distribution (ie distance moved vsprobability of occurrence) is then best predicted by leptokurtic dispersal models which arecharacterized by high peaks associated with the sedentary fish and longer tails associatedwith a lower proportion of mobile fish (Skalski amp Gilliam 2000 Rodriacuteguez 2002) Thismodelling framework has been applied to describe the general movement patterns offish and a predictive tool was recently developed based on the positive correlation ofmovement distances with four variables fish length aspect ratio of the caudal fin streamsize and duration of the study (Radinger amp Wolter 2014) At a local scale both abiotic andbiotic factors affect the degree of movement variability and dispersal rates For instancesome studies have reported an increase in exploratory behaviour (ie higher proportionsof mobile fish) with decreasing habitat heterogeneity (Albanese Angermeier amp Dorai-Raj2004Heggenes et al 2007) the presence of fish cover (eg woody debris and boulders) hasbeen related to a reduction in fish movement (Aparicio amp Sostoa 1999 Harvey Nakamotoamp White 1999) and population density has been positively linked to fish movement rates(Hesthagen 1988)

The mobility and dispersal patterns of brown trout Salmo trutta Linnaeus 1758 havebeen reported to be largely variable Some populations are mainly sedentary (Northcote1992Burrell et al 2000Knouft amp Spotila 2002) while others are dominated by individualsthat move extensively (Clapp Clark Jr amp Diana 1990 Meyers Thuemler amp Kornely 1992Ovidio et al 1998) The movement patterns of brown trout have been widely studiedwithin its native range in central and northern Europe as well as in its introduced rangein North America (see Rodriacuteguez 2002 and references therein) but the information onbrown trout movements in rivers flowing to the Mediterranean Sea in its southern nativerange is scarce There is only a study in the Rhocircne River focussed on the movements of age0+ individuals (Vatland amp Caudron 2015) Brown trout populations in the MediterraneanBasin are highly genetically differentiated from central and northern Europe populationsand are considered as distinct lineages (Bernatchez 2001 Cortey Pla amp Garciacutea-Mariacuten2004) These populations have experienced a marked decline in the last decades due tostream habitat degradation (Benejam et al 2016) overfishing (Almodoacutevar amp Nicola 2004)and introgressive hybridization with hatchery stocks (Aparicio et al 2005) Climate changescenarios for theMediterranean predict additional reductions in brown trout distributionalrange due to rising water temperatures (Almodoacutevar et al 2012) Consequently there is anurgent need to develop sound conservation andmanagement strategies for preserving thesepopulations With increasing river fragmentation quantifying the scale and magnitude ofMediterranean brown trout dispersal patterns and determining the factors that drive theseare necessary to optimize management planning

The overall purpose of this study was to examine the movement patterns of browntrout populations from three separate streams of a Mediterranean catchment of the IberianPeninsula and the influence of biotic and abiotic factors The three study streams are smallhave a high gradient and considerable habitat heterogeneity When all habitats required to

Aparicio et al (2018) PeerJ DOI 107717peerj5730 222

complete the cycle of a fishrsquos life-history are in spatial proximity lower movement distancesare expected (Albanese Angermeier amp Dorai-Raj 2004) Also structural habitat complexityincreases protection from predators refuge from disturbances (eg flooding) and preyavailability reducing territory size and fish mobility (Heggenes et al 2007 Zaacutevorka et al2015) The objectives of this study were to employmark-recapture approaches over a periodof 18ndash24 months to determine (1) the degree of mobility (ie proportion of stationaryand mobile individuals) and extent of displacement distances (2) seasonal variations inthe pattern of movements and (3) the effect of habitat characteristics and biotic factors ontrout movement

MATERIAL AND METHODSStudy areaThe marking and recapturing of individuals were conducted in three streams (NogueraPallaresa Flamisell and Noguera Vallferrera) of the Segre River basin (Catalonia NEIberian Peninsula) (Fig 1) The Segre is the largest catchment in the Southern Pyrenees(265 km long 22580 km2 of basin area and about 100 m3 sminus1 of average water flow)and is the main tributary of the Ebro River the river with the highest water flow in theIberian Peninsula (annual mean is about 426 m3 sminus1) discharging into the MediterraneanSea (for more details see Rovira Alcaraz amp Ibaacutentildeez 2012) A tributary of the Segre River isthe Noguera Pallaresa River (154 km of length and 371 m3 sminus1 of average flow) wherethree different reaches were selected (Fig 1) the main stem (4244primeN 058primeE hereafterNP) and two tributaries the Flamisell (4227primeN 059primeE FLM) and the Noguera Vallferrera(4234primeN 119primeE NV) The three reaches had a mean stream width less than 10 m showeda high gradient and channel morphology consisted primarily of pool-run-riffle sequencesunder a forest canopy Physical characteristics of study reaches are shown in Table 1 Thehydrological regime is snow-fed thus the highest flows generally occur in spring aftersnowmelt (Ebro Water Authority httpwwwchebroes) Brown trout was the only fishspecies present and populations belong to the Mediterranean lineage (Aparicio et al 2005)During the study sport fishing was closed in the study area

Sampling designStudy reaches were continuous and ranged from 2400 m long in the NP 1500 in theFLM to 1200 m long in the NV The reaches were selected on the basis of being wadeableand the absence of barriers to fish movement or significant tributaries within the studyboundaries Each reach was divided into 25ndash40 m-long sections (mean length 306 mplusmn 43SD) the boundaries of which usually coincided with habitat discontinuities in the channelmorphology (pool-run-riffle sequence) Brown trout were marked in October 1992 in NPand in March 1993 in NV and FLM Four subsequent mark-recapture sampling eventswere conducted in each reach NP was sampled in 1993 (May and October) and 1994 (Julyand October) and both NV and FLM reaches were sampled in 1993 (July and October)and 1994 (July and October) Autumn surveys (October) were performed just prior orat the beginning of the spawning season which in the study area occurs from the end of

Aparicio et al (2018) PeerJ DOI 107717peerj5730 322

Figure 1 Map of the study area (A) Location of the study streams Noguera Pallaresa (NP) Flamisell(FLM) and Noguera Vallferrera (NV) Location of the sampled stream reaches is highlighted with arectangle The photos show representative habitats in the streams sampled (B) NP (C) NV (D) FLMPhotographs by Enric Aparicio

Full-size DOI 107717peerj5730fig-1

October to early December Therefore movement data from October sampling includedthe possible fish displacements related to spawning Sampling sections were isolated withblock nets (mesh size 5 mm) and three-pass electrofishing removal was conducted in anupstream direction (500 V 10 A pulsed DC) Captured fish were anaesthetised (MS-222solution) measured for fork length (FL mm) and weighed (g) Sex was determinedexternally for mature individuals in autumn (ie October) surveys by the production of

Aparicio et al (2018) PeerJ DOI 107717peerj5730 422

Table 1 Stream features during the sampling period FLM Flamisell NP Noguera Pallaresa and NVNoguera Vallferrera Meanplusmn standard deviation when necessary is shown

Variable Stream

FLM NP NV

Length (m) 1500 2400 1200Mean elevation (m) 1250 1780 1175Mean slope (m kmminus1) 533 208 625Mean base flow (m3sminus1) 152 095 129Temperature range (C) 42ndash153 01ndash151 25ndash154pH range 73ndash89 74ndash88 67ndash81Conductivity range (microS cmminus1) 77ndash171 46ndash187 20ndash95Mean stream width (m) 497plusmn 139 434plusmn 236 680plusmn 109Mean water depth (cm) 994plusmn 466 2002plusmn 646 1714plusmn 277Mean current velocity (m sminus1) 066plusmn 026 059plusmn 027 070plusmn 014Mean bottom substrate ()

Boulders (gt256 mm) 3147plusmn 1382 3084plusmn 1874 4887plusmn 1267Cobble (65ndash256 mm) 3414plusmn 1189 3468plusmn 1517 3245plusmn 1095Gravel (2ndash65 mm) 1501plusmn 617 1489plusmn 1037 769plusmn 304Sand (01ndash2 mm) 1060plusmn 423 1077plusmn 1016 374plusmn 304Silt (lt01 mm) 345plusmn 255 441plusmn 454 204plusmn 113Organic 433plusmn 304 413plusmn 317 517plusmn 301

Mean fish density (Ind haminus1) 13459plusmn 13222 19014plusmn 9893 38625plusmn 15590Mean fish biomass (Kg haminus1) 675plusmn 598 1078plusmn 828 3447plusmn 1827Mean fish cover () 1546plusmn 881 1018plusmn 757 2247plusmn 652

milt or visible evidence of eggs beneath the body wall The adipose fin was clipped onall captured fish which constituted a permanent batch mark to distinguish recapturedindividuals

All fish larger than 120 mm FL (corresponding at least to age 1+ individuals seeFigs S1ndashS3) were tagged with uniquely coded Visible Implant Tags (VI-tags NorthwestMarine Technology Shaw Island Washington USA) in the adipose tissue posterior to eyeFish smaller than 120 mm were not tagged because the tissue available for placing the tagwas too thin thus leading to extremely low retention rates (Niva 1995) After handling fishwere allowed to recover from anaesthesia and released into the mid-point of the capturesection In recapture surveys each fishrsquos VI-tag code was read and those untagged fishlarger than 120 mm FL (ie captured for the first time or having lost the VI-tag) weretagged following the above procedure

Habitat variables were measured along transects perpendicular to the flow at 10-mintervals Water depth and mean water velocity (at 06 times total water depth) weremeasured every 1m along transects and averaged for each section For each transect thepercent cover of each substrate category (Table 1) was visually estimated averaged persection and an index of substrate coarseness was derived for each section following BainFinn amp Booke (1985) Fish cover was estimated as the percentage area of a section coveredby woody debris rocks or undercut banks

Aparicio et al (2018) PeerJ DOI 107717peerj5730 522

Permission for electrofishing and capture of S trutta individuals was approved by thecompetent authorities Departament de Medi Ambient i Habitatge de la Generalitat deCatalunya (current Departament drsquoAgricultura Ramaderia Pesca Alimentacioacute i MediNatural) (SF602) of the regional authorities of Catalonia

Data analysesFish movement (ie distance covered) was measured from the midpoint of the recapturesection to the midpoint of its previous capture section A movement distance of 0 m wasassigned to individuals captured and recaptured in the same section Thus the minimumdetectable movement was the distance between two or more sections (ie larger than25ndash40 m) Movement patterns were quantified using frequency distributions (two-tailed)of distances moved Positive values were assigned to upstream movements and negativevalues to downstream movements Mark-recapture studies are generally biased since longmovements are less frequently measured than short distances (Albanese Angermeier ampGowan 2003) To assess this source of bias all movement observations were weightedby determining the total possible movements sampled for each distance and weightingthe under-sampled distances (Porter amp Dooley 1993 Albanese Angermeier amp Gowan2003) The movement distributions were not significantly different after adjustment(KolmogorovndashSmirnov test P gt 040 for all comparisons) suggesting a good study design(Porter amp Dooley 1993) Therefore unweighted distributions were used for all analyses

Trout were classified as moving (individuals leaving the home section between samplingevents) and non-moving (sedentary individuals not leaving the home section) individualsDifferences in distance covered by moving fish among streams and sampling events wereanalysed with analysis of variance (two-way ANOVA) The kurtosis and skewness of thefrequency distribution of distances moved by brown trout were obtained with the packagemoments (version 014) for the program R (Komsta amp Novometsky 2015) and were usedas an indicator of individual level variation in movement behaviour (Skalski amp Gilliam2000) Dispersal range was estimated as the distance between the two most distant sectionsin which an individual was recorded and only fish captured at least in three samplingevents (elapsed time between first and last capture was 7 to 24 months) were includedin the analyses The frequency distribution of dispersal range was fitted to a two-groupexponential function (Rodriacuteguez 2002)

f (x)= pλs

2eminusλsx+

(1minusp

) λm2eminusλmx

where x is the distance covered λs correspond to the inverse of mean dispersal distancefor the stationary component λm correspond to the inverse of mean dispersal distance forthe mobile component p is the proportion of stationary individuals and thus 1minusp is theproportion of mobile individuals Parameters p λs and λm are estimated from movementdata (see Rodriacuteguez 2002) Estimated movement parameters were then compared to theexpectedmovement parameters for stream-resident brown trout obtained with the packagefishmove (version 03ndash3) for the program R which models dispersal in relation to speciesfish length aspect ratio of the caudal fin stream size and duration of the study (Radingeramp Wolter 2014)

Aparicio et al (2018) PeerJ DOI 107717peerj5730 622

In order to examine the relationship between fish growth rate and movement thespecific growth rate (SGR in dayminus1) for a given fish in the summer period was calculatedas SGR = (ln(final fork length)mdashln(initial fork length)) times100(days between captures)Differences in the growth ratendashfork length relationship between moving and non-movingindividuals were tested with analysis of covariance (ANCOVA) for each stream SoftwarePopPro 10 (Kwak 1992) was used to estimate population size per section and captureprobability per size class (ie trout smaller and larger than 120 mm) from the three-passelectrofishing data Fish density (individuals haminus1) and biomass (kg haminus1) per section wereestimated by dividing number and weight by the area sampled Variations in fish densityand biomass among streams were analysed with multiple analysis of variance (MANOVA)MANOVA is used when several dependent variables are measured on each sampling unitMANOVA compares the mean vectors of k groups whereas equality of the mean vectorimplies that the k means are equal for each variable if two means differ for just one variablethen we conclude that the mean vectors of the k groups are different (Sokal amp Rohlf 1995)We used η2 (eta squared) as a measure of effect size (ie importance of factors) Similarlyto r2 η2 is the proportion of variation explained for a certain effect

The effects of biotic and abiotic stream features of sampling sections on trout departureratio (defined as the proportion of individuals leaving section i from the total number ofrecaptures of trout initially marked in section i) were analysed with Generalized LinearModels (GLMs) We performed a global analysis including the three different streams andthen streams were analysed separately In each case an information-theoretic approachwas used to find the best approximating models (Burnham amp Anderson 2002) GLMs werebuilt including all possible combinations of environmental and biotic variables excludinginteractions due to the large number of variables included Two additional criteria wereused to define the set of candidate models those performing significantly better thanthe null model and with a variance inflation factor le5 in order to avoid multicollinearityeffects The second order Akaike Information Criterion (AICc) was used to assess the degreeof support for each candidate model AICc was rescaled to obtain1AICc values (1AICc=AICcimdashminimum AICc) since models with1AICcle 2 have the most substantial supportThe relative plausibility of each candidate model was assessed by calculating Akaikersquosweights (wi) which range from 0 to 1 and can be interpreted as the probability that agiven model is the best model in the candidate set Because no model was clearly the bestone (ie wi ge 09) model-average regression coefficients were calculated using the relativeimportance of each independent variable (Burnham amp Anderson 2002) Model-averagedcoefficients were compared with those from the full model to assess the impact of modelselection bias on parameter estimates All data analyses were performed with R softwareversion 332

RESULTSRecapture rates and VI-tag retentionA total of 13340 individuals were captured and fin-clipped during the study period and7050 of them were larger than 120 mm FL and tagged (NP 2201 FLM 1181 NV 3668)

Aparicio et al (2018) PeerJ DOI 107717peerj5730 722

Figure 2 Evolution of the percentage of brown troutgt120 mm fork length andle120 mm fork lengthcaptured for the first time (ie adipose fin not clipped) on each sampling event in the study streamsFLM Flamisell NP Noguera Pallaresa NV Noguera Vallferrera

Full-size DOI 107717peerj5730fig-2

Mean FL of tagged fish was 1651mmplusmn 347 SD in the NP 1628mmplusmn 381 SD in the FLMand 1830 mm plusmn 421 SD in the NV Capture probabilities estimated from electrofishingdepletion surveys ranged from 040 to 088 (micro= 065) for trout larger than 120 mm andfrom 014 to 049 (micro= 038) for individuals smaller than 120 mm The proportion of thenew trout individuals (ie fish not fin-clipped) larger than 120 mm FL captured for thefirst time was similar among streams and decreased sharply until the end of the study witha maximum reduction of about 60 between the first and the second sampling (Fig 2)The decline in the proportion of new individuals smaller than 120 mm FL captured in thesecond survey was less pronounced (Fig 2) probably because of recruitment and reducedcapturability Despite the high recapture rates the recovery percentage of VI-tag (ie fish

Aparicio et al (2018) PeerJ DOI 107717peerj5730 822

Table 2 Parameter estimates from the two-group exponential model to calculate the proportion ofstationary (p) andmobile (1minus p) individuals andmean dispersal range (1λ) of brown trout popula-tion FLM Flamisell NP Noguera Pallaresa and NV Noguera Vallferrera λ (mminus1) is the inverse of themean displacement range lower values of λ correspond to more mobile individuals

Stream Populationcomponent

Proportion()

Parameterestimate

Mean dispersalrange (m)

Stationary 875 λs= 002201 454FLM

Mobile 125 λm= 000185 5405Stationary 856 λs= 004184 239

NPMobile 144 λm= 000269 3717Stationary 711 λs= 004822 207

NVMobile 289 λm= 000436 2294

larger than 120 mm) was relatively low due to tag loss Overall mean VI-tag retention ratewas 450 but varied in relation to fish size Trout smaller than 180 mm FL at taggingshowed mean retention rates of 324 whereas the rate was 749 for trout larger than240 mm FL

Movements and dispersal patternsOverall movements of brown trout between consecutive sampling events were relativelyshort since the 534 (ranging between 481 and 544) of the total recaptures (N =3073) were in the same stream section of previous capture When fish moving out fromhome section were analysed separately most of the displacements were between contiguoussections and long-range movements were infrequent for instance only the 218 of themovements were over 400 m in NV being 237 in NP and 335 in FLM Distancecovered by fish moving out from home section was not significantly related to fork length(ANCOVA P gt 029 in all rivers) or lengthtimes season interaction (P gt 015) so an ANOVAwas used Distance covered by fish showed significant temporal variations in both theNP (ANOVA F3426 = 449 P = 0004 η2 = 024) and NV (F3795 = 945 P lt 00001η2= 026) streams with trout covering larger distances in autumn (Fig 3) but seasonaldifferences were not significant in the FLM stream (F3198= 108 P = 036)

In the three streams the distance-frequency distributions were highly leptokurtic andskewed to the right the Kurtosis values were 289 for FLM 196 for NV and 1446 forNP The two-group exponential model provided a good fit to the frequency distributionsof dispersal range (Fig 4) Based on the estimates of p the proportion of stationaryindividuals in the three streams ranged from 711 to 875 and thus the percentageof mobile individuals varied from 125 to 289 (Table 2) The mean dispersal rangecalculated as 1λ (see lsquoMethodsrsquo) ranged from 207 to 454 m for the stationary group andfrom 2294 to 5405 m for the mobile group (Table 2) These dispersal ranges were similarto that expected by the general model for stream resident brown trout estimated withfishmove package which calculates a mean movement distance of 251 m for the stationarygroup and 4283 m for the mobile group

Aparicio et al (2018) PeerJ DOI 107717peerj5730 922

Figure 3 Box-plot of absolute distance moved by brown trout individuals recaptured out fromhome section per sampling occasion in the study streams The box corresponds to the 25th and 75thpercentiles the dark line inside the box represents the median the error bars are the confidence intervalsat the 95 confidence level and the filled circle is the mean (A) Flamisell (B) Noguera Pallaresa (C)Noguera Vallferrera

Full-size DOI 107717peerj5730fig-3

Aparicio et al (2018) PeerJ DOI 107717peerj5730 1022

Figure 4 Brown trout frequency distributions of the dispersal range fitted with a two-groupexponential model in the study streams The solid lines are the linear regression fits for the estimatedstationary (open circles) and mobile (filled circles) components of the fish populations Distance classesare calculated from absolute values of distance moved y-axis is in logarithmic scale (A) Flamisell (B)Noguera Pallaresa (C) Noguera Vallferrera

Full-size DOI 107717peerj5730fig-4

Aparicio et al (2018) PeerJ DOI 107717peerj5730 1122

Influence of biotic and abiotic factors on movementsSeveral biotic and abiotic factors were associated with brown trout movement patternsOverall the information theoretic analysis of the candidate set of GLMs selected eightplausible models (ie1AICc lt 2) to explain variability in departure ratio (ie proportionof individuals leaving section i from the total number of recaptures of trout initially markedin section i) in relation to stream features The best explanatory variables (SP values inTable 3) were water depth fish cover and fish biomass Interestingly fish biomass waspreferentially selected over density despite the high correlation between them (Pearsonrsquosr = 094 N = 246 P lt 00001) All three variables (water depth cover and biomass) werenegatively related to departure ratio in contrast organic substrate water velocity substratecoarseness and fish density had a positive relationship with departure ratio Season had theweakest relationship with trout departure ratio However the lower correlation betweenobserved and predicted values and larger parameters bias indicated some differences amongstreams (Table 3 Table S1) When analysed separately by streams similar patterns wereobserved with some variations linked to specific stream habitat characteristics (Table 1Table 3) For instance in the FLM stream where higher departure ratios (ANOVAF2243 = 1167 P lt 00001) and lower biomass and density values (MANOVA Wilksrsquosλ= 0513 F4484 = 4792 P lt 00001) were reported physical features (ie fish coversubstrate coarseness and water depth) and fish biomass were the best explanatory variablesSeasonal variation in departure ratio confirmed previous observed results (Fig 3) thuswhile no seasonal differences were observed in the FLM stream autumn departure ratio(ie pre-spawn movements) was higher in the NV stream and lower in the NP streamIn the NV stream along with fish biomass and season the most important variable waswater velocity all negatively related to departure ratio In the NP stream fish density andbiomass showed a contrasting pattern (Table 3) which might be explained by differencesin fish length since NP fish were the smallest (ANOVA F23070= 18460 P lt 00001)

Moving brown trout were significantly larger than non-moving individuals in thethree study streams (P lt 0029 and η2 gt 016 in all cases) but distances moved were notsignificantly correlated with trout fork length (P gt 015 in all cases) In both FLM andNV streams specific growth rate of moving trout (microplusmn standard error 0132 plusmn 0006dayminus1 in the FLM and 0057 plusmn 0002 dayminus1 in the NV) was significantly higher thannon-moving trout (0103plusmn 0005 and 0050plusmn 0001 dayminus1) (ANCOVA F1242= 573 Plt 005 η2= 019 in FLM and F11262= 607 P lt 005 η2= 028 in NV) but no significantdifferences were observed in the NP stream (0048 plusmn 0002 dayminus1 for moving fish and0052 plusmn 0002 dayminus1 for non-moving fish) (F1327= 106 P = 031)

DISCUSSIONMovements and dispersal patternsBrown trout exhibited limited mobility and few fish performed long-range movementsBetween successive sampling events a significant proportion of fishmoved from their homesection but most of them moved less than 100 m and settled in contiguous sections (ieadjacent channel units) Individuals move in response to changing resource abundance and

Aparicio et al (2018) PeerJ DOI 107717peerj5730 1222

Table 3 Model averaged parameter estimates by means of an information theoretic approach fromGLMs analysis of the influence of stream features of samplingsections on brown trout departure ratio (see text for definition) FLM Flamisell NP Noguera Pallaresa and NV Noguera Vallferrera Model-averaged regression coef-ficients (β) are parameter coefficients averaged by model weight (wi) across all candidate models (1AICc lt 2) in which the given parameter occurs selection probabil-ity (SP) indicates the importance of an independent variable and parameter bias is the difference between the averaged estimates (β) and the full model coefficients Thenumber (N ) of candidate models (1AICc lt 2) and Pearsonrsquos correlation coefficient (r) between observed and model predicted values are also shown

Model parameter All rivers modelN = 8 r = 047

FLMModelN = 3 r = 079

NPModelN = 12 r = 062

NVModelN = 5 r = 072

β SP Bias β SP Bias β SP Bias β SP Bias

Intercept 74999 0237 minus73138 minus0227 4288 4772 91815 0030Mean Depth (cm) minus1442 1000 minus0021 minus2745 1000 minus0239 minus0123 0208 minus2289 2352 0224 0019Fish Cover () minus1109 1000 minus3069 minus1772 1000 0242 minus1339 1000 minus0396 minus0264 0366 minus0239Fish Biomass (kg haminus1) minus0017 0623 minus0782 minus0097 1000 minus0515 minus0025 0148 minus3881 minus0038 1000 minus1123Organic Substrate () 0118 0182 minus5555 1246 0126 minus2124 0390 0292 minus2289 0761 0464 0104Water Velocity (m sminus1) 0991 0153 minus10204 minus5704 0323 0867 5981 0318 minus0923 minus55562 1000 minus0214Fish Density (Ind haminus1) 0012 0147 minus29545 Not selected 0014 0891 minus1178 Not selectedSubstrate Coarseness 0305 0104 6702 46146 1000 0021 2951 0339 minus1425 4977 0212 0004Pre Spawn Season minus0125 0077 2550 Not selected minus9752 1000 0112 14314 1000 minus0131

Aparicio

etal(2018)PeerJDOI107717peerj5730

1322

quality thus movement extent should depend on the distance to a suitable habitat (Fauschet al 2002) with longer movement distances presumably occurring when suitable habitatsare widely spaced (Wiens 2001) Therefore the lowmobility exhibited by brown trout fromthis study could indicate that the different habitats required to complete its life cycle are inspatial proximity (Northcote 1992 Gowan et al 1994 Albanese Angermeier amp Dorai-Raj2004) Spatial habitat heterogeneity is usually higher in small streams than in largerrivers where habitat units are more widely spaced (Gorman amp Karr 1978) and habitatheterogeneity reduces territory size and mobility (Heggenes et al 2007) Consequentlylimited movements of nomore than a few hundredmeters are frequently reported for troutpopulations in small streams (Heggenes 1988 Knouft amp Spotila 2002) when compared tothose from large and long rivers where longer movements up to several kilometres havebeen observed (Clapp Clark Jr amp Diana 1990 Young 1994 Zimmer Schreer amp Power2010) Therefore the size of the study streams could have influenced the limited range ofmovements observed (Woolnough Downing amp Newton 2009)

Although there was substantial intra-population variation in movement behaviourthe studied populations were dominated by stationary individuals in concordance withmost of the studies on movement in stream salmonids (Rodriacuteguez 2002 and referencestherein) Our results also match the estimated movement parameters (ie proportion ofstationary and mobile individuals and mean dispersal distances) provided by predictivemodels for stream-resident brown trout (Radinger amp Wolter 2014) for both the stationaryand the mobile component The displacement range of the stationary component was lessthan 50 m and thus is considered as a restricted movement (Rodriacuteguez 2002) Despitebeing in low proportion mobile individuals are important since they are responsible forexchange between populations and successful colonization of new habitats (Vera et al2010 Kennedy Rosell amp Hayes 2012)

Many studies dealing with brown trout movements showed a seasonal increasein mobility due to upstream spawning migration and downstream movement foroverwintering (Clapp Clark Jr amp Diana 1990 Meyers Thuemler amp Kornely 1992 Ovidioet al 1998) Our results showed temporal differences in the distance covered by movingfish thus suggesting that seasonal changes in environmental conditions or the reproductivecycle of fish affected trout movements There was an increase in distances moved in autumnsurveys in both the NP and the NV stream just before the spawning period although thiscannot be considered a true spawning migration because movement did not involve mostof the population nor were distances moved long-range In streams where spawning sitesare located in or near the rearing habitats as occurred in the studied reaches where gravelbeds were widely distributed spawning-related movements may be minimal or involveshort distances (Northcote 1992 Nakamura Maruyama ampWatanabe 2002)

Movement data were based on recaptured (ie surviving) tagged fish thus severalpotential sources of bias could have affected our results For instance VI-tag insertion andadipose fin clipping may result in different behaviour or mortality rate compared to thenon-tagged fish However studies on salmonid species suggest that adipose fin clippingand VI tag insertion do not affect condition growth or survival (eg Bryan amp Ney 1994Shepard et al 1996) thus the behaviour of tagged fish is representative of the population

Aparicio et al (2018) PeerJ DOI 107717peerj5730 1422

Another potential source of bias could be related to the relatively low percentage of VI-tagsrecovered not only due to tag loss but also to natural mortality which is expected in anearly two years study Although it was not measured annual mortality for adult browntrout is estimated to be 43ndash72 in similar Pyrenean rivers (Gouraud et al 2000) Tag lossmay bias abundance estimates from mark-recapture methods (Frenette amp Bryant 1996)but movement estimates are not biased since differences in behaviour between tagged andtag-loss fish are not expected Therefore both mortality and tag loss only affect movementestimates by reducing total data gathered but this was avoided by maximizing samplesize increasing the number of fish tagged Trout leaving the sampling area probably alsoaccounted for a percentage of missed recaptures thus leading to an underestimation oflong-range movements (Gowan et al 1994) However the high proportion of recapturesof fin-clipped fish suggests that the emigration from the study area was proportionally lowcompared to the trout population monitored (Gowan et al 1994)

Relationship of habitat and biotic factors with brown troutmovementsFish movements are mediated by biotic and abiotic factors affecting individual fitness(Gowan et al 1994 Railsback et al 1999 Beacutelanger amp Rodriacuteguez 2002) Our results showedthat departure ratiowas not consistently linked to particular habitat features suggesting thatmovement behaviour probably does not depend on a single parameter but on a complexcombination and availability of several factors For instance fish inhabiting shallowersections in the FLM stream showed higher departure ratios than those from deeper watersAmong the study streams the FLM had the lowest mean water depth therefore deeperwaters which confer greater protection against predators (Lonzarich amp Quinn 1995) werea valuable resource motivating trout movement In the same sense sections with higherfish cover in the FLM and NP stream showed lower departure ratios probably because ofthe refugia provided from predators and fast currents (Harvey Nakamoto amp White 1999Aylloacuten et al 2014) Finally the negative effect of water velocity on departure ratio in theNV stream could be mediated by an increase in prey delivery rate in high velocity areas(Leung Rosenfeld amp Bernhardt 2009) thus promoting residency

Movements of stream fishes have been associated with fish size and growth ratesThere is evidence from previous studies that movement distances increase with fishlength particularly for trout larger than 300ndash400 mm FL (Meyers Thuemler amp Kornely1992 Young 1994 Quinn amp Kwak 2011) In the three study streams moving fish werelarger than non-moving trout but there was no relationship between distance movedand fish size Results are probably skewed by the scarcity of large fish since only fewindividuals (lt05) exceeded 300 mm FL or the fact that large dominant fish coulddisplace subordinate individuals thus leading a movement cascade of fish of all sizesand obscuring length-related patterns (Gowan amp Fausch 2002 Railsback amp Harvey 2002)Mobility is energetically expensive and increases risk of predation since fish have to movethrough unfamiliar space (Peacutepino Rodriacuteguez amp Magnan 2015) Nevertheless mobilityappeared to confer an advantage on trout growth rate Several authors have reported

Aparicio et al (2018) PeerJ DOI 107717peerj5730 1522

that when resources are scattered mobile fish maximize food supply and compensateenergetic costs of movement thus increasing growth rate (eg Hilderbrand amp Kershner2004 Zaacutevorka et al 2015)

Management implicationsMost of the remnant brown trout populations of the Mediterranean lineage are confinedto streams where many artificial in-stream barriers have been built for hydropowergeneration For example more than 400 weirs and dams are distributed in the SpanishPyrenean rivers (Espejo amp Garciacutea 2010) mainly for small-scale hydropower productionBarriers can restrict movement pathways among critical habitats for spawning feeding orrefuge (Dunham Vinyard amp Rieman 2011) thus stream longitudinal connectivity is a keyconsideration in the management and conservation of the brown trout The low rate ofaverage movement at population level reported here suggests that management of browntrout populations in headwater streams similar to those studied here should focus on theconserving high quality habitats whereas stream connectivity may not be as critical sincerelatively small stream length may provide sufficient habitat to meet all life-history needsHowever connectivity could become important when other anthropogenic impacts arepresent or the population verges the threshold of the minimum viable size (Hilderbrandamp Kershner 2011) Then the importance of medium and long distance movements maybe higher for population persistence by allowing fish to withstand locally unfavourableconditions or favouring the recolonization after the disturbance (Fausch et al 2009)

Finally since brown trout is an important game species in populations with limiteddispersal fishing regulations could exert a high influence on trout abundance Overfishingand depletion of fishery stocks seem more probable in stream reaches inhabited by troutpopulations with low mobility because substantial immigration of new individuals oreventual displacement of fish to safer areas (ie closed fishing reaches) are infrequentTo improve native trout abundance in streams with significant fishing pressure no-killregulations are advised or at least hardening existing regulations to avoid overexploitation

ACKNOWLEDGEMENTSAuthors wish to thank the many people who assisted to fieldwork especially to F CasalsMA Puig JM Olmo MJ Vargas J Malo C Franco and J Laborda We are grateful tothe anonymous reviewers for their feedback which was very helpful in improving themanuscript

ADDITIONAL INFORMATION AND DECLARATIONS

FundingThis research was supported by the Biodiversity Conservation Plan of ENDESA throughthe project PIE 121043-00-FECSA The funders had no role in study design data collectionand analysis decision to publish or preparation of the manuscript

Aparicio et al (2018) PeerJ DOI 107717peerj5730 1622

Grant DisclosuresThe following grant information was disclosed by the authorsENDESA PIE 121043-00-FECSA

Competing InterestsRafel Rocaspana is an employee of Gesna Estudis Ambientals SL Carles Alcaraz is anemployee of the IRTA (Institut de Recerca i Tecnologia Agroalimentagraveries)

Author Contributionsbull Enric Aparicio conceived and designed the experiments performed the experimentsanalyzed the data contributed reagentsmaterialsanalysis tools prepared figures andortables authored or reviewed drafts of the paper approved the final draftbull Rafel Rocaspana analyzed the data contributed reagentsmaterialsanalysis toolsauthored or reviewed drafts of the paper approved the final draftbull Adolfo de Sostoa and Antoni Palau-Ibars conceived and designed the experimentsperformed the experiments authored or reviewed drafts of the paper approved the finaldraftbull Carles Alcaraz analyzed the data contributed reagentsmaterialsanalysis tools preparedfigures andor tables authored or reviewed drafts of the paper approved the final draft

Animal EthicsThe following information was supplied relating to ethical approvals (ie approving bodyand any reference numbers)

The Departament de Medi Ambient i Habitatge de la Generalitat de Catalunya (currentDepartament drsquoAgricultura Ramaderia Pesca Alimentacioacute i Medi Natural) of the regionalauthorities of Catalonia provided authorization to conduct the research (SF602)

Data AvailabilityThe following information was supplied regarding data availability

The raw data are provided as Supplemental Files

Supplemental InformationSupplemental information for this article can be found online at httpdxdoiorg107717peerj5730supplemental-information

REFERENCESAlbanese B Angermeier PL Dorai-Raj S 2004 Ecological correlates of fish movement

in a network of Virginia streams Canadian Journal of Fisheries and Aquatic Sciences61857ndash869 DOI 101139f04-096

Albanese B Angermeier PL Gowan C 2003 Designing markmdashrecapture studiesto reduce effects of distance weighting on movement distance distributionsof stream fishes Transactions of the American Fisheries Society 132925ndash939DOI 101577T03-019

Aparicio et al (2018) PeerJ DOI 107717peerj5730 1722

Almodoacutevar A Nicola GG 2004 Angling impact on conservation of Spanish stream-dwelling brown trout Salmo trutta Fisheries Management and Ecology 11173ndash182DOI 101111j1365-2400200400402x

Almodoacutevar A Nicola GG Aylloacuten D Elvira B 2012 Global warming threatens thepersistence of Mediterranean brown trout Global Change Biology 181549ndash1560DOI 101111j1365-2486201102608x

Aparicio E Garciacutea-Berthou E Araguas RMMartiacutenez P Garciacutea-Mariacuten JL 2005Body pigmentation pattern to assess introgression by hatchery stocks in nativeSalmo trutta from Mediterranean streams Journal of Fish Biology 67931ndash949DOI 101111j0022-1112200500794x

Aparicio E Sostoa A 1999 Pattern of movements of adult Barbus haasi in a smallMediterranean stream Journal of Fish Biology 551086ndash1095DOI 101111j1095-86491999tb00743x

Aylloacuten D Nicola GG Parra I Elvira B Almodoacutevar A 2014 Spatio-temporal habitatselection shifts in brown trout populations under contrasting natural flow regimesEcohydrology 7569ndash579 DOI 101002eco1379

BainMB Finn JT Booke HE 1985 Quantifying stream substrate for habitatanalysis studies North American Journal of Fisheries Management 5499ndash500DOI 1015771548-8659(1985)5lt499QSSFHAgt20CO2

Beacutelanger G Rodriacuteguez MA 2002 Local movement as a measure of habitat quality instream salmonids Environmental Biology of Fishes 64155ndash164DOI 101023A1016044725154

Benejam L Saura-Mas S BardinaM Solagrave C Munneacute A Garciacutea-Berthou E 2016 Ecolog-ical impacts of small hydropower plants on headwater stream fish from individual tocommunity effects Ecology of Freshwater Fish 25295ndash306 DOI 101111eff12210

Bernatchez L 2001 The evolutionary history of brown trout (Salmo trutta L) inferredfrom phylogeographic nested clade and mismatch analyses of mitochondrial DNAvariation Evolution 55351ndash379 DOI 101111j0014-38202001tb01300x

Bryan RD Ney JJ 1994 Visible implant tag retention by and effects on condition of astream population of brook trout North American Journal of Fisheries Management14216ndash219 DOI 1015771548-8675(1994)014lt0216VITRBAgt23CO2

BurnhamKP Anderson DR 2002Model selection and multimodel inference a practicalinformation-theoretic approach New York Springer-Verlag

Burrell KH Isely JJ Bunnell Jr DB Van Lear DH Dolloff CA 2000 Seasonal move-ment of brown trout in a southern Appalachian river Transactions of the AmericanFisheries Society 1291373ndash1379DOI 1015771548-8659(2000)129lt1373SMOBTIgt20CO2

Clapp DF Clark Jr RD Diana JS 1990 Range activity and habitat of large free-rangingbrown trout in a Michigan stream Transactions of the American Fisheries Society1191022ndash1034 DOI 1015771548-8659(1990)119lt1022RAAHOLgt23CO2

Aparicio et al (2018) PeerJ DOI 107717peerj5730 1822

Cortey M Pla C Garciacutea-Mariacuten JL 2004Historical biogeography of MediterraneantroutMolecular Phylogenetics and Evolution 33831ndash844DOI 101016JYMPEV200408012

Dunham JB Vinyard GL Rieman BE 2011Habitat fragmentation and extinctionrisk of lahontan cutthroat trout North American Journal of Fisheries Management171126ndash1133 DOI 1015771548-8675(1997)017lt1126HFAEROgt23CO2

Espejo C Garciacutea R 2010 Agua y energiacutea produccioacuten hidroeleacutectrica en Espantildea Investiga-ciones Geograacuteficas 51107ndash129

Fausch KD Rieman BE Dunham JB YoungMK Peterson DP 2009 Invasion versusisolation trade-offs in managing native salmonids with barriers to upstream move-ment Conservation Biology 23859ndash870 DOI 101111j1523-1739200801159x

Fausch KD Torgersen CE Baxter CV Li HW 2002 Landscapes to riverscapes bridgingthe gap between research and conservation of stream fishes BioScience 52483ndash498DOI 1016410006-3568(2002)052[0483LTRBTG]20CO2

Frenette BJ Bryant MD 1996 Evaluation of visible implant tags applied to wild coastalcutthroat trout and dolly varden in Margaret Lake Southeast Alaska North AmericanJournal of Fisheries Management 16926ndash930DOI 1015771548-8675(1996)016lt0926EOVITAgt23CO2

Gorman OT Karr JR 1978Habitat structure and stream fish communities Ecology59507ndash515 DOI 1023071936581

Gouraud V Baran P Lim P Sabaton C 2000 Dynamics of a population of brown trout(Salmo trutta) and fluctuations in physical habitat conditionsmdashexperiments on astream in the Pyrenees first results In Cowx IG edManagement and ecology of riverfisheries Oxford Fishing News Books Blackwell Science 126ndash142

Gowan C Fausch KD 2002Why do foraging stream salmonids move during summerEnvironmental Biology of Fishes 64139ndash153 DOI 101023A1016010723609

Gowan C YoungMK Fausch KD Riley SC 1994 Restricted movement in residentstream salmonids a paradigm lost Canadian Journal of Fisheries and Aquatic Sciences512626ndash2637 DOI 101139f94-262

Harvey BC Nakamoto RJ White JL 1999 Influence of large woody debris and abankfull flood on movement of adult resident coastal cutthroat trout (Oncorhynchusclarki) during fall and winter Canadian Journal of Fisheries and Aquatic Sciences562161ndash2166 DOI 101139f99-154

Heggenes J 1988 Effect of experimentally increased intraspecific competition on seden-tary adult brown trout (Salmo trutta) movement and stream habitat choice Cana-dian Journal of Fisheries and Aquatic Sciences 451163ndash1172 DOI 101139f88-139

Heggenes J Omholt PK Kristiansen JR Sageie J Oslashkland F Dokk JG BeereMC 2007Movements by wild brown trout in a boreal river response tohabitat and flow contrasts Fisheries Management and Ecology 14333ndash342DOI 101111j1365-2400200700559x

Aparicio et al (2018) PeerJ DOI 107717peerj5730 1922

Hesthagen T 1988Movements of brown trout Salmo trutta and juvenile Atlanticsalmon Salmo salar in a coastal stream in northern Norway Journal of Fish Biology32639ndash653 DOI 101111j1095-86491988tb05404x

Hilderbrand RH Kershner JL 2004 Are there differences in growth and conditionbetween mobile and resident cutthroat trout Transactions of the American FisheriesSociety 1331042ndash1046 DOI 101577T03-0151

Hilderbrand RH Kershner JL 2011 Conserving inland cutthroat trout in small streamshow much stream is enough North American Journal of Fisheries Management20513ndash520 DOI 1015771548-8675(2000)020lt0513CICTISgt23CO2

Kennedy RJ Rosell R Hayes J 2012 Recovery patterns of salmonid populationsfollowing a fish kill event on the River Blackwater Northern Ireland FisheriesManagement and Ecology 19214ndash223 DOI 101111j1365-2400201100819x

Knouft JH Spotila JR 2002 Assessment of movements of resident stream brown troutSalmo trutta L among contiguous sections of stream Ecology of Freshwater Fish1185ndash92 DOI 101034j1600-06332002110203x

Komsta L Novometsky F 2015moments Moments cumulants skewness kurtosis andrelated tests Available at https cranr-projectorgwebpackagesmoments indexhtml

Kwak TJ 1992Modular microcomputer software to estimate fish population parame-ters production rates and associated variance Ecology of Freshwater Fish 173ndash75DOI 101111j1600-06331992tb00009x

Leung ES Rosenfeld JS Bernhardt JR 2009Habitat effects on invertebrate driftin a small trout stream implications for prey availability to drift-feeding fishHydrobiologia 623113ndash125 DOI 101007s10750-008-9652-1

Lonzarich DG Quinn TP 1995 Experimental evidence for the effect of depth andstructure on the distribution growth and survival of stream fishes Canadian Journalof Zoology 732223ndash2230 DOI 101139z95-263

Lucas MC Baras E 2001Migration of freshwater fishes Oxford Blackwell ScienceMeyers LS Thuemler TF Kornely GW 1992 Seasonal movements of brown trout in

northeast Wisconsin North American Journal of Fisheries Management 12433ndash441DOI 1015771548-8675(1992)012lt0433SMOBTIgt23CO2

Nakamura T Maruyama TWatanabe S 2002 Residency and movement of stream-dwelling Japanese charr Salvelinus leucomaenis in a central Japanese mountainstream Ecology of Freshwater Fish 11150ndash157DOI 101034j1600-0633200200014x

Niva T 1995 Retention of visible implant tags by juvenile brown trout Journal of FishBiology 46997ndash1002 DOI 101111j1095-86491995tb01404x

Northcote T 1992Migration and residence in stream salmonids-some ecologicalconsiderations and evolutionary consequences Nordic Journal of Freshwater Research675ndash17

Aparicio et al (2018) PeerJ DOI 107717peerj5730 2022

OvidioM Baras E Goffaux D Birtles C Philippart JC 1998 Environmental unpre-dictability rules the autumn migration of brown trout (Salmo trutta) in the BelgianArdennes Hydrobiologia 371372263ndash274 DOI 101023A1017068115183

PeacutepinoM Rodriacuteguez MA Magnan P 2015 Shifts in movement behavior of spawn-ing fish under risk of predation by land-based consumers Behavioral Ecology26996ndash1004 DOI 101093behecoarv038

Porter JH Dooley JL 1993 Animal dispersal patterns a reassessment of simple mathe-matical models Ecology 742436ndash2443 DOI 1023071939594

Quinn JW Kwak TJ 2011Movement and survival of brown trout and rainbow troutin an Ozark Tailwater river North American Journal of Fisheries Management31299ndash304 DOI 101080027559472011576204

Radinger J Wolter C 2014 Patterns and predictors of fish dispersal in rivers Fish andFisheries 15456ndash473 DOI 101111faf12028

Railsback SF Harvey BC 2002 Analysis of habitat-selection rules using an individual-based model Ecology 831817ndash1830DOI 1018900012-9658(2002)083[1817AOHSRU]20CO2

Railsback SF Lamberson RH Harvey BC DuffyWE 1999Movement rulesfor individual-based models of stream fish Ecological Modelling 12373ndash89DOI 101016S0304-3800(99)00124-6

Rasmussen JE Belk MC 2017 Individual movement of stream fishes linking ecologicaldrivers with evolutionary processes Reviews in Fisheries Science amp Aquaculture2570ndash83 DOI 1010802330824920161232697

Rodriacuteguez M 2002 Restricted movement in stream fish the paradigm is incomplete notlost Ecology 831ndash13 DOI 1023072680115

Rovira A Alcaraz C Ibaacutentildeez C 2012 Spatial and temporal dynamics of suspended loadat-a-cross-section the lowermost Ebro River (Catalonia Spain)Water Research463671ndash3681 DOI 101016jwatres201204014

Shepard BB Robison-Cox J Ireland SCWhite RG 1996 Factors influencing reten-tion of visible implant tags by westslope cutthroat trout in-habiting headwaterstreams of Montana North American Journal of Fisheries Management 16913ndash920DOI 1015771548-8675(1996)016lt0913FIROVIgt23CO2

Skalski GT Gilliam JF 2000Modelling diffusive spread in a heterogeneous populationa movement study with stream fish Ecology 811685ndash1700DOI 1018900012-9658(2000)081[1685MDSIAH]20CO2

Sokal RR Rohlf FJ 1995 Biometry the principles and practice of statistics in biologicalresearch New York Freeman

Vatland S Caudron A 2015Movement and early survival of age-0 brown troutFreshwater Biology 601252ndash1262 DOI 101111fwb12551

VeraM Sanz N HansenMM Almodoacutevar A Garciacutea-Mariacuten JL 2010 Population andfamily structure of brown trout Salmo trutta in a Mediterranean streamMarine andFreshwater Research 61672ndash681 DOI 101071MF09098

Aparicio et al (2018) PeerJ DOI 107717peerj5730 2122

Wiens JA 2001 The landscape context of dispersal In Clobert J Danchin E DhondtAA Nichols JD eds Dispersal New York Oxford University Press 97ndash109

Woolnough DA Downing JA Newton TJ 2009 Fish movement and habitat usedepends on water body size and shape Ecology of Freshwater Fish 1883ndash91DOI 101111j1600-0633200800326x

YoungMK 1994Mobility of brown trout in southcentral Wyoming streams CanadianJournal of Zoology 722078ndash2083 DOI 101139z94-278

Zaacutevorka L Aldveacuten D Naumlslund J Houmljesjouml J Johnsson JI 2015 Linking lab activitywith growth and movement in the wild explaining pace-of-life in a trout streamBehavioral Ecology 26877ndash884 DOI 101093behecoarv029

ZimmerM Schreer JF PowerM 2010 Seasonal movement patterns of CreditRiver brown trout (Salmo trutta) Ecology of Freshwater Fish 19290ndash299DOI 101111j1600-0633201000413x

Aparicio et al (2018) PeerJ DOI 107717peerj5730 2222