2016 fish assemblages in neotropical reservoirs colonization patterns, impacts and management

7/25/2019 2016 Fish Assemblages in Neotropical Reservoirs Colonization Patterns, Impacts and Management

1/11

Fisheries Research 173 (2016)2636

Contents lists available at ScienceDirect

Fisheries Research

journal homepage: www.elsevier .com/ locate / f ishres

Fish assemblages in Neotropical reservoirs: Colonization patterns,impacts and management

Angelo A. Agostinhoa,b,, Luiz C. Gomesa,b, Natlia C.L. Santosa,Jean C.G. Ortegaa,FernandoM. Pelicice c

a Programa de Ps-Graduaco em Ecologia de Ambientes Aquticos Continentais, Ncleode Pesquisa em Limnologia, Ictiologia e Aquicultura, Universidade

Estadual deMaring, Laboratrio de Ictiologia, Av.Colombo, 5.790, Bloco H-90, CEP: 87020-900Maring, PR, Brazilb Ncleo de Pesquisa em Limnologia, Ictiologia e Aquicultura,UniversidadeEstadual de Maring, Maring, PR, Brazilc Ncleo de Estudos Ambientais, Universidade Federal de Tocantins, Rua3, Quadra 17, Jardimdos Ips, CEP77500-000Porto Nacional, TO, Brazil

a r t i c l e i n f o

Article history:

Available online13May 2015

Keywords:

Freshwater fishDamimpactFishmanagementFish stockingFishpass

a b s t r a c t

Brazil has more than 700 large reservoirs distributed in all ofthe major river basins ofSouth America.Most dams were constructed to produce electricity. Although these reservoirs favor the development oflocal and regional economies, they seriously impact the aquatic biota. An unavoidable consequence isthe change in the composition and abundance ofspecies, with the proliferation ofsome and reductionor even local extinction ofothers. The intensity and nature ofthese changes are related to peculiaritiesofthe local biota and the location, morphometric and hydrological characteristics ofthe reservoir, damoperation and interactions with other uses ofthe basin, including other reservoirs. These impacts exhibitsubstantialspatiotemporalvariations.Thefillingphase ismarked byabruptandintense changesin thekeyattributesofaquatic habitats, followed bypredominantly heterotrophic processes, with possible thermalstratificationand anoxic conditions. Fish richness increases soon after filling and decreases in subsequentyears. Trophic depletion is expected, and diversity gradients are intensifiedtowardmore lentic stretches,the average length offish decreases, and the fish fauna becomes dominated by species with sedentarystrategies and/or parental care. The virtual absence ofspecieswith pre-adaptations to inhabit lentic areas

of large reservoirs leads to a concentration ofbiomass in shallow littoral areas. Long-distance migratoryspecies are themost affected,which include larger fishwithhighmarket value.Migratory species requiredifferent biotopes to fulfill their life cycles and strongly depend on the seasonal flood regime, which isaltereddue to damoperation. In this study, wediscuss the details ofthese trends aswell asthemitigationmeasures andmanagement actions that are practiced in Brazil. We conclude that these actions have notpromoted the conservation offish; on the contrary, some ofthem have generated additional impacts. Asa consequence, the conservation ofNeotropical fish and aquatic resources is severely threatened.

2015 Elsevier B.V. All rights reserved.

1. Introduction

Impoundments lead to extreme changes in fluvial habitats,transforming rivers into semi-lentic systems. Animals and plantsforwhichthesenewconditionsare restrictivewill have theirpopu-lations drastically reduced. However, species that can complete

Correspondingauthorat: UniversidadeEstadualde Maring,NcleodePesquisaem Limnologia, Ictiologia e Aquicultura, Programa de Ps-Graduaco em Ecologiade AmbientesAquticos Continentais, Laboratrio de Ictiologia, Av.Colombo,5.790,Bloco H-90, CEP: 87020-900Maring, PR,Brazil. Tel.: +5544 3011 4610.

E-mail addresses: [email protected], [email protected](A.A. Agostinho), [email protected](L.C.Gomes), [email protected](N.C.L. Santos), [email protected](J.C.G. Ortega), [email protected](F.M. Pelicice).

their life cycle in the new environment and take advantage of theavailable food resourceswill achieve their full potential for prolif-eration (Agostinho et al., 2007a). The nature of and intensitywithwhich thefluvialbiota is altered by impoundments arehighlyvari-able among reservoirs and must be studied case by case.

The literature demonstrates that even reservoirs arranged inseries in the same river, with unidirectional interactions fromupstream to downstream, show distinct peculiarities in relationto the colonization process and the organization of assemblages(Agostinho and Gomes, 1997; Petesse and Petrere, 2012). Thedegreeof alteration in thestructure anddynamics of the localbiotadepends on several local and regional factors, such asmorphome-tryofthecatchment,discharge,patternsofwatercirculation,depth,habitat structure, species pool, surface area, the design of the damand its operational procedures. Thus, a detailed understanding of

http://dx.doi.org/10.1016/j.fishres.2015.04.0060165-7836/2015 Elsevier B.V. All rights reserved.
http://localhost/var/www/apps/conversion/tmp/scratch_7/dx.doi.org/10.1016/j.fishres.2015.04.006http://www.sciencedirect.com/science/journal/01657836http://www.elsevier.com/locate/fishresmailto:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]://localhost/var/www/apps/conversion/tmp/scratch_7/dx.doi.org/10.1016/j.fishres.2015.04.006http://localhost/var/www/apps/conversion/tmp/scratch_7/dx.doi.org/10.1016/j.fishres.2015.04.006mailto:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]://crossmark.crossref.org/dialog/?doi=10.1016/j.fishres.2015.04.006&domain=pdfhttp://www.elsevier.com/locate/fishreshttp://www.sciencedirect.com/science/journal/01657836http://localhost/var/www/apps/conversion/tmp/scratch_7/dx.doi.org/10.1016/j.fishres.2015.04.006


2/11

A.A. Agostinho et al. / Fisheries Research 173 (2016) 2636 27

the context of a particular reservoir is paramount foreffectivemit-igation measures and/ormanagement actions for the conservationoffishpopulations (Weithman andHaas, 1982). Amanager should,based on local and regional studies, identify any alterations in thestructure of the local fish assemblage and take action to avoidirreversible losses of regional biological diversity and/or naturalresources as a consequence of river damming.

In general, the fish species most affected by impoundments arelarge in size, migrate and have high longevity(k-strategist). In con-trast, a massive proliferation of primarily small-sized sedentaryspecies (i.e. those that do not migrate) occurs, which have a highreproductive potential and short longevity (r-strategists) and forwhich the availability of food resources is high (Agostinho et al.,1999, 2008a; Hoeinghaus et al., 2009). Yet, sedentary species arealso affected by hydrological alterations and tend to redistributealong the river/reservoir gradient (Arajo et al., 2013). In the innerareas of large reservoirs, fish assemblages are profoundly alteredandcomposedofafewspecieswithpre-adaptationstoliveinsemi-lentic environments (Gomes and Miranda, 2001; Agostinho et al.,2007a).

Reservoirs are present in the main river basins in Brazil, andthe principal purpose is the production of electricity. Althoughreservoirs are widespread in the country, their distribution is nothomogeneous, e.g. the Upper Paran River has half of the totalimpounded area and is one of the most regulated rivers in theworld (Agostinho et al., 2008a). Even considering the specificityof the responseof thebiota to the impacts generatedby each reser-voir, some patterns can be described based on studies of dozensof reservoirs in Brazil. Therefore, the objective of this paper is toreview the patterns of fish fauna once a reservoir is formed. First,we described thevariation in fish assemblages over time, from thefilling of the reservoir to the periods in which environmental andbiotic conditions are rearranged and more stable. We categorizedthesevariationsintophases(heterotrophic,post-heterotrophicandtrophic equilibrium), considering predicted alterations in produc-tivity. Then, considering the phases, we described broad trendsin fish abundance, species richness, pre-adaptations to pelagic

environments, and variations in size and reproductive strategies.Finally, we evaluated management measures presently imple-mentedtomitigate impacts causedby reservoirs ontheNeotropicalfish fauna, and we discuss opportunities for improvement as wellas the existing knowledge gaps. As the Upper Paran River basin isthe most dammed in South America as well as the most studied,we used it as a model to achieve our goals every time an examplewas necessary.

2. Reservoirs and fish diversity

It is estimated that the number of large reservoirs (dams higher

than15m;World Commission onDams, 2000) in South America isgreater than one thousand, and around 50% of them are locatedwithin Brazilian territory (Fig. 1). Thirty-seven percent of thesereservoirs produce electricity. Although hydroelectric productionin dams started in Brazil at the end of the XIX century (MarmelosDam; Paraba do Sul River; 1889), most of the dams were con-structed in the second half of the XX century. With regard to thearea inundated by all reservoirs (>36,000km2), almost half of it(47%) is located in the Paran River, followed by the So Franciscoand Tocantins Rivers (Agostinho et al., 2007a). As potential areasfor the installation of new dams in these basins are depleted, thereis a motivation to extend the construction of dams to the Amazonbasin,especially in theMadeira, Tapajs andXingRivers (Castelloetal.,2013), inadditiontotheAndeantributaries(FinerandJenkins,

2012).

Ichthyofaunal monitoring surveys conducted in 77 reservoirs ofthemain river basins inBrazil(Agostinhoet al., 2007a) showedthatfishdiversity in the impoundedarea is very low. This study showedthat 85% of the reservoirs contain fewer than 40fish species; reser-voirs with more than 120 fish species are rare and usually young.Forty species can be considered very low if we consider that 80%of these reservoirs have areas greater than 10km2 and that a sin-glefloodplain lake ofmuch smaller dimensionscanharbor from 30species (Paran River basin; Oliveira et al., 2001) up to 99 species(AmazonRiver basin; Pouilly et al., 2004). In addition, streams andriversin theNeotropical regionusuallypresent hundreds ofspecies(Lowe-McConnell, 1999; Agostinho et al., 2007b), e.g. a stream lessthan 10km long had 108 species (Cancela Stream; Cuiab Riverbasin;Mendeset al., 2008). However, species richness in reservoirsvaries with their surface area, age and, primarily, the basin wheretheyare located. Thus, reservoirs located in theAmazonbasinwithareas greater than 500km2 and less than15years old containmorespecies than other reservoirs of similar dimensionsand agethat arelocated in other Neotropical basins. For example, more than 200fish species were found in the So Salvador Reservoir, TocantinsRiver (104km2; Amazon basin), in the first years after impound-ment (Limnobios, 2014). In contrast, 34 species were recorded inSegredo Reservoir (85 km2; Iguacu River; Agostinho and Gomes,1997) and107inItaipuReservoir(1350km2; Agostinhoetal.,1992)in a similar time lag. Furthermore, in Capivara Reservoir (576km2;Paranapanema River; Orsi and Britton, 2014) and Salto Osrio(63km2; Iguacu River; Baumgartner et al., 2006), both impound-ments are older than 30 years, were recorded 41 and 23 species,respectively. In fact, there is a consistent decrease in species rich-ness over time (Mol et al. , 2007; Orsi and Britton, 2014), i.e. thenumber of species averages 20 in Neotropical reservoirs older than20 years (Agostinho et al. , 2007a). This conspicuous decline inspecies richness is the resultof environmentalfilters that graduallyremovepre-existingfluvial species; thenewassemblages arecom-posed basicallyof species that present pre-adaptations to thrive instanding waters, with lower dependence on fluvial environmentsand habitat heterogeneity (Gomes andMiranda, 2001).

3. Variation in fish abundance

The large release of nutrients resulting from the decompositionoforganicmatter in theflooded areaduringa reservoirs earlyyearsand the subsequent reduction of nutrients result in wide fluctua-tions in production throughout a reservoirs history. The nutrientinput increases theproduction of all trophic levels during a periodknownasthetrophicupsurgeperiod (KimmelandGroeger,1986;Kimmel et al., 1990). This heterotrophic period begins in the fill-ing phase, which is marked by rapid and profound alterations inthewaters physical andchemical characteristics. Duringthefillingphase, vertical patterns resulting from the expansion of the water

column, lentic characteristics and thermal stratification, whichaffect the sedimentation rate, nutrient cycling and the distributionof the biota, are added to the predominant transport vector of theriver phase. Thehigh concentration of nutrients initiallydue to thepulses of litter decomposition and the release of nutrients fromthe inundated soil, followed by the decomposition of the leaves ofthe inundated vegetation (Cunha-Santinoet al., 2013), may lead tostressful conditions forthe aquatic biota (e.g. lowconcentrationsofdissolved oxygen, thermal stress, and low pH), especially near thebottom (Agostinho et al., 2008a).

For example, studies conducted in Corumb Reservoir (locatedin the Upper Paran River basin) showed a sharp increment inprimary production after an initial period of increased watertransparency (Secchi depth)due to sedimentation. Thus,the phyto-

plankton productivity that was below 0.17mgO2

l1

in the first 10


3/11

28 A.A. Agostinho et al. / Fisheries Research 173 (2016) 2636

Fig. 1. Map showing thedistribution of hydroelectric reservoirs in themain river basins of Brazil (Paran River basin is highlighted dueto thehigh numberof dams).

days after the reservoir began filling, reached 0.89mgO2 l1 after39 days (Agostinho et al., 1999). Therefore, due to this increase inprimary productivity, it is common a marked increase in the fishabundance in the inner areas of the reservoir (Fig. 2). The increasein nutrient concentration is also responsible for the intense pro-liferation of floatingmacrophytes, as verified in Tucuru Reservoir,located in the Tocantins River basin (Tundisi, 1994). The presenceof ananoxic layer during filling is also anevent common to tropicalreservoirs and can last for months or even years. However, afterfilling, a reduction in fertility due to the loss of organic matter byoxidation, sedimentation, biological assimilation and exportationis common (Cunha-Santino et al., 2013); this period is known asthe post-heterotrophic or depression period (Fig. 2).

Once thephaseofhighproductivity isover, fish species begin toadjust to thenew environment (Petrere, 1996). The highfish abun-dance verified during this phase tends to decrease in the reservoirover time (Fig. 2). This decrease will continue until the reservoirreaches a certain trophic equilibrium (at an unknown time), after

which the abundance of fish tends to be less variable but usuallyhigher than in the river before the dam was constructed.

To exemplify this decrease in productivity over time, we usedsample data collected in the Itaipu Reservoir (Paran River basin)from 1983 (one year after filling in 1982) to1997 (15 years afterfilling). In this reservoir, there was a clear temporal decrease infishabundance in number (catch per unit effort, CPUEnumber ofindividuals) and inweight (CPUEkg), especially in themore inter-nalareas (lacustrine zone).Notethat thedecreasedabundancewasfrom two- to four-fold in number and weight, respectively (Fig. 3).A clear decrease in fish abundance was noted for all zones (it wasless noticeable in thetransitionalzone),but a sharper decreasewasobserved in the inner areas (lacustrine zone) of the reservoir. Fishbiomass showed the same trends (Fig. 3). These results demon-strate that the degree of the impact of the impoundment on fishabundanceor biomass hasa longitudinal gradient.Orsi andBritton(2014) also reported a sharp decline in native fish abundance 40years after the formation of Capivara Reservoir, which involved


4/11


Fig. 2. Trends in fish abundance (catch per unit of effortCPUE) over time inNeotropical reservoirs (modified fromPetrere, 1996; Agostinho et al., 2007a).

Fig. 3. Variations in the catch per unit effort (CPUE; individuals and weight for1000m2 of gillnet in 24h) in thelongitudinal gradient of theItaipuReservoir from1983 (one year after the impoundment) to 1997 (white circles: CPUE in numbers;black circles: CPUE inweight).

the loss of 27 species; at present, non-native opportunistic speciesdominate the assemblage in Capivara.

4. Fish colonization during the filling phase

During reservoir filling, the patterns of vertical colonization offishareassociated with thermal stratification, an increase in depthand a sharp decrease or even the virtual absence of dissolved oxy-gen. All these factors impose changes on fishdistribution patterns.The increase inwater volume and the reduction inwater flow leadtoanincreaseintheareaavailableforcolonization(Agostinhoetal.,2008a;Wang et al., 2013). The lack of oxygen in the deeper strata(in the bathypelagic zone) may lead species to disperse in vertical

andhorizontal directionsor even upstream, far from the lacustrine

environment. Therefore, the changes in the physical and chemicalcharacteristics of the water due to the beginning of reservoir fill-ing may act as environmental filters, selecting for ecological traitssuch as trophic guilds, reproductive strategies and an alterationof the affinity for habitats (fidelity), which determine the successof colonization by a particular species. Species that successfullycolonize a reservoirhave theability to searchfor adequate environ-ments, such as lotic tributaries or even the littoral areas (Agostinhoet al., 2007a), aswell as those that developstrategies different fromthose exhibited in theprevious lotic environment (Kubecka, 1993).Specieswith pre-adaptationsto live in lacustrineenvironments areobviouslyselectedtocompose thenew assemblages (FernandoandHolcik, 1991;Gomes and Miranda, 2001).

Colonization during the filling of the Salto Caxias (IguacuRiver Basin) and Corumb (upper Paran River Basin) reser-voirs exemplifies the patterns of occupation of the new floodablearea. To study this phenomenon, we categorized the fish speciesaccording to habitat preference (benthonic, bentho-pelagic andpelagic) and capture location in relation to the longitudinal gra-dient (riverine and lacustrine zones) and habitat (littoral, pelagicor bathypelagic). During the filling phase of these reservoirs,a greater abundance of benthonic species was observed in thelittoral (Salto Caxias Reservoirtwo-way ANOVA; Interaction,ZonesHabitat Type; F2, 6 =5.99, p=0.037; Fig. 4a) and pelagichabitats (Corumbtwo-wayANOVA; Interaction ZonesHabitatType; F1, 8 =7.90, p=0.023; Fig. 4b), respectively. In the riverinezone, as expected, we found the opposite pattern, with benthonicspecies occupying deeper strata. A similar pattern was observedfor bentho-pelagic fish, with a greater capture rate in the littoral(F2, 6 =5.98, p=0.037; Fig. 4c) and pelagic (F1, 8 =14.00, p=0.006;Fig. 4d) areas. In contrast, pelagic species were captured in verylow abundance in both reservoirs and did not show any differ-ences in abundance according to the zone and habitat type (allpossible results withp>0.05). This result is due to the existence offewpelagic species in theNeotropical region (Gomes andMiranda,2001; Arajo et al., 2013).

The results presented for Salto Caxias and Corumb reveal low

habitat fidelity for benthic species in the inner part of the reser-voirs during the filling phase. In fact, fish use a habitat accordingto physiological convenience,which depends primarilyon thecon-centration of dissolved oxygen and water temperature (Prchalovet al., 2009); their vertical distribution is apparently driven byrestrictions related to thermal and dissolved oxygen stratifica-tion in the reservoir. This stratification, during the first year afterimpoundment, can lead to a chaotic pattern of species occupy-ing habitats in which they were previously not abundant (e.g.benthonic species abundant in the littoral or pelagic zonesof reser-voirs). Yet, the reassembly of fish species following changes inenvironmental conditions may occur within the first years afterimpoundment, creating newdiversity patterns along the reservoir(Arajo et al., 2013).

After thefillingof thereservoirand thebeginningof damopera-tion, critical conditionsof dissolvedoxygenmaypersist, dependingon the extension of the anoxic layer and the vertical position ofthe water intake for turbines and spillway. These conditions maylead to a narrow oxygenated layer, resulting in instability due towind and temperature changes, which can culminate in fish mor-talityconcentratednearthemarginsorthesurface(Agostinhoetal.,1999).

5. Heterotrophic and trophic equilibriumphases

There is evidence from several Neotropical reservoirs that thespecies richness increases immediately after the filling phase

(Fig. 5a and b). This increase in species richness is followed by an


5/11


Fig. 4. Mean catch per unit of effort (CPUE; log10x+1 transformed; vertical lines are the standard errors) for the species categorized as benthonic (a andb), bentho-pelagic(c and d) and pelagic (e and f) among the distinct reservoir zones (riverine and lacustrine) and types of habitats (littoral, pelagic and bathypelagic). (a), (c) and (e) Fillingphase of the Salto Caxias Reservoir (1998); (b), (d) and (f) Filling phase of the Corumb Reservoir (1996). Note that in Corumb, samples were taken only from the pelagicand bathypelagic regions.

increase in the abundanceof fish (Fig. 5c), which is common duringthetrophicupsurgeperiod.However,themagnitudeof theincreaseinabundancevariesamongspeciesinanewreservoir,andthedom-inance of certain species with regard to abundance (low evenness)causes a continuous decreasing in the species diversity measuredby the Shannon index (Fig. 5a).

An increase inspecies richness isexpectedduringfilling becausedifferent biotopes, such as wetlands, isolated lakes, lakes perma-

nently or seasonally connected to the river channel and adjacenttributaries (river, streams and creeks), are incorporated into thenewenvironment.Speciesassociatedwith these habitats areincor-porated into the fish fauna of the reservoir, which consequentlyincreases species richness. The number of species present in arecent reservoir should not be much lower than the sum of thepreviously existing species in the flooded habitats. However, thistendency for high richness does not last long (Fig. 5a). The rea-sons for its decrease have beenpreviously discussed and appear tobe related to environmental filters, species sorting and the accom-modation of the fish fauna to the new environment, in additionto trophic depletion and the absence of truly lacustrine species(Agostinhoet al., 2007a). Theoretically, thedrop in species richnessresults from the movement of fish out of the reservoir (upstreamor tributaries) in search of better conditions to complete their life

cycle(Lowe-McConnell,1999;Agostinhoet al., 2007b;Arajoetal.,2013; Franssen and Tobler, 2013).

Abundance follows a similar trend. Upon filling of a reservoir,fish abundance increases (Fig. 5c) due to the high input of terres-trial organic matter, which leads to increased food availability inthe entire reservoir, especially for omnivorous, herbivorous andinsectivorous species. The proliferation of these species causes anincrease in food availability for piscivores. However, at the end of

theheterotrophicphase(see Fig.2), theabundanceof fishdecreasesfollowing the decrease in primary production (see Section 3).

6. Constraints andpre-adaptations

The virtual absence of natural lakes in Brazil (excepting thoseassociated with fluvial corridors) and the consequent scarcity ofspecies with pre-adaptations to occupy open areas of reservoirs,allied with the longitudinal gradients related to the processes oftransport and deposition (e.g. transparency and nutrient loads),leads to a heterogeneous pattern of the occupation of the newenvironment. The most important characteristics of truly pelagicspecies are their short food chains, high fecundity, pelagic adap-tations, and short life cycle, as exhibited by the Clupeiformes

Stolothrissa tanganicae (in Africa) and Dorosoma cepedianum (in


6/11


Fig. 5. Variations in the Shannon diversity index ((a) before and twoperiods afterthe formation of Jordo Reservoir; numbers in brackets are species richness andevenness), species richness ((b) before and twoperiods after the formation of fourreservoirs in the Upper Paran River basin) and abundance ((c) catch per uniteffortCPUE ind. 1000m2 gillnet in 24h before and after the formation of fourreservoirs in theUpperParanRiverbasin).ModifiedfromAgostinho et al., (2007a).

North America) (Gomes and Miranda, 2001). Thus, in reservoirs,the colonization success of the species depends on their pre-adaptations. In theUpper ParanRiver,Gomes andMiranda (2001)described that, among the220 species theyanalyzed, only approx-imately 5% were considered as lacustrine adapted (i.e. Plagioscionsquamosissimus, Hypophthalmus edentatus).

In general, the pelagic areas of the Upper Paran Reservoirsare inhabited by few fish species, such as the piscivores P. squa-mosissimus (a sciaenid introduced from the Amazon basin) andRhaphiodon vulpinus, and the planktivoresH. edentatus andHemio-dus orthonops (the latter was recently introduced through the fishpassage at Itaipu Dam;Julio et al., 2009; Agostinho et al., 2015).For example, the success ofP. squamosissimus may be attributed

to its reproductive strategy (Agostinho et al., 1999). This speciesproduces small, pelagic (FonteneleandPeixoto, 1978), andbuoyanteggs, spawned in several batches early in the reproductive period(matching with food availability), and the larvae are also pelagic(Nakatanietal.,1993). Otherimportantcharacteristicsaremorpho-logicalandrelatedtodietandfoodcapture(Mrona andVigouroux,2012). Therefore, the lownumber and unevendistribution of largeanddeepnaturallakesintheNeotropicalregionledtothecompleteabsence of a truly pelagic and deep bottom-dwelling species thatarepre-adaptedto occupyopenareasof large reservoirs. Ingeneral,species that successfully colonize reservoirs are those that inhabitshallowfloodplainlakes,whichusually occupythe littoral regionofreservoirs (Casattiet al., 2003;Pelicice et al., 2005;Agostinhoet al.,2007a). Thus, the greatest abundance of fish species and diversity

are found in the littoral region.

Fig. 6. Spatial (longitudinal, lateral and vertical gradients) and temporal gradientsin species richness and abundance of fish in the Itaipu Reservoir (Riv= Riverine;Tra=Transitional; Lac= LacustrineSource: Agostinho et al., 1999).

This pattern can be verified in the Itaipu Reservoir five (1987)and15 (1997) yearsafter itsformation (Fig.6). Species richness andtheabundanceoffishwere considerablyhigher in the littoral zone,and this pattern tended to increaseover time. After15 years, 64outofthe67speciescapturedingillnetsintheItaipuwereinthelittoralzone, whereas in the pelagic and bathypelagic zones, this numberwas 22 and 20, respectively (Fig. 6). The proportions of abundanceamongthezoneswere19.4:1.1:1.0,respectively(Fig.6).Moreover,the riverine zone, inwhich theprocesses of transportpredominateover the depositional processes, has higher species richness but

is not the most productive zone (Kimmel et al., 1990; Agostinhoet al. , 2007a). The upper third of the Itaipu Reservoir harbors allof the species recorded in the two more internal thirds in addi-tion to those typical to the lotic stretch upstream (the river). Thehigher similarities in flow with the original river, the lower depth,the input of allochthonous matter, and predator attraction due thehigherabundanceofpreyspecies inrelationto theupstreamstretchmay explain this pattern (Agostinho et al., 2007a; Arajo et al.,2013).

Reproduction, dueto itsmoreconservativenature, imposes lim-itationsontheoccupation ofa newreservoir by theriverfishfauna;it is probably the main constraint limiting fish fauna reassembly.In reservoirs, it is expected that species with higher plasticity inthe selection of spawning sites havemore success in the coloniza-

tionoftheseenvironments.Speciesthatdemandparticularhabitats(e.g. tributaries) or environmental triggers (e.g. hydrological varia-tion) may not complete the reproductive process, mainly duringthe years following the impoundment. Medeiros et al. (2014)reported failed reproduction for Hemiodontidae after the forma-tion of Lajeado Reservoir, Tocantins River, an event that changedenergy allocation patterns for these species. However, most ofthe species that inhabit reservoirs search for lateral tributaries,upstreamstretchesor other loticareas forspawning,indicating thedependenceon riverinehabitats to complete their life cycles. In thefirst years after the formation of a reservoir, internal fecundationappears to be a successful strategy. However, in older reser-voirs, species with more elaborate reproductive strategies (usuallycichlids with complex mating choice, nest-building and parental

care) have greater occupation success, along with small-sized


7/11


8/11


Fig. 9. Decadetendenciesof thedominant species in thelandings of theartisanal fisheryconducted in theParan River beforeand after theformationof theItaipu Reservoir.The years presented arefive years beforethe reservoirwasformed andfiveand 15 years after theformation of thereservoir (adapted from Agostinhoet al., 2007a).

the implemented actions; (iii) the eminentlypoliticalnatureof themanagement decisions,which should be essentially technical; (iv)insufficient knowledge about the problems to be solved,which ledto a lack of clarity in the objectives of management actions; and(v) the nave belief that impacts caused by impoundments can bereversed orminimizedwithsimplemanagementactions or copiedfrom other part of theworld.

Concerns with impoundment impacts on migratory speciespopulations led to the recommendation to construct fish passagesin the dams. This action attempted to facilitate the transit of fishto their spawning or feeding sites or the dispersal of juvenile fishto downstream stretches of the basin. The installation of thesepassages was mandatory for decades (Decree 4390 of 1928), andit is still mandatory in some Brazilian states. However, the con-struction requirements and the use of standardized protocols toinstall fish passages, whose performance depends on the interac-tion between their technical characteristics and the nature of thelocal ichthyofauna, were at high risk of failure, wasting financialresources, effort and opportunity (Agostinho et al., 2002; Peliciceand Agostinho, 2008; Pompeu et al., 2012). In fact, some fish pas-sages were built immediately upstream of natural barriers thatfish historically did not cross (Charlier, 1957). Other passageswere highly selective, allowing the passage of large numbers ofsedentary species and restricting the passage of the migratoryspecies (Agostinhoetal.,2007b;Makrakisetal.,2007a). In addition,invasive species previously limited by natural barriers have been

reported to be greatly dispersed along river channels, as the caseof the Itaipu Dam in the Paran River, where the Piracema Canal islocated (Julio et al., 2009; Makrakis et al., 2007b; Agostinho et al.,2015). In this case, the Itaipu Reservoir, filled in 1982, covered theSeteQuedasFalls, whichwasthe limit forthe distributionofseveralspecies, separating twodistinct ichthyofaunaprovinces, the upperand the middle Paran (Bonetto, 1986). After the Itaipu Reservoirwascompletely filled, several species were able to reach the upperpart of the basin (Julio et al., 2009). Some of these species becameabundantandreplacednativecongenericspecies( Agostinho,2003;Alexandre et al., 2004). During the following 20 years, approxi-mately 17 species remained restricted to the stretch downstreamfrom the Itaipu Dam. However, after the Canal de Piracema (anatural-likefishpassage) started operation,other specieswereable

to reach the reservoir and dispersed to the upstream stretches ofthewatershed(Makrakisetal., 2007a; Julio et al., 2009; Vituleet al.,2012). An emblematic example wasH. orthonops, absent from theupper Paran River Basin. The invasion of this species was note-worthy for both its fast colonization of the new environment andfor its abundance, reaching approximately 8% of the total catchat the upstream plain, in less than five years (Agostinho et al.,2015).

However, effectivemonitoring of theperformance of these pas-sages began only in the 2000s; even though fishways have beeninstalledforacentury.Yet,mostofthecurrentstudiesarerestrictedto monitoring species in the fish passage (fish ladders and fishelevators) with no consideration of the availability of adequatehabitats for the species in the upstream stretches or in the region.Tagging studies on fish movements are recent (Hahn et al., 2007;Fontes et al. , 2012; Wagner et al., 2012), and these studies havenot eliminated the controversies regarding the adequacy and theefficiency of fish passages. The controversial aspect on the fishpassage issue refers mainly to its simplicity and convenience formanagement programs that are mandatory in Brazil, but with lowsignificance for conservation, i.e. aiding recruitment of migratoryfish (Pelicice and Agostinho, 2008; Pompeu et al., 2012).

Other relevant aspects that should be considered in discussionsonfishpassages are their high selectivity (Agostinho et al., 2007d),the difficulty in controlling which species go through the passage(Pompeu et al., 2012), and the absence of downstream movement

of adults and their offspring (Agostinho et al., 2007c, 2011; Suzukiet al., 2011; Pelicice andAgostinho, 2012). Solutions to the existingbottlenecks concerning recommendationsof fishpassages as a tooltomitigate impacts onmigratory species must address the follow-ingissues:(i)whetherthepassagesareefficienttoattractfishandtoallow free movements; (ii) whether the reservoirs represent a bar-rier to downstreammovement of adult fish or to the drift of theireggs and larvae (Agostinho et al., 2007c; Pelicice et al., in press),(iii) whether long-distance migratory species have distinct behav-iors and the swimming ability to be attracted to and to overcomethe water flow in a fish passage; (iv) whether the passage is safe,with low rates of injury or predation (Agostinho et al., 2012), (v)whetherclear objectivesexist (e.g. geneticand/ordemographic) tojustify the use of fish passes, and (vi) whether the regional context


9/11


(i.e. distribution of critical habitats) supports theuseof fishways toachieve conservation goals (Pelicice andAgostinho, 2008; Pompeuet al., 2012). While these aspects areneglected, the decision on theconstruction of fishpassageswill remainnebulous,with the risk ofcausing further impacts and complicatingalternativeconservationefforts.

Another controversial management action has been fish stock-ing. The first stocking initiatives were conducted with non-nativespecies in northeastern Brazil and were successful in producingself-sustaining populations and improving fishing yields (Paivaet al., 1994). This experience, especially with non-native species,spread to other regions of Brazil and was the main fisherymanagement activity conducted by the Brazilian fishery-relatedinstitutions and by power companies. Until 1990, non-nativespecieswereemphasizedinstockingprogramsdevelopedinsouth-easternandsouthernBrazil.Someofthestockednon-nativespecieswere successful colonizers and are currently widespread in manybasins (e.g. silver croaker P. squamosissimus; peacock basses Cichlaspp.).Other specieswere successful in some reservoirs,where theyappear in high abundance (e.g. tilapiasmainlyOreochromis niloti-cus, oscar Astronotus ocellatus, and freshwater sardine Triportheusangulatus). Although stockingactivitiescurrentlyemphasizenativespecies, monitoring of commercial (artisanal) fishery outputsshows that stocking has not been efficient and might even repre-sent an additional source of impacts (Agostinho et al., 2004, 2010).There are no relationship between stocking efforts and captures(landed fish) in artisanal fisheries in the reservoirs of southeast-ernandsouthernBrazil, where stockingwasmore intense. Some ofthestocked specieswere never captured in thefisheries (AES-Tiet,2007).

Stocking programswere historically conducted based on a pre-carious knowledge of the system to be managed, of the speciesto be released, and of the real need for the action. In addition,inexperience regarding how to conduct stocking (which species,the necessary quantities, the appropriate location, the size of thefish, and the time of release, among others) led to the practiceof trial and error. Furthermore, stocking has been conducted

without monitoring and without learning from the past actions,which can be helpful in avoiding future mistakes (Gomes et al.,2004; Agostinho et al., 2007a; Pelicice et al., 2009). For example,knowledge of the carrying capacity of the receptor environmentand the size of wild stocks are fundamental assumptions in stock-ing for supplementation (Cowx, 1999), and such knowledge hasbeen ignored in the stocking programs conducted in Brazil. Addi-tionally, the processes used to rear fish for stocking programs arefrequently the sameused toproduce fish for farming; in fact,fishesfor both purposes have been reared together in fish farms. Thus,by ignoring the genetic quality of the brooders and other possi-ble negative impacts on natural populations, stocking became apotentialandconstant threatto localpopulations andto thefisheryitself, although such consequenceswere never empirically studied

(Agostinho et al., 2010). It should be highlighted that stocking pro-grams are supported by society, based on the nave belief that fishpopulations were impacted, declined and must be recomposed inthe reservoirs; stocking, in this sense, is a valid compensation ormitigationmeasure toaddress the impactof impoundmentsonfishdiversity and fishery resources (Agostinho et al., 2007a). This is acommon-sense explanation used in legislative initiatives to makestockingmandatory in theentirecountry,recentlyapprovedin LawProject 5989-09 (Lima et al., 2012; Pelicice et al., 2014). Ideally,stocking strategies should consider (i) the need for stocking, basedondetailed information about the environment, the target species,and the intensity of the exploration; (ii) an understanding of theprocesses that drove thewild stock to depletion; (iii) the establish-ment of clear and quantifiable objectives; (iv) the capacity to rear

and distribute fries with a genetic quality equivalent to the wild

stock (Flagg and Nash, 1999); (v) the compatibility of the quantityof fish, size, place and timing of the release with the distributionand structureof natural populations (Molony et al., 2003); and (vi)monitoring of the stocking and wild populations. In fact,monitor-ing programs should be an integral and indissoluble component ofstocking, and the results obtained should be the base for adjustingor evenhalting the procedures.

Aquaculture, in a strict sense, is not considered a managementactivity destined to mitigate impoundment impacts. However,aquaculture has been conducted under the argument that it mini-mizes fishing pressure on wild stocks, either by the involvementof the fishermen in production activities (farming) or by a reduc-tionin the demand for wildfish (Agostinhoet al., 2007a). Althoughconsidered an important food production activity, aquaculture, asany other production method, affects the environment with anintensity that varies according to the type (intensive or extensive)and the species farmed. Such impacts are evident in the intensivefarming conducted in caging nets installed in Brazilian reservoirs,whichhasreceivedsubsides from thegovernmentalfinancialagen-cies related to fishproduction (Agostinho et al., 2008b; Lima et al.,2012).Althoughfishfarmingincageshasnotbeenadequatelymon-itored, preliminary studies already indicate some distortions withregard to proposed objectives, conflicts among users, profitability,introductionofspecies,and aquacultureasa sourceofwaterqualitydegradation (Agostinho et al., 2007a; Strictar-Pereira et al., 2010;Azevedo-Santos et al., 2011; Pelicice et al., 2014). Given the com-mon occurrence of escapes in aquaculture, the use of non-nativespecies was prohibited in reservoirs where these species were notestablished. However, this restriction was removed by a FederalDecree, which provided the status of native to several speciesfrom other continents (i.e. tilapias species), as a mean of fosteringaquaculture in large reservoirs (Vitule et al., 2012; Pelicice et al.,2014). This decision may increase non-native dispersion acrossSouth American basins; it is well known that aquaculture is themain sourceof non-nativespecies toNeotropical reservoirs (Ortegaet al., 2015). In addition, fish farming of native species in cagesconducted by traditional fishermen has not been promising due to

the high costs of production, difficulties in commercialization, andsmall-scale production stemming from the investment capacity ofthe fishermen. Regardless of these negative points, aquaculturein public waters (reservoirs) may be environmentally sustainableandmay promote social development,generating incomeand jobs.However, such a systemrequires a programwith ample interactionwith other activities related to fishery resources, created with rig-orous planning and sustained by technical studies on production,impacts and marketing (Agostinho et al., 2008b). Unfortunately,aquaculture in Brazilian reservoirs does not follow these highenvironmental standards, and constitutes an additional source ofdisturbancetowildfreshwaterfish(Agostinhoetal.,2007a;Peliciceet al., 2014).

Control of the fishery activity is an ongoing alternative toman-

aging reservoirs in Brazil. However, there are also huge practicalandconceptual difficulties to overcome.In general, fishingin reser-voirs hasalready begun in theheterotrophic phase just after filling,when the harvest is high. In this phase, a great number of fisher-men engage in the activity. Thus, the fishermen who traditionallyfished in the river are included among the unemployed peoplewho worked in the construction of the dam and the farmers wholost part of their land to the impoundment and who need a com-plementary source of income for subsistence. With the naturaldecrease in the harvest after the trophic upsurge period, fishstocks do not support the fishery pressure, causing poverty inthe area. This type of fishery is not characterized by initial plan-ning, and control becomes virtually impossible due to the highdemand fora scarce resource.Overfishing is constant andactssyn-

ergistically with other disturbances such as those resulting from


10/11
http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0115http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0115http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0115http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0115http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0115http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0115http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0115http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0115http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0115http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0115http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0115http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0115http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0115http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0115http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0115http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0115http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0115http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0115http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0110http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0110http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0110http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0110http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0110http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0110http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0110http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0110http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0110http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0110http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0110http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0110http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0110http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0110http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0110http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0110http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0110http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0110http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0110http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0110http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0110http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0110http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0110http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0110http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0105http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0105http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0105http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0105http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0105http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0105http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0105http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0105http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0105http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0105http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0105http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0100http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0100http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0100http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0100http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0100http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0100http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0100http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0100http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0100http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0100http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0100http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0100http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0100http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0100http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0100http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0100http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0100http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0100http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0100http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0100http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0100http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0100http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0100http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0095http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0095http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0095http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0095http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0095http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0095http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0095http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0095http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0095http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0095http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0095http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0095http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0095http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0095http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0095http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0095http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0095http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0095http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0095http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0095http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0095http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0095http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0095http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0095http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0090http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0090http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0090http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0090http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0090http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0090http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0090http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0090http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0090http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0090http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0090http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0090http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0090http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0090http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0090http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0090http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0090http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0090http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0090http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0085http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0085http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0085http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0085http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0085http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0085http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0085http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0085http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0085http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0085http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0085http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0085http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0085http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0085http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0085http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0085http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0085http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0085http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0085http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0085http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0085http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0085http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0085http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0085http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0085http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0085http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0085http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0080http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0080http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0080http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0080http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0080http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0080http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0080http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0080http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0080http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0080http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0080http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0080http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0080http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0080http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0080http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0080http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0080http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0080http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0080http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0080http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0080http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0080http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0080http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0080http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0375http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0075http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0075http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0075http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0075http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0075http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0075http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0075http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0075http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0075http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0075http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0075http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0075http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0075http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0075http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0075http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0075http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0075http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0075http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0075http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0075http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0075http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0075http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0075http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0075http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0070http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0070http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0070http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0070http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0070http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0070http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0070http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0070http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0070http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0070http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0070http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0070http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0070http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0070http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0070http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0070http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0070http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0070http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0070http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0070http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0070http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0070http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0065http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0065http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0065http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0065http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0065http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0065http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0065http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0065http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0065http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0065http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0065http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0065http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0065http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0065http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0065http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0060http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0060http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0060http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0060http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0060http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0060http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0060http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0060http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0060http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0060http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0060http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0060http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0060http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0060http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0060http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0060http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0060http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0060http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0055http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0055http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0055http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0055http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0055http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0055http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0055http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0055http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0055http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0055http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0055http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0055http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0055http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0055http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0055http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0055http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0055http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0055http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0050http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0050http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0050http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0050http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0050http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0050http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0050http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0050http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0050http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0050http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0050http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0050http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0050http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0050http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0050http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0050http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0050http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0050http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0050http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0050http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0050http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0050http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0050http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0050http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0045http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0045http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0045http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0045http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0045http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0045http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0045http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0045http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0045http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0045http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0045http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0045http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0045http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0045http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0045http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0040http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0040http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0040http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0040http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0040http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0040http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0040http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0040http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0040http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0040http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0040http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0040http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0040http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0040http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0040http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0035http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0035http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0035http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0035http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0035http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0035http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0035http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0035http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0035http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0035http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0035http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0035http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0035http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0035http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0035http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0035http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0035http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0035http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0035http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0035http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0035http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0030http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0030http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0030http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0030http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0030http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0030http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0030http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0030http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0030http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0030http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0030http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0030http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0025http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0025http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0025http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0025http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0025http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0025http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0025http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0025http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0025http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0025http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0025http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0025http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0025http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0370http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0370http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0370http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0370http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0370http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0370http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0370http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0370http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0370http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0370http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0370http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0370http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0370http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0370http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0370http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0370http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0370http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0370http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0370http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0370http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0370http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0020http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0020http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0020http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0020http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0020http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0020http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0020http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0020http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0020http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0020http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0020http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0020http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0020http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0020http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0015http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0010http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0010http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0010http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0010http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0010http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0010http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0010http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0010http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0010http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0010http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0365http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0005http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0005http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0005http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0005http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0005http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0005http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0005http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0005http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0005http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0005http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0005http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0005http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0005http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0005http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0005http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0005http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0005http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0005http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0005http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0005http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0005http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0005http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0005


11/11
http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0360http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0360http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0360http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0360http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0360http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0360http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0360http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0360http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0360http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0360http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0360http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0360http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0360http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0360http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0360http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0360http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0360http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0360http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0360http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0360http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0360http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0360http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0355http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0355http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0355http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0355http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0355http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0355http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0355http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0355http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0355http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0355http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0355http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0355http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0355http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0355http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0355http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0355http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0355http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0355http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0350http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0345http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0345http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0345http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0345http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0345http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0345http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0345http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0345http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0345http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0345http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0345http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0345http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0345http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0345http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0345http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0345http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0345http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0345http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0345http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0345http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0345http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0345http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0345http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0345http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0345http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0345http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0345http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0340http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0340http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0340http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0340http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0340http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0340http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0340http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0335http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0335http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0335http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0335http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0335http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0335http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0335http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0335http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0335http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0335http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0335http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0335http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0335http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0335http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0335http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0335http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0335http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0335http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0335http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0335http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0335http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0335http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0335http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0330http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0330http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0330http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0330http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0330http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0330http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0330http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0330http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0330http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0330http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0330http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0330http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0330http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0330http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0330http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0330http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0330http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0330http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0330http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0330http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0330http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0330http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0330http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0330http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0330http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0330http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0325http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0325http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0325http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0325http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0325http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0325http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0325http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0325http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0325http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0325http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0325http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0325http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0325http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0325http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0325http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0325http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0325http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0325http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0325http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0325http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0325http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0325http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0325http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0325http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0325http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0320http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0320http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0320http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0320http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0320http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0320http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0320http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0320http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0320http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0320http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0320http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0320http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0320http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0320http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0320http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0320http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0320http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0320http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0320http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0320http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0320http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0320http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0320http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0320http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0320http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0315http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0315http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0315http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0315http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0315http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0315http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0315http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0315http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0315http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0315http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0315http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0315http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0315http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0315http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0315http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0315http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0315http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0315http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0315http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0310http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0310http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0310http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0310http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0310http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0310http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0310http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0310http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0310http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0310http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0310http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0310http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0310http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0310http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0310http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0310http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0310http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0310http://refhub.elsevier.com/S0165-7836(15)00129-0/sbref0310http://r

2016 fish assemblages in neotropical reservoirs colonization patterns, impacts and management

Documents