Non-native minnows cause much larger negative effects than trout on littoral macroinvertebrates of high mountain lakes

Abstract

Despite being naturally fishless, the widespread introduction of trout and minnows is threatening the conservation of high mountain lakes all over the world. Previous studies have reported that amphibians quickly disappear after trout introduction, followed by many conspicuous invertebrates. Here, we have studied the effects of minnows versus trout on the littoral macroinvertebrate community of 54 high mountain lakes from the Pyrenees, covering a gradient of environmental characteristics. The relative importance of fish compared to other variables in explaining macroinvertebrate communities was assessed using distance-based redundancy analysis (dbRDA) and multivariate regression tree (MRT) to find the main environmental thresholds. Both dbRDA and MRT approaches revealed that minnow density was the most important variable negatively determining community structure, followed by aquatic macrophytes, which increased taxa richness. The occurrence and abundance of relevant taxa was analysed in relation to fish densities and other environmental factors using binomial and gamma generalized linear models (GLM). GLM suggested that trout had an impact on the distribution of swimmer taxa and caused declines in the abundance of conspicuous clinger and burrower taxa. Minnows restricted the occurrence of more taxa than trout and negatively affected a wide variety of body sizes and functional groups. Indeed, we found that minnows were responsible for a dramatic biodiversity loss in the littoral macroinvertebrate community. The fast spread of minnows in high mountain areas is of great concern for the conservation of lake macroinvertebrates. Urgent measures to stop minnow introductions are strongly recommended.

Full text

Biological Conservation 272 (2022) 109637
Available online 1 July 2022
0006-3207/© 2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Non-native minnows cause much larger negative effects than trout on
littoral macroinvertebrates of high mountain lakes
Víctor Osorio
a
,
*
, María ´
Angeles Puig
a
, Teresa Buchaca
a
, Ibor Sab´
as
a
, Alexandre Mir´
o
a
,
Federica Lucati
a
,
b
, Jongmo Suh
a
, Quim Pou-Rovira
c
, Marc Ventura
a
a
Integrative Freshwater Ecology Group (IFE), Centre for Advanced Studies of Blanes (CEAB), CSIC, Acc´
es Cala Sant Francesc 14, Blanes 17300, Catalonia, Spain
b
Department of Political and Social Sciences, Universitat Pompeu Fabra (UPF), Ramon Trias Fargas 25-27, Barcelona 08005, Catalonia, Spain
c
Sorell´
o, Estudis al Medi Aqu`
atic, Parc Cientíc de la Universitat de Girona (UdG), Girona 17003, Catalonia, Spain
ARTICLE INFO
Keywords:
Invasive species
Exotic sh
Alpine lakes
Biodiversity loss
ABSTRACT
Despite being naturally shless, the widespread introduction of trout and minnows is threatening the conser-
vation of high mountain lakes all over the world. Previous studies have reported that amphibians quickly
disappear after trout introduction, followed by many conspicuous invertebrates. Here, we have studied the ef-
fects of minnows versus trout on the littoral macroinvertebrate community of 54 high mountain lakes from the
Pyrenees, covering a gradient of environmental characteristics. The relative importance of sh compared to other
variables in explaining macroinvertebrate communities was assessed using distance-based redundancy analysis
(dbRDA) and multivariate regression tree (MRT) to nd the main environmental thresholds. Both dbRDA and
MRT approaches revealed that minnow density was the most important variable negatively determining com-
munity structure, followed by aquatic macrophytes, which increased taxa richness. The occurrence and abun-
dance of relevant taxa was analysed in relation to sh densities and other environmental factors using binomial
and gamma generalized linear models (GLM). GLM suggested that trout had an impact on the distribution of
swimmer taxa and caused declines in the abundance of conspicuous clinger and burrower taxa. Minnows
restricted the occurrence of more taxa than trout and negatively affected a wide variety of body sizes and
functional groups. Indeed, we found that minnows were responsible for a dramatic biodiversity loss in the littoral
macroinvertebrate community. The fast spread of minnows in high mountain areas is of great concern for the
conservation of lake macroinvertebrates. Urgent measures to stop minnow introductions are strongly
recommended.
1. Introduction
Invertebrate declines have been documented worldwide and are
equally or more severe than in vertebrates (Cowie et al., 2022). More-
over, there exist important spatial and taxonomic biases in biodiversity
assessments that are evidenced by the underrepresentation of this ani-
mal group in the IUCN Red List (only 2 % of the 1,305,250 species were
assessed in 2018; Eisenhauer et al., 2019). Particularly, insects have
been reported to suffer large losses in terrestrial and aquatic groups, and
across ecological guilds, which could cascade onto ecosystem func-
tioning and human well-being (S´
anchez-Bayo and Wyckhuys, 2019;
Wagner, 2020).
Among the biomes of the planet, freshwaters have suffered the
strongest decline in biodiversity, as proved by the recent decline in the
number of vertebrate species (81 % in freshwaters compared to 38 % in
terrestrial and 36 % in marine biomes; WWF, 2016). This great decline is
a result, on a global scale, of factors related to habitat destruction,
pollution, water-level reduction and invasive species among others
(Collen et al., 2014). Fish are among the most common animal invaders
threatening freshwaters (Hulme et al., 2009). Nowadays, many high
mountain lakes around the world have suffered from the introduction of
trout and other sh species for angling purposes (Wiley, 2003; Mir´
o and
Ventura, 2013; Ventura et al., 2017). Along with trout, minnows (mainly
Phoxinus genus) have also expanded considerably throughout European
mountains during the 20th century due to their use as live bait by an-
glers (Museth et al., 2007) and remain as the only sh when trout
disappear (Mir´
o and Ventura, 2015). The widespread presence of min-
nows has been documented not only in the north European lakes of
* Corresponding author.
E-mail addresses: vosorio@ceab.csic.es (V. Osorio), ventura@ceab.csic.es (M. Ventura).
Contents lists available at ScienceDirect
Biological Conservation
journal homepage: www.elsevier.com/locate/biocon
https://doi.org/10.1016/j.biocon.2022.109637
Received 24 February 2022; Received in revised form 5 June 2022; Accepted 14 June 2022

Biological Conservation 272 (2022) 109637
2
Scotland and Norway (Maitland and Campbell, 1992; Museth et al.,
2007), but also in the Pyrenees (Mir´
o and Ventura, 2015) and the Alps
(Tiberti et al., 2020).
The introduction of trout in high mountain lakes has been respon-
sible for the decline or elimination of native fauna in many regions,
including amphibians, large crustaceans, and conspicuous littoral mac-
roinvertebrates (Knapp et al., 2001; Tiberti et al., 2014; Mir´
o et al.,
2018; Toro et al., 2020). As a result, the direct predation of more visible
taxa can produce a series of indirect ecological effects (trophic cascade)
affecting the entire ecosystem (Schindler et al., 2001; Eby et al., 2006).
On the other hand, the effects of minnow introduction have scarcely
been studied in comparison with those of trout. Regardless of their
smaller size, minnows share a similar diet with trout, thus becoming the
top predators when established in shless lakes (Hesthagen et al., 1992).
They seem to have a strong impact on both amphibians and zooplankton
communities (Schabetsberger et al., 2006; Mir´
o et al., 2018), and the
evidence from boreal lakes suggest that a negative effect may be ex-
pected on macroinvertebrates, as minnows were found to feed on con-
spicuous taxa (Museth et al., 2010). However, there are no studies
describing their effects on the conservation of the macroinvertebrate
community of high mountain lakes.
In comparison with other freshwater ecosystems, high mountain
lakes have littoral macroinvertebrate communities with relatively low
species richness (Füreder et al., 2006). Their taxa, though, play an
important role in lake food-web stability via their diverse feeding stra-
tegies and the magnied importance of the littoral zone in small lakes
(high perimeter to surface area ratio) (S´
anchez-Hern´
andez et al., 2015).
These communities comprise species well adapted to harsh environ-
mental conditions, which are sensitive to local and global anthropogenic
impacts. Indeed, groups like chironomids and caddisies have been
recognised as good indicators of climatic change in mountain areas
(ˇ
Ciamporov´
a-Zat'oviˇ
cov´
a et al., 2010). The current study aimed to
compare the effects of minnows versus trout introduction on the littoral
macroinvertebrate community of high mountain lakes. Specically, our
objectives were: (i) to determine the relative importance of non-native
sh introductions in explaining the macroinvertebrate community
compared to other environmental factors, (ii) to evaluate how differ-
ently minnow and trout densities affect both macroinvertebrate com-
munity composition and abundance, and (iii) to ascertain whether the
impact of minnows on the conservation of macroinvertebrate commu-
nity is equally important as that of trout.
2. Materials and methods
2.1. Study area
We surveyed 54 high mountain lakes and ponds from the Central-
Eastern Pyrenees (0◦52′E–1◦22′E, 42◦46′–42◦31′N), located within the
area of the Aigüestortes i Estany de Sant Maurici National Park and the
Alt Pirineu Natural Park, which were both designated as Sites of Com-
munity Interest by the European Commission's Natura 2000 conserva-
tion program (Fig. A.1). In addition, high mountain lakes are protected
habitats listed in the European Directive 92/43/EEC for the conserva-
tion of natural habitats and of wild fauna and ora (Council of the Eu-
ropean Communities, 1992). The lakes sampled were relatively small
and shallow water bodies (area range: 0.03–8 ha, median area: 1.9 ha;
maximum depth range: 0.6–51.4 m, median maximum depth: 7.1 m). A
more detailed description of Pyrenean lake characteristics is provided in
the online Appendix A.
We distributed the lakes into three nominal categories according to
the sh species present in them: naturally shless lakes (n =22), lakes
with trout only (n =15) and lakes with minnows (n =17), of which six
had both trout and minnows. The most common trout species found in
the lakes was Salmo trutta (n =14), whereas Salvelinus fontinalis (n =4)
and Oncorhynchus mykiss (n =4) were less frequent. The coexistence of
two trout species, S. trutta and O. mykiss, was observed in only two lakes
(Lakes Llastra and Coveta). Selection of lakes was made to compose a set
of sites as similar as possible to each other in their physical and
geographic characteristics but at the same time, covering a gradient of
small to medium size. As a result, the categories comprised similar
replicates in terms of lake morphology and aquatic habitats availability
for the different sh categories.
2.2. Macroinvertebrate samples
Lakes were sampled once during July–August to make collected data
comparable over the years, with 26 lakes sampled in 2014, 6 in 2015, 11
in 2016, 8 in 2017, and 3 in 2018. The one-time sampling approach is
thought to successfully characterize the summer littoral macro-
invertebrate community, as most mountain species have relatively long
life cycles and larval development periods. As a result, species compo-
sition and richness change little over the summer (Knapp et al., 2001;
Tiberti et al., 2014). Littoral macroinvertebrates were sampled from
representative shore-accessible habitats at ca. 80 cm depth using a d-
frame net with a mesh size of 250
μ
m, employing commonly used
sampling techniques (de Mendoza and Catalan, 2010; Tiberti et al.,
2014). Collected material was pooled in a unique composite semi-
quantitative sample per lake, and preserved in the eld in absolute
ethanol. Sampling methods are extensively described in the Supporting
Information. Macroinvertebrates were counted under the stereo micro-
scope (Olympus SZH, Japan) and determined to the lowest practical
taxonomic unit (typically genus or species), with the exception of
Hydrachnidia, Oligochaeta and Palpomyiini (Ceratopogonidae) speci-
mens, that were counted only. For Chironomidae, the used taxonomic
resolution was tribe or subfamily rank. Abundance per sample was
expressed as total individuals m
−2
preceding any statistical treatment.
2.3. Environmental data
The environmental descriptors measured to understand the factors
shaping littoral macroinvertebrate community included water chemis-
try, morphometric, physical and biological lake variables, as well as
catchment and geographical descriptors.
Water chemistry samples were taken from the lakes' outlet, or sub-
supercially at the deepest point from an inatable boat when there
was no active outlet. Laboratory analyses of pH, conductivity, alkalinity,
total nitrogen, dissolved inorganic nitrogen and total phosphorus were
performed following the procedures described by Ventura et al. (2000).
Water samples for chlorophyll-a (Chl-a) were taken from the deepest
point of the lake at a depth of 1.5-times the Secchi depth, using a
UWITEC-type sampling equipment. In those lakes where the Secchi disk
bottomed out the samples were taken between 1 and 2 m above the
sediment. From the collected samples, a known volume of water (be-
tween 1.5 and 2 L) was ltered using a manual vacuum pump and GF/F
lters (47 mm in diameter). The lters were wrapped in aluminium foil
and kept cold before freezing (−20 ◦C) within 3–6 h. Chlorophyll-a was
determined using the spectrophotometric method of Jeffrey and Hum-
phrey (1975).
Fish density in lakes was measured by the installation of multi-mesh
gill nets (CEN 14757:2015). The number of nets placed in each lake was
decided depending on its area (from 2 to 6 nets in a lake between 0.4 and
6 ha, distributed among different depths). They were kept in the lakes
ca. 20–24 h, before counting and weighting all sh captured. The
catches of trout (Salmonidae family) and minnows (Phoxinus sp.) in gill
nets were used to produce the standardized values of catch per unit
effort (CPUE) and biomass per unit effort (BPUE). CPUE was expressed
as individuals net
−1
day
−1
for trout and minnows, whereas BPUE was
expressed as kg net
−1
day
−1
. The density of minnows expressed as in-
dividuals net
−1
day
−1
can be roughly converted to individuals trap
−1
day
−1
using the regression equation Y =0.77X +39.22 (R
2
=0.32),
which resulted from tting a linear model on data available from 18
lakes with minnows (data not shown).
V. Osorio et al.

Biological Conservation 272 (2022) 109637
3
Due to unavailability of data, CPUE and BPUE estimates of three
lakes (namely Bassa Cloto de Baix-Minnow, Estanyet 4 Gargolhes
Inferior-Trout, and L`
ossa Mig-Minnow/Trout) were obtained from other
similar lakes or averaging over similar lakes, according to expert
criteria. Bassa Cloto de Baix values were obtained by averaging over the
rest of lakes with minnows. The values of Estanyet 4 Gargolhes Inferior
came from Gargolhes Inferior, while the ones of L`
ossa Mig came from
Xic, as both pairs of lakes share similar size and habitat diversity.
Water transparency was measured by estimating the light extinction
coefcient in the water column (Kd; m
−1
) from the Secchi disk depth
measurement. In lakes where the Secchi disk reached the bottom, a
constant extinction coefcient of 0.2 m
−1
was used (Buchaca, 2005). It
was not possible to measure Secchi disk depth at Lake L`
ossa Mig, where
a Kd value estimated from the similar Lake Obago d'Amunt was used.
Lake maximum depth, littoral substrate composition, littoral macro-
phytes percentage, and percentage of lithological and vegetation units in
the catchment were measured in the eld. The abundance of amphibians
along the shoreline (ind. m
−1
) was estimated by visual encounter sur-
veys (Mir´
o et al., 2018). Additionally, the density of animal manure was
counted in a 20 m wide area along the shoreline, obtaining a rough
estimate of local livestock pressure (excrements m
−1
). Additional
morphometric parameters (altitude, lake area, total catchment area) and
geographical coordinates (easting and northing) were determined using
cartographic information.
2.4. Statistical methods
We used an exploratory non-parametric Kruskal-Wallis test and post-
hoc tests with Bonferroni correction (level of signicance was
α
=0.017)
to examine whether the richness of macroinvertebrates taxa differed
among the three categories, i.e. shless, trout, and minnow lakes. Lakes
with both trout and minnows were included within minnow category
since minnows are always present in the littoral zone either with or
without trout. The same approach was used to compare for alpha, beta,
and gamma diversity measures in the samples, which were calculated on
rareed subsets of lakes (30 permutations of nine randomly chosen lakes
at a time) and employing the adaptation of common Shannon entropy
index suggested by Jost (2007).
To shed light on the main patterns of variation in the community, we
employed distance-based redundancy analysis (dbRDA) ordination
(Legendre and Anderson, 1999). The response matrix consisted of Bray-
Curtis distances on log-transformed taxa abundance. As a preliminary
step, the explanatory variables that could potentially inuence the
community were classied into four groups, describing: i) spatial vari-
ation (space-dbRDA), which comprised 22 Moran's eigenvector maps
(MEMs) with signicant positive spatial autocorrelation and generated
employing relative neighbourhood graph connectivity (Peres-Neto et al.,
2006); ii) catchment properties (catchment-dbRDA), including 10 de-
scriptors of catchment features, in addition to lake altitude and lake
geomorphology, iii) in-lake habitat (inlake-dbRDA), consisting of 12
variables describing water physicochemical condition, littoral substrate,
and littoral macrophytes dominance, and; iv) vertebrate interaction
(vertebrate-dbRDA), composed by 6 variables describing predator and
competitive interaction with vertebrates, among which CPUE and BPUE
of trout and minnows, as well as tadpoles and newts abundance.
Using each group of explanatory variables, we rstly produced four
dbRDA models, namely space-dbRDA, catchment-dbRDA, inlake-
dbRDA, and vertebrate-dbRDA. All variables were normalised and
standardised before entering the ordination algorithm. Searching for
parsimony and preventing the problem of ination of the overall type I
error, we ran forward selection with the Blanchet et al. (2008) double
stopping criterion on the above-mentioned models (P <0.05, 9999
randomizations). The variables selected were incorporated into the full
dbRDA model (full-dbRDA), that was examined using variation parti-
tioning analysis by means of partial dbRDA to estimate pure and shared
effects of groups of variables by computing adjusted R
2
. Ultimately, a
forward selection applied on the full-dbRDA produced the parsimonious
dbRDA (pars-dbRDA), which was explored for linear dependencies
employing variance ination factors (VIF).
To highlight local structures in macroinvertebrate community, we
used a multivariate regression tree (MRT) on log-transformed taxa
abundance (Bray-Curtis distance, 1000 multiple cross-validations)
(De'Ath, 2002). Explanatory variables in the pars-dbRDA were passed
to MRT as candidates to be explored. The number of clusters included in
the tree were dened by the solution that minimized the cross-validated
relative error (CVRE). We subsequently determined taxa signicantly
associated to groups of sites or combinations of those by means of the
indicator species analysis suggested by De C´
aceres et al. (2010).
We investigated the impact of sh density (minnows: Minnow_CPUE;
trout: Trout_CPUE) on invertebrate taxa representative of different body
sizes and functional groups according to their propensity and frequency
of movement within the habitat (Table A.1) by generalised linear models
(GLM). To control for additional sources of variability, we also consid-
ered other biologically relevant variables identied in pars-dbRDA, such
as littoral macrophyte cover, altitude and lake area. The probability that
a taxon was found in the samples was assessed by a binomial GLM (logit
link function) on presence-absence data. Subsequently, the non-zero
observations from the dataset were modelled using a zero truncated
gamma GLM (log link), assessing the relevant variables driving its
abundance when the taxon was present in lakes. Those taxa with low
occurrences were aggregated to a higher taxonomic rank (always within
the same functional group; Table A.1) to ensure sufcient cases. If a
taxon was found in >48 of the 54 lakes (e.g. ubiquitous groups such as
Oligochaeta and some Chironomidae taxa), only the gamma GLM was
tted, since there was not enough variation in the response variable to
run the binomial model. All GLMs were initially tted including the
following predictors and interaction terms: Minnow_CPUE, Trout_CPUE,
macrophyte cover, altitude, area, Minnow_CPUE ×macrophyte cover,
Trout_CPUE ×macrophyte cover, Minnow_CPUE ×Trout_CPUE.
Macrophyte cover was previously transformed using the inverse-sine
function, and the square root transformation was applied on Trout_C-
PUE and Minnow_CPUE. A backward stepwise selection based on AIC
(Akaike's Information Criterion) was used to nd the best subset of terms
for each taxon and GLM model separately. We determined the relative
weight of terms by dropping them from GLM models and evaluating the
resultant change in deviance.
Statistical analyses were conducted within the R 4.1.0 statistical
environment (R Core Team, 2021). R packages used are detailed in the
Supporting Information. The level of signicance adopted for all ana-
lyses was
α
=0.05.
3. Results
3.1. Taxa diversity
A total of 77 macroinvertebrate taxa were found in 54 lakes. The
order Diptera was the most diverse and abundant group (64 % of mac-
roinvertebrate abundance), followed by Bivalvia, Ephemeroptera,
Megaloptera, Hemiptera, and Coleoptera taxa. Oligochaeta were largely
dominated by Tubicidae taxa. The list of all identied taxa and their
frequency of occurrence in the sampled lakes are presented in Table A.1
of the Supporting Information.
There were signicant differences in taxa richness among lake cat-
egories (
χ
2
=11.2, P <0.005). The richness of shless lakes (16.3 ±2.5)
was signicantly different from that of minnow lakes (11.7 ±5.3).
However, trout lakes (14.7 ±3.9) did not signicantly differ from the
other categories. Signicant differences arose among lake categories
regarding alpha, beta, and gamma diversities (Fig. 1). Fishless lakes had
signicantly higher values of the three indices, whereas minnow lakes
showed the lowest values of alpha and gamma diversity.
V. Osorio et al.

Biological Conservation 272 (2022) 109637
4
3.2. Overall community analysis
Five MEM variables (MEM-3, 13, 14, 19, 20) signicantly described
the spatial autocorrelation in the community (Table 1). The catchment-
related variables best explaining the community were the percentage of
meadows cover, the ratio total catchment/lake area (a surrogate of
water renewal time), lake area, and lake altitude. Among in-lake vari-
ables, selected variables were littoral macrophyte cover, algal biomass
(chlorophyll-a), dissolved inorganic nitrogen, total phosphorus and
water conductivity. The most relevant vertebrate-related variables were
the measures of catch per unit effort of minnows and trout, so the group
was henceforth called sh-related variables.
The full-dbRDA was composed by the 16 variables listed above,
explaining 54.3 % of macroinvertebrate variance (adjusted R
2
=0.38).
The variation partitioning analysis revealed that sh-related variables
had the highest fraction of pure effects, doubling the variation uniquely
attributed to spatial and catchment inuence, followed by in-lake vari-
ables (Table 1). In-lake variables showed the highest shared variance,
most of it shared with sh variables (Fig. A.3 in Supporting Informa-
tion). Most of the explanatory power of MEM variables (70.7 %) was
shared with variables of other nature. The forward selection algorithm
reduced the number of variables from the full-dbRDA to 10, which
comprised the pars-dbRDA: Minnow_CPUE (33.9 % of exclusive vari-
ance), total catchment/lake area (14.5 %), lake area (7.9 %), macro-
phyte cover (6.9 %), total phosphorus (6.6 %), altitude (6.5 %),
conductivity (6.2 %), chlorophyll-a (6.0 %), MEM20 (5.8 %), and
Trout_CPUE (5.7 %) (Fig. 2a). Summary statistics of all but MEM vari-
ables are shown in Table A.2 in the Supporting Information. This model
signicantly accounted for 44.3 % of macroinvertebrate variance
(pseudo-F =3.4, P =0.001), with the rst three axes being signicant
(dbRDA1: pseudo-F =14.1, dbRDA2: pseudo-F =6.8, dbRDA3: pseudo-
F =4.5; P =0.001, for all three). The variables Minnow_CPUE, Chlo-
rophyll-a, and MEM20 were negatively correlated with the rst axis (r =
−0.7, −0.3, and −0.3, respectively). The second axis was negatively
correlated with macrophyte cover, total phosphorus and chlorophyll-a
(r = − 0.6, −0.6 and −0.4, respectively), but positively correlated with
lake area (r =0.4) and altitude (r =0.3). The third axis accounted for
13.1 % of the constrained variance, with total catchment/lake area and
Trout_CPUE as negatively correlated variables (r = − 0.6 and −0.4,
respectively), and altitude positively correlated (r =0.4).
A tree with four ‘leaves’ resulted from the MRT analysis (Fig. 2b).
Minnow density was responsible for the rst split, with a threshold of
27.6 individuals net
−1
day
−1
. Macrophyte cover created two successive
splitting nodes with similar threshold values (19 % and 18.5 % of cover).
We found four taxa indicating signicant association with lakes in leaves
A, B, and C (Megaloptera, Bivalvia, and Chironomidae taxa). Five in-
dicator taxa were shared between the sites with highest abundance of
macrophytes (i.e. leaves B and C), including Hydrachnidia and Odonata
taxa. Five taxa were exclusively associated with leaf B (Corixidae, Hir-
udinea, Odonata), and two with leaf C (Diptera and Coleoptera taxa). No
taxon was signicantly associated with leaf D.
3.3. Factors explaining macroinvertebrates occurrence
A total of 14 taxa across different body sizes and functional groups
were analysed by GLM. For nine of them, minnows had a signicantly
negative effect on their probability of occurrence, whereas four taxa
were affected by trout (Fig. A.4). Exceptions were Gastropoda,
Fig. 1. Comparative boxplots of alpha (a), beta (b) and gamma (c) diversity indices among lake categories (Kruskal-Wallis
χ
2
=69.6,
χ
2
=16.8, and
χ
2
=71.8,
respectively, and P <0.001 each). Diversity measures were calculated on rareed subsets of lakes (30 permutations of nine randomly chosen lakes at a time) and
employing the adaptation of common Shannon entropy index suggested by Jost (2007). Asterisks indicate the level of statistical signicance of post-hoc tests with
Bonferroni correction: ***P ≤0.001; ns, not signicant (P >0.05).
V. Osorio et al.

Biological Conservation 272 (2022) 109637
5
Hirudinea, Hydrachnidia, and Odonata, which appeared unaffected by
sh. The density of both minnows and trout was found to decrease the
probability of occurrence of the coleopteran genus Hydroporus along
with Dytiscidae and Corixidae families. The dytiscid genus Agabus was
not found in any lake with sh presence, and both Boreonectes (Dytis-
cidae) and Arctocorisa (Corixidae) genera were absent in lakes with
minnows. The occurrence of Sialidae (Megaloptera) was negatively
affected by minnows. Minnow density was the only signicant and
negative variable in the models of Baetidae (Ephemeroptera), Sphaer-
iidae (Bivalvia), Polycentropodidae (Trichoptera), Macropelopiini, and
Procladiini (both Chironomidae). Trout density was a negative predictor
for Limnephilidae (Trichoptera).
Macrophyte cover increased the likelihood of occurrence of Odonata,
Hirudinea, and Gastropoda, but had a negative effect on Limnephilidae.
We found a positive interaction between minnows and macrophytes for
Sialidae and Limnephilidae (the impact of minnow density on their
probability of occurrence was less severe in lakes with high macrophyte
cover) and between trout and macrophytes for Dytiscidae. See
Tables A.3–A.21 for detailed GLM results and Fig. A.4 for visualization
of taxa response (Supporting information).
For those macroinvertebrates signicantly affected by sh, we
compared the estimated sh density at which their probability of
occurrence was decreased to 0.5 (Fig. 3). The relative sensitivity to
minnow density was variable among taxa, but organisms differing in
body size and functional group were impacted at low minnow densities,
such as large and mid-sized clingers, large swimmers, and large and mid-
sized burrowers. Although signicantly affected, some burrower taxa
like Sphaeriidae and Macropelopiini were particularly threatened at
higher minnow densities. By contrast, trout density was mainly
responsible for the impact on large swimmers and clingers.
3.4. Factors explaining macroinvertebrates abundance
The factors controlling the abundance of macroinvertebrates were
assessed for 19 taxa (including ve groups for which the probability of
occurrence was not studied due to their ubiquitousness). Nine taxa were
negatively affected by minnows, including large clingers, along with
large, mid-sized, and small burrowers. Four taxa were negatively
affected by trout, mainly large and mid-sized clingers, but also large and
small burrowers (see Fig. 4 for relevant taxa). A negative relationship
with minnows was found for Sialidae, Macropelopiini, Orthocladiinae
(Chironomidae), Limnephilidae, Procladiini, non-Tanytarsini Chirono-
minae (Chironomidae), Polycentropodidae, Hydrachnidia, and Tany-
tarsini (Chironomidae). The impact of minnows on Oligochaeta was
marginally signicant. Oligochaeta, Sialidae, Baetidae and Limnephili-
dae displayed declines in abundance related to trout. Nevertheless, trout
density had a positive effect on the abundance of the chironomids
Orthocladiinae and Pentaneurini.
Macrophyte cover was a positive predictor of the abundance of
Corixidae, non-Tanytarsini Chironominae, Pentaneurini, Tanytarsini,
Hydrachnidia, Procladiini, Gastropoda, Baetidae, Sphaeriidae, and Oli-
gochaeta. Signicant effects of lake altitude were detected for Corixidae
and Tanytarsini, which were more abundant at higher altitudes. The
abundance of Macropelopiini and Hydrachnidia increased with lake
area, whereas Baetidae showed a negative relationship with that vari-
able. Negative interactions were found between minnows and trout for
Pentaneurini (the impact of minnow density on its abundance was more
severe in lakes with low trout density), and between trout and macro-
phyte cover for Procladiini (the impact of trout density on its abundance
was more severe in lakes with low macrophyte cover). For further details
on GLM results, see Tables A.3–A.21 in the Supporting Information.
4. Discussion
In this study, non-native minnows and trout were shown to play a
strong role on shaping the littoral macroinvertebrate communities in
high mountain lakes. Variation partitioning analysis indicated that the
variation uniquely attributable to the density of sh was larger than the
one explained by either spatial autocorrelation, catchment features, or
in-lake habitat descriptors. Altitude-related and habitat variables have
been usually dened as the most important factors affecting the mac-
roinvertebrate distribution in high mountain lakes (Fjellheim et al.,
2000; de Mendoza and Catalan, 2010; de Mendoza et al., 2015). Yet, in
agreement with our results, the presence of predatory sh can also
control the occurrence and abundance of many macroinvertebrate taxa
and even surpass environmental and spatial factors in their inuence on
the community (Schilling et al., 2009; de Mendoza et al., 2012).
Trout have been described to threaten a variety of animal groups in
high mountain lakes (e.g. Knapp et al., 2001; Tiberti et al., 2014; Mir´
o
et al., 2018). The effects of the presence of introduced trout on macro-
invertebrate communities of naturally shless lakes have been described
previously across geographically disparate areas (Knapp et al., 2001;
Tiberti et al., 2014; Toro et al., 2020). One contribution of this study, has
been to take into account the effects of trout density (Figs. 3, 4 and A.4).
In accordance with Knapp et al. (2001) and Tiberti et al. (2014), we
found that trout caused a decline in the abundance of conspicuous Tri-
choptera, Ephemeroptera, Coleoptera, Hemiptera and Oligochaeta taxa.
Conversely, Chironomidae groups such as Orthocladiinae and Penta-
neurini beneted from higher trout densities. Evidence of enhanced
emergence of chironomids caused by trout is reported in the literature
(Tiberti et al., 2016). It seems very likely that the selective predation of
trout on large-bodied and predator taxa is involved in the higher
abundances of certain chironomids from interspecies competition and
predation within the littoral invertebrate community (Mousavi et al.,
2002), but other factors like an increase in nutrient concentration could
also be involved (Leavitt et al., 1994).
Our results suggest that minnows have a much larger negative effect
Table 1
Forward selection of variables in distance-based redundancy analysis (dbRDA)
for spatial, catchment, in-lake and vertebrate variables explaining littoral mac-
roinvertebrate community (P <0.05, 9999 randomizations). Adjusted R
2
of
variables selected from each model to be incorporated in full-dbRDA is indicated
in parentheses. Variation partitioning among groups of variables in full-dbRDA
(total adj R
2
=0.384) is displayed below. Abbreviations stand for: TC/LA =total
catchment area/lake area, Alt =altitude, Mac_cov =percentage of littoral
macrophyte cover, Chla =chlorophyll-a, TP =total phosphorus concentration,
DIN =dissolved inorganic nitrogen concentration, Cond =water conductivity,
Minnow_CPUE =CPUE of minnows, Trout_CPUE =CPUE of trout.
Spatial
variables
(space-
dbRDA)
Catchment
variables
(catchment-
dbRDA)
In-lake
variables
(inlake-
dbRDA)
Vertebrate
variables
(vertebrate-
dbRDA)
Selected variables (adj R
2
)
MEM-3
(0.037**)
Meadow
(0.051**)
Mac_cov
(0.093***)
Minnow_CPUE
(0.071***)
MEM-20
(0.03*)
TC/LA
(0.028*)
Chla
(0.032**)
Trout_CPUE
(0.04**)
MEM-14
(0.03**)
Alt (0.028*) TP (0.028**)
MEM-13
(0.023*)
Area (0.026*) DIN (0.023*)
MEM-19
(0.02*)
Cond
(0.021*)
Variation partitioning on full-dbRDA
Total (adj
R
2
)
0.140 0.106 0.176 0.157
Pure (adj
R
2
)
0.041 0.041 0.064 0.080
Shared
(adj R
2
)
0.099 0.065 0.112 0.077
Asterisks indicate the level of statistical signicance associated with each vari-
able: *P ≤0.05; **P ≤0.01 and ***P ≤0.001.
V. Osorio et al.

Biological Conservation 272 (2022) 109637
6
(a)
(b)
Minnow
CPUE ≥ 27.6
(Split 1, 27.2%)
(Split 2, 7.5%) (Split 3, 11.%)
Residual: 0.543
CVRE: 0.879
SE: 0.13
Sigara (0.81)
Arctocorisa (0.79)
Erpobdella (0.73)
Enallagma (0.66)
Pyrrhosoma (0.58)
Hydrachnidia (0.83)
Aeshna (0.73)
Gerris (0.66)
Chrysops (0.55)
Culicoides (0.48)
Cloeon (0.75)
Pisidium (0.95)
Sialis (0.93)
Macropelopiini (0.89)
Procladiini (0.83)
Palpomyiini (0.82)
Donacia (0.48)
Macropelopiini (0.89)
Procladiini (0.83)
n = 4 n = 9
n = 9n = 32
(A) (B)
(C) (D)
Cloeon (0.75)
Pisidium (0.95)
Sialis (0.93)
Pisidium (0.95)
Sialis (0.93)
Macropelopiini (0.89)
Procladiini (0.83)
Aeshna (0.73)
Gerris (0.66)
Chrysops (0.55)
Culicoides (0.48)
Hydrachnidia (0.83)
< 18.5%≥ 18.5%
≥ 19%< 19%
Macrophyte
cover
Macrophyte
cover Macrophyte
cover
Macrophyte
cover
< 27.6. Minnow
CPUE
−2.0 −1.5 −1.0 −0.5 0.0 0.5 1.0
−1.5
−1.0
−0.5
0.0
0.5
1.0
1.5
db−RDA1 41.4%
db−RDA2 19.9%
TP
Chla
MEM-20 TC/LA
Cond
Area
Alt
Trout_CPUE
Mac_cov
Minnow_CPUE
TP
TC/LA
MEM-20
Chla
Trout_CPUE
Cond
Alt
Area
Mac_cov
Fishless
Trout
Minnow
Minnow_CPUE
Fig. 2. (a) Ordination diagram showing the linear combination of variables describing the main gradients of variation in the littoral macroinvertebrate community
(Bray-Curtis distances on log-transformed taxa abundance) along the rst two axes of pars-dbRDA. A principal component analysis (PCA) of variables in pars-dbRDA
is included as a subplot. (b) Multivariate regression tree (MRT) run on macroinvertebrate dataset (Bray-Curtis distances on log-transformed taxa abundance, 1000
multiple cross-validations) and using explanatory variables in pars-dbRDA as candidates to be explored. Percentage of constrained variance related to each split of the
tree is displayed in parentheses. Taxa signicantly associated with leaves are included along with their indicator value. Minnow CPUE is expressed as individuals
net
−1
day
−1
. See Table 1 for a description of variable abbreviations.
V. Osorio et al.

Biological Conservation 272 (2022) 109637
7
on the littoral macroinvertebrate community than trout. Both dbRDA
and MRT approaches revealed that minnow density was by far the most
important single variable determining community structure, with a
threshold of 27.6 individuals net
−1
day
−1
(that corresponds to approx-
imately 60.5 individuals trap
−1
day
−1
) beyond which the community
was affected signicantly. Only scarce literature exists on the role of
minnows in high mountain lakes and their effects on littoral macro-
invertebrates have remained so far unexplored. Previous studies in
subalpine lakes revealed that minnow diet included Cladocera and large
taxa, such as Gammaridae, Ephemeroptera, Plecoptera, and Trichoptera
(Hesthagen et al., 1992; Museth et al., 2010). Yet, disagreeing results
describing an increasing abundance of Ephemeroptera, Plecoptera and
Trichoptera taxa in lakes with minnows have also been reported (Næstad
and Brittain, 2010).
Macroinvertebrate communities in lakes with minnows were
composed of a very limited number of taxa, and dominated by small
burrower organisms with semi-sessile habits. Chironomidae and Oli-
gochaeta showed a substantial increase in their relative abundance in
lakes with minnows (92 % of individuals counted in the samples on
average), but simply as a result of the absence of other potential com-
munity members. Indeed, minnows restricted the occurrence of more
taxa than trout, and negatively affected a wide variety of functional
groups (i.e. swimmers, clingers, and burrowers). In contrast to trout,
high minnow densities did not benet any taxon, but rather caused
population declines of Trichoptera and inconspicuous taxa, like
Hydrachnidia and most of Chironomidae groups. Only Corixidae and
Dytiscidae were both affected by minnows and trout, as these large and
mobile insects are generally very vulnerable to visual predators (Tate
and Hershey, 2003). The high densities in which minnows often occur
may certainly be one of the reasons to their exacerbated impact on the
community. The shore-based shoals of minnows perform a faster
location of patchy food than solitary trout and their efciency increases
with shoal size (Pitcher et al., 1982). Moreover, when living in sympatry
with trout, minnows are conned to the shallow parts of the lake to
avoid trout predation. Under these circumstances, minnows that would
be otherwise consuming pelagic invertebrates may shift towards a
heavier predation on littoral benthic fauna (Museth et al., 2010). Trout
has a less exible feeding strategy than minnows and is clearly inu-
enced by the size frequency distribution of potential prey (S´
anchez-
Hern´
andez and Amundsen, 2015).
Aquatic macrophyte cover was the only signicant factor explaining
the occurrence of groups like Hirudinea, Gastropoda, and Odonata. In
fact, irrespective of minnow and trout density, we found higher taxa
richness in lakes with more abundant macrophytes (Fig. A.5 of the
Supporting Information). Aquatic vegetation is of great importance in
providing food resources as well as increased habitat heterogeneity to
macroinvertebrates (Carlisle and Hawkins, 1998). As a source of both
competitive and predatory refuges, we observed that higher values of
macrophyte cover increased the likelihood of occurrence and the
abundance of taxa vulnerable to sh predation. Even inconspicuous taxa
such as Oligochaeta, Sphaeriidae, and most chironomids, were favoured
by submerged vegetation. On the other hand, Limnephilidae taxa were
more common in lakes with low macrophyte dominance. This fact can
probably be attributed to the higher availability of mineral grains
required by dominant species of Limnephilidae (e.g. Annitella pyrenaea)
to build their larval cases in lakes with low macrophyte dominance.
The results of this study show that non-native minnows pose a very
serious threat for the conservation of littoral macroinvertebrates, being
responsible for a dramatic biodiversity loss. Other studies have shown
that conservation implications of minnow introduction extend far
beyond the macroinvertebrate communities, as so far the survival of
macrophyte populations and amphibians is also compromised by
Polycentropodidae
(L,C)
Baetidae
(M,C)
Hydroporus
(M,Sw, Dytiscidae)
Procladiini
(M,B, Chironomidae)
Corixidae
(L,Sw)
Dytiscidae
(L,Sw)
Sialidae
(L,B)
Sphaeriidae
(M,B)
Macropelopiini
(M,B, Chironomidae)
020406080 0 5101520
Limnephilidae
(L,C)
Minnow CPUE
(individuals net−1 day−1)
Trout CPUE
(individuals net−1 day−1)
1.7
1.6
2.5
4.0
6.7
6.9
7.9
39.8
75.3
3.8
5.7
8.2
13.0
Dytiscidae
(L,Sw)
Hydroporus
(M,Sw, Dytiscidae)
Corixidae
(L,Sw)
Fig. 3. Estimated sh density at which macroinvertebrate probability of occurrence was equal to 0.5. Only macroinvertebrates signicantly affected by either
minnows or trout are presented. Fish density estimations and their standard errors are displayed (see Supporting Information for statistical procedures). The
threshold of minnow CPUE suggested by MRT analysis (Fig. 2b) is indicated with a thick vertical line. The body size and functional group of taxa are codied as: L =
large, M =mid-sized, S =small, Sw =swimmer, C =clinger, and B =burrower.
V. Osorio et al.

Biological Conservation 272 (2022) 109637
8
minnows. Gacia et al. (2018) described the massive uprooting of Pyr-
enean quillwort meadows related to minnow presence. The predation on
primary consumers, together with the release of nutrients and increased
turbidity due to the benthic scavenging activity of shoals, were proposed
as the underlying causes. Amphibian declines are the rst consequence
of trout introduction (Knapp et al., 2001; Mir´
o et al., 2018), but minnow
presence is also equally unfavourable for amphibians as is trout. Am-
phibians disappear from lakes when trout are introduced, but do not
return when trout disappear and minnows remain as the only sh spe-
cies present in the lake (Mir´
o et al., 2018). Mir´
o and Ventura (2015) did
not nd any lake where minnows disappeared once established in the
Pyrenees, whereas pre-existing trout populations have disappeared after
minnow introductions in twenty lakes and ponds. Some of these lakes
had brown trout introduced centuries ago while others were stocked
recently. In addition, Tiberti et al. (2022) found that there is a strong
negative relationship between trout and minnows, indicating that
minnows have a direct role in the decrease of trout abundance. This
suggests that minnows are a bigger conservation threat due to their
higher resistance to adverse environmental conditions and their greater
impact on littoral invertebrates than trout.
The conservation of invertebrates is at least as threatened as that of
vertebrates worldwide (S´
anchez-Bayo and Wyckhuys, 2019).
2200 2300 2400 2500
0
10
20
30
40
50
60
70
Altitude (m a.s.l.)
Corixidae abundance
(indv. m-2)
Macrophytes (% littoral cover)
0
10
20
30
40
0306090
D = 20.9%D = 96.1%
Minnow CPUE
0
2
4
6
8
10
0 50 100
0
2
4
6
8
10
0 5 10 15
D = 62.6% D = 34.1%
Limnephilidae abundance
(indv. m-2)
Trout CPUE
90
0
20
40
60
80
Macrophytes (% littoral cover)
03060
Baetidae abundance
(indv. m-2)
D = 23.2%
0
2
4
6
8
10
12
Trout CPUE
02040
D = 8.2%
0123456
0
5
10
15
20
Area (ha)
D = 16.9%
Sialidae abundance
(indv. m-2)
Large swimmer Large clinger
Mid-sized clinger
(indv. m-2)
Macrophytes (% littoral cover)
D = 42.4%
Mid-sized burrower
0
200
400
600
800
Trout CPUE
0204060
0
5
10
15
20
Minnow CPUE
0 50 100 200
D = 31.7 %D = 36.0 %
Orthocladiinae abundance
(indv. m-2)
Macropelopiini abundance
(indv. m-2)
0
1
2
3
4
5
0 50 100
D = 79.5%
Minnow CPUE
02468
5
10
15
20
25
Area (ha)
D = 19.2%
Mid-sized burrower Small burrower
0
1
2
3
4
5
6
7
02040
Large burrower
0
1
2
3
4
5
6
7
0 50 100
Minnow CPUE
D = 26.3%
Trout CPUE
D = 73.6%
0
20
40
60
80
100
03060
Macrophytes (% littoral cover)
Large crawler
Gastropoda abundance
(indv. m-2)
D = 81.6 %
Sphaeriidae abundance
0
50
100
150
200
03060
Fig. 4. Estimated effect of signicant variables (P <0.05) on the abundance of relevant macroinvertebrate taxa obtained by gamma GLM. The contour of the shaded
areas corresponds to ±2 SE (95 % CI) relative to the main estimate, and hatch marks at the top and the bottom are descriptors of the frequency of data points along
the gradient of continuous variables (above and below the main estimate, respectively). The increase in deviance (“D”, equivalent to variance for this type of an-
alyses) resulting from dropping the selected variable from the model is indicated. Minnow and trout CPUE are expressed as individuals net
−1
day
−1.
V. Osorio et al.

Biological Conservation 272 (2022) 109637
9
Invertebrate groups such as Coleoptera and Bivalvia, which were
heavily impacted by minnows in this study, include very sensitive and
endangered species that rely on the integrity of these habitats for their
survival (Council of the European Communities, 1992; Verdú et al.,
2011). The recovery of high mountain macroinvertebrate communities
has been shown to be possible after trout eradication (Knapp et al., 2001;
Tiberti et al., 2019). Although hard to accomplish, restoration of lakes
invaded by minnows may be an effective conservation measure for
impacted macroinvertebrate communities, as already proved for pro-
tected amphibians (Mir´
o et al., 2020).
4.1. Conclusions
We found that the negative effects of non-native minnow introduc-
tion prevailed over environmental and spatial factors in determining the
littoral macroinvertebrate community of high mountain lakes. Minnows
had a much larger negative effect on macroinvertebrates than trout and
were linked to the elimination of many taxa from the studied ecosys-
tems. Our results demonstrate substantial local effects at the lake scale,
but uncontrolled expansion of minnows could also lead to regional and
biogeographic changes in native aquatic biodiversity. As sh introduc-
tion makes water bodies unsuitable for the survival of vulnerable taxa,
fragmentation and isolation of their populations is expected to occur. We
conclude that given the ongoing and fast spread of non-native minnows
in mountain ranges worldwide, there is an urgent need for incorporating
restrictive criteria about trout stocking and the abolition of the use of
minnows as live bait for angling. In addition, we suggest to implement
restoration measures to remove non-native sh, especially minnows in
early stages of introduction. In those lakes where minnows are already
introduced and where conservation is not the main goal, we also
recommend to establish more sustainable shing methods such as sh-
ing without death. This should prevent the disappearance of trout, and
therefore a larger deterioration of the lake.
Funding
Economic support was provided by the European Commission LIFE
+LIMNOPIRINEUS (LIFE13 NAT/ES/001210), LIFE RESQUE ALPYR
(LIFE20 NAT/ES/00347), BiodivERsA FISHME (BiodivRestor-280) and
FUNBIO (RTI2018-096217-B-I00) funded by MCIN/AEI/10.13039/
501100011033 and BIOOCULT (2413/2017) funded by MTERD/OAPN
and by “ERDF A way of making Europe”.
CRediT authorship contribution statement
Víctor Osorio: Conceptualization, Methodology, Formal analysis,
Investigation, Data curation, Writing – original draft, Writing – review &
editing, Visualization. María ´
Angeles Puig: Conceptualization, Inves-
tigation, Writing – review & editing, Supervision, Funding acquisition.
Teresa Buchaca: Conceptualization, Investigation, Data curation,
Writing – review & editing, Funding acquisition. Ibor Sab´
as: Investi-
gation, Data curation, Writing – review & editing. Alexandre Mir´
o:
Investigation, Data curation, Writing – review & editing. Federica
Lucati: Investigation, Data curation, Writing – review & editing.
Jongmo Suh: Investigation, Data curation, Writing – review & editing.
Quim Pou-Rovira: Conceptualization, Investigation, Data curation,
Writing – review & editing, Funding acquisition. Marc Ventura:
Conceptualization, Investigation, Writing – review & editing, Supervi-
sion, Project administration, Funding acquisition.
Declaration of competing interest
The authors declare that they have no known competing nancial
interests or personal relationships that could have appeared to inuence
the work reported in this paper.
Acknowledgements
We would like to acknowledge the eld and shing teams and all the
people who helped us during the eld work, including Berta P´
erez and
Eva Docampo. Also, we want to sincerely thank the valuable assistance
given in the eld work by the staff of the Pyrenean protected areas of
Aigüestortes i Estany de Sant Maurici National Park and Alt Pirineu
Natural Park.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.biocon.2022.109637.
References
Blanchet, F.G., Legendre, P., Borcard, D., 2008. Forward selection of explanatory
variables. Ecology 89, 2623–2632. https://doi.org/10.1890/07-0986.1.
Buchaca, T., 2005. Pigments indicadors: estudi del senyal en estanys dels Pirineus i de la
seva aplicaci´
o en paleolimnologia. University of Barcelona, Barcelona, Spain. http://
hdl.handle.net/10803/1433. Accessed 05 November 2021.
Carlisle, D.M., Hawkins, C.P., 1998. Relationships between invertebrate assemblage
structure, 2 trout species, and habitat structure in Utah mountain lakes. J. N. Am.
Benthol. Soc. 17, 286–300. https://doi.org/10.2307/1468332.
ˇ
Ciamporov´
a-Zat'oviˇ
cov´
a, Z., Hamerlík, L., ˇ
Sporka, F., Bituˇ
sík, P., 2010. Littoral benthic
macroinvertebrates of alpine lakes (Tatra Mts) along an altitudinal gradient: a basis
for climate change assessment. Hydrobiologia 648, 19–34. https://doi.org/10.1007/
s10750-010-0139-5.
Collen, B., Whitton, F., Dyer, E.E., Baillie, J.E.M., Cumberlidge, N., Darwall, W.R.T.,
Pollock, C., Richman, N.I., Soulsby, A.-M., B¨
ohm, M., 2014. Global patterns of
freshwater species diversity, threat and endemism. Glob. Ecol. Biogeogr. 23, 40–51.
https://doi.org/10.1111/geb.12096.
Council of the European Communities, 1992. Council Directive 92/43/EEC of 21 May
1992 on the conservation of natural habitats and of wild fauna and ora. Off. J. Eur.
Union 206, 7–50. https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?
uri=CELEX:31992L0043&from=EN. Accessed 05 November 2021.
Cowie, R.H., Bouchet, P., Fontaine, B., 2022. The sixth mass extinction: fact, ction or
speculation? Biol. Rev. https://doi.org/10.1111/brv.12816.
De'Ath, G., 2002. Multivariate regression trees: a new technique for modeling
species–environment relationships. Ecology 83, 1105–1117. https://doi.org/
10.1890/0012-9658(2002)083[1105:MRTANT]2.0.CO;2.
De C´
aceres, M., Legendre, P., Moretti, M., 2010. Improving indicator species analysis by
combining groups of sites. Oikos 119, 1674–1684. https://doi.org/10.1111/j.1600-
0706.2010.18334.x.
de Mendoza, G., Catalan, J., 2010. Lake macroinvertebrates and the altitudinal
environmental gradient in the Pyrenees. Hydrobiologia 648, 51–72. https://doi.org/
10.1007/s10750-010-0261-4.
de Mendoza, G., Rico, E., Catalan, J., 2012. Predation by introduced sh constrains the
thermal distribution of aquatic Coleoptera in mountain lakes. Freshw. Biol. 57,
803–814. https://doi.org/10.1111/j.1365-2427.2012.02746.x.
de Mendoza, G., Ventura, M., Catalan, J., 2015. Environmental factors prevail over
dispersal constraints in determining the distribution and assembly of Trichoptera
species in mountain lakes. Ecol. Evol. 5, 2518–2532. https://doi.org/10.1002/
ece3.1522.
Eby, L.A., Roach, W.J., Crowder, L.B., Stanford, J.A., 2006. Effects of stocking-up
freshwater food webs. Trends Ecol. Evol. 21, 576–584. https://doi.org/10.1016/j.
tree.2006.06.016.
Eisenhauer, N., Bonn, A., Guerra, C.A., 2019. Recognizing the quiet extinction of
invertebrates. Nat. Commun. 10, 1–3. https://doi.org/10.1038/s41467-018-07916-
1.
Fjellheim, A., Boggero, A., Halvorsen, G.A., Nocentini, A.M., Rieradevall, M., Raddum, G.
G., Schnell, Ø.A., 2000. Distribution of benthic invertebrates in relation to
environmental factors. A study of European remote alpine lake ecosystems. Int.
Vereinigung Theor. Angew. Limnol. Verhandlungen 27, 484–488. https://doi.org/
10.1080/03680770.1998.11901278.
Füreder, L., Ettinger, R., Boggero, A., Thaler, B., Thies, H., 2006. Macroinvertebrate
diversity in Alpine lakes: effects of altitude and catchment properties. Hydrobiologia
562, 123–144. https://doi.org/10.1007/s10750-005-1808-7.
Gacia, E., Buchaca, T., Bernal-Mendoza, N., Sab´
as, I., Ballesteros, E., Ventura, M., 2018.
Non-native minnows threaten quillwort populations in high mountain shallow lakes.
Front. Plant Sci. 9, 329. https://doi.org/10.3389/fpls.2018.00329.
Hesthagen, T., Hegge, O., Skurdal, J., 1992. Food choice and vertical distribution of
European minnow, Phoxinus phoxinus, and young native and stocked brown trout,
Salmo trutta, in the littoral zone of a subalpine lake. Nord. J. Freshw. Res. 67, 72–76
http://hdl.handle.net/2077/48954. Accessed 05 November 2021.
Hulme, P.E., Nentwig, W., Pyˇ
sek, P., et al., 2009. In: Drake, J.A. (Ed.), DAISIE Handbook
of Alien Species in Europe, Invading Nature. Springer Series in Invasion Ecology.
Springer, Dordrecht, the Netherlands. https://doi.org/10.1007/978-1-4020-8280-1_
1.
Jeffrey, S.W., Humphrey, G.F., 1975. New spectrophotometric equations for determining
chlorophylls a, b, c1 and c2 in higher plants, algae and natural phytoplankton.
V. Osorio et al.

Biological Conservation 272 (2022) 109637
10
Biochem. Physiol. Panz. 167, 191–194. https://doi.org/10.1016/S0015-3796(17)
30778-3.
Jost, L., 2007. Partitioning diversity into independent alpha and beta components.
Ecology 88, 2427–2439. https://doi.org/10.1890/06-1736.1.
Knapp, R.A., Matthews, K.R., Sarnelle, O., 2001. Resistance and resilience of alpine lake
fauna to sh introductions. Ecol. Monogr. 71, 401–421. https://doi.org/10.1890/
0012-9615(2001)071[0401:RAROAL]2.0.CO;2.
Leavitt, P.R., Schindler, D.E., Paul, A.J., Hardie, A.K., Schindler, D.W., 1994. Fossil
pigment records of phytoplankton in trout-stocked alpine lakes. Can. J. Fish. Aquat.
Sci. 51, 2411–2423. https://doi.org/10.1139/f94-241.
Legendre, P., Anderson, M.J., 1999. Distance-based redundancy analysis: testing
multispecies responses in multifactorial ecological experiments. Ecol. Monogr. 69,
1–24. https://doi.org/10.1890/0012-9615(1999)069[0001:DBRATM]2.0.CO;2.
Maitland, P.S., Campbell, R.N., 1992. Freshwater Fishes of the British Isles. Harper
Collins, London, UK. ISBN,9780007406623.
Mir´
o, A., O’Brien, D., Tom`
as, J., Buchaca, T., Sab´
as, I., Osorio, V., Lucati, F., Pou-
Rovira, Q., Ventura, M., 2020. Rapid amphibian community recovery following
removal of non-native sh from high mountain lakes. Biol. Conserv. 251, 108783
https://doi.org/10.1016/j.biocon.2020.108783.
Mir´
o, A., Sab´
as, I., Ventura, M., 2018. Large negative effect of non-native trout and
minnows on Pyrenean lake amphibians. Biol. Conserv. 218, 144–153. https://doi.
org/10.1016/j.biocon.2017.12.030.
Mir´
o, A., Ventura, M., 2015. Evidence of exotic trout mediated minnow invasion in
Pyrenean high mountain lakes. Biol. Invasions 17, 791–803. https://doi.org/
10.1007/s10530-014-0769-z.
Mir´
o, A., Ventura, M., 2013. Historical use, shing management and lake characteristics
explain the presence of non-native trout in Pyrenean lakes: implications for
conservation. Biol. Conserv. 167, 17–24. https://doi.org/10.1016/j.
biocon.2013.07.016.
Mousavi, S.K., Sandring, S., Amundsen, P.-A., 2002. Diversity of chironomid assemblages
in contrasting subarctic lakes-impact of sh predation and lake size. Arch.
Hydrobiol. 154, 461–484. https://doi.org/10.1127/archiv-hydrobiol/154/2002/
461.
Museth, J., Borgstrøm, R., Brittain, J.E., 2010. Diet overlap between introduced
European minnow (Phoxinus phoxinus) and young brown trout (Salmo trutta) in the
lake, Øvre Heimdalsvatn: a result of abundant resources or forced niche overlap?
Hydrobiologia 642, 93–100. https://doi.org/10.1007/978-90-481-9388-2_9.
Museth, J., Hesthagen, T., Sandlund, O.T., Thorstad, E.B., Ugedal, O., 2007. The history
of the minnow Phoxinus phoxinus (L.) in Norway: from harmless species to pest.
J. Fish Biol. 71, 184–195. https://doi.org/10.1111/j.1095-8649.2007.01673.x.
Næstad, F., Brittain, J.E., 2010. Long-term changes in the littoral benthos of a Norwegian
subalpine lake following the introduction of the European minnow (Phoxinus
phoxinus). Hydrobiologia 642, 71–79. https://doi.org/10.1007/978-90-481-9388-2_
7.
Peres-Neto, P.R., Legendre, P., Dray, S., Borcard, D., 2006. Variation partitioning of
species data matrices: estimation and comparison of fractions. Ecology 87,
2614–2625. https://doi.org/10.1890/0012-9658(2006)87[2614:VPOSDM]2.0.CO;
2.
Pitcher, T.J., Magurran, A.E., Wineld, I.J., 1982. Fish in larger shoals nd food faster.
Behav. Ecol. Sociobiol. 10, 149–151. https://doi.org/10.1007/BF00300175.
R Core Team, 2021. A Language And Environment for Statistical Computing. R
Foundation for Statistical Computing, Vienna, Austria https://www.R-project.org.
S´
anchez-Bayo, F., Wyckhuys, K.A.G., 2019. Worldwide decline of the entomofauna: a
review of its drivers. Biol. Conserv. 232, 8–27. https://doi.org/10.1016/j.
biocon.2019.01.020.
S´
anchez-Hern´
andez, J., Amundsen, P.-A., 2015. Trophic ecology of brown trout (Salmo
trutta L.) in subarctic lakes. Ecol. Freshw. Fish 24, 148–161. https://doi.org/
10.1111/eff.12139.
S´
anchez-Hern´
andez, J., Cobo, F., Amundsen, P.-A., 2015. Food web topology in high
mountain lakes. PLoS One 10, e0143016. https://doi.org/10.1371/journal.
pone.0143016.
Schabetsberger, R., Grill, S., Hauser, G., Wukits, P., 2006. Zooplankton successions in
neighboring lakes with contrasting impacts of amphibian and sh predators. Int.
Rev. Hydrobiol. 91, 197–221. https://doi.org/10.1002/iroh.200610867.
Schilling, E.G., Loftin, C.S., Huryn, A.D., 2009. Effects of introduced sh on
macroinvertebrate communities in historically shless headwater and kettle lakes.
Biol. Conserv. 142, 3030–3038. https://doi.org/10.1016/j.biocon.2009.08.003.
Schindler, D.E., Knapp, R.A., Leavitt, P.R., 2001. Alteration of nutrient cycles and algal
production resulting from sh introductions into mountain lakes. Ecosystems 4,
308–321. https://doi.org/10.1007/s10021-001-0013-4.
Tate, A.W., Hershey, A.E., 2003. Selective feeding by larval dytiscids (Coleoptera:
Dytiscidae) and effects of sh predation on upper littoral zone macroinvertebrate
communities of arctic lakes. Hydrobiologia 497, 13–23. https://doi.org/10.1007/
s10021-001-0013-4.
Tiberti, R., Bogliani, G., Brighenti, S., Iacobuzio, R., Liautaud, K., Rolla, M., von
Hardenberg, A., Bassano, B., 2019. Recovery of high mountain alpine lakes after the
eradication of introduced brook trout Salvelinus fontinalis using non-chemical
methods. Biol. Invasions 21, 875–894. https://doi.org/10.1007/s10530-018-1867-0.
Tiberti, R., Buchaca, T., Cruset, E., Iacobelli, L., Maini, M., Ribelli, F., Pou-Rovira, Q.,
Ventura, M., 2022. Minnow introductions in mountain lakes result in lower salmonid
densities. Biol. Invasions. https://doi.org/10.1007/s10530-022-02783-w.
Tiberti, R., Buscaglia, F., Armodi, M., Callieri, C., Ribelli, F., Rogora, M., Tartari, G.,
Bocca, M., 2020. Mountain lakes of Mont Avic Natural Park: ecological features and
conservation issues. J. Limnol. 79, 43–58. https://doi.org/10.4081/
jlimnol.2019.1923.
Tiberti, R., Rolla, M., Brighenti, S., Iacobuzio, R., 2016. Changes in the insect emergence
at the water–air interface in response to sh density manipulation in high altitude
lakes. Hydrobiologia 779, 93–104. https://doi.org/10.1007/s10750-016-2801-z.
Tiberti, R., von Hardenberg, A., Bogliani, G., 2014. Ecological impact of introduced sh
in high altitude lakes: a case of study from the European Alps. Hydrobiologia 724,
1–19. https://doi.org/10.1007/s10750-013-1696-1.
Toro, M., Granados, I., Rubio, ´
A., Camacho, A., Sab´
as, I., Ventura, M., Montes, C., 2020.
Response of the aquatic invertebrate community to the eradication of an exotic
invasive sh 30 years after its introduction into an Iberian alpine lake. Aquat. Sci.
82, 1–21. https://doi.org/10.1007/s00027-020-00728-w.
Ventura, M., Camarero, L., Buchaca, T., Bartumeus, F., Livingstone, D.M., Catalan, J.,
2000. The main features of seasonal variability in the external forcing and dynamics
of a deep mountain lake (Red´
o, Pyrenees). J. Limnol. 59, 97–108. https://doi.org/
10.4081/jlimnol.2000.s1.97.
Ventura, M., Tiberti, R., Buchaca, T., Bu˜
nay, D., Sab´
as, I., Mir´
o, A., 2017. Why should we
preserve shless high mountain lakes? In: Catalan, J., Ninot, J.M., Aniz, M. (Eds.),
High Mountain Conservation in a Changing World. Springer International
Publishing, Dordrecht, the Netherlands, pp. 181–205. https://doi.org/10.1007/978-
3-319-55982-7_8.
Verdú, J.R., Numa, C., Galante, E. (Eds.), 2011. Atlas y Libro Rojo de los Invertebrados
amenazados de Espa˜
na (Especies Vulnerables). Direcci´
on General de Medio Natural
y Política Forestal. Ministerio de Medio Ambiente, Medio rural y Marino, Madrid,
Spain. ISBN, 9788480147941.
Wagner, D.L., 2020. Insect declines in the Anthropocene. Annu. Rev. Entomol. 65,
457–480. https://doi.org/10.1146/annurev-ento-011019-025151.
Wiley, R.W., 2003. Planting trout in Wyoming high-elevation wilderness waters.
Fisheries 28, 22–27. https://doi.org/10.1577/1548-8446(2003)28[22:PTIWHW]
2.0.CO;2.
WWF, 2016. Living Planet Report 2016. Risk And Resilience in a New Era. WWW
International, Gland, Switzerland. ISBN, 9782940529407.
V. Osorio et al.