Content uploaded by Virgilio Hermoso
Author content
All content in this area was uploaded by Virgilio Hermoso on Nov 03, 2017
Content may be subject to copyright.
1
Freshwater conservation in a fragmented world: dealing with barriers in a
systematic planning framework
Virgilio Hermoso1,2, Ana Filipa Filipe3,4, Pedro Segurado5, Pedro Beja3,4
1 Centre Tecnològic Forestal de Catalunya, Crta. Sant Llorenç de Morunys, Km 2. 25280,
Solsona. Lleida, Spain.
2 Australian Rivers Institute and Tropical Rivers and Coastal Knowledge, National
Environmental Research Program Northern Australia Hub, Griffith University, Nathan,
Queensland, 4111, Australia.
3 CIBIO/InBIO, Centro de Investigação em Biodiversidade e Recursos Genéticos da
Universidade do Porto, Campus Agrário de Vairão, R. Padre Armando Quintas, 4485-661
Vairão, Portugal.
4 CEABN/InBIO, Centro de Ecologia Aplicada, Instituto Superior de Agronomia, Universidade
de Lisboa, Tapada da Ajuda, 1349-017 Lisboa, Portugal
5 Centro de Estudos Florestais, Instituto Superior de Agronomia, Universidade de Lisboa 1349-
017 Lisboa, Portugal.
Running title: Conservation in disconnected systems.
Word count: 4488
Corresponding author: Virgilio Hermoso
email: virgilio.hermoso@gmail.com
Tlf: (34) 671832489
2
Abstract
1. Disruption of longitudinal connectivity poses one of the most important threats to the
persistence of freshwater biodiversity worldwide. Longitudinal connectivity plays a key role by
facilitating ecological processes, such as migrations or energy transfer along river networks. For
this reason, effective conservation of freshwater biodiversity is highly dependent on our
capacity to maintain all processes associated with connectivity. Freshwater protected areas are
commonly affected by disruptions of connectivity due to anthropogenic activities and recent
approaches to addressing connectivity when identifying priority areas have overlooked the
limitations that human perturbations pose to connectivity.
2. Here, a novel approach is presented to address this issue by accounting for the spatial
distribution of barriers in Marxan, a commonly used tool for conservation planning. This
approach is first tested on a simulated example and then applied to the identification of priority
areas for the conservation of freshwater vertebrates in the Iberian Peninsula (Spain and
Portugal).
3. When using this new approach, the number of disrupted connections within priority areas can
be significantly reduced at no additional cost in terms of area needed, which would help
maintain connectivity among populations of species with low-medium migratory needs.
4. Given the widespread occurrence of barriers in the study region, the improvement in
connectivity within priority areas also resulted in the selection of river reaches closer to the
headwaters and the river mouth. Focusing on both extremes of the longitudinal gradient might
compromise the effectiveness of conservation efforts for long-distance migratory species, such
as the European eel. This inevitably means that additional management measures, like barrier
removal or construction of fish passages, would be necessary to ensure these species may
complete their life cycles.
5. The method demonstrated here could be applied to other regions where connectivity is
compromised.
3
Keywords: amphibians, connectivity, dam, fish, Iberian Peninsula, Marxan, , reptiles, river,
systematic planning.
4
Introduction
Longitudinal connectivity plays a key role in freshwater ecosystems by facilitating ecological
processes, such as migrations or the transport of energy and materials along river networks
(Ward, 1989; Pringle, 2001). A multitude of anthropogenic activities have, however, deeply
modified the natural connectivity of running waters; these include barriers such as in-channel
dams, weirs or culverts (Januchowski-Hartley et al., 2013; Rolls, Ellison, Faggotter, & Roberts,
2013; Zarfl, Lumsdon, Berlekamp, Tydecks, & Tockner, 2015), often associated with alteration
of natural flows (Bunn, & Arthington, 2002), habitat degradation and modification of water
quality (Dudgeon et al., 2006). Alterations to natural connectivity increase habitat fragmentation
and cause impacts such as disruption of gene flow (Heggenes & Røed 2006; Roberts,
Angermeier, & Hallerman, 2013) or complete blockage of migratory routes essential to the
survival of diadromous fish or aquatic mammals (Choudhary et al., 2012; Maceda-Veiga, 2013;
Segurado, Branco, Avelar, & Ferreira, 2015). For example, Clavero & Hermoso (2015)
estimated that the European eel, a migratory species that was once widespread throughout the
whole Iberian Peninsula, has lost over 80% of its original distribution range mainly due to the
massive dam construction since the 1950s. Segurado et al., (2015) also reported considerable
habitat losses in Portugal of 58% and 72% for sea lamprey (Petromyzon marinus) and allis shad
(Alosa alosa), respectively, during the last century. In-stream barriers might also compromise
the persistence of other not strictly aquatic taxa such as amphibians or reptiles that use riparian
corridors to disperse along streams due to the low habitat resistance offered by this pathway
(Grant, Lowe, & Fagan, 2007; Grant, Nichols, Lowe, & Fagan, 2010). Modification of natural
flow patterns caused by dams can also result in habitat loss for these taxa (e.g., Hunt et al.,
2013; Eskew, Price, & Dorcas, 2012), with consequences, for example, on breeding success
(e.g., Lind, Welsh Jr, & Wilson, 1996) and reductions of connectivity among populations
(Fagan, 2002). For all those reasons, effective conservation of freshwater biodiversity is highly
dependent on our capacity to maintain all ecological and evolutionary processes linked to
stream connectivity.
5
Despite a later start than in other realms, there have been considerable advances in the field of
conservation planning applied to freshwater systems in the last decade (Collier, 2011; Hermoso,
Abell, Linke, & Boon 2016). The attention has mainly focused on addressing the singularities of
these systems derived from their dendritic nature (Peterson et al., 2013), with particular
connectivity constraints that make effective conservation planning and implementation a
challenge (Linke, Turak, & Nel, 2011; Collier, 2011). There are novel methods available to
integrate different aspects of connectivity in the identification of priority areas for freshwater
biodiversity conservation and enhance their effectiveness, such as longitudinal, lateral or
temporal connectivity (Moilanen, Leathwick, & Elith, 2008; Hermoso, Linke, Prenda, &
Possingham, 2011; Hermoso, Kennard, & Linke 2012a; Hermoso, Ward, & Kennard, 2012b).
However, most of these advances and conservation efforts have overlooked the limitation that
human perturbations pose to connectivity, both structural (i.e., physical relationships among
habitat patches irrespective of behavioural response of organisms) and functional (ease with
which individuals can move within the river-landscape). As a consequence, freshwater protected
areas are commonly affected by disconnection from upstream or downstream areas imposed by
dams and other infrastructures (e.g., Hermoso, Filipe, Segurado, & Beja, 2015a; Thieme et al.,
2016). Even when fish passages are present, they are often inefficient in facilitating movement
(e.g., fish ladders), and consequently the dam often restricts the gene flow between fish
populations located up- and downstream of the barrier, leading to population isolation
(Esguícero, & Arcifa, 2010). Therefore, there is an urgent need for novel approaches to help
decision-makers and stakeholders address the constraints to the effectiveness of protected areas
for the conservation of freshwater biodiversity.
There are two main alternatives to address the problem of fragmentation caused by dams. One is
to selectively remove barriers or implementing effective passages (e.g., fish ladders) to enhance
connectivity within and among existing protected areas. Optimization of barrier removal has
been explored in the last decade, aiming to identify priority sets of barriers that should be
removed or made more permeable (e.g., installing fish passages) to maximise gains in
6
connectivity (e.g., O’Hanley, Wright, Diebel, Fedora, & Soucy, 2013; Zheng, & Hobbs, 2013;
Branco, Segurado, Santos, & Ferreira, 2014). However, barrier removal can be expensive
(Zheng, & Hobbs, 2013) or unfeasible due to socioeconomic constraints (Arthington, Naiman,
McClainme, & Nilsson, 2010), so this management option is often unrealistic. The second
option is to account for extant or planned barriers and allocate conservation effort to areas that
are less prone to their impacts. In this way it might be possible to improve the effectiveness of
priority conservation areas for freshwater biodiversity, while reducing potential conflicts with
stakeholders by addressing key socioeconomic constraints. However, this option has yet to be
explored from a conservation planning perspective.
Here, a method for integrating information on the spatial distribution of barriers in a systematic
planning framework to enhance the effectiveness of conservation efforts for freshwater
biodiversity is demonstrated. The software Marxan (Ball et al., 2009) is used to identify priority
areas for conservation while accounting for existing disconnections. Different alternatives to
address disconnections on a simulated case study are first tested, and then one of them is used to
demonstrate how it can be implemented on a real case study using aquatic vertebrates and
existing dams in the Iberian Peninsula. The aim of this method is to reduce habitat
fragmentation of priority areas for conservation, both within each area and from upstream/
downstream areas. Accounting for the spatial distribution of barriers would help compensate for
the impacts derived from these infrastructures and ultimately the effectiveness of conservation
efforts for freshwater biodiversity.
Methods
Addressing disconnections in Marxan
In this study the software Marxan (Ball, Possingham, & Watts, 2009) was used to find an
optimal set of planning units to represent all conservation features at the minimum cost, while
7
incorporating spatial constraints, such as longitudinal connectivity. The formulation of the
mathematical problem addressed in Marxan is as follows:
Eq. 1
Eq. 2
where xi is a control variable that takes a value of 1 when the subcatchment ci is selected and 0
otherwise; cvi1,i2 represents a penalty for missing connections in the solution (e.g., when
subcatchment i1 is in the solution but i2 is not); aij is the representation of species j achieved by
selecting subcatchment i and tj is the total representation target aimed for each species j.
Marxan uses a heuristic optimisation algorithm to minimize an objective function (Eq. 3) that
comprises Eq 1 and Eq 2, and includes the cost of planning units in the solution and other
penalties for not achieving the conservation target for all the conservation features (Feature
Penalty, weighted by Species’ Penalty Factor, SPF) and spatial constraints (connectivity
weighted by a Connectivity Strength Modifier, CSM). This latter parameter in Marxan allows
users to incorporate parameters that accommodate the spatial particularities derived from the
dendritic structure and connectivity inherent to freshwater ecosystems and their threats (see
Hermoso et al., 2011).
Eq. 3.
To test the potential use of Marxan to address existing disconnections in a freshwater setting, a
simple case was simulated with a linear structure composed of ten consecutive planning units,
together representing a hypothetical subcacthment, flowing from a headwater planning unit to
an outlet (Fig. 1a). This structure was used to build a connectivity file as usually done for
featuresunitsplanning PenaltytyConnectiviCSMPenaltyFeatureSPFCostfunctionObjective
8
Marxan applications in freshwater environments (Hermoso et al., 2011; Esselman et al., 2013).
This file differs from terrestrial and marine boundary files as it contains all the longitudinal
connections between planning units, instead of shared boundaries as normally done in other
realms. Penalties in the boundary file are distance weighted according to the distance between
planning units along the river network following equation 4.
Connectivity penalty=dij(km)-1/2 Eq. 4
where dij is the distance between planning units i and j (Fig. 1b). In this example, we assumed
planning units to have a regular shape with a total length of river within each cell of 1 km. For
example the penalty for including planning unit 6 but not 8 in the solution would be 0.71
(penalty=2-1/2, where the value 2 is the distance between planning units 6 and 8). This penalty
decays exponentially with distance between planning units, so the farther two planning units are
apart, the lower the penalty that would apply if not selected together.
The occurrence of two longitudinal barriers, located in planning units 4 and 7, was also
simulated, (Fig. 1b). These barriers break the continuity of the system with, restricting or
outright eliminating the connection between close planning units such as 6 and 8, depending on
the strength of the barrier effect. To address these disconnections in Marxan, the boundary file
was modified by incorporating a new parameter in the calculation of the connectivity penalties.
In this case, the existence of a barrier between each combination of planning units connected
along the river network was accounted for by introducing a new discount factor (R) that
modifies the distances used in Eq. 2, as follows:
dij´=dij+(R/dij) Eq. 5
where dij´ is a modified distance between planning units i and j, R is the discount factor for the
barrier effect and dij as above (Fig. 1b). This new dij´ was used to recalculate the connectivity
penalties as above. In this way, the connectivity penalty between planning units separated by a
barrier would decrease, indicating that the connection is weaker than expected from the true
geographic distance between units. Factor R can be adjusted to account for the strength or
9
imperviousness of each particular barrier. For example, large R values would indicate stronger
disconnections due to the absence of fish passages, while smaller R values would indicate that
the barrier is more permeable (e.g., road crossings or culverts which some species might be able
to pass over; e.g. Jauchowski-Hartley et al., 2012). For the sake of simplicity, we kept R
constant for both barriers in this case study (R=1), so the discount factor was just based on the
distance between each pair of planning units (the more distant they were from the barrier, the
lower the decay due to the effect of the barrier). In the example in Fig. 1, the penalty for the
connection between planning units 6 and 8 would decrease to 0.35 [0.71*(1/2)], so this
connection would be considered less relevant. To test the effect of this new boundary file we
simulated the distribution of three conservation features, which occurred in all planning units
and assumed equal cost for all planning units. Marxan was then ran with a CSM=1, SPF=10 and
a representation target=5 (50% of each species´ distribution) under three alternative planning
approaches: i) as if no barriers existed (Fig 1c; similar to Hermoso et al., 2011); ii) locking out
planning units where barriers occur (Fig 1d); and iii) using the new approach to discount for
broken connections (Fig 1e).
Application to the Iberian Peninsula
The methodology detailed above was applied to the identification of priority areas for
conservation of freshwater biodiversity in the Iberian Peninsula, under the constrains of existing
dams. The Iberian Peninsula (Spain and Portugal, excluding islands) is located in south-western
Europe, covering a total area of approximately 583,000 km2 and spans across four freshwater
ecoregions (Abell et al., 2008). The study focused on the entire Iberian Peninsula to account for
(1) the role of connectivity along river networks for freshwater ecosystems and biodiversity
(e.g., migration of organisms, flow of energy and matter) across a whole distinctive
biogeographical unit, and (2) the propagation of disconnection effects due to existing dams in
both countries. This area is also especially relevant for a study like this as it holds one of the
10
most severely regulated river systems in the world (Grill et al., 2015), and it is home to some of
the most highly threatened freshwater biodiversity in Europe (Maceda-Veiga, 2013).
The most up-to-date information on the spatial distribution for 93 freshwater-related species,
including 63 fish species, 25 amphibians and 5 semi-aquatic reptiles was compiled (Appendix
S1) The occurrence datasets of aquatic amphibians and reptiles from most recent atlases at a 10-
km grid cell resolution were merged (Spain: Pleguezuelos, Márquez, & Lizana, 2002; Portugal:
Loureiro, Ferrand de Almeida, Carretero, & Paulo, 2010). Fish data for Portugal was based on
the database built in Filipe et al., (2009) and in the Carta Piscícola
(http://www.cartapiscicola.org/), at sampling site resolution (Rogado et al., 2005), whereas data
for Spain was derived from the most recent atlas at a 10-km grid cell resolution (Doadrio,
2002). Recent sampling carried out by the authors were used to update these datasets. The final
database comprised the most comprehensive information on freshwater species occurrences for
these taxonomic groups in the Iberian Peninsula, with 49,463 occurrence records within 5,938
10-km grid cells. To make the analyses sounder for freshwater ecosystems, the information
originally reported in 10-km grid cells was translated into subcatchments. A total of 19,854
subcatchments, each including the portion of river length between two consecutive nodes or
river connections and its contributing area (Length= 7.7 ± 4.8 km, Area= 29.12 ± 23.5 km2;
Average ± SD) were delineated from a 90 m digital elevation model (sourced from the SRTM
90m Digital Elevation Database v4.1; Jarvis, Reuter, Nelson, & Guevara, 2008) in ArcGIS 10.1
(ESRI 2011). The grid cells and subcatchments were then intersected and a species was
assumed to be present in a subcatchment whenever the grid cell occupied more than 50% of the
subcatchment. The spatial distribution of each species was then visually inspected to ensure that
occurrences had not been assigned to the wrong hydrological catchment from grid cells
overlapping two neighbour catchments. This resulted in 180,584 occurrence records (species x
subcatchmnets) for the whole Iberian Peninsula.
The distribution of large dams was sourced from the Global Reservoir and Dam (GRanD)
database (Lehner et al., 2011), which contain georeferenced locations for all reservoirs with a
11
storage capacity > 100 Hm3. This dataset contained the location of 305 dams in the Iberian
Peninsula. Although there are a number of smaller reservoirs (<100 Hm3) and other barriers to
longitudinal connectivity in the Iberian Peninsula, these were not considered because
comprehensive georeferenced information was lacking and large dams represent the main
obstacles to connectivity along rivers. An arbitrarily large and constant discount factor (R=500)
was used for the analyses, given that there was no reliable data available on the relative
permeability of each dam. This large value assumes that the permeability of barriers included in
this study was always low, as all of them were large dams that strongly constrain movements
along waterlines of the species considered. Given the special interest of this study in testing the
effect of this new boundary file, a constant cost for all subcatchments was used (cost=1), as well
as constant CSM=1 and SPF=10 for the two different boundary files mentioned previously:
traditional (Hermoso et al., 2011; see Fig 1c) and accounting for disconnections. A
representation target of 200 subcatchments was set for both scenarios, though this arbitrary
threshold would need to be fine-tuned in relation to properly defined conservation objectives in
practical applications of the approach. Nevertheless, this conservation target was considered
reasonable because it represents the whole range for 30 species, and >25% of the range for 50%
of the species considered. Marxan was ran 100 times (1 million iterations each) and retained the
best solution over those runs for further analyses.
Three alternative measures were used to evaluate how the modified boundary file affected the
conservation planning solutions. First, the number of disrupted connections in both directions
(up- and downstream) between all priority subcatchments included in the best solution was
measured for each planning scenario, testing the expectation that there should be less
disruptions when using the modified boundary file. A connection between two subcatchments
was considered to be disrupted whenever there was a dam between two priority subcatchments
that were otherwise connected longitudinally along the stream network. To account for
differences between planning scenarios in the number of subcatchments included in the best
solution, the number of disrupted connections was standardised by the number of priority
12
subcatchments selected. This was done because the larger the number of subcatchments in the
solution, the larger the number of connections involved. Second, for each planning scenario the
average distances (±SE) of subcatchments included in the best solution to their headwaters and
the mouth of the river was also measured, to check whether the spatial location of priority
subcatchments moved within the stream network across planning scenarios to account for
existing barriers. Finally, the number of dams within subcatchments included in the best
solution were also counted for each planning scenario.
Results
Effect of different approaches to address disconnections
For the hypothetical river configuration, the use of different approaches to dealing with
disconnections had significant effects on the selection of priority areas. Planning units selected
when using the traditional boundary file were clumped towards headwaters (Fig. 1c). When
planning units containing dams were locked out of the selection process, the units selected were
also clumped towards the headwaters, with the exception of unit 7 that was not available in this
case (Fig. 1d). In both these alternatives, the best solutions included disrupted connections
between adjacent planning units , which would compromise connectivity within priority areas
for conservation. Only when using the modified boundary file the disrupted connections
between the planning units selected were minimised (Fig. 1e). In this latter case, the planning
units selected were located at both extremes of the longitudinal gradient (i.e., either close to
headwater or to the mouth). Given the large constraints to connectivity imposed by the presence
of dams in the simulated example, there were still disrupted connections in the solution (e.g.,
connections between planning units in the mouth and headwaters). However, there were no
adjacent planning units selected with broken connections due to dams, which would help
maximise connectivity within selected areas.
13
Figure 1. Simulated case study of ten subcatchments (a) and spatial allocation of two barriers
(b). Penalties to missed connectivity in Marxan were derived from the distance between
subcatchments in a) and modified to account for the disruption to connectivity caused for
barriers in b). For the sake of simplicity only the upstream connections are shown here, although
downstream connections were equally accounted for. * Indicates those upstream connections to
subcatchment number 6 that were affected by a dam, located in subcatchment number 7.
Solutions from three alternative planning scenarios are also shown: c) by using the standard
connectivity measure, d) by locking subcatchments that contain barriers out, and e) by using the
modified connectivity measure to account for the spatial distribution of barriers. Open cells in c-
e indicate subcatchments not selected, close cells indicate subcatchments that were selected by
Marxan and bold cells indicate those that contained a dam.
Application to the Iberian Peninsula
Accounting for existing barriers in the identification of priority areas for conservation of
freshwater biodiversity in the Iberian Peninsula had a significant effect on all the parameters
evaluated. Although the numbers of subcatchments selected in both planning approaches were
14
similar (N=4859 and N=4910, there were significant differences (Cohen´s kappa=0.69,
p<0.001) in the spatial allocation of priority subcatchmnets when using either the traditional or
the modified boundary file accounting for dams (Fig. 2). As for the spatial allocation of these
priority subcatcments within the stream network, these moved towards the extremes of the
longitudinal gradient when using the modified boundary file. This is highlighted when
comparing the spatial location of the priority subcatchments selected under either the traditional
or the modified approach, which in the latter case were significantly closer to the headwaters
(F=6.51, p=0.01) and the river mouth (F=6.45, p=0.01) (Fig 3). In addition, the number of
disrupted connections between priority subcatchments was lowest when using the modified
approach (Fig. 4), and the number of dams within priority subcatchments was lower in the
modified (N = 59) versus the traditional (N = 79) planning solution.
15
Figure 2. Spatial distribution of subcatchments selected by Marxan under two alternative
planning scenarios: a) by using the standard connectivity measure and b) by using the modified
connectivity measure to account for the spatial distribution of large dams in the Iberian
Peninsula. The map represents the best solution after 100 runs with 1 million iterations each.
Grey subcatchments represent those selected in both solutions; red subcatchments represent
those selected in the solution with the standard connectivity; green subcatchments represent
those selected in the solution with the modified connectivity.
Discussion
This study demonstrates how to incorporate information on existing barriers in a conservation
planning framework to help minimise impacts derived from these infrastructures, and ultimately
enhance the effectiveness of conservation efforts for freshwater biodiversity. The loss of
connectivity is a critical threat to the conservation of freshwater biodiversity (Humphries &
Winemiller 2009; Vörösmarty et al., 2010; Lierman et al., 2012), but has not been adequately
addressed yet in a conservation prioritisation context. As a consequence, freshwater protected
areas are generally affected by disconnections from their upper and lower catchments (e.g.,
Hermoso et al., 2015a; Thieme et al., 2016), and freshwater species, in particular migratory fish,
continue to decline in those areas (Dudgeon et al., 2006; McRae, Freeman, & Deinet, 2014).
16
Figure 3. Average ± SE distance to headwater and river mouth across subcatchments selected by
Marxan under the two alternative planning scenarios: when using the traditional connectivity
measure (black bars) and when accounting for the spatial distribution of large dams (white
bars). There were significant differences in the average location of subcatchments selected
(F=6.51, p=0.01; F=6.45, p=0.01 for headwater and river mouth respectively).
By explicitly accounting for the location of dams in the prioritisation process, the connectivity
within and among priority areas can be improved. In our application to the Iberian Peninsula,
we found that the number of disrupted longitudinal connections among priority subcatchments
could be reduced to just 1/3 of those that would occur in solutions disregarding dams. This was
achieved at no additional cost in terms of the amount of area selected in relation to the solution
provided by the traditional approach. However, there were changes in the spatial location of the
subcatchments selected, which tended to be closer to either the river mouths or especially to
headwaters. This happened because of the widespread occurrence of dams in medium-low
reaches of Iberian rivers, which made it difficult to achieve fully connected priority areas. The
weight applied to connectivity ensured that these areas hosted long reaches completely
connected (see Fig. 2), which would help maintain connectivity among populations of species
that undergo short to moderate migratory movements to complete their life cycle. However,
focusing on both extremes of the longitudinal gradient might compromise the effectiveness of
conservation efforts for long-distance migratory species, such as the European eel (Feunteun,
2002). In fact, under the limitations imposed by the widespread occurrence of barriers there
would be a shortage of medium reaches selected for conservation management to connect upper
and lower reaches. This inevitably means that additional management measures, like barrier
removal or construction of fish passages, would be necessary to ensure these long-distance
migratory species may complete their life cycles. In this sense, a combination of the
methodology demonstrated here and other approaches to identifying key barriers to be managed
17
would be necessary to enhance the overall effectiveness of conservation efforts in river
networks.
Figure 4. Number of upstream and downstream broken connections among subcatchments
selected by Marxan under two alternative planning scenarios. The number of broken
disconnections were standardised by the number of subcatchments in each solution. A
connection between two subcatchments selected by Marxan was considered broken when they
were longitudinally connected along the river network but a dam was present between them.
The number of broken connections was standardised by the number of subcatchmnets selected
under the two alternative scenarios.
The maintenance of ecological processes in freshwater ecosystems require whole catchment
scale planning, as we demonstrate here, to ensure issues related to longitudinal connectivity are
adequately addressed. This might entail in many cases transboundary collaboration between
different regional and/ or national governments (e.g., Dolezsai, Sály, Takács, Hermoso, & Eros,
2015), which is a case in point in the Iberian Peninsula, where many of the largest catchments
have the headwaters in Spain and the mouth in Portugal. Efforts have been made to overpass
transboundary problems of rivers (e.g., Correia & da Silva 1997), as the bilateral agreement
“Albufeira Convention” between Spain and Portugal for assuring minimum flows and
communication needed during floods and droughts (López-Moreno et al., 2009). But additional
18
efforts are needed to maximize and assure persistence of freshwater biodiversity, as without
whole catchment planning the effectiveness of conservation efforts in one country could be
compromised by the lack of action in the other (e.g., Anaecypris hispanica, an endangered
species according with IUCN, has isolated populations in each side or the border and needs
concerted planning and actions). The maintenance of ecological processes related to
longitudinal connectivity might, however, result in large areas to be managed for conservation
(e.g., Filipe at al., 2004). This could compromise the feasibility of conservation plans in these
systems, at least under traditional management schedules based on strict reservation. However,
more flexible approaches based on multiple management zones (see Abell, Allan, & Lehner,
2007) is gaining the attention in this field for its capacity to accommodate the demand of large
areas in need of management under specific regimes (e.g., Hermoso, Cattarino, Kennard, &
Linke, 2015b; Hermoso, Filipe, Segurado, & Beja, 2016). For example, among all
subcatchments selected in this study, not all need to be strictly protected, as some have been
selected to ensure connectivity between populations, for example. Those subcatchments
highlighted for connectivity could be under a special management regime to ensure connectivity
is effective all year (e.g., no barriers) or temporally, whenever it is needed (e.g., adequate
environmental flows to facilitate migrations during spawning periods).
The adequate consideration of barriers in conservation planning will need good inventories of
their spatial location, permeability characteristics and potential impact on biodiversity
(Januchowski-Hartley, Diebel, Doran, & McIntyre, 2012; Januchowski-Hartley et al., 2013;
Mantel, Rivers-Moore, & Ramulifho, 2017). Here, the spatial location of large dams in the
Iberian Peninsula was used for the sake of demonstration and consistency in data quality across
the whole study area. However, there is a multitude of other smaller infrastructures, such as
weirs and ponds, culverts or road passages, that can seriously compromise the maintenance of
key ecological processes, with severe cumulative effects (Alexandre & Almeida 2010). Not all
of them, however, pose impassable barriers or compromise longitudinal connectivity, and then
do not necessarily need to be accounted for in planning exercises. Discerning the relative impact
19
of these barriers, and not only mapping them, seems an enormous effort that could be supported
by novel modelling approaches (e.g., Januchowski-Hartley et al., 2012). The approach that we
demonstrate here can accommodate the information on the impact of different barriers by
modifying the R parameter specifically for each of them according to their characteristics (e.g.,
dam high, reservoir size, effectiveness of fish passage for fish, and lateral riparian corridors for
amphibians and reptiles) and potential impact on biodiversity. Finally, the approach
demonstrated here to account for the effect of barriers in conservation planning to be useful in
other realms and habitats, from terrestrial to marine, affected by similar connectivity problems.
Conclusions
Given the widespread modification to longitudinal connectivity of the world´s streams, and our
limited socio-economic and technical capacity to address the problem, the approach we present
here could help minimise impacts derived from these infrastructures and ultimately to enhance
the effectiveness of conservation efforts for freshwater biodiversity. The approach demonstrated
here could be an alternative or complementary to other approaches more focused on identifying
barriers to be removed or made more permeable (e.g., O´Hanley et al., 2013; Zheng, & Hobbs,
2013; Branco et al., 2014). These other alternatives are commonly constrained by financial
and/or technical capacity to implement the recommendations that arise from the prioritisation
analyses. Moreover, the effectiveness of traditional efforts to mitigate the impact of barriers,
such as fish passages or riparian corridors, at re-establishing connectivity between isolated
populations is limited in many cases (Esguícero & Arcifa, 2010). The approach demonstrated in
this study does not propose methods to act directly on barriers, but accommodate instead the
identification of priority areas for conservation efforts to the spatial distribution of existing
barriers. In this way, the potential socio-economic impact of our recommendations which could
be minimised and help enhance implementation of conservation in this realm. This is especially
relevant under the current context of accelerated rate of new dam construction, which is
expected to increase in the near future worldwide (Zarfl et al., 2015). Proactive planning taking
20
into account future barriers is urgently needed (Winemiller et al., 2016) to ensure that
freshwater ecosystems, biodiversity and the services they provide do not continue the declining
path they have experienced in the last decades (McRae et al., 2014).
Acknowledgements
We acknowledge funding support provided by the Spanish Government through a Ramon y
Cajal contract (RYC-2013-13979) to VH. AFF was supported by the FRESHING Project
funded by FCT and COMPETE (PTDC/AAG-MAA/2261/2014 – POCI-01-0145-FEDER-
016824). PS was supported by MARS project funded under the 7th EU Framework Programme
(Contract No.: 603378). PB was supported by EDP Biodiversity Chair.
21
References
Abell, R., Allan, J.D. and Lehner, B. (2007). Unlocking the potential of protected areas for
freshwaters. Biological Conservation, 134, 48–63.
Abell, R., Thieme, M.L., Revenga, C., Bryer, M., Kottelat, M., Bogutskaya, N., …
Wikramanayake, E. (2008). Freshwater ecoregions of the world: a new map of
biogeographic units for freshwater biodiversity conservation. BioScience, 58, 403e414.
Alexandre, C.M., & Almeida, P.R. (2010). The impact of small physical obstacles on the
structure of freshwater fish assemblages. River Research and Applications, 26, 977–994.
Arthington, A.H., Naiman, R.J., McClainme, & Nilsson, C. (2010). Preserving the biodiversity
and ecological services of rivers: new challenges and research opportunities. Freshwater
Biology, 55, 1–16.
Ball, I.R., Possingham, H.P., & Watts, M. (2009). Marxan and relatives: Software for spatial
conservation prioritisation. Pages 185-195 in Moilanen A., Wilson K.A., Possingham H.P.,
Editors. Spatial conservation prioritisation: Quantitative methods and computational tools.
Oxford University Press, Oxford, UK.
Branco, P., Segurado, P., Santos, J.M., & Ferreira, M.T. (2014). Prioritizing barrier removal to
improve functional connectivity of rivers. Journal of Applied Ecology, 51, 1197–1206.
Bunn, S.E., & Arthington A.H. (2002). Basic principles and ecological consequences of altered
flow regimes for aquatic biodiversity. Environmental Management, 30, 492–507.
Choudhary, S., Dey, S., Dey, S., Sagar, V., Nair, T., Kelkar, N. (2012). River dolphin
distribution in regulated river systems: implications for dry-season flow regimes in the
Gangetic basin. Aquatic Conservation: Marine and Freshwater Ecosystems, 22, 11–25.
Clavero, M., & Hermoso V. (2015). Historical data to plan the recovery of the European eel.
Journal of Applied Ecology, 52, 960–968.
22
Collier, K.J. (2011). The rapid rise of streams and rivers in conservation assessment. Aquatic
Conservation: Marine and Freshwater Ecosystems, 21, 397–400.
Correia, F.N., & da Silva, J.E. (1997). Transboundary Issues in Water Resources. Pages 315-
334 in Conflict and the Environment, Gleditsch NP, Editor. Springer Netherlands.
Doadrio, I. (2002). Atlas y Libro Rojo de los Peces Conti-nentales de Espaňa. Direccioón
General de Conservación de la Naturaleza. Museo Nacional de Ciencias Naturales, Madrid.
Dolezsai, A., Sály, P., Takács, P., Hermoso, V., & Eros, T. (2015). Restricted by borders: trade-
offs in transboundary conservation planning for large river systems. Biodiversity and
Conservation, 24, 1403–1421.
Dudgeon, D., Arthington, A.H., Gessner, M.O., Kawabata, Z.I., Knowler, D.J., Lévêque, C., …
Sullivan, C.A. (2006). Freshwater biodiversity: importance, threats, status and conservation
challenges. Biological Reviews, 81, 163–182.
Esguícero, A.L., & Arcifa, M.S. (2010). Fragmentation of a Neotropical migratory fish
population by a century-old dam. Hydrobiologia, 638, 41-53.
Eskew, E.A., Price, S.J., & Dorcas, M.E. (2012). Effects of River-Flow Regulation on Anuran
Occupancy and Abundance in Riparian Zones. Conservation Biology, 26, 504–512.
Esselman, P.C., Edgar, M., Breck, J., Hay-Chmielewski, E., & Wang, L. (2013). Riverine
connectivity, upstream influences, and multi-taxa representation in a conservation area
network for the fishes of Michigan, USA. Aquatic Conservation: Marine and Freshwater
Ecosystems, 23, 7–22.
ESRI. (2011). ArcGIS Desktop: Release 10. Redlands, CA: Environmental Systems Research
Institute.
Fagan, W.F. (2002). Connectivity, fragmentation, and extinction risk in dendritic
metapopulations. Ecology, 83, 3243-3249.
23
Feunteun, E. (2002). Management and restoration of European eel population (Anguilla
anguilla): an impossible bargain. Ecological Engineering, 18, 575-591.
Filipe, A.F., Marques, T.A., Tiago, P., Ribeiro, F., Da Costa, L.M., Cowx, I.G., & Collares-
Pereira, M.J. (2004). Selection of priority areas for fish conservation in Guadiana River
Basin, Iberian Peninsula. Conservation Biology, 18, 189-200.
Filipe, A.F., Araujo, M.B., Doadrio, I., Angermeier, P.L., & CollaresPereira, M.J. (2009).
Biogeography of Iberian freshwater fishes revisited: the roles of historical versus
contemporary constraints. Journal of Biogeography, 36, 2096-2110.
Grant, E.H.C., Lowe, W.H., & Fagan, W.F. (2007). Living in the branches: population
dynamics and ecological processes in dendritic networks. Ecology Letters, 10, 165–175.
Grant, E.H.C., Nichols, J.D., Lowe, W.H., & Fagan, W.F. (2010). Use of multiple dispersal
pathways facilitates amphibian persistence in stream networks. Proceedings of the National
Academy of Sciences, 107, 6936–6940.
Grill, G., Lehner, B., Lumsdon, A.E., MacDonal, G.K., Zarfl, C., Reidy Liermann, C. (2015).
An index-based framework for assessing patterns and trends in river fragmentation and flow
regulation by global dams at multiple scales. Environmental Research Letters, 10,
doi:10.1088/1748-9326/10/1/015001.
Heggenes, J., & Røed, K.H. (2006). Do dams increase genetic diversity in brown trout (Salmo
trutta)? Microgeographic differentiation in a fragmented river. Ecology of Freshwater Fish,
15, 366-375.
Hermoso, V., Cattarino, L., Kennard, M.J., & Linke, S. (2015b). Catchment zoning for
freshwater conservation: refining plans to enhance action on the ground. Journal of Applied
Ecology, 52, 940-949.
Hermoso V, Filipe AF, Segurado P, Beja P. 2015a. Effectiveness of a large reserve network in
protecting freshwater biodiversity: a test for the Iberian Peninsula. Freshwater Biology 60:
698-710.
24
Hermoso, V., Filipe, A.F., Segurado, P., Beja, P. (2016). Catchment zoning to unlock
freshwater conservation opportunities in the Iberian Peninsula. Diversity and Distributions,
22, 960-969.
Hermoso, V., Kennard, M.J., & Linke S. (2012a). Integrating multi-directional connectivity
requirements in systematic conservation planning in freshwater systems. Diversity and
Distributions, 18, 448-458.
Hermoso, V., Linke, S., Prenda, J., Possingham, H.P. (2011). Addressing longitudinal
connectivity in the systematic conservation planning of fresh waters. Freshwater Biology,
56, 57–70.
Hermoso, V., Ward, D.P., Kennard, M.J. (2012b). Using water residency time to enhance
spatio-temporal connectivity for conservation planning in seasonally dynamic freshwater
ecosystems. Journal of Applied Ecology, 49, 1028-1035.
Hermoso, V., Abell, R., Linke, S., & Boon, P. (2016). The role of protected areas for freshwater
biodiversity conservation: challenges and opportunities in a rapidly changing world. Aquatic
Conservation: Marine and Freshwater Ecosystems, 26, 3–11.
Humphries, P.L. & Winemiller, K.O. (2009). Historical impacts on river fauna, shifting
baselines and challenges for restoration. Bioscience, 59, 673–684.
Hunt, S.D., Guzy, J.C., Price, S.J., Halstead, B.J., Eskew, E.A., & Dorcas, M.E. (2013).
Responses of riparian reptile communities to damming and urbanization. Biological
Conservation, 157, 277-284.
Januchowski-Hartley, S.R., Diebel, M., Doran, P.J., & McIntyre, P.B. (2012). Predicting road
culvert passability for migratory fishes. Diversity and Distributions, 20, 1414–1424.
Januchowski-Hartley, S.R., McIntyre, P.B., Diebel, M., Doran, P., Infante, D.M., Joseph, C.,
Allan, J.D. (2013). Restoring aquatic ecosystem connectivity requires expanding inventories
of both dams and road crossings. Frontiers in Ecology and the Environment, 11, 211-217.
25
Jarvis, A., Reuter, H.I., Nelson, A., & Guevara, E. (2008). Hole-filled SRTM for the globe
Version 4. Available from the CGIAR-CSI SRTM 90m Database http://srtm.csi.cgiar.org
(accessed 15 July 2015).
Lehner, B., Reidy Liermann, C., Revenga, C., Vorosmarty, C., Fekete, B., Crouzet, P., …
Wisser, D. (2011). Global Reservoir and Dam Database, Version 1 (GRanDv1): Dams,
Revision 01. Palisades, NY: NASA Socioeconomic Data and Applications Center (SEDAC).
http://dx.doi.org/10.7927/H4N877QK. Accessed 1 August 2016.
Lind, A.J., Welsh Jr., H.H., & Wilson, R.A. (1996). The effects of a dam on breeding habitat
and egg survival of the foothill yellow-legged frog (Rana boylii) in Northwestern
Calfifornia. Herpetological Review, 27, 62-66.
Linke, S., Turak, E., & Nel, J. (2011). Freshwater conservation planning: the case for systematic
approaches. Freshwater Biology, 56, 6–20.
López‐Moreno, J.I., Vicente‐Serrano, S.M., Beguería, S., García‐Ruiz, J.M., Portela, M.M., &
Almeida, A.B. (2009). Dam effects on droughts magnitude and duration in a transboundary
basin: The Lower River Tagus, Spain and Portugal. Water Resources Research, 45, DOI:
10.1029/2008WR007198.
Loureiro, A, Ferrand de Almeida, N., Carretero, M.A., & Paulo, O.S. (2010). Atlas dos Anfíbios
e Répteis de Portugal. 2ª edição, Instituto da Conservação da Natureza e da Biodiversidade,
Lisbon.
Maceda-Veiga, A. (2013). Towards the conservation of freshwater fish: Iberian Rivers as an
example of threats and management practices. Review in Fish Biology and Fisheries, 23; 1–
22.
Mantel, S.K., Rivers‐Moore, N., & Ramulifho, P. (2017). Small dams need consideration in
riverscape conservation assessments. Aquatic Conservation: Marine and Freshwater
Ecosystems, DOI: 10.1002/aqc.2739.
26
McRae, L., Freeman, R., Deinet, S. (2014). The Living Planet Index in: Living Planet Report
2014: species and spaces, people and places [McLellan, R., Iyengar, L., Jeffries, B. and N.
Oerlemans (Eds)]. WWF, Gland, Switzerland.
Moilanen, A., Leathwick, J., Elith, J. (2008). A method for freshwater conservation
prioritization. Freshwater Biology, 53, 577–592.
O’Hanley, J.R., Wright, J., Diebel, M., Fedora, K.A., & Soucy, C.L. (2013). Restoring stream
habitat connectivity: A proposed method for prioritizing the removal of resident fish passage
barriers. Journal of Environmental Management, 125, 19e27.
Peterson, E.E., Hoef, J.M.V., Isaak, D.J., Falke, J.A., Fortin, M.-J., Jordan, C.E., … Wenger,
S.J. (2013). Modelling dendritic ecological networks in space: an integrated network
perspective. Ecology Letters, 16, 707-719.
Pleguezuelos, J.M., Márquez, R., & Lizana, M. (2002). Atlas y Libro Rojo de los Anfibios y
Reptiles de España. Dirección General de Conservación de la Naturaleza-Asociación
Herpetologica Española. Madrid, Spain.
Pringle, C.M. (2001). Hydrologic connectivity and the management of biological reserves: a
global perspective. Ecological Applications, 11, 981–998.
Roberts, J.H., Angermeier, P.L., Hallerman, E.M. (2013). Distance, dams and drift: what
structures populations of an endangered, benthic stream fish? Freshwater Biology, 58, 2050-
2064.
Rogado, L., Alexandrino, P.J., Almeida, P.R., Alves, M.J., Bochechas, J., Cortes, R.V., ...
Magalhães, M.F. (2005). Peixes. Livro Vermelho dos Vertebrados de Portugal. Pages 63–
114 in Cabral MJ, Almeida J, Almeida PR, Dellinger T, Ferrand de Almeida N, Oliveira ME,
Palmeirim JM, Queirós AI, Rogado L, Santos-Reis M, Editors. Instituto da Conservação da
Natureza, Lisbon.
Rolls, R.J., Ellison, T., Faggotter, S., & Roberts, D.T. (2013). Consequences of connectivity
alteration on riverine fish assemblages: potential opportunities to overcome constraints in
27
applying conventional monitoring designs. Aquatic Conservation: Marine and Freshwater
Ecosystems, 23, 624–640.
Segurado, P., Branco, P., Avelar, A.P., Ferreira, M.T. (2015). Historical changes in the
functional connectivity of rivers based on spatial network analysis and the past occurrences
of diadromous species in Portugal. Aquatic Sciences, 77, 427–440.
Thieme, M.L., Sindorf, N., Higgins, J., Abell, R., Takats, J.A., Naidoo, R., Barnett, A. (2016).
Freshwater conservation potential of protected areas in the Tennessee and Cumberland River
Basins, USA. Aquatic Conservation: Marine and Freshwater Ecosystems, 26, 60–77.
Vörösmarty, C.J., McIntyre, P.B., Gessner, M.O., Dudgeon, D., Prusevich, A., Green, P., …
Davies, P.M. (2010). Global threats to human water security and river biodiversity. Nature,
467, 555–561.
Ward, J.V. (1989). The four-dimensional nature of lotic ecosystems. Journal of the North
American Benthological Society, 8, 2–8.
Winemiller, K.O., McIntyre, P., Castello, L., Fluet-Chouinard, E., Giarrizzo, T., Nam, S., …
Sáenz, L. (2016). Balancing hydropower and biodiversity in the Amazon, Congo, and
Mekong. Science, 6269, 128-129.
Zarfl, C., Lumsdon, A.E., Berlekamp, J., Tydecks, L., & Tockner, K. (2015). A global boom in
hydropower dam construction. Aquatic Sciences, 77, 161-170.
Zheng, P.Q., & Hobbs, B.F. (2013). Multiobjective portfolio analysis of dam removals
addressing dam safety, fish populations and cost. Journal of Water Resources Planning and
Management, 139, 65-75.
