ArticlePDF Available

An accessible optimisation method for barrier removal planning in stream networks

Authors:

Abstract

Barriers associated to human infrastructure are a widespread impact in freshwater ecosystems worldwide, disrupting connectivity along river networks and key processes. Restoration of connectivity has risen in the last decade, with thousands of dams, weirs and culverts removed. Spatial optimisation methods can help inform decision on what barriers to remove to maximise gain in connectivity under limited budgets. However, current optimisation approaches rely on programming skills that are not easily accessible to stakeholders, which restrict the use of these methods. We demonstrate how Marxan, a publicly available tool, can be used to prioritise the allocation of barrier removal projects. We mapped the distribution of >900 barriers in the Tagus River (Iberian Peninsula) and 29 freshwater fish species with different movement abilities and needs. We assessed the passability of each barrier by all species and relative removal cost. We then identified priority barriers for removal to increase connectivity of populations of all species simultaneously. We tested two alternative scenarios: i) locking out barriers assesses as non-removable for their high strategic value or removal cost and ii) making all barriers available for removal. We found that connectivity recovery targets could be achieved by removing a small proportion of barriers, and avoiding large infrastructure. However, for some species, large recovery targets could only be achieved by removing some of these large infrastructures at high increases in cost. We also found some spatial differences in the recovery value of particular barriers for improving upstream and downstream connectivity. Our study demonstrates how to use a robust optimisation approach in an accessible tool, to address the complexity of prioritisation exercises commonly faced by stakeholders when deciding where to invest in barrier removal projects. This will improve decision-making for river connectivity restoration through a transparent, reproducible, and better-informed approach than traditional opportunistic or ranking-based approaches.
Hermoso, V., Clavero, M., Filipe, A.F. (2020). An accessible optimisation method for barrier
removal planning in stream networks. Science of the Total Environment, 752, 141943
1
An accessible optimisation method for barrier removal
planning in stream networks
Virgilio Hermoso1,2, Miguel Clavero3, Ana Filipa Filipe4,5,
1 Centre de Ciència i Tecnologia Forestal de Catalunya (CTFC), Solsona, Lleida (Spain)
2 Australian Rivers Institute, Griffith University, Nathan, Queensland (Australia)
3 Estación Biológica de Doñana (EBD-CSIC), Sevilla (Spain)
4 CIBIO/InBio, Centro de Investigação em Biodiversidade e Recursos Genéticos, Instituto
Superior de Agronomia, Universidade de Lisboa, Lisboa, Portugal
5 Centro de Estudos Florestais, Instituto Superior de Agronomia, Universidade de Lisboa,
Lisboa, Portugal
Hermoso, V., Clavero, M., Filipe, A.F. (2020). An accessible optimisation method for barrier
removal planning in stream networks. Science of the Total Environment, 752, 141943
2
Abstract
Barriers associated to human infrastructure are a widespread impact in freshwater ecosystems
worldwide, disrupting connectivity along river networks and key processes. Restoration of
connectivity has risen in the last decade, with thousands of dams, weirs and culverts removed.
Spatial optimisation methods can help inform decision on what barriers to remove to maximise
gain in connectivity under limited budgets. However, current optimisation approaches rely on
programming skills that are not easily accessible to stakeholders, which restricts the use of these
methods.
We demonstrate how Marxan, a publicly available tool, can be used to prioritise the allocation
of barrier removal projects. We mapped the distribution of >900 barriers in the Tagus River
(Iberian Peninsula) and 29 freshwater fish species with different movement abilities and needs.
We assessed the passability of each barrier by all species and relative removal cost. We then
identified priority barriers for removal to increase connectivity of populations of all species
simultaneously. We tested two alternative scenarios: i) locking out barriers assesses as non-
removable for their high strategic value or removal cost and ii) making all barriers available for
removal.
We found that connectivity recovery targets could be achieved by removing a small proportion
of barriers, and avoiding large infrastructure. However, for some species, large recovery targets
could only be achieved by removing some of these large infrastructures at high increases in cost.
We also found some spatial differences in the recovery value of particular barriers for
improving upstream and downstream connectivity.
Our study demonstrates how to use a robust optimisation approach in an accessible tool, to
address the complexity of prioritisation exercises commonly faced by stakeholders when
deciding where to invest in barrier removal projects. This will improve decision-making for
river connectivity restoration through a transparent, reproducible, and better-informed approach
than traditional opportunistic or ranking-based approaches.
Keywords: connectivity, dam, fish, Marxan, prioritisation, weir.
Hermoso, V., Clavero, M., Filipe, A.F. (2020). An accessible optimisation method for barrier
removal planning in stream networks. Science of the Total Environment, 752, 141943
3
Introduction
The disruption of connectivity caused by artificial barriers in rivers is considered a major threat
to freshwater ecosystems at the global scale and a priority for future management to halt the
decline of freshwater biodiversity (Tickner et al., 2020). River damming has severely
fragmented river ecosystems worldwide, with only one third of them lacking large dam
regulation (Grill et al., 2019). The magnitude of the problem escalates when smaller
infrastructure, like small hydropower dams, weirds or culverts, are considered (Lange et al.,
2019; Jones et al., 2019; Januchowski-Hartley et al., 2013). The cumulative effect of thousands
of these small infrastructures can cause even greater impact than the implementation of a single
large infrastructure on freshwater ecosystems and their biodiversity.
Despite the deep impacts caused by artificial barriers, freshwater ecosystems have shown rapid
response to barrier removal (O’Connor et al., 2015; Roni et al., 2002), recovering pre-damming
geomorphological conditions, and fish diversity and abundance soon after the restoration of
natural connectivity (Catalano et al., 2007; Gardner et al. 2013). For this reason, the design and
implementation of barrier removal projects has received increasing attention in the last two
decades (McManamay et al., 2019; Bellmore et al. 2017; East et al., 2015). For example, over
1000 dams have been removed in the USA (O’Connor et al., 2015). The European Union has
recently committed to restoring connectivity in 25,000 km of rivers in the next decade as part it
its 2030 Biodiversity Strategy (EC, 2020).
However, barrier removal projects can be expensive and can involve important opportunity
costs associated to the socio-economic benefits of dams and reservoirs (e.g., Kraft et al., 2019).
Therefore, restoration of natural connectivity needs to be carefully planned to maximise the
recovery benefit while minimising socio-economic trade-offs (Zheng & Hobbs, 2013; Erős et
al., 2018), ideally at the whole catchment scale (Hermoso et al., 2012). Although opportunistic
decision-making scales can leverage local interests (Neeson et al., 2015), it is only at large
scales that the cumulative impact of barriers can be tackled in an efficient way (O’Hanley &
Tomberlin 2005; Kemp & O’Hanley 2010; Segurado et al., 2013). Multiple approaches have
been proposed for prioritising barrier removal projects at large scales, many of which use
scoring approaches based on physical, economic, or ecological attributes of each barrier (Kraft
et al., 2019). Despite being simple and straightforward, this approach overlooks cumulative
benefits from removing multiple barriers at a time and, therefore, hinders efficiency (Kemp &
O´Hanley, 2010). Spatial optimisation approaches have gained attention recently (O´Hanley,
2011, O´Hanley et al., 2013) as they are able to deal with the complexity of exploring the
multitude of possible combinations of barrier removals, and their joint costs and benefits (see
Kemp and O´Hanley, 2010 for a comprehensive review on scoring vs. optimisation approaches).
However, most of these optimisation methods are complex and require mathematical and
Hermoso, V., Clavero, M., Filipe, A.F. (2020). An accessible optimisation method for barrier
removal planning in stream networks. Science of the Total Environment, 752, 141943
4
programming skills that are frequently inaccessible to stakeholders and decision-makers. There
is thus an urgent need for new tools that help address the complex decision-making with robust
optimisation methods, but more accessible to stakeholders (McManamay et al., 2019; King et
al., 2017).
In this study we demonstrate how to use Marxan, a freely accessible spatial prioritisation tool
commonly used in conservation planning assessments, for prioritising barrier removal projects
at the catchment scale. We show how to integrate commonly reported needs when planning
barrier removal, such as i) barrier passability and native species distribution -to calculate
potential benefits associated to the removal of each barrier-; ii) feasibility of removal (e.g., large
dams that could not be realistically removed because of their strategic value) and relative costs
for those that could be removed; iii) different types of species movements along the catchment
(e.g., upstream for anadromous or upstream-downstream for potamodromous species); iv) and
spatial relationships among barriers along the river network, to address connectivity in the
dendritic structure of rivers. We use the Tagus River catchment in the Iberian Peninsula as a
case study. We mapped the distribution of more than 900 barriers including a broad range of
infrastructure types, from large dams to small weirs, assessing their passability for different fish
species and their relative removal cost. We then identified a minimum set of barriers needed to
achieve re-connection targets, at minimum cost, for all freshwater species inhabiting the
catchment, including catadromous species (e.g., the European eel, Anguilla anguilla),
anadromous species (e.g. shads, Alosa alosa and Alosa fallax), and potamodromous species
(e.g., four Luciobarbus barbel species). The user friendly tool, publicly available, and guidance
provided in this study warrants the accessibility to a broad group of stakeholders in need of
robust decision-making.
Methods
Study area and data compilation
The Tagus River is the largest in the Iberian Peninsula, with more than 1,000 km in length and
an average annual flow of 300 m3×s-1and one of the largest catchments in the European Atlantic
coast, with approximately 80,000 km2.
We gathered information on the distribution of 29 fish species (Table 1) across the whole Tagus
River catchment at 100 km2 resolution, based on the Portuguese Red Book Project (Rogado et
al., 2005) and the most recent atlas from Spain (Doadrio, 2002) and a database built by Filipe et
al. (2009), representing the most complete information on the distribution for the catchment
(Hermoso et al., 2015b). This dataset included distribution for species with a broad range of
biological life cycles and swimming capacities (Table 1). We assessed the capacity of each
Hermoso, V., Clavero, M., Filipe, A.F. (2020). An accessible optimisation method for barrier
removal planning in stream networks. Science of the Total Environment, 752, 141943
5
species to pass different types of barriers based on their swimming and jumping abilities (see
also Rincon et al., 2017), and classified them into three broad categories: high, medium and low
(Table 1).
We identified and mapped the distribution of barriers that were observable by overlapping the
HydroSHEDS river network (Lehner et al., 2008) and Google Earth Pro 7.3 for the Tagus River
catchment. The HydroSHEDS river network was built from a 15 arcsec resolution digital
elevation model and covers all permanent reaches and a large proportion of temporal streams in
the catchment with a minimum upstream area of 25 km2 (Lehner et al., 2008). We then snapped
the location of each barrier to the nearest river in ArcGIS 10 (ESRI, 2011) to ensure spatial
coherence of barrier locations along the river network for further analyses. Each of these
barriers was further assessed for their relative passability by each type of species previously
described, and the relative removal cost. Given the lack of information on effectiveness of
different treatment costs (e.g., construction of fish ladders), we just considered a potential action
for each barrier that implied its complete removal. Under this planning scenario, we could
assume that once treated all movement restrictions to those species that were affected by a given
barrier would disappear due to natural stream connectivity restoration. Each barrier was
classified into a broad passability class, according to estimates of the height of the
infrastructure, presence of water spilling over it, or presence of breaks that could facilitate some
species to pass (Supplementary Table 1). In this way, barriers were classified in one of four
categories, as passable by all species, passable by species with at least medium or high
capacities, or completely impassable.
Given the lack of accurate estimates of removal cost we classified barriers into four broad
categories: low, medium, high, very high. Barriers were classified into one of these cost-classes
according to criteria related to size, construction material and conservation status. In this way,
the smaller structures, like small weirs built with not consolidated material gathered from the
river bed were classified in the low cost class; larger weirs, or other infrastructure, like gauging
stations build with concrete were classified in the medium cost class; small dams were mainly
included in the high cost class; and large dams and hydropower stations, were included in the
very high cost class. We then translated this qualitative assessment into a semi-quantitative
value using the low cost class as the reference for cost-units, and applied a logarithmic increase
in cost-units as we escalated in categories. In this way, removal cost of a small weir would be
one cost-unit, while removal of a large dam would be 1000 cost-units. These broad estimates of
cost were used with demonstration purposes only, and further assessments on costs for each
barrier would be needed to better inform a realistic barrier removal plan for this catchment. We
also classified barriers as removable or not, to further account for potential constraints of
restoration plans to tackle removal of large or strategic infrastructure which could not be
Hermoso, V., Clavero, M., Filipe, A.F. (2020). An accessible optimisation method for barrier
removal planning in stream networks. Science of the Total Environment, 752, 141943
6
realistically implemented. Under the not-removable category we included mainly large dams
and hydropower stations with a high opportunity cost associated.
Spatial prioritisation of barrier removal
We used the software Marxan (Ball et al., 2009) to identify a set of priority barriers for
improving connectivity between populations of all fish species in the Tagus River catchment.
Marxan is a spatial planning tool, commonly used for identifying priority areas for conservation
of biodiversity, aiming to identify a minimum set of areas that cover the distribution of all
conservation features under consideration (e.g., species, habitats, ecoregions, or ecosystem
services) at minimum cost. To do that, Marxan uses a heuristic optimisation algorithm to
minimize an objective function (3) that includes the cost of planning units in the solution and
other penalties for not achieving the desired spatial coverage (conservation targets) for all the
conservation features and spatial constraints, such as connectivity among selected planning
units (Hermoso et al., 2011).
In our case, each barrier was treated as a planning unit. Each barrier had, therefore, a potential
benefit and cost associated, like spatial planning units in traditional Marxan applications. The
benefit associated to each barrier, if selected for removal, was the length of each species´
distribution that would be reconnected to surrounding river reaches from which they were
previously isolated by the barrier (Fig. 2). We accounted for the species-specific estimate of
barrier passability when calculating this benefit. Whenever a barrier was considered passable for
a group of species, we did not consider that these species would benefit from the removal of that
barrier and, therefore, were not accounted for when calculating as a benefit if the barrier was
removed. In this way, we avoided overestimating benefits of barrier removals and focus on the
identification of barriers that were truly effective.
We considered asymmetrical benefits of barrier treatment in both directions, upstream and
downstream movement (Rincon et al., 2017). To do that, for each species we created two
pseudo-species, one for each direction of movement, and calculated the benefit for each pseudo-
species independently. So, for a barrier that was considered impassable for a given species, the
removal would have an upstream benefit, measured as the length of river upstream of the barrier
occupied by the species, and a downstream benefit, as the length of the river downstream of the
barrier occupied by the species (Fig. 3). We set independent targets for each pseudo-species,
avoiding in this way, double accounting a reach as upstream and downstream benefit for the
same species if two consecutive barriers were treated. For the sake of demonstration, we used
the same passability values in both directions. So, if a barrier was considered passable by a
species, it was equally passable upstream and downstream. We accounted for these different
Hermoso, V., Clavero, M., Filipe, A.F. (2020). An accessible optimisation method for barrier
removal planning in stream networks. Science of the Total Environment, 752, 141943
7
mobility needs when setting reconnection targets, using larger targets for highly mobile pseudo-
species than pseudo-species that do not move much at the planning scale. The reconnection
target was expressed as the proportion of the distribution of each pseudo-species currently
affected by barriers that would benefit from barrier removal (e.g., length of river affected by a
given barrier that would benefit from barrier removal). In this way, we set a 50% target for long
migratory species (anadromous, catadromous and amphidromous), a 25% target for
potamodromous species that display medium-long migrations and a 10% target for resident
species that only perform short distance movements (Table1). Reconnection targets should,
therefore, be interpreted as proportion of the distribution of each species currently disconnected
that would benefit from the removal of barriers.
Therefore, our optimisation problem was:

    

 

  
where, xi is a control variable that takes a value of 1 when the barrier i is selected and 0
otherwise; i belongs to the group of m barriers in the Tagus River catchment; ci is the cost-units
of removing barrier i; ai is the benefit for each pseudo-species j provided by each barrier if it
was removed i (measured as the length of river occupied by each species in the river segment
either upstream or downstream to each barrier i in our case);  is normally the penalty for
missing the connection between a given pair of planning units (i1 and i2) in the solution, and
weighted by b, a connectivity strength modifier (CSM); and tj is the target for each pseudo-
species.
We used the connectivity penalty feature in Marxan to aggregate barrier removal projects along
the river network as a way to maximise connectivity of fish populations. We used Hermoso et
al. (2011) recommendations on how to address longitudinal connectivity in Marxan for river
applications. In our case, we built a connectivity matrix containing all connections of barriers
along the river network (Fig. 2). In traditional Marxan applications in rivers, these connections
have a penalty associated that is calculated as a function of the inverse of the distance between
each pair of planning units. In this way, the penalty in the objective function for not including a
close neighbour is higher than the penalty for not including a long-distance planning unit. This
helps achieving cluster of connected planning units and, therefore, longitudinal connectivity
along groups of planning units selected (Hermoso et al., 2011). In our case, instead of distance
between planning units as normally done, we used the number of barriers in between each
pairwise combination of barriers as a penalty. We calculated the penalty between each pair of
Hermoso, V., Clavero, M., Filipe, A.F. (2020). An accessible optimisation method for barrier
removal planning in stream networks. Science of the Total Environment, 752, 141943
8
barriers as the inverse of the squared number of barriers in between plus one (to avoid 0s for
contiguous barriers) [penalty=1/(number of barriers in between +1)2]. Therefore, for two
contiguous barriers, the penalty was one, while for two barriers with one another in between the
penalty was 0.25. In this way, we tried to avoid potential conflicts between the benefit and
connectivity in Marxan´s objective function. Distant, but consecutive barriers along a river
reach could pose a long stretch of river with continuous habitats for fish (high benefit), but low
connectivity penalty if calculated based on distance and, therefore, low priority for connectivity.
This would end up in contradictory decisions, such as selecting barriers that release long river
reaches, but not their contiguous neighbours. By only considering the number of barriers in
between, we made them virtually “closer” to each other, and therefore, fostered the selection of
consecutive barriers regardless their distance. We calibrated the CSM as recommended in
Ardron et al. (2010), resulting in a value of 20 (Supplementary Fig. 1).
Under these premises the objective function that we tried to minimise was as follows:

    

 


where there are n pseudo-species under consideration; SPFj is a Species Penalty Factor or
weighting factor that applies for not achieving the desired representation target for each pseudo-
species j; H(s) is a Heaviside function that takes a value of 0 when s/tj≤0 and 1 otherwise; s is
the shortfall in targets not achieved and is measured as tj-representation achieved; the ratio s/tj
equals 1 when the pseudo-species j is not represented within the solution and approaches 0 as
the level of representation approaches the target amounts (tj). We used a constant SPF=100 for
all pseudo-species to ensure they all achieved the desired targets. With this configuration we ran
Marxan 100 times, 10 million iterations each across all analyses and kept for subsequent
comparison the best solution out of those 100.
Barrier removal planning scenarios
We explored three different barrier removal planning scenarios. In a first scenario (all barriers
scenario, hereafter) we identified priority barriers for removal accounting for both, upstream and
downstream movements, and no limitations to potential barriers to be removed (all barriers
where available, regardless their status). A second scenario replicated the previous, but locking
out all barriers that had been assessed as not available for removal (Locked-out barriers
scenario, hereafter). A third set of scenarios replicated the locked-out barriers scenarios but
considering only either upstream or downstream movement needs respectively. For these
scenarios, we set reconnection targets for downstream and upstream pseudo-species to 0
respectively.
Hermoso, V., Clavero, M., Filipe, A.F. (2020). An accessible optimisation method for barrier
removal planning in stream networks. Science of the Total Environment, 752, 141943
9
We carried out a sensitivity analysis on targets across all scenarios by applying a multiplier
factor to the general targets mentioned above for pseudo-species. In total we ran the analyses for
39 different targets, from 0.1 to 2 at 0.05 increments. To evaluate the effect of locking out
barriers on the reconnection benefit and efficiency of solutions, we compared results between
all-barriers and locked-out barriers scenarios. The benefit of a given solution was measured as
the proportion of the total distribution of each species that did not have a barrier disconnecting it
(Eq. 1).



 
where,
 is the total length occupied by each species in the catchment and  is the
length of a given species that would be reconnected after removal of barrier i. So, for example,
if we had a species separated in 4 disconnected populations by three barriers of equal length
each and one of the barriers was removed, the overall connectivity for that species would be
50% of the total potential if all three barriers were removed.
To assess the relative value of different barriers either for the improvement of either upstream or
downstream mobility, we checked for spatial differences in the set of barriers selected under the
upstream and downstream movement scenarios respectively. We would expect the spatial
allocation of priority barriers for removal to vary depending on the scenario tested.
Results
Half of the 934 barriers identified in the Tagus River catchment were assessed as impassable by
any species (N=471), and an additional 28.4% passable only by species with high capacity
(N=256). From the remaining 21.1%, 127 barriers were assessed as passable by species with
medium capacity, 71 barriers as passable by all species. This translated into an overall
connectivity assessment across all species of 46% (Table 1).
The number of barriers selected increased linearly with the targets from only 23 at the lowest
target level to 380 to the highest under the locked-out barriers scenario (Fig. 4). The gain in
connectivity, however, did not follow the same pattern with a two-fold increase in connectivity
after target 0.55 that was not related to a similar increase in the number of barriers selected (Fig.
4). Despite the overall gain in connectivity, some of the pseudo-species did not achieve the
desired targets under the locked-out scenario for the highest target levels (Supplementary Fig.
2), although all pseudo-species were close to the desired target, with average achievement over
90% (Supplementary Fig. 2).
Hermoso, V., Clavero, M., Filipe, A.F. (2020). An accessible optimisation method for barrier
removal planning in stream networks. Science of the Total Environment, 752, 141943
10
There was a high similarity in both the number of barriers selected and their spatial allocation,
under the all-barriers and the locked-out barriers scenarios, regardless the target used
(Supplementary Fig. 3). Although all barriers were available for removal under the all-barriers
scenario, only two large dams were selected for removal in the all-barriers scenario (Fig. 5),
probably due to the high cost associated to large infrastructures (1000 cost units).
The selection of barriers considered as non-removable under the all-barriers scenario, increased
with targets, from none of them selected for low targets, but up to 17 large barriers selected
under the largest target (Supplementary Fig. 4). All pseudo-species achieved the desired targets
across all target levels in this case. Therefore, the lack of target achievement under the locked-
out scenario was related to the impossibility to remove some barriers. However, the increase in
the number of large barriers, assessed as non-removable, selected in solutions posed a steep
increase in cost of solutions for increasing targets under the all-barriers scenario (Supplementary
Fig. 4).
The similarity of solutions under the upstream and downstream scenarios increased with
increasing targets (Supplementary Fig. 5). Barriers selected under the upstream scenario were
mainly located in the headwaters of the Sorraia River and the upper Tagus River catchment. In
most of cases, these barriers were allocated in rivers upstream non-removable barriers. On the
other hand, barriers selected under the downstream scenario were more dispersed across the
catchment and mainly allocated close to headwaters free from non-removable barriers
(Supplementary Fig. 5).
Discussion
We have demonstrated how to use Marxan (Ball et al., 2009), a freely available spatial
prioritisation tool commonly used for conservation planning, to prioritise the allocation of
barrier removal projects across a whole catchment. We integrated different features claimed
important when deciding where to invest in barrier removal, such as i) the relative passability of
each barrier by different species (Rincon et al., 2017), ii) the directionality of fish movements
across barriers; iii) the simultaneous benefits for multiple species associated to the removal of
each barrier; iv) the costs associated to removals; v) the feasibility of removal; and vi) the
connectivity of barriers within the river network. Some of these features had been addressed
before individually, but to our knowledge this is the first study that integrates them all in a
single prioritisation exercise to make the analysis and results more comprehensive. In this way,
we have overcome traditional limitations of assessments of barriers removal based on simple
structural connectivity, typically focused on creating the longest possible river reach connected,
not informed by the distribution of species in them (O´Hanley, 2011; Segurado et al., 2013) or
Hermoso, V., Clavero, M., Filipe, A.F. (2020). An accessible optimisation method for barrier
removal planning in stream networks. Science of the Total Environment, 752, 141943
11
exercises based on a single species (Kuby et al., 2005; Branco et al., 2014; Ioannidou and
O´Hanley, 2019). More importantly, our study demonstrates how to use a robust optimisation
approach to address the complexity of prioritisation exercises commonly faced by stakeholders
when deciding where to invest in barrier removal projects, only approached before through
complex self-coded optimisation algorithms.
River restoration has received increasing attention over the last decades as a response to the
poor conservation status of freshwater ecosystems (Roni et al., 2008; Hermoso et al., 2012). In
this context barrier removal is gaining momentum, with thousands of dams and weir ved
(O’Connor et al., 2015), or the commitment made by the EU to restore connectivity in at least
25,000 km of river in its new Biodiversity Strategy 2030. Barrier removal is also a priority to
ensure that freshwater ecosystem services and biodiversity do not continue declining under the
impacts of climate change. The combination of barriers and declining flows and water
availability compromise longitudinal connectivity among populations of fish and other strict
freshwater species (Branco et al., 2017), or completely impede the completion of the life cycles
of migratory fish (Clavero & Hermoso, 2015). The hierarchical geometry of freshwater
ecosystem makes them different from other spatially structured habitats and more sensitive to
fragmentation (Campbell-Grant et al., 2007). For species that are restricted to network branches,
population stability and local extinction risk are highly sensitive to connections among branches
(Labonne et al., 2008). Therefore, maximizing the longitudinal connectivity is vital to improve
the likelihood of metapopulation persistence in freshwater systems (Fagan 2002). However, the
strategic value of freshwater resources for human development (Holland et al., 2015) in
combination with limited resources available for river restoration (Hermoso et al., 2012) forces
stakeholders to carefully plan how to invest these resources. Decisions on how to use these
limited resources are normally not easy to make when planning for cumulative impacts at whole
catchment scale. The magnitude of the problem increases when accounting for multiple species
or the spatial dependencies among barriers (O’Hanley & Tomberlin, 2005; Branco et al., 2014;
Eros et al., 2018). For this reason, decision-making in this context must be supported by
optimisation methods (Eros et al., 2018). However, the use of optimisation approaches proposed
so far rely on programming skills and solvers like CPLEX (www.cplex.com), not easily
accessible to most of stakeholders. Thereby the need for new approaches that give stakeholders
the opportunity to access robust optimization routines. Marxan is a freely available tool with
training materials that, in combination with the guidelines provided in this study, should
facilitate future applications of optimisation methods to inform decision-making when planning
barrier removal, and overcome the issues associated to ad-hoc or opportunistic decisions
(Mckay et al., 2017; Hermoso et al., 2012). Marxan does require some data preparation that can
Hermoso, V., Clavero, M., Filipe, A.F. (2020). An accessible optimisation method for barrier
removal planning in stream networks. Science of the Total Environment, 752, 141943
12
also be carried out in freely accessible GIS like QGIS (QGIS.org, 2020), or spreadsheets [see
Game and Grantham (2008) for more detail on how to prepare Marxan´s input files].
Marxan has the potential to deal with prioritisation problems more complex than the ones used
here for demonstration, with 934 planning units (barriers) and 60 conservation features (pseudo-
species), like the prioritisation of barrier removal in larger catchments or with more obstacles
mapped (e.g., Januchowski et al., 2013; Jones et al., 2019). We were also able to tackle spatial
dependencies among barriers within the objective function, usually reported as a difficult
feature to code within a programming environment and important limitation (McManamay et
al., 2019). This tool has been used to inform decision-making in more complex scenarios with
2-3 orders of magnitude of more planning units and species with optimal results at no expenses
of quality of solutions or processing time. Although Marxan uses a heuristic optimisation
algorithm, solutions have been found to be optimal or close to the optimal when compared to
optimisation algorithms based on integer linear programming (Beyer et al., 2012).
For the particular case study in the Tagus River catchment, we found that there was not a direct
translation between the number of barriers removed and the improvement of connectivity. We
found a tipping point at target 0.55, when for a little increase in the number of barriers selected
for removal there was disproportionate improvement in connectivity across species. Previous
studies had found similar results with large increments in connectivity for the first few barriers
removed, at low cost, but quickly reaching a plateau after which small improvements in
connectivity could only be achieved at large expenses (e.g., O´Hanley et al., 2013). Despite this
overall gain in connectivity at increasing targets, we found that some pseudo-species did not
reach the desired target at the highest target levels under the locked-out scenario. The
impossibility to achieve the targets for these species was due to the presence of barriers that
were assessed as non-removable, since all pseudo-species did achieve the targets when all
barriers were made available under the all-barriers scenario, although at a high cost. Large
targets achievement under the all-barriers scenario could only be achieved by removing large
dams, that came to a high cost. Target achievement for all species under the all-barriers scenario
showed a 3.5-fold increase in cost for the highest target level compared to the locked-out
scenario. We also found different priorities when planning for improvement of connectivity for
upstream or downstream movements. Barriers close to headwaters of large river sections free
from non-removable barriers, several of them connecting these headwaters with the estuary and
allowing movements for long migratory species (e.g., eels or shads) or amphidromous species
(e.g., mullets). On the other hand, some barriers close to the headwaters in tributaries upstream
from non-removable barriers were selected when prioritising upstream movements, for
potamodromous species (e.g., barbels). The combination of both, the removal of barriers that
connect the estuary with a river network before unavailable for migratory species in the lower
Hermoso, V., Clavero, M., Filipe, A.F. (2020). An accessible optimisation method for barrier
removal planning in stream networks. Science of the Total Environment, 752, 141943
13
part of the catchment and barriers closer to the headwaters in the upper part of the catchment,
respond to the different ecological needs in the study area. Although the full reconnection from
mouth to headwaters was not attainable, given the presence of several large dams that
realistically would be difficult to remove, the solutions reported here enhance access to areas
important for completing life cycles of the species considered.
Our study demonstrates how a freely accessible source of information, such as Google Earth can
be used for identifying not only large barriers, such as dams, but also smaller infrastructure like
ponds, weirs or road passages. Future prioritisation exercises can also use publicly available
datasets of barriers like the Global Reservoir and Dam (GRanD; Lehner et al., 2011) or datasets
with smaller structures (Januchowsky-Hartley et al., 2013), which can also benefit from citizen
science (e.g., Amber in the EU; Garcia de Leaniz et al., 2018). The dataset that we used in this
exercise was compiled using standardised criteria across the whole catchment (e.g.,
Supplementary Table 2). However, the estimates of passability, removability or removal cost of
barriers in this study were used for demonstration purposes and further evaluation with ground
truth validation (Bourne et al., 2011) or modelling (Januchowsky-Hartley et al., 2014) would be
needed to make more robust decisions. Upstream passability depends on barriers´ physical
structure such as height, slope or upper sill, while downstream passability related to
characteristics of spillways or pipes, when present (Rincon et al., 2017), and these features
could only be roughly estimated in this study. We also included a semiquantitative estimate of
removal cost that could be further refined field estimates based on the barrier location,
dimension or material. Investments associated to barrier removal projects, however, go beyond
the cost of the work needed to remove the physical structure and additional opportunity costs
derived from the socio-economic benefit (e.g., energy production, water retention for human
uses or recreational opportunities; Eros et al., 2018; Kraft et al., 2019) or even ecological
benefits (e.g., retention of spread of invasive species; Hermoso et al., 2015a). There are
unavoidable trade-offs between the recovery of natural connectivity for native species and these
other socio-economic or ecological benefits that need to be addressed when planning for barrier
removal (Eros et al., 2018). These trade-offs could be addressed in Marxan through the cost
parameter in the objective function. For example, if a barrier was important for reducing the
spread of invasive species, a high cost could be applied to the removal of such barrier. An
alternative approach to addressing multiple objectives can be explicitly done in Marxan with
Zones, also freely available, as demonstrated in Hermoso et al. (2018) in a planning exercise for
multiple objectives in a freshwater context. In this way, the multiplicity of interests could be
considered in the spatial prioritisation of barriers with different objectives, including the
maintenance of barriers for different purposes (e.g., hydropower production or confinement of
invasive species) and the removal of those that are necessary for the improvement of
Hermoso, V., Clavero, M., Filipe, A.F. (2020). An accessible optimisation method for barrier
removal planning in stream networks. Science of the Total Environment, 752, 141943
14
connectivity. To achieve this, each barrier would have different benefits associated according to
the objectives pursued, and a compromise needs to be found on what to do with each barrier to
achieve all objectives. Finally, the distribution of species used in this study depicts current areas
of occupancy that might have already been affected by habitat fragmentation and, therefore,
limiting our capacity to identify barriers as important simple due to underestimated areas of
occupancy. For example, the eel only occupies a small portion of its original distribution in the
catchment (Clavero & Hermoso, 2015). Therefore, future planning work would benefit from
better estimate of historical distribution of species along the catchment (e.g., Clavero et al.,
2018; Duarte et al., 2018).
Conclusions
With this exercise we aimed to provide new methods and tools to a broad range of stakeholders
who often lack programming skills, in order to help them to overcome the knowledge barrier
that limits the implementation of robust approaches for barrier removal planning. This will
improve decision-making for river connectivity restoration through a transparent, reproducible,
and better-informed approach.
Hermoso, V., Clavero, M., Filipe, A.F. (2020). An accessible optimisation method for barrier
removal planning in stream networks. Science of the Total Environment, 752, 141943
15
References
Ardron, J.A., Possingham, H.P., & Klein, C.J. (eds), 2010. Marxan Good Practices Handbook,
Version 2. Pacific Marine Analysis and Research Association, Victoria, BC, Canada. 165
pages. www.pacmara.org.
Bellmore, J.R., Duda, J.J., Craig, L.S., Greene, S.L., Torgersen, C. E., Collins, M. J., Vittum,
K., 2017. Status and trends of dam removal research in the United States. WIREs Water
4,e1164.
Bourne, C.M., Kehler, D.G., Wiersma, Y.F., Cote, D. (2011). Barriers to fish passage and
barriers to fish passage assessments: the impact of assessment methods and assumptions on
barrier identification and quantification of watershed connectivity. Aquat. Ecol. 45, 89403.
Branco, P., Segurado, P., Santos, J.M., & Ferreira, M.T., 2014. Prioritizing barrier removal to
improve functional connectivity of rivers. J. App. Ecol. 51, 11971206.
Campbell-Grant, E.H., Lowe, W.H., & Fagan, W.F. (2007). Living in the branches: Population
dynamics and ecological processes in dendritic networks. Ecol. Lett. 10, 165175.
Catalano, M.J., Bozek, M.A., Pellett, T.D., 2007. Effects of dam removal on fish assemblage
structure and spatial distributions in the Baraboo River, Wisconsin. N. Am. J. Fish. Manage.
27, 519530.
Clavero, M., Hermoso, V., 2015. Historical data to plan the recovery of the European eel. J.
App. Ecol. 52, 960968.
Clavero, M., Ninyerola, M., Hermoso, V., Filipe, A.F., Pla, M., Villero, D., Brotons, L.,
Delibes, M., 2017. Historical citizen science to understand and predict climate-driven trout
decline. P. Roy. Soc. B. 284, 20161979.
Duarte, G., Moreira, M., Branco, P., da Costa, L., Ferreira, M. T., Segurado, P., 2018. One
millennium of historical freshwater fish occurrence data for Portuguese rivers and streams.
Sci. Data. 5, 180163.
East, A.E., Pess, G.R., Bountry, J.A., Magirl, C.S., Ritchie, A.C., Logan, J.B., Randle, T.J.,
Mastin, M.C., Minear, J.T., Duda, J.J., Liermann, M.C., McHenry, M.L., Beechie, T.J.,
Shafrothi, P.B., 2015. Large-scale dam removal on the Elwha River, Washington, USA:
River channel and floodplain geomorphic change. Geomorphology 228,765786.
EC (2020). EU Biodiversity Strategy for 2030 Bringing nature back into our lives. COM(2020)
380 final. Available at: https://ec.europa.eu/info/files/communication-eu-biodiversity-
strategy-2030-bringing-nature-back-our-lives_en (last visited 21st May 2020).
Hermoso, V., Clavero, M., Filipe, A.F. (2020). An accessible optimisation method for barrier
removal planning in stream networks. Science of the Total Environment, 752, 141943
16
Erős, T., O’Hanley, J.R., Czeglédi, I., 2018. A unified model for optimizing riverscape
conservation. J. App. Ecol. 55, 18711883.
ESRI (2011). ArcGIS Desktop: Release 10. Redlands, CA: Environmental Systems Research
Institute.
Fagan, W.F., Unmack, P.J., Burgess, C., Minckley, W.L., 2002. Rarity, fragmentation and
extinction risk in desert fishes. Ecology 83, 32503256.
Game, E.T., Grantham, H.S., 2008. Marxan User Manual: For Marxan version 1.8.10.
University of Queensland, St. Lucia, Queensland, Australia, and Pacific Marine Analysis and
Research Association, Vancouver, British Columbia, Canada.
Garcia de Leaniz, C., Belletti, B., Bizzi, S., Segura, G., Borger, L., Jones, J., Olivo del
Amo, R., Wanningen, H., Tummers, J., Kerr, J., Kemp, P., van de Bund, W., 2018. The
importance of having a good database for restoring river connectivity: The AMBER barrier
atlas in Europe. In: Brink, K., Gough, P., Royte, J., Schollema, P.P., Wanningen, H. (Eds.),
From Sea to Source 2.0. Protection and Restoration of Fish Migration in
RiversWorldwide.World FishMigration Foundation, Groningen, pp. 142145.
Gardner, C., Coghlan, S., Zydlewski, J., Saunders, R., 2013. Distribution and abundance of
stream fishes in relation to barriers: implications for monitoring stream recovery after barrier
removal. River Res. App. 29, 6578.
Grill, G., Lehner, B., Thieme, M., Geenen B., Tickner D., Antonelli F., Babu, S., Borrelli, P.,
Cheng, L., Crochetiere, H., Ehalt Macedo, H., Filgueiras, R., Goichot, M., Higgins, J.,
Hogan, Z., Lip, B., McClain, M. E., Meng, J., Mulligan, M., Nilsson, C., Olden, J. D.,
Opperman, J. J., Petry, P., Reidy Liermann, C., Sáenz, L., Salinas-Rodríguez, S., Schelle, P.,
Schmitt, R. J. P., Snider, J., Tan, F., Tockner, K., Valdujo, P. H., van Soesbergen, A., Zarfl,
C., 2019. Mapping the world’s free-flowing rivers. Nature 569, 215221.
Hermoso, V., Cattarino, L., Linke, S., Kennard, M.J., 2018. Catchment zoning to enhance co-
benefits and minimise trade-offs between ecosystem services and freshwater biodiversity
conservation. Aquat. Conserv. 28, 1004-1014.
Hermoso, V., Januchowski-Hartley, S.R., Linke, S., 2015a. Systematic planning of
disconnection to enhance conservation success in a modified world. Sci. Total Environ. 536,
10381044.
Hermoso, V., Filipe, A.F., Segurado, P., Beja, P., 2015b. Filling gaps in a large reserve network
to address freshwater conservation needs. J. Environ. Manage. 161, 358-365.
Hermoso, V., Clavero, M., Filipe, A.F. (2020). An accessible optimisation method for barrier
removal planning in stream networks. Science of the Total Environment, 752, 141943
17
Hermoso, V., Linke, S., Possingham, H..P., Prenda, J., 2011. Addressing longitudinal
connectivity in the systematic conservation planning of fresh waters. Freshwater Biol. 56,
57-70.
Hermoso, V., Pantus, F., Olley, J., Linke, S., Mugodo, J., Lea, P., 2012. Systematic planning for
river rehabilitation: Integrating multiple ecological and economic objectives in complex
decision-making. Freshwater Biol. 57, 1-9.
Holland, R.A., Scott, K.A., Flörke, M., Brown, G., Ewers, R.M., Farmer, E., Kapos, V.,
Muggeridge, A., 2015. Global impacts of energy demand and global water resources of
nations. P. Natl. Acad. Sci. 112, 6707-6716.
Ioannidou, C., O’Hanley, J.R., 2019. The importance of spatiotemporal fish population
dynamics in barrier mitigation planning. Biol. Conserv. 231, 67-76.
Januchowski-Hartley, S.R., Diebel, M., Doran, P.J., McIntyre, P.B., 2014. Predicting road
culvert passability for migratory fishes. Divers. Distrib. 20, 14141424.
Januchowski-Hartley, S.R., McIntyre, P.B., Diebel, M., Doran, P.J., Infante, D.M., Joseph, C.
Allan, J.D., 2013. Restoring aquatic ecosystem connectivity requires expanding inventories
of both dams and road crossings. Front. Ecol. Environ. 11, 211217.
Jones, J., Börger, L., Tummers, J., Jones, P., Lucas, M., Kerr, J., Kemp, P., Bizzi, S.,
Consuegra, S., Marcello, L., Vowles, A., Belletti, B., Verspoor, E., Van de Bund, W.,
Gough, P., Garcia de Leaniz, C., 2019. A comprehensive assessment of stream fragmentation
in Great Britain. Sci. Total Environ. 673, 756762.
Kemp, P. S., O’Hanley, J. R., 2010. Procedures for evaluating and prioritizing the removal of
fish passage barriers: a synthesis. Fisheries Manag. Ecol. 17, 297322.
King, S., O’Hanley, J.R., Newbold, L.R., Kemp, P.S., Diebel, M.W., 2017. A toolkit for
optimizing fish passage barrier mitigation actions. J. App. Ecol. 54, 599611.
Kraft, M., Rosenberg, D.E., Null, S.E., 2019. Prioritizing stream barrier removal to maximize
connected aquatic habitat and minimize water scarcity. J. Am. Water Resour. Ass. 55, 382
400.
Kuby, M.J., Fagan, W.F., ReVelle, C.S., Graf, W.L., 2005. A multiobjective optimization model
for dam removal: an example trading off salmon passage with hydropower and water storage
in the Willamette basin. Adv. Water Resou. 28, 84555.
Labonne, J., Virginie, R., Parisi, B., Gaucherel, C., 2008. Linking dendritic network structures
to population demogenetics: The downside of connectivity. Oikos 117, 14791490.
Hermoso, V., Clavero, M., Filipe, A.F. (2020). An accessible optimisation method for barrier
removal planning in stream networks. Science of the Total Environment, 752, 141943
18
Lange, K., Wehrli, B., Åberg, U., Bätz, N., Brodersen, J., Fischer, M., Hermoso, V., Reidy
Liermann, C., Schmid, M., Wilmsmeier, L., Weber, C., 2019. Small hydropower goes
unchecked. Front. Ecol. Environ. 17, 256-258.
Lehner, B., Verdin, K., Jarvis, A., 2008. New global hydrography derived from spaceborne
elevation data. Eos Transactions 89, 9394.
Lehner, B., Reidy Liermann, C., Revenga, C., Vorosmarty, C., Fekete, B., Crouzet, P., Doll, P.,
Endejan, M., Frenken, K., Magome, J., Nilsson, C., Robertson, J.C., Rodel, R., Sindorf, N.,
Wisser, D.,.2011. Global Reservoir and Dam Database, Version 1 (GRanDv1): Dams,
Revision 01. Palisades, NY: NASA Socioeconomic Data and Applications Center (SEDAC).
Mckay, S.K., Cooper, A.R., Diebel, M.W., Elkins, D., Oldford, G., Roghair, C., Wieferich, D.,
2017. Informing watershed connectivity barrier prioritization decisions: a synthesis. River
Res. App. 33, 847862.
McManamay, R.A., Perkin, J.S., Jager, H.I., 2019. Commonalities in stream connectivity
restoration alternatives: an attempt to simplify barrier removal optimization. Ecosphere 10,
e02596.
O’Connor, J., Duda, J. Grant, G., 2015. 1000 dams down and counting. Science 348, 496497.
O’Hanley, J., 2011. Open rivers: barrier removal planning and the restoration of free-flowing
rivers. J. Environ. Manage. 92, 31123120.
O’Hanley, J.R., Tomberlin, D., 2005. Optimizing the removal of small fish passage barriers.
Environ. Model. Assess. 10, 8598.
O’Hanley, J.R., Wright, J., Diebel, M., Fedora, M., Soucy, C., 2013. Restoring stream habitat
connectivity: a proposed method for prioritizing the removal of resident fish passage
barriers. J. Environ. Manage. 125, 1927.
QGIS.org (2020). QGIS Geographic Information System. Open Source Geospatial Foundation
Project. http://qgis.org
Rincón, G., Solana-Gutiérrez, J., Alonso, C., Saura, S., García de Jalón, D., 2017. Longitudinal
connectivity loss in a riverine network: accounting for the likelihood of upstream and
downstream movement across dams. Aquat. Sci. 79, 573585.
Roni, P., Hanson, K., Beechie, T., 2008. Global review of the physical and biological
effectiveness of stream habitat rehabilitation techniques. N. Am. J. Fish. Manage. 28, 856
890.
Hermoso, V., Clavero, M., Filipe, A.F. (2020). An accessible optimisation method for barrier
removal planning in stream networks. Science of the Total Environment, 752, 141943
19
Roni, P., Beechie, T.J., Bilby, R.E., Leonetti, F.E., Pollock, M.M., Pess, G.R., 2002. A review
of stream restoration techniques and a hierarchical strategy for prioritizing restoration in
Pacific Northwest watersheds. N. Am. J. Fish. Manage. 22, 120.
Schiermeier, Q., 2018. Dam removal restores rivers. Nature 557, 290-291.
Segurado, P., Branco, P., Ferreira, M.T., 2013. Prioritizing restoration of structural connectivity
in rivers: a graph based approach. Landscape Ecol. 28, 12311238.
Tickner, D.; Opperman, J.; Abell, R.; Acreman, M.; Arthington, A.H.; Bunn, S.E., 2020.
Bending the curve of global freshwater biodiversity loss - An emergency recovery plan.
Bioscience 70, 330-342.
Zheng, P.Q., Hobbs, B.F., 2013. Multiobjective portfolio analysis of dam removals addressing
dam safety, fish populations, and cost. J. Water Res. Plan. Manage. 139, 6575.
Hermoso, V., Clavero, M., Filipe, A.F. (2020). An accessible optimisation method for barrier
removal planning in stream networks. Science of the Total Environment, 752, 141943
20
Table 1. List of species included in this study, their type of biological cycle and capacity to
overcome obstacles, based on their swimming and jumping abilities. The current connectivity
status of extant populations under the distribution of barriers in the catchment and their
passability, as assessed for this study, is also shown.
Species
Type of
biological cycle
Capacity to pass
obstacles*
Current
status
Achondrostoma arcasii
Resident
Medium
0.49
Achondrostoma oligolepis
Resident
Medium
0.52
Alosa alosa
Anadromous
High
0.28
Alosa fallax
Anadromous
High
0.30
Anguilla anguilla
Catadromous
High
0.35
Atherina boyeri
Resident
Low
0.50
Chelon labrosus
Amphidromous
Medium
0.51
Cobitis calderoni
Resident
Low
0.49
Cobitis paludica
Resident
Low
0.50
Cobitis vettonica
Resident
Low
0.50
Dicentrarchus labrax
Amphidromous
Medium
0.55
Iberochondrostoma lemmingii
Resident
Low
0.50
Iberochondrostoma lusitanicum
Resident
Low
0.50
Iberochondrostoma olissiponensis
Resident
Low
0.59
Lampetra fluviatilis
Anadromous
Medium
0.50
Lampetra planeri
Resident
Low
0.50
Liza ramada
Amphidromous
Medium
0.51
Luciobarbus bocagei
Potamodromous
High
0.39
Luciobarbus comizo
Potamodromous
High
0.42
Luciobarbus steindachneri
Potamodromous
High
0.46
Parachondrostoma miegii
Potamodromous
High
0.37
Petromizon marinus
Anadromous
Medium
0.52
Pomatoschistus microps
Resident
Low
0.50
Pomatoschistus minutus
Resident
Low
0.50
Pseudochondrostoma polylepis
Potamodromous
High
0.39
Salmo trutta
Potamodromous
High
0.34
Squalius alburnoides
Resident
Low
0.50
Squalius castellanus
Resident
Medium
0.35
Squalius pyrenaicus
Resident
Medium
0.49
Data sources: Fishbase (www.fishbase.de) and Carta Piscícola Española
(http://www.sibic.org/carta-piscicola-espanola)
Hermoso, V., Clavero, M., Filipe, A.F. (2020). An accessible optimisation method for barrier
removal planning in stream networks. Science of the Total Environment, 752, 141943
21
Figure 1. Example of different types of barriers identified in the Tagus River catchment,
covering from large dams, assessed as non-removable due to the high socio-economic impact
and removal cost (A); smaller dams, assessed as not passable by any species, but removable at a
high cost (B); different types of weirs, from consolidated hard structures (C), to accumulation of
material (D), with different degrees of passability and cost; other structures, such as gauging
stations (E); or small barriers with smaller removal costs (F).
Hermoso, V., Clavero, M., Filipe, A.F. (2020). An accessible optimisation method for barrier
removal planning in stream networks. Science of the Total Environment, 752, 141943
22
Figure 2. Spatial distribution of the 934 barriers identified in the Tagus River catchment (SW
Iberian Peninsula). Barriers that were assessed as non-removable, such as large dams and
hydropower stations, are identified with a triangle.
Hermoso, V., Clavero, M., Filipe, A.F. (2020). An accessible optimisation method for barrier
removal planning in stream networks. Science of the Total Environment, 752, 141943
23
Figure 3. Conceptual representation of fragmentation of populations of two fish species by three
barriers in a hypothetical river network. Barriers #1 and #3 are completely impassable to both
species, while barrier #2 is only not passable for species B. Therefore, when estimating benefits
of treating barriers #1 and #3, the length of rivers connected for both species would be
accounted for, as a measure of the gain of accessible habitat that allows species to move or use
resources. The benefit of treating barrier #2 would only be the river length increase for species
B. The length of each species in the river reaches connected after the barrier treatment were
considered in the estimates of benefits. For example, treating barrier #3 would have a different
benefit for species A, that only occupies part of the connected reaches, and species B, that
occupies all connected reaches.
Hermoso, V., Clavero, M., Filipe, A.F. (2020). An accessible optimisation method for barrier
removal planning in stream networks. Science of the Total Environment, 752, 141943
24
Figure 4. Number of barriers selected and average gain in connectivity across the 30 species
considered in this study (Table 1) for different reconnection targets. The dotted line separates
solutions before and after the tipping point identified at target 0.75.
Hermoso, V., Clavero, M., Filipe, A.F. (2020). An accessible optimisation method for barrier
removal planning in stream networks. Science of the Total Environment, 752, 141943
25
Figure 5. Spatial distribution of barriers selected in Marxan´s best solution for the gain
connectivity target of 0.5 under the locked-out barriers scenario (back dots), barriers that were
considered as no removable and, therefore, locked out from analyses (triangles), and barriers not
selected (white dots). The only two locked out barrier that were selected under all barriers
scenario are pointed with an arrow. A complete comparison of barriers selected under both
scenarios can be checked in Supplementary Fig. 3.
... It can be viewed as excessively prescriptive (McKay et al., 2020) and tends to ignore local knowledge (Fox et al., 2016), which may antagonize some stakeholders (Sneddon et al., 2017) and make communication of results difficult. It also requires a high degree of mathematical and computer programing expertise, although open source spatial planning software, such as Marxan (Hermoso et al., 2021), and special purpose decision support systems, such as OptiPass (O'Hanley, 2014) and the River Infrastructure Planning (RIP) tool should facilitate more mainstreaming use of optimization in barrier removal programs. Other downsides include the fact that (1) small changes to budgets and project cost can result in markedly different solutions since there is no guarantee that solutions will be nested (O'Hanley, 2011); (2) the quality of solutions tends to be heavily reliant on the availability of complete and accurate barrier location data; and (3) recommended solutions may require cooperation of multiple barrier owners, which may or may not be easy to achieve. ...
... For example, in the Willamette River, USA, removing just 8 % of barriers would reconnect 52 % of the basin (Kuby et al., 2005). Several studies have shown that the removal of certain key barriers can result in disproportionately high gains in connectivity (Hermoso et al., 2021), but that benefits eventually top out (O'Hanley et al., 2013). ...
... Barrier removal planning must also contend with uncertainties related to the potential spread of invasive species (Cooper et al., 2021;Hermoso et al., 2021;Jones et al., 2021b;Muha et al., 2021) and with future demands for water resources (Baumgartner et al., 2021;Duarte et al., 2021;Radinger and García-Berthou, 2020;Tickner et al., 2020). Many would argue that the answer to resolving issues around uncertainty is to gather more data before making a decision. ...
Article
Barrier removal can be an efficient method to restore river continuity but resources available for defragmenting rivers are limited and a prioritization strategy is needed. We review methods for prioritizing barriers for removal and report on a survey asking practitioners which barrier prioritization methods they use. Opportunities for barrier removal depend to a large extent on barrier typology, as this dictates where barriers are normally located, their size, age, condition, and likely impacts. Crucially, river fragmentation depends chiefly on the number and location of barriers, not on barrier size, while the costs of barrier removal typically increase with barrier height. Acting on many small barriers will often be more cost-efficient than acting on fewer larger structures. Barriers are not randomly distributed and a small proportion of barriers have a disproportionately high impact on fragmentation, therefore targeting these ‘fragmentizers’ can result in substantial gains in connectivity. Barrier prioritization methods can be grouped into six main types depending on whether they are reactive or proactive, whether they are applied at local or larger spatial scales, and whether they employ an informal or a formal approach. While mathematical optimization sets the gold standard for barrier prioritization, a hybrid approach that explicitly considers uncertainties and opportunities is likely to be the most effective. The effectiveness of barrier removal can be compromised by inaccurate stream networks, erroneous barrier coordinates, and underestimation of barrier numbers. Such uncertainties can be overcome by ground truthing via river walkovers and predictive modelling, but the cost of collecting additional information must be weighed against the cost of inaction. To increase the success of barrier removal projects, we recommend that barriers considered for removal fulfill four conditions: (1) their removal will bring about a meaningful gain in connectivity; (2) they are cost-effective to remove; (3) they will not cause significant or lasting environmental damage, and (4) they are obsolete structures. Mapping barrier removal projects according to the three axes of opportunities, costs, and gains can help locate any ‘low hanging fruit.’
... Additionally, given the demonstrated importance of hosts to the persistence of M. margaritifera, it is also urgent to take actions towards the conservation of the affiliated fish species already threatened by climate change (Clavero et al., 2017). Ensuring that fish hosts are able to reach suitable environments, by reversing the current trends of habitat loss and fragmentation, is a foremost priority (Hermoso et al., 2021). In conclusion, our study provides strong evidence for proposing the generalised use of biotic information about hosts in addition to purely environmental variables to model the distribution of freshwater mussels, as well as for other species with important obligatory biotic interactions. ...
Article
The freshwater pearl mussel Margaritifera margaritifera has been suffering major population declines in Europe. This endangered species is a host specialist and exclusively requires salmonid species (Salmo trutta and Salmo salar) to complete its life cycle. In theory, obligatory biotic interactions should deserve special conservation attention, because the loss or massive decline of fish hosts may elicit the extirpation of their affiliated species. While many threats disturbing M. margaritifera are similarly affecting salmonids, climate change is particularly alarming, with the potential to significantly change the fish‐mussel dynamics. To evaluate the importance of including the occurrence of fish hosts for predicting the distribution of M. margaritifera in Europe, three datasets were used to build species distribution models (SDMs) with a maximum entropy (MaxEnt) approach: (1) environmental variables (ENV); (2) probability of fish hosts occurrence (FH); and (3) environmental variables and probability of fish hosts occurrence (ENV + FH). We identified the environmental variables that better explain M. margaritifera distribution and modelled its current and future distribution under a suite of climate change scenarios. Furthermore, projections were used to evaluate the adequacy of current networks of European protected areas in covering the suitable habitats for M. margaritifera. Results showed that incorporating data about fish hosts into M. margaritifera SDMs avoided the overprediction of geographical projections and, to a minor extent, improved model performance (area under the curve: ENV = 0.851; FH = 0.848; ENV + FH = 0.867). The distribution range of M. margaritifera in Europe is expected to contract in all future timeframes and emission scenarios considered. Forecasts point to large contractions particularly in central and southern Europe and lowland regions. The European network of protected areas fails to protect 69% of the current and 66%–67% of the future predicted M. margaritifera distribution. This study clearly illustrates the importance of including mussel–fish hosts interactions for accurately predicting M. margaritifera's distribution. The response of M. margaritifera to environmental variables highlights its vulnerability to the higher temperatures, particularly in southern Europe. While predictions indicate large contractions in M. margaritifera's distribution as a result of future climate change, the current European network of protected areas fails to safeguard M. margaritifera. This work provides strong evidence for proposing the generalised use of biotic information about hosts in addition to purely environmental variables to model the distribution of freshwater mussels, as well as for other species with obligatory biotic interactions. Building SDMs such as those discussed here can inform political decision‐making about the likely scenarios for species occurrence in future decades, the requirements needed for an effective conservation strategy, and the regions where conservation should be a priority.
... Inland fishery resources are important protein sources of human lives and their socioeconomic importance has been rapidly increasing. 1,2 Life histories of inland fishery resources, especially fishes migrating along rivers, are critically affected by transverse hydraulic structures like dams and weirs because they physically prevent the fishes from ascending/descending. 3,4 These physical barriers not only serve as obstructions of the migration processes but also disturb fish assemblages in both their upstream and downstream reaches of rivers, 5 resulting in fragmentation of fish habitats. 6 Environmental improvement schemes, such as barrier removal and flow adaptation, have been considered previously to mitigate the human-induced impacts. ...
Article
Full-text available
In inland fisheries, transporting fishery resource individuals from a habitat to spatially apart habitat(s) has recently been considered for fisheries stock management in the natural environment. However, its mathematical optimization, especially finding when and how much of the population should be transported, is still a fundamental unresolved issue. We propose a new impulse control framework to tackle this issue based on a simple but new stochastic growth model of individual fishes. The novel growth model governing individuals' body weights uses a Wright–Fisher model as a latent driver to reproduce plausible growth dynamics. The optimization problem is formulated as an impulse control problem of a cost–benefit functional constrained by a degenerate parabolic Fokker–Planck equation of the stochastic growth dynamics. Because the growth dynamics have an observable variable and an unobservable variable (a variable difficult or impossible to observe), we consider both full‐information and partial‐information cases. The latter is more involved but more realistic because of not explicitly using the unobservable variable in designing the controls. In both cases, resolving an optimization problem reduces to solving the associated Fokker–Planck and its adjoint equations, the latter being nontrivial. We present a derivation procedure of the adjoint equation and its internal boundary conditions in time to efficiently derive the optimal transporting strategy. We finally provide a demonstrative computational example of a transporting problem of Ayu sweetfish Plecoglossus altivelis altivelis based on the latest real data set.
... Additionally, given the demonstrated importance of hosts to the persistence of M. margaritifera, it is also urgent to take actions towards the conservation of the affiliated fish species already threatened by climate change (Clavero et al., 2017). Ensuring that fish hosts are able to reach suitable environments, by reversing the current trends of habitat loss and fragmentation, is a foremost priority (Hermoso et al., 2021). In conclusion, our study provides strong evidence for proposing the generalised use of biotic information about hosts in addition to purely environmental variables to model the distribution of freshwater mussels, as well as for other species with important obligatory biotic interactions. ...
Article
Many river networks in southern Europe are intermittent. In summer, the surface flow is zero and many streams become isolated pools. In this study, 128 dry season pools were studied covering first- to fourth-order streams on the Degebe River network (south Portugal). The aim of the study was to identify pool types based on environmental drivers and conditions and fish assemblages. In summer, dry streambed area exceeded 50% in all reaches and 95% in headwater sections. The pool features were primarily shaped by their location in the river network, which determined the pool morphology and the structure of fish assemblages. Pool sizes increased from upstream to downstream, as did species richness and diversity. Pools in upstream reaches were dominated by small native fishes while the larger-sized individuals tended to occupy deeper, larger, and more persistent pools. Smaller pools in downstream reaches were dominated by non-native species, which may be related to habitat preferences and minimization of negative interactions between native and non-native species. Because dry season pools represent key habitats in intermittent streams, conservation programs should be designed to reduce human pressures and improve hydromorphological heterogeneity and water quality, taking into account the natural patterns of pool types at regional and local scales.
Article
Full-text available
A robust assessment of the American eel (Anguilla rostrata) stock, required to guide conservation efforts, is challenged by the species' vast range, high variability in demographic parameters and data inadequacies. Novel ideas and underutilised resources that may assist both analytic assessments and spatially oriented modelling include (1) species and environmental databases; (2) mining of data from scattered sources; (3) infilling of data gaps by spatial analysis; (4) age estimation from measurements of DNA methylation; evaluation of eel abundance by (5) larval, (6) glass-bottom boat, (7) net enclosure and (8) eDNA surveys; (9) accounting for dam-induced habitat increases in eel watercourse modelling; (10) spatially oriented modelling with and without temporal components; (11) geographically nested modelling of glass eel recruitment; (12) spawner per recruit modelling and (13) life cycle modelling to examine larval allocation effects. Eel biologists are too few to gather the required assessment data across all of the species’ range. Public posting of electrofishing and eDNA metabarcoding data sets and the use of machine learning techniques to comprehensively inventory small dams will help meet some data needs. These approaches address only a small proportion of the assessment challenges that face American eels. Worldwide collaboration amongst Anguilla scientists is a key enabler of progress towards stock assessment goals.
Article
Full-text available
In dryland rivers, flow intermittency means fish populations are often subjected to drought disturbance. The viability of these fish populations depends on the availability of waterhole refuges for individuals to survive drought (resistance) and the ability of surviving fish to repopulate the rivers by recruitment and dispersal once flow returns (resilience). In this study we combined remote-sensed mapping of the locations of waterholes that lasted through an extreme drought in the northern Murray Darling Basin, Australia, with an assessment of the impacts of in-stream barriers on limiting the opportunities for fish to move and repopulate after drought. We found that at the peak of this 2018–2020 drought, the worst on record for some rivers and the most spatially synchronous recorded across the region, waterholes were few and generally small – representing only 11% of the total river channel network. All the fish in the region that survived the drought were concentrated into this limited waterhole refuge habitat. Even small instream structures, such as minor weirs, caused large reductions in the opportunities for fish to move between river segments when there is flow. Almost all the 104 instream structures assessed reduced long-term fish movement opportunities, measured as days with discharge greater than calculated barrier drown out thresholds, by more than 70% and up to 100%, when compared to opportunities for movement if the barrier was not present. This large impact from small instream barriers is a consequence of flow intermittency and is likely to reduce fish population resilience and impact the capacity of fish populations to recover after drought. Combining information on the risks posed by limited refuge habitat availability during drought and from reduced movement opportunity following drought allowed us to identify river segments where these combined threats are the greatest risk to viability of local fish populations. Considering the spatial arrangements of these risks provides a means to systematically prioritize mitigation measures such as weir removal to improve fish movement opportunities and local management of key waterholes to increase drought resistance. The approach used here provides a guide for assessing and prioritizing the management of fish population viability risks from drought and fragmentation by barriers in any non-perennial river setting.
Article
Ecological restoration is increasingly being upscaled to larger spatial scales of 10s to 100 s of kilometres. Yet, the complex logistics and high costs of ecological restoration mean that actions must be placed strategically at local scales of 10s of meters to maximise ecological benefits and reduce socio‐economic costs. Despite the purported use of systematic planning tools for allocating restoration effort, the uptake and implementation of data‐driven restoration planning and ecological goal setting remains poor in many restoration programs. Here we demonstrate how the sequential workflows of systematic conservation planning can be translated to restoration at two spatial scales to enhance estuarine fisheries in eastern Australia. We select estuaries where restoration is feasible and recommended based on quantitative regional ecological goals (i.e. regional scale prioritisation), and then identify potential restoration sites at smaller spatial scales within estuaries based on the principles of spatial ecology to ensure that the success and benefits of restoration are maximised (i.e. local scale prioritisation). At the regional scale, we identified four levels of restoration priorities (very high, high, intermediate, and low) using quantitative ecological goals and the current ecological understanding of each system. At the local scale, we used spatially explicit Bayesian belief networks to identify sites that maximise restoration outcomes based on the environmental niche of habitat‐forming species and the spatial configuration of habitats that maximises their use by fish. We show that using systematic frameworks can become an essential tool to optimise restoration investments at multiple scales as efforts upscale globally. This article is protected by copyright. All rights reserved.
Article
Biodiversity offsetting is a popular conservation tool to reduce the impact of human activities. This is especially relevant in freshwater ecosystems, under the increasing threat posed by the development of infrastructure to store freshwater or produce energy that break longitudinal connectivity and modify the structure and functioning of these systems. We demonstrate how to plan offset of connectivity loss in rivers derived from the construction of new barriers, by using the Tagus River (Iberian Peninsula) as a model. We simulate the construction of new barriers, measure the impact they would have on connectivity for each species individually, and identify an optimal set of existing barriers that should be removed to counterbalance the loss of connectivity caused for all species collectively. We found that loss in connectivity could be offset for most of species when a single new barrier was simulated at a time, by removing a small number of existing barriers. However, there was a group of species with very restricted ranges that could undergo irreversible loss of connectivity even when all existing barriers were made available as an offset option. The list of species that could not be offset and the cost of barrier removals increased as the number of new barriers simulated increased. The approach presented here could be used to plan offset actions for other types of impacts in freshwater systems or elsewhere, or to assess the vulnerability of particular species or processes to potential future impacts by identifying the boundaries of development that can be offset.
Article
Full-text available
Despite their limited spatial extent, freshwater ecosystems host remarkable biodiversity, including one third of all vertebrate species. This biodiversity is declining dramatically: globally, wetlands are vanishing three times faster than forests, and freshwater vertebrate populations have fallen more than twice as steeply as terrestrial or marine populations. Threats to freshwater biodiversity are well documented but co-ordinated action to reverse the decline is lacking. We present an Emergency Recovery Plan to “bend the curve” of freshwater biodiversity loss. Priority actions include: 1) accelerating implementation of environmental flows; 2) improving water quality; 3) protecting and restoring critical habitats; 4) managing exploitation of freshwater ecosystem resources, especially species and riverine aggregates; 5) preventing and controlling non-native species invasions; and 6) safeguarding and restoring river connectivity. We recommend adjustments to targets and indicators for the Convention on Biological Diversity and the Sustainable Development Goals, and roles for national and international state and non-state actors. *** This paper has been accepted for publication in BioScience. A link to the BioScience version will follow in due course ***
Article
Full-text available
It's renewable but not sustainable. We argue that the unchecked development promoting new small hydropower plants should be replaced by a new paradigm that builds on three points: (1) Small hydropower plants must be subject to the same environmental regulations as large hydropower plants because both are associated with ecological threats and high socioeconomic costs. (2) Regardless of their size, the development of hydropower plants needs to be guided by policies requiring long-term planning and assessment at the basin scale because impacts will propagate over decades and cumulatively add up at the basin scale. (3) Governments, legislative bodies, international funding agencies, and private investors should revise their subsidy programs to consider the true ecological and socioeconomic costs and benefits of small hydropower plants. Most are not economically viable without subsidies.
Article
Full-text available
Free-flowing rivers (FFRs) support diverse, complex and dynamic ecosystems globally, providing important societal and economic services. Infrastructure development threatens the ecosystem processes, biodiversity and services that these rivers support. Here we assess the connectivity status of 12 million kilometres of rivers globally and identify those that remain free-flowing in their entire length. Only 37 per cent of rivers longer than 1,000 kilometres remain free-flowing over their entire length and 23 per cent flow uninterrupted to the ocean. Very long FFRs are largely restricted to remote regions of the Arctic and of the Amazon and Congo basins. In densely populated areas only few very long rivers remain free-flowing, such as the Irrawaddy and Salween. Dams and reservoirs and their up- and downstream propagation of fragmentation and flow regulation are the leading contributors to the loss of river connectivity. By applying a new method to quantify riverine connectivity and map FFRs, we provide a foundation for concerted global and national strategies to maintain or restore them.
Article
Full-text available
The insights that historical evidence of human presence and man-made documents provide are unique. For example, using historical data may be critical to adequately understand the ecological requirements of species. However, historical information about freshwater species distribution remains largely a knowledge gap. In this Data Descriptor, we present the Portuguese Historical Fish Database (PHish–DB), a compilation of 2214 records (557 at the basin scale, 184 at the sub-basin scale and 1473 at the segment scale) resulting from a survey of 194 historical documents. The database was developed using a three-scale approach that maximises the inclusion of information by allowing different degrees of spatial acuity. PHish database contains records of 25 taxonomical groups and covers a time span of one millennium, from the 11th until the 20th century. This database has already proven useful for two scientific studies, and PHish further use will contribute to correctly assess the full range of conditions tolerated by species, by establishing adequate benchmark conditions, and/or to improve existing knowledge of the species distribution limits.
Article
Full-text available
• Integrating ecosystem services (ESs) in landscape planning can help to identify conservation opportunities by finding co‐benefits between biodiversity conservation and the maintenance of regulating and cultural ecosystem services. The adequate integration of ESs needs careful consideration of potential trade‐offs, however, especially between provisioning services and biodiversity conservation (e.g. the potentially negative consequences of agricultural water extraction within areas important for the maintenance of biodiversity). These trade‐offs have been overlooked in systematic spatial planning to date, especially in freshwater systems. • marxan with zones was used to identify priority areas for the conservation of freshwater biodiversity (139 species of freshwater fish, turtles, and waterbirds) and the provision of freshwater ESs in the Daly River, northern Australia. Four different surrogates for ESs were mapped, including those potentially incompatible with conservation goals (i.e. groundwater provision for agriculture and recreational fisheries) and those that are more compatible with conservation (i.e. flood regulation by riparian forests; provision of perennial water). The spatial allocation of multiple management zones was prioritized: (i) three conservation zones, aiming to represent freshwater biodiversity and compatible ESs to enhance co‐benefits; and (ii) two production zones, where access to provisioning ESs could be granted. The representation of ESs obtained when using the multi‐zoning approach was compared with that achieved with a single management zone approach. The comparison was performed across different representation targets. • Different results were found with low and high targets for ESs. With low targets (<25% of all ESs), the multi‐zoning approach achieved up to 53% more co‐benefits than the single‐zone approach. With high targets (>25% of all ESs), the trade‐offs avoided were more evident, with up to 56% less representation of incompatible ESs within conservation zones. • Multi‐zone planning could help decision makers respond better to the increasingly complex catchment management context, caused by an increasing demand for provisioning services and a diminishing availability of resources, as well as manage and plan for challenges in other realms facing similar problems.
Article
Full-text available
Spatial prioritization tools provide a means of finding efficient trade-offs between biodiversity protection and the delivery of ecosystem services. Although a large number of prioritization approaches have been proposed in the literature, most are specifically designed for terrestrial systems. When applied to river ecosystems, they often fail to adequately account for the essential role that landscape connectivity plays in maintaining both biodiversity and ecosystem services. This is particularly true of longitudinal connectivity, which in many river catchments is highly altered by the presence of dams, stream-road crossings, and other artificial structures. 2.We propose a novel framework for coordinating river conservation and connectivity restoration. As part of this, we formulate an optimization model for deciding which subcatchments to designate for ecosystem services and which to include in a river protected area (RPA) network, while also deciding which existing river barriers to remove in order to maximize longitudinal connectivity within the RPA network. In addition to constraints on the size and makeup of the RPA network, the model also considers the suitability of sites for conservation, based on a biological integrity index, and connectivity to multiple habitat types. We demonstrate the usefulness of our approach using a case study involving four managed river catchments located in Hungary. 3.Results show that large increases in connectivity-weighted habitat can be achieved through targeted selection of barrier removals and that the benefits of barrier removal are strongly depend on RPA network size. We find that (i) highly suboptimal solutions are produced if habitat conservation planning and connectivity restoration are done separately and (ii) RPA acquisition provides substantially greater marginal benefits than barrier removal given limited resources. 4.Synthesis and applications. Finding a balance between conservation and ecosystem services provision should give more consideration to connectivity restoration planning, especially in multi-use riverscapes. We present the first modelling framework to directly integrate and optimize river conservation and connectivity restoration planning. This framework can help conservation managers to account better for connectivity, resulting in more effective catchment scale maintenance of biological integrity and ecosystem services delivery. This article is protected by copyright. All rights reserved.
Article
Artificial barriers are one of the main threats to river ecosystems, resulting in habitat fragmentation and loss of connectivity. Yet, the abundance and distribution of most artificial barriers, excluding high-head dams, is poorly documented. We provide a comprehensive assessment of the distribution and typology of artificial barriers in Great Britain, and estimate for the first time the extent of river fragmentation. To this end, barrier data were compiled from existing databases and were ground-truthed by field surveys in England, Scotland and Wales to derive a correction factor for barrier density across Great Britain. Field surveys indicate that existing barrier databases underestimate barrier density by 68%, particularly in the case of low-head structures (<1 m) which are often missing from current records. Field-corrected barrier density estimates ranged from 0.48 barriers/km in Scotland to 0.63 barriers/km in Wales, and 0.75 barriers/km in England. Corresponding estimates of stream fragmentation by weirs and dams only, measured as mean barrier-free length, were 12.30 km in Scotland, 6.68 km in Wales and 5.29 km in England, suggesting the extent of river modification differs between regions. Our study indicates that 97% of the river network in Great Britain is fragmented and <1% of the catchments are free of artificial barriers.
Article
In this study, we propose a novel framework combining spatially explicit population viability analysis and optimization for prioritizing fish passage barrier mitigation decisions. Our model aims to maximize the equilibrium population size, or alternatively minimize the extinction risk, of a target fish species subject to a budget on the total cost of barrier mitigation. A case study involving a wild coho salmon (Oncorhynchus kisutch) population from the Tillamook basin, Oregon, USA is used to illustrate the benefits of our approach. We consider two different spawning adult dispersal patterns, river and reach level homing, as well as straying. Under density dependent population growth, we find that homing behavior type has a significant effect on barrier mitigation decisions. In particular, with reach homing, our model produces virtually the same population sizes as a more traditional barrier prioritization procedure designed to maximize accessible habitat. With river homing, however, we find that it is not necessary to remove all barriers in order to maximize equilibrium population size. Indeed, a stochastic version of our model reveals that removing all barriers actually results in a marginal increase in quasi-extinction risk. We hypothesize that this is due to a population thinning effect of barriers, resulting in a surplus of recruits in areas of low spawner density. Our findings highlights the importance of considering spatiotemporal fish population dynamics in river connectivity restoration planning. By adding greater biological realism, models such as ours can help conservation managers to more strategically allocate limited resources, resulting in both cost savings and improved population status for a focal species.
Article
Movement within stream corridors is a basic life history requirement of many aquatic organisms. Barrier removal in streams has become a common practice in the United States aimed to restore organism dispersal and meet conservation objectives; however, there are social and economic costs to the removal of barriers. Accordingly, tools to prioritize barrier removal, particularly optimization techniques, can be used to evaluate cost‐benefit trade‐offs. Many of these techniques, however, require programming experience and are not available to natural resource managers. Furthermore, conservation objectives vary considerably depending on the life histories of organisms under consideration, and these opposing objectives, in conjunction with variant socioeconomic costs, will influence optimization solutions, specifically which barriers to remove. To promote the use of optimization tools, straightforward and open‐access platforms are needed to support use by managers, while also providing general approaches for holistic basin‐scale connectivity restoration. Herein, we use two case studies, White Oak Creek (small watershed) and the Roanoke River Basin (large basin), to explore the divergent outcomes stemming from different conservation objectives and socioeconomic costs used to prioritize barrier removal. We conducted optimization modeling using a widely accessible platform along with an open‐access solver plug‐in to support a wide variety of conservation objectives. We used simple approaches to find commonalities in barriers identified for removal among divergent conservation objectives and provide alternative (i.e., hybrid removal‐passage) strategies for approaching habitat restoration for diverse aquatic communities while increasing social benefits (i.e., hydropower energy). As expected, different conservation objectives aimed to support varied species life histories (e.g., diadromy, large‐river vs. small‐river potamodromy) have very different effects on optimization solutions. In both case studies, however, commonalities in solutions were identified through clustering groups of barriers into general connectivity restoration strategies. Furthermore, strategy types for a given barrier could be predicted with ≥72% accuracy using only four metrics. This suggests that optimization results can be simplified into general standards to support adoption of sustainable basin connectivity criteria strategies. Our framework provides a flexible and open‐access approach to conduct relatively complex optimization modeling for stream barrier prioritization, while examining potential for agreement among divergence conservation objectives.