Linking habitats for multiple species

doi:10.1016/j.envsoft.2012.08.001

Environmental Modelling & Software

Volume 40, February 2013, Pages 336-339

https://doi.org/10.1016/j.envsoft.2012.08.001 Get rights and content

Abstract

The establishment of linkages between habitats is of great importance to avert the detrimental impacts of land fragmentation and climate change on biodiversity. Linkages need to be cost-efficient, and should account for specific dispersal requirements of species. Since cost-efficient linkages defined independently for each individual species are more costly than linkages optimised for multiple species, there is need for methods specifically designed to retrieve efficient linkages for multiple species. MulTyLink (Multiple Type Linkages) is a C++ open source program that defines cost-efficient linkages free of barriers for the species considered, and that allows species-specific dispersal requirements to be considered. Here we present, discuss and illustrate the algorithms used by MulTyLink to identify cost-efficient linkages for multiple species.

Introduction

Habitat fragmentation is a key driver of biodiversity loss (Brooks et al., 2002; Hanski, 2005). Linking once connected natural habitats is imperative to maintain biological diversity (Luque et al., 2012). However, given the limited resources available for conservation and the existing conflicts between conservation and exploitative uses, location-allocation efficiency is a desirable property of any conservation plan (Pressey and Nicholls, 1989).

Graph theory is recognised as being a convenient framework to incorporate spatial criteria into conservation planning (Urban and Keitt, 2001; Fall et al., 2007), and a number of methods use graph theory for identifying cost-efficient areas to promote connectivity between habitats. Some of these methods address connectivity and species' representation simultaneously. Examples are the selection of areas to achieve representation targets for species, forming a unique contiguous network (Önal and Briers, 2005; Cerdeira et al., 2005; Fuller et al., 2006; Cerdeira et al., 2010), or permitting less strict forms of spatial coherence, selecting areas that are spatially clustered but not necessarily connected by contiguous corridors (Önal and Briers, 2002, 2003; Alagador and Cerdeira, 2007). Other methods solely involve connecting existing sets of reserves or habitats. Table 1 reports open-source software implementing these latter approaches.

Some of the reported applications deliver a unique (or a small set of) “best” linkage(s) (i.e., the areas promoting connectivity between habitats), while others produce a quantitative evaluation of how suitable (or probable) each area is for the established connectivity goals.

None of the above methods was specifically designed to efficiently link habitats occupied by multiple species, presenting distinct distribution patterns, i.e., occurring in distinct sets of suitable isolated habitats, and having different dispersal constraints, i.e., distinct dispersal distances and dissimilar suitable areas to disperse. One size fits all would not be a valid approach since areas suitable for the dispersal of some species may be barriers for others.

Although the existing methods for identification of linkages can be used for multiple species (by merging individual species solutions into a single solution), the resulting configuration would most probably be far from a minimum cost (or area) solution. To overcome this limitation Lai et al. (2011) used graphs to take into account species dispersal specificities and proposed procedures which they applied to trace links for populations of wolverines and Canadian lynx in Western Montana. Alagador et al. (2012) also developed an approach to this problem, which they applied to link Iberian Peninsula protected areas, clustered in four environmental-similar classes.

We introduce MulTyLink, a software to efficiently link the habitats occupied by multiple species with distinct distributions and/or dispersal requirements, and specifically designed to deal with large data sets. MulTyLink constructs a graph for each group (of “similar”) species, taking into account the areas acting as barriers and the dispersal capacities of these species. When selecting areas for a group of species in a graph, MulTyLink deems the possibility of using these areas for other groups, thus reducing costs and the number of selected areas. After having obtained a solution ensuring the connection of all the habitats for each group of species, the last step of MulTyLink's procedure consists of removing all areas whose removal would not affect the required linkages for all species in analysis.

Section snippets

Methods

It is assumed that the study region is divided into cells. Cells that for whatever reason are not suitable for conservation action are filtered out from the analysis, leaving a set T of candidate cells for usage as linkage units.

For each (group of “similar”) species, k (k = 1, 2, …, m), a subset of cells T_k, called terminal cells of type k, represent the habitats where the species occurs and that need to be linked. A distance threshold, d_k is used to define adjacency rules between cells (i.e.,

Program description

MulTyLink is a dialog-based application integrating mapping capabilities and visualization routines, that implements Type by Type and Grasp to identify cost-efficient linkages for multiple species.

MulTyLink requires an input file with geographical information for mapping purposes and the identification of terminal cells, friction values (cells with friction values greater than a given threshold are excluded from consideration), costs associated to non-terminal cells and, for each type k, the

Example

Here we exemplify how MulTyLink operates using a simple example where the habitats occupied by three threatened reptile species (Lacerta bilineata, the Western green lizard; Lacerta schreiberi, the Iberian emerald lizard; and Coronella austriaca, the Smooth snake) within Iberian Peninsula protected areas are to be linked.

Terminals were defined as those cells, among 2310 10^′ × 10^′ (c. 18 km lat. × 15 km long.) cells representing the Iberia Peninsula, where any of the three species occurs and

Acknowledgements

RB, JOC and DA were supported by the Portuguese Foundation for Science and Technology (FCT): RB was funded by the project PEst-OE/EGE/UI0491/2011 under the FEDER/POCI Programme; JOC and DA were funded through the project PEst-OE/AGR/UI0239/2011 under FEDER/POCI, and project PTDC/AAC-AMB/113394/2009, and DA had financial support from a FCT post-doctoral fellowship (SFRH. BPD.51512.2011). MBA is funded through the EC FP6 ECOCHANGE project (GOCE-CT-13 2006-036866) and acknowledges the Spanish

References (30)

D. Alagador et al.
Designing spatially-explicit ecological reserve networks in the presence of mandatory sites
Biological Conservation
(2007)
M.B. Araújo et al.
Ensemble forecasting of species distributions
Trends in Ecology and Evolution
(2007)
J.O. Cerdeira et al.
Species specific connectivity in reserve-network design using graphs
Biological Conservation
(2010)
T. Fuller et al.
Incorporating connectivity into conservation planning: a multi-criteria case study from central Mexico
Biological Conservation
(2006)
T. Fuller et al.
LQGraph: a software package for optimizing connectivity in conservation planning
Environmental Modelling and Software
(2006)
R.L. Pressey et al.
Efficiency in conservation evaluation: scoring versus iterative approaches
Biological Conservation
(1989)
S. Saura et al.
Conefor sensinode 2.2: a software package for quantifying the importance of habitat patches for landscape connectivity
Environmental Modelling and Software
(2009)
D. Alagador et al.
Linking like with like: optimising connectivity between environmentally-similar habitats
Landscape Ecology
(2012)
M.B. Araújo et al.
Climate change threatens European conservation areas
Ecology Letters
(2011)
T.M. Brooks et al.
Habitat loss and extinction in the hotspots of biodiversity
Conservation Biology
(2002)

C. Carroll et al.

Use of linkage mapping and centrality analysis across habitat gradients to conserve connectivity of gray wolf populations in western North America

Conservation Biology

(2012)

J.O. Cerdeira et al.

Connectivity in priority area selection for conservation

Environmental Modelling and Assessment

(2005)

D. Du et al.

Steiner Tree Problems in Computer Communication Networks

(2008)

A. Fall et al.

Spatial graphs: principles and applications for habitat connectivity

Ecosystems

(2007)

T.A. Feo et al.

Greedy randomized adaptive search procedures

Journal of Global Optimization

(1995)

Cited by (30)

A vertex-separator-based integer linear programming formulation for the partitioned Steiner tree problem
2023, Computers and Operations Research
Given an undirected graph $G$ with a cost function on vertices, a collection of subgraphs of $G$ such that in each subgraph, there are some distinguished vertices called terminals, the Partitioned Steiner Tree Problem (PSTP) asks for a minimum cost vertex set such that, in each of the given subgraph $G_{i}$ , the graph induced by the vertex set spans the terminal set in $G_{i}$ . The PSTP generalizes the well-known Steiner tree problem and has important applications in computational sustainability, network design, and social network analysis. However, for solving the PSTP, conventional integer programming approaches based on single-commodity flow, multi-commodity flow and subtour elimination integer linear programs, suffer from low computational efficiency due to a substantial number of variables. In this paper, we propose a compact vertex-separator-based integer linear programming formulation with much fewer variables. Enhancing inequalities are also studied for tightening the formulation. We further investigate a branch-and-cut algorithm, a local-branching heuristic algorithm, and a hybrid algorithm combining them. In experiments where both public real-world and synthetic graphs are used, our hybrid algorithm outperforms all conventional approaches, especially for large graphs with more than ten thousand vertices. Further tests also validate the effectiveness of the proposed formulation and enhancing inequalities.
Operations research applicability in spatial conservation planning
2022, Journal of Environmental Management
A large fraction of the current environmental crisis derives from the large rates of human-driven biodiversity loss. Biodiversity conservation questions human practices towards biodiversity and, therefore, largely conflicts with ordinary societal aspirations. Decisions on the location of protected areas, one of the most convincing conservation tools, reflect such a competitive endeavor. Operations Research (OR) brings a set of analytical models and tools capable of resolving the conflicting interests between ecology and economy. Recent technological advances have boosted the size and variety of data available to planners, thus challenging conventional approaches bounded on optimized solutions. New models and methods are needed to use such a massive amount of data in integrative schemes addressing a large variety of concerns. This study provides an overview on the past, present and future challenges that characterize spatial conservation models supported by OR. We discuss the progress of OR models and methods in spatial conservation planning and how those models may be optimized through sophisticated algorithms and computational tools. Moreover, we anticipate possible panoramas of modern spatial conservation studies supported by OR and we explore possible avenues for the design of optimized interdisciplinary collaborative platforms in the era of Big Data, through consortia where distinct players with different motivations and services meet. By enlarging the spatial, temporal, taxonomic and societal horizons of biodiversity conservation, planners navigate around multiple socioecological/environmental equilibria and are able to decide on cost-effective strategies to improve biodiversity persistence under complex environments.
Behind the scenes of landscape connectivity: Optimization, thresholding
2022, Environmental Modelling and Software
Citation Excerpt :
The silent assumption may be that adding connectivities automatically yields best-for-all results (e.g. Koen et al., 2014; Miranda et al., 2021). In (Pereira et al., 2017), the problem was properly recognized and then converted into finding good connectivity for different sets of habitat patches (one set per agent), but solution of this problem was only heuristic (KP algorithm), similar approach was in (Brás et al., 2013). Optimization by seeking among sets of patches (combinatorial) is more difficult than by graph weight function.
Optimization is underlying technique of many graph-theoretic algorithms, including shortest paths and minimum spanning tree and hence – hidden component of landscape connectivity methods. Optimization solutions have different selectivity but it remains unexplored. Indeterminate solutions happen without warning and may be spatially biased. Errors are hidden. These issues may undermine connectivity measures and conclusions.
Two optimization targets may be distinguished: mobile agent and connectivity provider, with different requirements. Review of how are they handled in literature finds them entangled and often misspecified, owing to inconclusive results, especially pronounced when conservation and explanatory goals are mixed. I categorize graph structures against targets and show how initial choice of structure introduces implicit optimization. Allowing for stochasticity of mobile agent resolves omniscience problem.
Lastly, I advocate against creating landscape graphs with thresholds in favor of complete planar graphs. Thresholding is impractical, biased by design and yields non-optimal solutions due to simplistic partitioning.
Modelling patch mosaic connectivity and ecological corridors with GraphScape
2020, Environmental Modelling and Software
Citation Excerpt :
At last, in the fourth group (CIRCUITSCAPE software) a least-cost path is replaced by more general least-cost flow moving across “landscape circuitry". Some of the applications deliver a unique (or a small set of) “best” linkage(s) (i.e., the areas promoting connectivity between habitats), while others produce a quantitative evaluation of how suitable (or probable) each area is for the established connectivity goals (Brás et al., 2013). However, in the light of the above review, it is worth recalling that one of the first papers to analyze the usefulness of the graph theory in landscape ecological studies (Cantwell and Forman, 1993) was focused on evaluating landscape configuration, with less attention devoted to evaluating landscape connectivity.
Out of many approaches to landscape connectivity, we choose patch-mosaic, flow-neutral framework for modelling. Our new software system GraphScape transforms a vector mosaic map into landscape graph, finds shortest paths and combines them into minimum spanning tree. The result is an optimal network of connections for a given set of patches. To capture mobile agent properties, we use two types of resistance assigned to patches and transfer resistance assigned to combinations of patch classes. Another alternative is step-model, when transfers occur as if on unweighted graph. Resulting MST is rendered by linear skeleton (binary), Transit Density measure assignable to patches (continuous), and summary statistics. A novelty is compete workflow from vector-based map to complex graph structures. We describe concepts and operation of software. GraphScape can be used for ecological corridors planning, identification of key areas of corridors overlap, identification of the role of particular mobility factors, assessment of patch isolation.
Applying ant colony algorithm to identify ecological security patterns in megacities
2019, Environmental Modelling and Software
Citation Excerpt :
Comparatively speaking, the latter is more accepted due to its focusing on quantification and functionality. Although there are many methods to identify ecological corridors (Rouget et al., 2006; Parks et al., 2013; Bhowmik et al., 2015; Brás et al., 2013), the minimum cumulative resistance (MCR) model is the most widely applied because it can simulate the process of ecological flow with considering landscape heterogeneity and connectivity (Kong et al., 2010; Hepcan and Özkan, 2011; Wang et al., 2008). However, MCR model can effectively delineate the direction of ecological corridors, but cannot provide any information about the range of ecological corridors (Zhang et al., 2017b).
Ecological security patterns composed of ecological sources and corridors provide an effective approach to conserving natural ecosystems. Although the direction of ecological corridors has been identified in previous studies, the precise range remains unknown. To address this crucial gap, ant colony algorithm and kernel density estimation were applied to identify the range and restoration points of ecological corridors, which is important for natural conservation and ecological restoration. In this case study of Beijing City, ecological sources were identified based on habitat importance and landscape connectivity. The results showed that, in total 3119.65 km² of ecological land had been extracted as ecological sources, which were mainly located in the northern, northwestern and northeastern mountainous areas. The identified key ecological corridor covered an area of 198.86 km², with 567.30 km² for potential ecological corridors, both connecting the ecological sources. 34 key points were also identified with priority in restoring ecological corridors.
Climate change and opposing spatial conservation priorities for anuran protection in the Brazilian hotspots
2019, Journal for Nature Conservation
Citation Excerpt :
The present study addressed some important conservation features and targets to propose the prioritization scheme for anurans in the AF and CER hotspots, such as the anthropogenic impacts across the study area, the prioritization of areas having unique species composition and also those that shelter threatened species. Nonetheless, a maximized spatial conservation planning that would certainly help in achieving the 17% target should consider additional biological metrics and/or corridors among the selected areas that maintain some specific population and community dynamics (e.g., Alagador et al., 2012; Brás et al., 2013). To do so, the inclusion of a maximum number of taxa and the incorporation of functional and phylogenetic diversity (e.g. Campos et al., 2017; Carvalho et al., 2017), as well as the consideration of future deforestation scenarios into more precise modelling procedures for those species highly dependent on the habitat structure (e.g., Gouveia, 2016; Triviño, Thuiller, Cabeza, Hickler, & Araújo, 2011), are important approaches that address the multiple facets of biological diversity.
In conservation biogeography, the process of spatial conservation prioritization (SCP) aims to select areas that meet biodiversity targets at a minimum set coverage. Here, we propose a SCP scheme for the highly endemic and diverse anuran fauna of the Atlantic Forest (AF) and Cerrado (CER) South American hotspots under different climate change scenarios. Specifically, we make use of predicted anuran occurrences, built for baseline and future (2050 and 2070) time slices, and address biological and conservation metrics to identify potential priority regions for anuran conservation over time using the software MARXAN. Considering each time slice separately, the percentage area needed for total anuran representation varies at magnitudes of 9.8–10.66% for the AF and 6.4–8.8% for the CER. Pooling all time slices together in the selected conservation network, the identified spatial priorities account for 15.56% and 13.25% of the total AF and CER areas respectively. However, we identified opposing strategies for the anuran spatial conservation prioritization in the AF and CER over the different time periods; the increasing of priority cells across time considering the potential species redistribution under climate change in the AF, and the selection of fewer priority cells in the future than the identified for the baseline climate in the CER. The southeastern AF coast was identified as a priority area for amphibian conservation in this hotspot, as well as some other smaller areas in the northern and southern regions. Priority areas identified in the CER, although patchy distributed across the hotspot, are found in specific central-northern, western, and southeastern regions. The different conservation strategies identified in the present SCP emphasize the need for establishing different conservation efforts according to a sequential scheduling of priority areas that optimizes the long-term conservation goals.

View all citing articles on Scopus

View full text

Short communicationLinking habitats for multiple species

Abstract

Introduction

Section snippets

Methods

Program description

Example

Acknowledgements

Biological Conservation

Trends in Ecology and Evolution

Biological Conservation

Biological Conservation

Environmental Modelling and Software

Biological Conservation

Environmental Modelling and Software

Linking like with like: optimising connectivity between environmentally-similar habitats

Landscape Ecology

Climate change threatens European conservation areas

Ecology Letters

Habitat loss and extinction in the hotspots of biodiversity

Conservation Biology

Use of linkage mapping and centrality analysis across habitat gradients to conserve connectivity of gray wolf populations in western North America

Conservation Biology

Connectivity in priority area selection for conservation

Environmental Modelling and Assessment

Steiner Tree Problems in Computer Communication Networks

Spatial graphs: principles and applications for habitat connectivity

Ecosystems

Greedy randomized adaptive search procedures

Journal of Global Optimization

Short communication
Linking habitats for multiple species