A Composite Network Approach for Assessing Multi-Species Connectivity: An Application to Road Defragmentation Prioritisation

Luca Santini, Santiago Saura & Carlo Rondinini

One of the biggest challenges in large-scale conservation is quantifying connectivity at broad geographic scales and for a large set of species. Because connectivity analyses can be computationally intensive, and the planning process quite complex when multiple taxa are involved, assessing connectivity at large spatial extents for many species turns to be often intractable. Such limitation results in that conducted assessments are often partial by focusing on a few key species only, or are generic by considering a range of dispersal distances and a fixed set of areas to connect that are not directly linked to the actual spatial distribution or mobility of particular species. By using a graph theory framework, here we propose an approach to reduce computational effort and effectively consider large assemblages of species in obtaining multi-species connectivity priorities. We demonstrate the potential of the approach by identifying defragmentation priorities in the Italian road network focusing on medium and large terrestrial mammals. We show that by combining probabilistic species graphs prior to conducting the network analysis (i) it is possible to analyse connectivity once for all species simultaneously, obtaining conservation or restoration priorities that apply for the entire species assemblage; and that (ii) those priorities are well aligned with the ones that would be obtained by aggregating the results of separate connectivity analysis for each of the individual species. This approach offers great opportunities to extend connectivity assessments to large assemblages of species and broad geographic scales.


Fig 2. (a) Amount of suitable habitat (node weight), (b) Road density (used for obtaining the link weights), (c) restoration priority as given by varPC values (cells where actions to mitigate the barrier effect of roads would yield the highest benefit) according to the cumulative results (sum of individual species restoration priorities), and (d) restoration priority according to the best performing composite network (composite network F).

Read full publication

Posted in Lab

Setting population targets for mammals using body mass as a predictor of population persistence

Jelle P. Hilbers, Luca Santini, Piero Visconti, Aafke M. Schipper, Cecilia Pinto, Carlo Rondinini, and Mark A.J. Huijbregts

Conservation planning and biodiversity assessments need quantitative targets to optimize planning options and assess the adequacy of current species protection. However, targets aiming at persistence require population-specific data, which limits their use in favor of fixed and non-specific targets, likely leading to unequal distribution of conservation efforts among species. Here we propose a method to derive equitable population targets, which are quantitative targets of population size that ensure equal probabilities of persistence across a set of species, and can be easily inferred from species-specific traits. We applied population dynamics models across a range of life-history traits representative for mammals, and estimated minimum viable population targets intrinsically related to species body mass. Our approach provides a compromise between pragmatic non-specific targets, and detailed context-specific estimates of population viability for which only limited data is available. It enables a first estimation of species-specific population targets based on a readily available trait, and thus allows setting equitable targets for population persistence in large-scale and multispecies conservation assessments and planning.

Read full publication.

Posted in Lab

Global mammal beta diversity shows parallel assemblage structure in similar but isolated environments

Caterina Penone, Ben G. Weinstein, Catherine H. Graham, Thomas M. Brooks, Carlo Rondinini, S. Blair Hedges, Ana D. Davidson & Gabriel C. Costa

The taxonomic, phylogenetic and trait dimensions of beta diversity each provide us unique insights into the importance of historical isolation and environmental conditions in shaping global diversity. These three dimensions should, in general, be positively correlated. However, if similar environmental conditions filter species with similar trait values, then assemblages located in similar environmental conditions, but separated by large dispersal barriers, may show high taxonomic, high phylogenetic, but low trait beta diversity. Conversely, we expect lower phylogenetic diversity, but higher trait biodiversity among assemblages that are connected but are in differing environmental conditions. We calculated all pairwise comparisons of approximately 110 × 110 km grid cells across the globe for more than 5000 mammal species (approx. 70 million comparisons). We considered realms as units representing geographical distance and historical isolation and biomes as units with similar environmental conditions. While beta diversity dimensions were generally correlated, we highlight geographical regions of decoupling among beta diversity dimensions. Our analysis shows that assemblages from tropical forests in different realms had low trait dissimilarity while phylogenetic beta diversity was significantly higher than expected, suggesting potential convergent evolution. Low trait beta diversity was surprisingly not found between isolated deserts, despite harsh environmental conditions. Overall, our results provide evidence for parallel assemblage structure of mammal assemblages driven by environmental conditions at a global scale.


Hypothesis framework and expected mapped results. We expect trait and phylogenetic beta diversity to be coupled in most cases (bottom left and top right). Dimensions of beta diversity can be decoupled when assemblages are located in contrasting environments within a realm because of limited historic isolation and environmental filtering (top left) or in similar environments of different realms because of convergent structure of assemblages in similar environmental con- ditions (bottom right). Mechanisms corresponding to each combination of high and low beta diversity dimensions are in italics. Colours in maps highlight expected median beta diversity for specific examples.


Read full publication.

Posted in Lab

Assessing the Cost of Global Biodiversity and Conservation Knowledge

Diego Juffe-Bignoli , Thomas M. Brooks, Stuart H. M. Butchart, Richard B. Jenkins, Kaia Boe, Michael Hoffmann, Ariadne Angulo, Steve Bachman, Monika Böhm, Neil Brummitt, Kent E. Carpenter, Pat J. Comer, Neil Cox, Annabelle Cuttelod, William R. T. Darwall, Moreno Di Marco, Lincoln D. C. Fishpool, Bárbara Goettsch, Melanie Heath, Craig Hilton-Taylor, Jon Hutton, Tim Johnson, Ackbar Joolia, David A. Keith, Penny F. Langhammer, Jennifer Luedtke, Eimear Nic Lughadha, Maiko Lutz, Ian May, Rebecca M. Miller, María A. Oliveira-Miranda, Mike Parr, Caroline M. Pollock, Gina Ralph, Jon Paul Rodríguez, Carlo Rondinini, Jane Smart, Simon Stuart, Andy Symes, Andrew W. Tordoff, Stephen Woodley, Bruce Young and Naomi Kingston

Knowledge products comprise assessments of authoritative information supported by standards, governance, quality control, data, tools, and capacity building mechanisms. Considerable resources are dedicated to developing and maintaining knowledge products for biodiversity conservation, and they are widely used to inform policy and advise decision makers and practitioners. However, the financial cost of delivering this information is largely undocumented. We evaluated the costs and funding sources for developing and maintaining four global biodiversity and conservation knowledge products: The IUCN Red List of Threatened Species, the IUCN Red List of Ecosystems, Protected Planet, and the World Database of Key Biodiversity Areas. These are secondary data sets, built on primary data collected by extensive networks of expert contributors worldwide. We estimate that US$160 million (range: US$116–204 million), plus 293 person-years of volunteer time (range: 278–308 person-years) valued at US$ 14 million (range US$12–16 million), were invested in these four knowledge products between 1979 and 2013. More than half of this financing was provided through philanthropy, and nearly three-quarters was spent on personnel costs. The estimated annual cost of maintaining data and platforms for three of these knowledge products (excluding the IUCN Red List of Ecosystems for which annual costs were not possible to estimate for 2013) is US$6.5 million in total (range: US$6.2–6.7 million). We estimated that an additional US$114 million will be needed to reach pre-defined baselines of data coverage for all the four knowledge products, and that once achieved, annual maintenance costs will be approximately US$12 million. These costs are much lower than those to maintain many other, similarly important, global knowledge products. Ensuring that biodiversity and conservation knowledge products are sufficiently up to date, comprehensive and accurate is fundamental to inform decision-making for biodiversity conservation and sustainable development. Thus, the development and implementation of plans for sustainable long-term financing for them is critical.

Read full publication.

Posted in Lab

Using habitat suitability models to scale up population persistence targets

Moreno Di Marco, Luca Santini, Piero Visconti, Alessio Mortelliti, Luigi Boitani & Carlo Rondinini

ARC Setting operational targets for the protection of species is crucial for identifying conservation priorities and for monitoring conservation actions’ effectiveness. The use of quantitative targets for global species conservation has grown in the past ten years as a response to the commitment of reducing extinction rates established by the Convention on Biological Diversity. We reviewed the use of conservation targets in global scale conservation analyses, and found that most of the publications adopted species representation targets, corresponding to an amount of area to be protected. We found no work adequately targeting species’ persistence, i.e. the complement to species extinction risk. Despite the adoption of pragmatic population targets, consisting in a number of individuals to be protected, has been recently proposed for global species conservation, the use of these targets at the species level is not always warranted. Pros and cons of using population persistence targets for species conservation have been discussed, yet the fundamental issue of how to scale these targets from populations to species is still unresolved. We discuss the process of “scaling up” population persistence targets to the species level using habitat distribution models, and test our approach in a case study on the European ground squirrel (Spermophilus citellus). We identified three main steps to be followed: (i) definition of a population target, (ii) characterisation of the species’ populations by means of a habitat suitability model, and (iii) definition of a scaled species target. An up-scaled species target should include multiple conditions reflecting species persistence (number, size, location of the populations to be protected), uniqueness (e.g. evolutionary potential) and representativeness (e.g. presence in different ecosystems). Adopting scaled up species persistence targets within conservation planning approaches can allow protected area plans to give the highest contribution to reducing global species extinction risk.

Distribution range of Spermophilus citellus. Suitable habitat (coloured area) is surrounded by a potential dispersal matrix (shaded area) within the species range (in light grey). Areas smaller than the defined target area are reported in dark green, while clusters of suitable habitat larger than the target area are reported in random colours (with different colours indicating different clusters).  

Read full publication.

Posted in Lab

Assessing the umbrella value of a range-wide conservation network for jaguars

Daniel Thornton, Kathy Zeller, Carlo Rondinini, Luigi Boitani, Kevin Crooks, Christopher Burdett, Alan Rabinowitz and Howard Quigley.

Umbrella species are employed as conservation short-cuts for the design of reserves or reserve networks. However, empirical data on the effectiveness of umbrellas is equivocal, which has prevented more widespread application of this conservation strategy. We perform a novel, large-scale evaluation of umbrella species by assessing the potential umbrella value of a jaguar (Panthera onca) conservation network (consisting of viable populations and corridors) that extends from Mexico to Argentina. Using species richness, habitat quality, and fragmentation indices of ~1500 co-occurring mammal species, we show that jaguar populations and corridors overlap a substantial amount and percentage of high-quality habitat for co-occurring mammals and that the jaguar network performs better than random networks in protecting high-quality, interior habitat. Significantly, the effectiveness of the jaguar network as an umbrella would not have been noticeable had we focused on species richness as our sole metric of umbrella utility. Substantial inter-order variability existed, indicating the need for complementary conservation strategies for certain groups of mammals. We offer several reasons for the positive result we document, including the large spatial scale of our analysis and our focus on multiple metrics of umbrella effectiveness. Taken together, our results demonstrate that a regional, single-species conservation strategy can serve as an effective umbrella for the larger community and should help conserve viable populations and connectivity for a suite of co-occurring mammals. Current and future range-wide planning exercises for other large predators may therefore have important umbrella benefits.


A map of the jaguar conservation network for the Americas. The network consists of Jaguar Conservation Units (JCUs; in black), which maintain viable populations of jaguars, and jaguar corridors (in gray), linking the JCUs.


Read full publication.

Posted in Lab

Toward quantification of the impact of 21st‐century deforestation on the extinction risk of terrestrial vertebrates

Łukasz Tracewski, Stuart H.M. Butchart, Moreno Di Marco, Gentile F. Ficetola, Carlo Rondinini, Andy Symes, Hannah Wheatley, Alison E. Beresford, and Graeme M. Buchanan

Conservation actions need to be prioritised, often taking into account species’ extinction risk. The International Union for Conservation of Nature (IUCN) Red List provides an accepted objective framework for the assessment of extinction risk, but field data to apply the IUCN Red List criteria are often limited. Information collected through remote sensing can inform these assessments, and forests are perhaps the best-studied habitat type for use in this approach. Using an open-access 30 m resolution map of tree cover and its change between 2000 and 2012, the extent of forest cover and loss within the distributions of 11,186 forest-dependent amphibians, birds and mammals worldwide was assessed. Sixteen species have experienced sufficiently high rates of forest loss to be considered at elevated extinction risk under Red List criterion A, owing to inferred rapid population declines. This number would increase to 23 if data deficient species (i.e., those with insufficient information previously to apply the Red List criteria) were included. Some 484 species (855 if data deficient species are included) may be considered at elevated extinction risk under Red List criterion B2, owing to restricted areas of occupancy resulting from little forest cover remaining within their ranges. This would increase the proportion of species of conservation concern by 32.8% for amphibians, 15.1% for birds and 24.7% for mammals. Central America, the Northern Andes, Madagascar, the Eastern Arc forests in Africa and the islands of South-East Asia are hotspots for these species. The analyses illustrate the utility of satellite imagery for global extinction risk assessment and measurement of progress towards international environmental agreement targets. We highlight areas for which subsequent analyses could be performed on satellite image data in order to improve our knowledge of extinction risk of species.


Number of species potentially qualifying for a higher International Union for Conservation of Nature Red List threat category: (a) amphibians, (b) birds, (c) mammals, and (d) all species combined. Data deficient species are excluded.

Read full publication.

Posted in Lab

Analysing biodiversity and conservation knowledge products to support regional environmental assessments

Thomas M. Brooks, H. Resit Akçakaya, Neil D. Burgess, Stuart H.M. Butchart, Craig Hilton-Taylor, Michael Hoffmann, Diego Juffe-Bignoli, Naomi Kingston, Brian MacSharry, Mike Parr, Laurence Perianin, Eugenie C. Regan, Ana S.L. Rodrigues, Carlo Rondinini, Yara Shennan-Farpon & Bruce E. Young.

Two processes for regional environmental assessment are currently underway: the Global Environment Outlook (GEO) and Intergovernmental Platform on Biodiversity and Ecosystem Services (IPBES). Both face constraints of data, time, capacity, and resources. To support these assessments, we disaggregate three global knowledge products according to their regions and subregions. These products are: The IUCN Red List of Threatened Species, Key Biodiversity Areas (specifically Important Bird & Biodiversity Areas [IBAs], and Alliance for Zero Extinction [AZE] sites), and Protected Planet. We present fourteen Data citations: numbers of species occurring and percentages threatened; numbers of endemics and percentages threatened; downscaled Red List Indices for mammals, birds, and amphibians; numbers, mean sizes, and percentage coverages of IBAs and AZE sites; percentage coverage of land and sea by protected areas; and trends in percentages of IBAs and AZE sites wholly covered by protected areas. These data will inform the regional/subregional assessment chapters on the status of biodiversity, drivers of its decline, and institutional responses, and greatly facilitate comparability and consistency between the different regional/subregional assessments.


Proportion of species, by Red List Category, in comprehensively assessed groups on The IUCN Red List of Threatened Species (Version 2015-2) occurring in each IPBES region (a) and subregion (b); and proportion of endemic species, by Red List Category, in comprehensively assessed groups on The IUCN Red List of Threatened Species (Version 2015-2) occurring in each IPBES region (c) and subregion (d). The vertical red lines show the best estimate for the proportion of extant species considered threatened (CR, EN and VU) if Data Deficient species are Threatened in the same proportion as data-sufficient species. The numbers to the right of each bar represent the total number of species assessed and in parentheses the best estimate of the percentage threatened. CR, critically endangered; DD, data deficient; EN, endangered; EW, extinct in the wild; EX, extinct; LC, least concern; NT, near threatened; VU, vulnerable.

Read full publication.


Posted in Lab

Connectivity of the global network of protected areas

Luca Santini, Santiago Saura & Carlo Rondinini.

Millennia of human activity have drastically shaped the Earth’s surface confining wildlife in ever more rare and sparse habitat fragments. Within the strategic Plan for Biodiversity 2011–2020, Aichi Target 11 aims at the expansion of the current protected area (PA) system and the maintenance and improvement of its connectivity. This study aims at providing the first overview of the functionality of the PA networks across the six continents at different dispersal distances relevant for terrestrial mammals.


We used a graph theory approach to assess the connectivity of PA networks of different continents across a wide range of dispersal distances. We assessed the connectivity of country-level PA networks, the connectivity of con- tinental PA networks and the contribution of country-level PA networks to continental connectivity.
Results National and continental networks are characterized by very different spatial arrangements that translate into different levels of connectivity, ranging from networks where the reachable area is mostly determined by structural connectivity within PAs (e.g. Africa) to networks where connectivity mostly depends on animal dispersal among PAs (e.g. Europe). PA size correlates positively with connectivity for all species, followed by PA number; dispersal contributes less and positively interacts with number of PAs.

Main conclusions
Continental networks perform worse than national networks. Transboundary connectivity is often weak and should be improved, especially for countries that are important in promoting continental connectivity. Increasing PA coverage and size is a good strategy to improve multispecies connectivity.


Percentage of reachable area (ECAnorm) for the protected area networks within world countries. (a) represents ECAnorm for the lowest dispersal distance considered (177 m). (b) represents the difference in ECAnorm between the lowest and the maximum dispersal distance considered (99.58 km), thus indicating the sensitivity to dispersal distance of each country’s network.

Read full publication.

Posted in Lab

Synergies and trade-offs in achieving global biodiversity targets

Moreno Di Marco, Stuart H. M. Butchart, Piero Visconti, Graeme M. Buchanan, Gentile F. Ficetola and Carlo Rondinini.

Following their failure to achieve a significant reduction in the global rate of biodiversity loss by 2010, world governments adopted 20 new ambitious Aichi biodiversity targets to be met by 2020.There is growing recognition that efforts to achieve one particular biodiversity target can contribute to achieving others, yet little attention is given to the fact that different targets may require conflicting solutions. Consequently, there is a risk that lack of strategic thinking might result, once again, in a failure to achieve governmental commitments to biodiversity conservation. We illustrate this dilemma by focusing on Aichi Target 11. This requires an expansion of terrestrial protected area coverage, which could also contribute to reducing the loss of natural habitats (Target 5), reducing human-induced species decline and extinction (Target 12), and maintaining global carbon stocks (Target 15). We consider the potential impact of expanding protected areas to mitigate global deforestation and the consequences for the distribution of suitable habitat for >10000 species of forest vertebrates (amphibians, birds and mammals). We found that expanding protected areas toward locations with the highest deforestation rates (Target 5) or the highest potential loss of aggregate species’ suitable habitat (Target 12) would result in partially different protected area network configurations (overlapping with each other by ca. 73%). Moreover, the latter approach would contribute to safeguarding ca. 30% more global carbon stocks (measures as tons/ha) than the former. Further investigation of synergies and trade-offs between targets would shed light on these and other complex interactions, such as the interaction between reducing overexploitation of natural resources (Targets 6, 7), controlling invasive alien species (Target 9) and preventing extinctions of native species (Target 12). Synergies between targets must be identified and secured soon and trade-offs must be minimized, before the options for co-benefits are reduced by human pressures.


Current protected areas and areas with highest deforestation impact. The maps represent: (a) current extension of the protected area network; (b) areas of synergy, i.e. places where highest forest loss correspond to highest habitat loss aggregated across species; (c) areas of forest trade-off, where highest forest loss is expected, but not highest loss of aggregate species habitat; (d) areas of species trade-off, where highest loss of aggregate species habitat is expected, but not highest forest loss. See Supporting Information for a color version of the map.

Read full publication.

Posted in Lab