Optimality in prioritizing conservation projects

The resources available to safeguard biodiversity are limited, and so funding must be allocated cost-effectively. To achieve this, conservation projects—such as threatened species recovery projects, or pest management projects—are often prioritized using algorithms. Conventionally, prioritizations have been generated by ranking projects according to their cost-effectiveness and selecting the top projects within a budget, or using backwards heuristic algorithms which iteratively remove projects until a budget is met. Yet such algorithms may not deliver optimal solutions. We investigated the performance of exact algorithms, a class of algorithms that guarantee optimality, compared with conventional algorithms for project prioritization. Specifically, we conducted a simulation study, involving 40 conservation projects, 50 species, and 80 management actions, and a case study involving recovery projects for 62 of New Zealand’s threatened bird species. In each of these studies, we generated prioritizations by (i) exact algorithms, (ii) ranking projects using cost-effectiveness, (iii) a backwards heuristic algorithm, and (iv) randomly funding projects. After generating the prioritizations, we evaluated their performance. We found that exact algorithms outperform conventional algorithms for project prioritization. In the simulation study, both the ranking and backwards heuristic algorithms returned solutions that were highly suboptimal when compared with solutions by exact algorithms. In the case study, both conventional algorithms returned solutions that would be expected to result in the needless loss of millions of years of avian evolutionary history due to poor planning. Furthermore, conventional algorithms returned solutions with large amounts of unallocated funding—providing little guidance for decision makers. Despite concerns that exact algorithm solvers require an inordinate amount of time, the longest run in either study took less than three minutes. Our results suggest that conservation agencies could benefit enormously from exact algorithms. To help make exact algorithms more accessible, we developed the oppr R package (https://CRAN.R-project.org/package=oppr) which can use open-source and commercial exact algorithm solvers to identify optimal solutions for a range of objectives and constraints. Our findings suggest that conservation plans could be substantially improved using exact algorithms, which could potentially save millions of dollars and lead to more species being saved from extinction.

[1]  Andrew Reeson,et al.  Benefits of integrating complementarity into priority threat management , 2015, Conservation biology : the journal of the Society for Conservation Biology.

[2]  Emmanuel Paradis,et al.  ape 5.0: an environment for modern phylogenetics and evolutionary analyses in R , 2018, Bioinform..

[3]  H. Possingham,et al.  Biodiversity gains from efficient use of private sponsorship for flagship species conservation , 2015, Proceedings of the Royal Society B: Biological Sciences.

[4]  Hugh P. Possingham,et al.  Balancing phylogenetic diversity and species numbers in conservation prioritization, using a case study of threatened species in New Zealand , 2014 .

[5]  Daniel P Faith,et al.  Threatened Species and the Potential Loss of Phylogenetic Diversity: Conservation Scenarios Based on Estimated Extinction Probabilities and Phylogenetic Risk Analysis , 2008, Conservation biology : the journal of the Society for Conservation Biology.

[6]  Alex James,et al.  Global variation in terrestrial conservation costs, conservation benefits, and unmet conservation needs , 2003, Proceedings of the National Academy of Sciences of the United States of America.

[7]  A. Land,et al.  An Automatic Method for Solving Discrete Programming Problems , 1960, 50 Years of Integer Programming.

[8]  Volker Loeschcke,et al.  The optimization of biodiversity conservation , 1995 .

[9]  I. Chades,et al.  Priority Threat Management for biodiversity conservation: A handbook , 2018, Journal of Applied Ecology.

[10]  Hugh P. Possingham,et al.  Effectiveness of alternative heuristic algorithms for identifying indicative minimum requirements for conservation reserves , 1997 .

[11]  Kerrie A. Wilson,et al.  Mixed policies give more options in multifunctional tropical forest landscapes , 2017 .

[12]  Liana N. Joseph,et al.  Evaluating Trade‐Offs between Target Persistence Levels and Numbers of Species Conserved , 2016 .

[13]  S. Trewick,et al.  Population structure and biogeography of Hemiphaga pigeons (Aves: Columbidae) on islands in the New Zealand region , 2011 .

[14]  Tobias Achterberg,et al.  Mixed Integer Programming: Analyzing 12 Years of Progress , 2013 .

[15]  Hugh P Possingham,et al.  Priority threat management of invasive animals to protect biodiversity under climate change , 2015, Global change biology.

[16]  Kevin J. Gaston,et al.  Optimisation in reserve selection procedures—why not? , 2002 .

[17]  A. Møller,et al.  Using the BirdTree.org website to obtain robust phylogenies for avian comparative studies: A primer. , 2015, Current zoology.

[18]  Harry John Betteley Birks,et al.  How to maximize biological diversity in nature reserve selection: Vascular plants and breeding birds in deciduous woodlands, western Norway , 1993 .

[19]  Jeet Sukumaran,et al.  DendroPy: a Python library for phylogenetic computing , 2010, Bioinform..

[20]  Manuela M. P. Huso,et al.  A comparison of reserve selection algorithms using data on terrestrial vertebrates in Oregon , 1997 .

[21]  C. Thompson,et al.  Optimal investment in conservation of species , 2008 .

[22]  J. Orestes Cerdeira,et al.  Flexibility, efficiency, and accountability: adapting reserve selection algorithms to more complex conservation problems , 2000 .

[23]  Gaurav Vaidya,et al.  Avibase – a database system for managing and organizing taxonomic concepts , 2014, ZooKeys.

[24]  Hugh P. Possingham,et al.  Optimality in reserve selection algorithms: When does it matter and how much? , 1996 .

[25]  Neil D. Burgess,et al.  Heuristic and optimal solutions for set-covering problems in conservation biology , 2003 .

[26]  S. Polasky,et al.  Integrating economic costs into conservation planning. , 2006, Trends in ecology & evolution.

[27]  Jorge Soberón,et al.  Global Mammal Conservation: What Must We Manage? , 2005, Science.

[28]  Hugh P. Possingham,et al.  Solving conservation planning problems with integer linear programming , 2016 .

[29]  R. Fuller,et al.  Reconciling recreational use and conservation values in a coastal protected area , 2016 .

[30]  W. Jetz,et al.  The global diversity of birds in space and time , 2012, Nature.

[31]  H. Possingham,et al.  Tax Shifting and Incentives for Biodiversity Conservation on Private Lands , 2018 .

[32]  Liana N. Joseph,et al.  Optimal Allocation of Resources among Threatened Species: a Project Prioritization Protocol , 2009, Conservation biology : the journal of the Society for Conservation Biology.

[33]  A. Baker,et al.  Explosive ice age diversification of kiwi , 2016, Proceedings of the National Academy of Sciences.

[34]  Les G. Underhill,et al.  Optimal and suboptimal reserve selection algorithms , 1994 .

[35]  Kendall R. Jones,et al.  Poor ecological representation by an expensive reserve system: Evaluating 35 years of marine protected area expansion , 2018, Conservation Letters.

[36]  Atte Moilanen,et al.  Assessing replacement cost of conservation areas: how does habitat loss influence priorities? , 2009 .

[37]  Ted K. Ralphs,et al.  The Symphony Callable Library for Mixed Integer Programming , 2005 .

[38]  János D. Pintér,et al.  Global Optimization: Software, Test Problems, and Applications , 2002 .

[39]  Lynn A. Maguire,et al.  Endangered species recovery: A resource allocation problem , 2018, Science.

[40]  W. Probert,et al.  A large-scale application of project prioritization to threatened species investment by a government agency , 2018, PloS one.