Introduction to Marxan Spatial Planning Software Heather Coleman Science Advisor, PacMARA Leif Olson Landscape Ecologist, O2 Planning + Design Linking Science with Decision Making
1. Introduction to Systematic Multi-Objective (Conservation & Resource Use) Planning Linking Science with Decision Making
Agenda Today 8:30 8:50 Setting the context for Planning Introductions, purpose and overview of the course 8:50 9:40 Key concepts in systematic conservation planning Brief history, how tools can be a useful for systematic planning 9:40 11:00 Network planning with decision support tools Marxan and Marxan with Zones concepts, role in planning processes Interactive planning activity I Examples of systematic conservation planning tools 10:00 10:15 Refreshment 11:00 12:00 Interactive planning activity II Explanation of example exercise and Zonae Cogito 12:00 1:00 Lunch 13:00 13:45 How does Marxan find good solutions? Simulated Annealing demonstration and other technical concepts 13:45 15:00 Continue planning activity II 15:00 15:30 Refreshment 15:30 16:00 Marxan case study Marxan s role in land use planning in Alberta, Canada Leif Olson 16:00 16:30 Setting goals, objectives, and targets Why targets can be the most difficult part of the whole process 16:30 17:30 Continue planning activity II
Agenda Tomorrow 8:30 8:45 Review of Day 1 activities, progress, and questions 8:45 9:30 Incorporating cost values into Marxan analyses How cost layers have been developed and examples 9:30 10:00 Interactive planning activity III Calibration, configuration editor, and cluster analysis 10:00 10:15 Refreshment 10:15 11:00 Working with Marxan results and stakeholder communication Good practices, pitfalls, and lessons learned Using Zonae Cogito and tools with stakeholders 11:00 12:00 Discussion Sharing experiences in marine planning Opportunities to work together 12:00 13:30 Lunch
The Problem Land use Tourism Oil &Gas Mariculture Coastal Defence Ports & Navigation Military Activities Culture Conservation Dredging & Disposal Fishing Submarine Cables Slide: Nicole Schaefer Renewable Marine Mineral Extraction Energy Recreation
In a Nutshell Humans are increasingly influencing the Earth s ecosystems (age of the Anthropocene ) Human activities have varying impacts and compatibilities (and are changing) Some habitats / species are more vulnerable to some activities Decisions are increasingly multicriteria, requiring spatial 11 efficiency Image: Natalie Ban
Conflicted History of Park Selection Rock & ice: Land nobody wanted Freaks of nature Charismatic animals Playgrounds in the wilderness Slide adapted from Emily Gonzoles, CACR, UBC
Could Existing Parks be False Friends? We found that despite spanning less than 4% of South Australian state waters, locking in the existing ad hoc marine reserves presented considerable opportunity costs. Even with representation targets set at 50%, more than half of South Australia s existing marine reserves were selected randomly or less in efficient marine reserve systems. Hence, ad hoc marine reserve systems are likely to be inefficient and may compromise effective conservation of marine biodiversity. Stewart, R.R., et al. 2003. Opportunity cost of ad hoc marine reserve design decisions: An example from South Australia. MEPS 253: 25-38.
Access to Full Information Improves Decisions Shared fishing data could have improved outcomes in California California Marine Life Protection Act Initiative (limited fishing data) Marxan Solutions limited fishing data Marxan full fishing data Fishers Conservation Mixed Group Final Decision Klein et al. 2008. Conservation Letters.
A More Systematic Approach to Planning A more systematic approach to locating and designing reserves has been evolving and this approach will need to be implemented if a large proportion of today s biodiversity is to exist in a future of increasing numbers of people and their demands on natural resources. Margules & Pressey 2000. Nature
What is Systematic Conservation Planning? Conservation planning: guides decisions about reserve location, configuration and management Protected Areas: An area of land and/or sea especially dedicated to the protection of biological diversity, and of natural and associated cultural resources, and managed through legal or other effective means (IUCN, 1994) Systematic Conservation Planning: process of making conservation decisions in a way that is efficient, repeatable, transparent & equitable
What is Systematic Conservation Planning? Structured step-wise approach Developing goals, objectives, targets Identifying existing gaps Identifying (possible) conservation sites Mapping (possible) conservation networks With feedback, revision, reiteration For SCP to be successful, there are many other ingredients, common to most public processes, such as participation, funding, governance, etc. We will come back to this later
Systematic Conservation Planning Steps 1. Scope and cost 2. Identify and involve stakeholders 3. Identify goals 4. Compile data 5. Set conservation targets 6. Assess existing conservation areas 7. Select new conservation areas 8. Implement conservation action 9. Maintain and monitor Slide Courtesy of Bob Pressey
Key Principles of SCP 1. Comprehensiveness See: Kukkala & Moilanen 2013. Biological Reviews.
Key Principles of SCP 1. Comprehensiveness The ideal is to sample every kind of biodiversity In practice, this is not possible so we try to include data that cover: species (and genes) habitats composition structure ecological processes function ecological regions biogeography
2. Adequacy Key Principles of SCP
2. Adequacy Key Principles of SCP Protecting enough to ensure persistence of biodiversity features How much is enough? (difficult question!) Usually addressed with targets Considering threats / habitats outside protected areas will influence how much space / representation is needed inside
Key Principles of SCP 3. Representativity
Key Principles of SCP 3. Representativity Sampling across the full range of variation of each feature (e.g., species or habitats) Capturing areas that have high variability within targeted features, as well as high productivity / biomass
4. Efficiency Key Principles of SCP
4. Efficiency Key Principles of SCP Achieving objectives for minimal cost Cost can be defined in terms of: acquisition cost ($) operational costs ($) opportunities lost (for users and industries) social values (local importance ) political (gain or loss of credibility / votes) A combination of the above or others
Establishing Sites in a Network Brings New Considerations Existing protected areas usually have to be factored in, even if they are not ideal Special places ( jewels ) recognised for their unique / irreplaceable ecology Threats in some places are more pressing than in others Achievable? (financially, legally, mandates) Broadly supported? (now and/or in the future?)
Multiple Criteria Decision-making A story: some boats are hard to build Multiple criteria problems are not just limited to EBSAs or protected areas. Imagine you want to buy a boat that is: fairly fast (>10 knots, say) fairly comfortable (allowing weekend trips, sleeping on board overnight) fairly economical (both to buy and to operate)
Finding a boat on just these 3 criteria should be easy, right? Even though the request sounds reasonable, boat sales people just shake their heads sadly No such boat exists, they tell you Instead, they show you the following
You can have fast & comfortable But that s not very affordable. Or Ben Bawden
You can have fast & economical But that s not very comfortable. Or
You can have economical & comfortable But that s not very fast.
The Moral of this Story: One site will not fit all expectations Some multiple criteria problems are very hard to solve, and compromises, if they exist, will make everyone unhappy It is better for parties to reconsider the criteria one by one for a given place If the criteria are truly irreconcilable, it is a better to focus on some criteria in one place, and the other criteria in another Design a network of complementary sites
Multi-criteria Decision Making Site selection / optimisation tools Sorting through more than five or six GIS layers is complicated tools can help But where data are scarce, Marxan-like tools are inappropriate and will favour the few areas with data The local expert approach: In data-scarce situations it is better to use discussions with locals and other experts to sort through multi-criteria problems When the tools work, they really help: when data and technical know-how are available, these tools are very powerful and are the preferred approach to dealing with multiple criteria Tools can evaluate scenarios too: can be used to check revisions against agreed targets and objectives
Properties of Tools like Marxan for Systematic Multi-Objective Planning Data-driven Directed at explicit objectives identifying existing gaps Efficient Explicit, transparent, repeatable Flexible Credit: Bob Pressey in Ardron et al 2010, UNEP report on amsp
Summary of Systematic Planning Multiple human uses and conservation interests compete for space and resources Ad hoc planning processes have set up paper parks and false friends Systematic planning process involve a specific set of steps (some involving tools) to promote informed resource use and/or conservation decision making
Pause for Questions? Comments? Discussion
2. Marxan as a Tool for Systematic Multi-Objective Planning Linking Science with Decision Making
Request for the Day Get me a reasonable amount of every mappable feature while minimizing the crankiness of other users in a system of sanctuaries that are relatively clumped! Credit: Natalie Ban
Why use tools? Tools can assist in complicated problems beyond human intuition or conventional approaches Save time Guide user through processes Stop reinventing the wheel Reduce the need for some human expertise Explore a wider range of alternatives Document decisions Integrating sector tools can help integrate planning across sectors Slide adapted from Sarah Carr, EBM Tools Network
What tools don t do Provide answers (decision support not decision makers) Replace need for intensive human interaction and collaboration Overcome people problems of politics, turf, mistrust, and technophobia Come with all the data they need Replace need for project-specific analyses Slide adapted from Sarah Carr, EBM Tools Network
What is the problem that Marxan & MwZones is trying to solve? Biodiversity features, targets Opportunity cost
Selecting conservation areas Establish the question formally Typical conservation area selection problems: The minimum set problem = capture some minimum representation of features for the least cost The maximum coverage problem = capture as much representation of features as possible with a fixed budget
Minimum set problem in Marxan 1. To minimize: The Marxan objective is: a) Total Cost of the reserve network b) Total Boundary of the network 2. While meeting all targets (i.e., minimizing penalties for not adequately representing features)
Marxan Why and How A tool to support spatial planning Minimize: Total cost and total boundary of a network, while meeting all (conservation) targets Selects multiple good solutions: sets of possible sites that together meet objectives Objectives are met in an efficient manner
Steps in a Marxan Analysis 1. Identify goals and objectives 2. Identify features, map best available data 3. Divide the area into planning units 4. Tally amount of each feature in each PU 5. Set targets for each feature 6. Run Marxan analysis, calibrate parameters 7. Map outputs from desired scenarios Reiterate
The Marxan score Combined Planning Unit Cost (efficiency) + Combined Boundary Length (clumping) + Combined Target Shortfall (penalty for not achieving conservation targets) = Marxan Score
Example of Marxan scoring PU area = 1 km x 1km Boundary length = 1 km BLM = 1.5 Target for all 3 species = represented at least once SPF = 10 Source: Modified from Smith, 2004
Example of Marxan scoring 32 Total PU cost = 4 Boundary cost = 12*BLM SPF = 10 16 Total PU cost = 4 Boundary cost = 8*BLM SPF not applied TOTAL SCORE Source: Modified from Smith, 2004
Interactive Planning Activity I Objectives: Represent target amount for features 1, 2, 3 Minimize summed cost of selected area Consider spatial configuration by trying to selected planning units that are adjacent to other selected planning units
Advanced: Consider clumping Count the number of outside edges BLM = 100 Each edge counts * 100 12 edges * 100 = 1200 8 edges * 100 = 800 46
Compute target gap and cost www.aproposinfosystems.com/media/marxan-demo.html 47
You can start now You have 10-15 min to find the best solution. Good luck!! 48
Marxan results Lowest cost solution = 1775
Marxan results Lowest cost clumped solution = 3140
Marxan outputs Best Solution (configuration with the lowest score) Solution 1 Solution 2 Solution 3 Solution 4 Solution 5 Score = 12 Score = 15 Score = 20 Score = 8 Score = 10 Source: Modified from Smith, 2004
Marxan outputs Selection frequency (or summed solution) Solution 1 Solution 2 Solution 3 Solution 4 Solution 5 Numbers represent how many times each PU was selected Source: Modified from Smith, 2004
Now consider More features (a few hundred?) More spatial constraints The problem gets so large that it is impossible to find a good solution in reasonable amount of time 53
Site selection Marxan with Zones is often the best choice for multiobjective planning/zoning if you have the data! Otherwise, check out also: C-Plan (Pressey, Univ. of Queensland / James Cook U.) Zonation (Moilanen, Univ. of Helsinki) ConsNet (Ciarleglio, U. of Texas)
Seasketch.org - not optimisation, but could be combined with Marxan
Example application: California Objectives: Identify a network of marine reserves Represent biological and physical diversity Minimize impact to commercial and recreational fishers
Planning units 1 NM 2 planning units (CA Department of Fish and Game management units) Calculated how much of each feature is in each PU Marxan provides potential solutions for meeting conservation goals at a minimum cost to fisheries Morro Bay
Conservation features Examples: Rocky reefs Kelp beds Estuaries Bird colonies Breeding sites Seamounts Canyons Pinnacles Choices guided by legislation, science advisory team, and data availability
Socio-economic cost of reservation Cost is not spatially homogenous! In this example, cost is measured as: Recreational Fishing Effort MARXAN O Recreational fishing effort Commercial fishing effort Fishing Effort 0-5 6-25 26-50 51-240 Planning Unit Inclusion 0-10 11-20 21-30 31-40 41-50 51-60 61-70 71-80 81-90 91-100 Goal: Minimize socio-economic impact
Spatial compactness of reserves Planning Unit Inclusion 0-10 11-20 21-30 31-40 41-50 51-60 61-70 71-80 81-90 91-100 0 10 20 40 60 80 Miles a. BLM = 0 b. BLM = 0.0001 c. BLM = 1 Adjusting the BLM
Marxan outputs solutions vs sums Solution 1 Solution 2 Summed Solution Credit: L. Kircher
Source: Airame 2005, Ban & Klein 2009, Linke et al. 2011 Cluster analysis
Suumary: Marxan is a tool that Addresses core Systematic Conservation Planning principles (e.g., representation, cost efficiency, spatial constraints, complementarity) Identifies multiple good solutions, even to large problems Selects areas in a systematic, repeatable and transparent manner Is free! Decision-support tool, not a decision-maker!
Fears and misconceptions Black box mysterious how it works Precludes expert / other stakeholder input Only works if data are perfect Only applicable to strict reserves Technically demanding (a bit, but many learn!) Using costs is not biologically pure All of these are incorrect! 65
Keep in mind Most issues arise because of communication challenges, not technical issues Computational capacity / algorithms rarely limit conservation planning lack of clear objectives do Marxan will always produce an answer, but without clear goals and objectives, it may not be the right one Many complexities can be added later (don t do it all at once) such as zoning, risks, temporal dynamics 66
Pause for Questions? Comments? Discussion