Research Document - 2004/082
Evaluation of site selection methodologies for use in marine protected area network design
By Evans, S.M.J., G. Jamieson, J. Ardron, M. Patterson, S. Jessen
Executive Summary
This report identifies and compares different methodologies used for the selection of (candidate) marine protected areas (mpas), termed areas of interest (AOIs). It is hoped that this will provide DFO with the necessary information to evaluate which selection methodology would be most effective in furthering it’s mpa objectives within the IM framework.
Choosing the most appropriate methodology depends on the underlying goal for establishing the set of marine protected areas. Clearly defining the purpose and the overall conservation goal is an important first step that must not be overlooked.
There are two main approaches to selecting AOIs; scoring/weighting (non-systematic) and systematic.
Scoring methods assign a rank of relative importance to all sites based on some user-defined criteria and then add those sites with the highest rank to an existing reserve. The product from this type of reserve site selection is not able to identify how each site relates to the others in the system beyond it’s ‘score’ which is not indicative of what is being captured by the sites. While the objective nature of a scoring selection process is preferred to subjective or opportunistic decision-making, it is not very rigorous, it is not able to efficiently select a set of complementary sites and does not have the spatial capacity to create a network.
Systematic methods of reserve selection make use of algorithm-based decision support tools. Systematic selection of mpas is based on the concept of ‘complementarity’ in which new sites contain features that are not currently captured in the reserve system and thus augment the overall diversity and representivity of the system. Of the systematic methods there are 4 main types of algorithms used; linear integer programming (ILP), simple iterative algorithms (heuristics), iterative simulated annealing and explicitly spatial population based models.
The advantage of the ILP methodology over other complementarity methods is its ability to find an optimal solution. However, if there are too many constraints or the problem is too complex (non-linear) this method will often fail to produce a solution. Thus it is best applied when there are only a few constraints to be optimised.
Heuristics are much faster than the ILP methods, but may arrive at a solution which is considerably less efficient than the theoretical minimum. These programs can manage conservation problems comprised of large datasets and several constraints. In some cases spatial constraints can be incorporated into the method via additional programming.
The simulated annealing method is considered superior to the other methodologies for selecting priority areas for conservation reviewed here. This algorithm can produce multiple solutions for a given scenario unlike heuristics which only provide one solution. It can produce more efficient solutions compared to heuristics in terms of minimising total area needed to meet the desired conservation objectives. There is a random component of this algorithm that allows for the search of the ‘global minima’.
The last systematic method reviewed in this paper, explicitly spatial programs, specifically address the issue of species persistence through the application of environmental variable models (those which influence the distribution of biodiversity) or metapopulation models that will direct the selection of a ‘connected’ set of sites. These programs can only select sites for a limited number of species and require detailed data sets regarding either environmental parameters or species population dynamics. Thus, they are often most appropriately applied at smaller scales for which this type of data exists, or as a post-selection tool (see section 3.3) to choose among candidate sites in the development of a network that ensures a particular species persistence.
This report also reviewed two case specific applications of the systematic algorithms to identify priority areas for conservation currently being used in Canada. These projects, by Living Oceans Society and World Wildlife Fund Canada are highlighted with regard to their potential applicability to DFO.
Upon review of the methodologies we recommend that DFO consider the use of a site selection methodology in its IM program. From our analysis we concluded that MARXAN (a software package which employs simulated annealing) would be most appropriate tool to assist DFO in furthering its mandate and MPA objectives under the Oceans Act.
Other recommendations include;
- multi-scale planning in MPA network design;
- perform analyses with multiple MPA objectives and datasets
- determine if MPA networks created using multiple agency mandates requires less area than performing the analyses separately specific for each agency
- understand the usefulness of the various frameworks and approaches to applying Marxan, especially those ongoing in Canada;
- current selection analyses in Canada can provide DFO with compiled data and information on lessons learned in applying MARXAN and developing ecological planning frameworks for both coasts
- facilitate further analyses ecological attributes for cells need to be defined in terms of parameters reflective of criteria used by different agencies to rationalise their mandates to establish mpas;
- A need, and role for DFO to undertake a pilot selection analysis within the Strait of Georgia;
Although spatial optimisation offers a powerful solution to MPA network design and while these programs make a contribution to improving rigour, transparency and efficiency of what is a complex process, they only contribute to part of the process. Other decision support tools (such as GIS and Delphic approaches – see Lewis et al. 2003) may need to be employed when fine-tuning boundaries, developing zoning plans, or when choosing among candidate sights that are of interest to several stakeholder groups.
Accessibility Notice
This document is available in PDF format. If the document is not accessible to you, please contact the Secretariat to obtain another appropriate format, such as regular print, large print, Braille or audio version.
- Date modified: