SecSel, a new software tool for conservation prioritization that is applicable to ordinal-scale data for multiple biodiversity features

It is apparent from Fig 4 that the output of SecSel specifies for what feature(s) a site was selected in the conservation area network, and users can identify the management needed for the site. Genericity: Planning tools should solve a variety of problems encountered in practice, using data on any set of biodiversity surrogates, from any type of ecosystem and geograph- ical location. SecSel is not designed to solve specific problems. Any type of feature can be handled, as long as the conservation value of the local units of a feature is evaluated on at least an ordinal scale. SecSel assumes that the focal area consists of multiple ‘sites’. The sites are not necessarily of the same size, nor are they arranged in grids. A set of isolated ‘sites’ such as frag- ments of ecosystems is acceptable. Modularity: The whole system can be a module, and part of the system can be substi- tuted. The conflict resolution part, cost calculation part, and compactness evaluation part are implemented as separate class objects. These parts can be substituted by user-defined python objects. All input and output data are in the form of a simple text file. Data can be passed to and from other systems easily. Stakeholder involvement Conservation planning tools are not expected to provide a final solution to be handed to practi- tioners for implementation. Rather, the tools are expected to be used repeatedly because they reflect stakeholders’ interests, values, and opinions. In the process of using conservation-plan-

ning tools, stakeholders can be involved in three stages; a priori involvement before the use of the tool, interactive involvement during the use of the tool, and a posteriori involvement after the use of the tool [18]. With SecSel, stakeholders can be involved in preparation of data (a pri- ori involvement), the repetition of calculations while settings are manipulated (interactive involvement), and interpretation of the results of calculations (a posteriori involvement). In the following sections, we describe in more detail how stakeholders can be involved in each stage. A priori involvement. Selection of features to be conserved/used is the first step of a priori stakeholder involvement. For conservation of biodiversity features, options are numerus. The feature may be ecosystem types or some taxa. For the conservation of taxa, a decision should be made about which taxonomic group and taxonomic level is to be considered. Furthermore, the feature may be the target of conservation by itself, or a surrogate of something else that is the real conservation target. Stakeholders should be involved in selecting features to be han- dled with conservation tools. The criteria for evaluating the features is another point to be considered before the collec- tion of data. Acquiring detailed and/or quantitative data involves cost and effort. The trade-off of the cost and the quality of data should be considered, and an appropriate balance should be found among stakeholders. Stakeholders should reach an agreement about the appropriate conservation/utilization tar-

get for individual features. However, the target may be manipulated interactively, as described later in this section. If some of the sites of the focal area are designated as protected areas beforehand, SecSel starts site selection with those sites and then adds more. It is up to stakeholders to decide with which sites to start. They may be currently protected areas, or areas of special interest for rea- sons such as historical, social, or scientific value. Stakeholders are expected to provide a wide range of viewpoints. Interactive involvement. The high computational efficiency of SecSel is due to the heuris- tic algorithm that supports interactive stakeholder involvement. Stakeholders can try various settings and compare results. Settings to be changed include targets for individual features, PLOS ONE SecSel, a new software tool for conservation prioritization applicable to ordinal-scale data PLOS ONE | https://doi.org/10.1371/journal.pone.0247737 July 23, 2021 16 / 19