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

Resolution of conflicts When there are conflicting features within a site, Secsel determines which one of the conflict- ing features is to be ignored prior to site selection. To determine which feature is to be ignored, the dispensabilities of the local units of the individual feature toward achieving the target are compared. To enable this comparison, a dispensability index, S, is defined as follows: S ¼ ðb 1Þ = N where N is the target of the feature, and b is the position of the local unit within all local units of a feature in the descending order of value. If there is more than one local unit with the same rank, b is calculated as follows: b ¼ k þ ð1 þ mÞ = 2 where k is the number of local units of the feature with higher ranks, and m is the number of local units with the same rank as the unit in the focal site. For example, if the top four local units have the same rank, all units are assigned b equal to the mean of 1–4, i.e., 2.5. If there are five rather than four tied local units, b is 3.0. A larger b results in a larger dispensability index, S. When more local units are available as substitutes for the one in the focal site, the local unit becomes more dispensable. The value of S varies inversely with the target, N. The implication of a larger N is that the local unit in the focal site is less dispensable. Even when the b is the same among features, the one with the larger N requires more local units with high ranks, and the local unit in the focal site is thus less dispensable.

After the S values of conflicting local units within a focal site have been calculated, the local unit with the larger S is ignored. Because the omission of a local unit affects the S values of other units of the feature, the S values of remaining units are recalculated. The fewer local units that remain, the less dispensable will be the remaining local units. Outputs The output of SecSel is a set of selected sites listed in the order of selection. Additional infor- mation includes the features by which each site contributed to the achievement of targets. When conflicts among features are taken into account, the information used to ignore some features is provided. This information is crucial to deciding conservation/utilization practices within selected sites when there are local units of conflicting features. In addition, SecSel outputs the frequency with which each site is included in the selected area in multiple runs. There are some stochastic processes in the SecSel site selection. First, ties can occur when deciding which site contributes most to achieving targets. One of the sites that is tied is chosen randomly. Such ties occur frequently, especially when there is a low resolution of the local unit evaluation. It should be noted that low resolution leads to flexibility with respect to site selection, and it widens the range of variation of site selection in the process of tie breaking. There is another stochastic process in the resolution of conflict. When there is a tie in the dispensability index among local units with conflicting features, one of the units is

randomly given priority. Ties can also occur in site selection during consideration of cost and scattering. A result of these stochastic processes is that site selection can vary between replicated runs of SecSel. The range of variation depends on the nature of the given data. The greater the prob- ability of ties, the more likely that the variation of selected sites is large. SecSel automatically repeats the user-specified number of runs. The frequency with which each site is included in the selected area in multiple runs may be an indicator of the 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 8 / 19