Development of a combined marine and terrestrial biodiversity indicator: research

2. Biodiversity indicators – a brief introduction

4. Many authors have attempted to define the essential and desirable qualities required from indicators (generally) and biodiversity indicators (specifically). For instance, Dale & Beyeler (2001) stated that ecological indicators should meet the following criteria: be easily measured; be sensitive to stresses on the system; respond to stress in a predictable manner; be anticipatory; predict changes that can be averted by management actions; be integrative; have a known response to disturbances, anthropogenic stresses, and changes over time; and have low variability in response (e.g. not show large fluctuations due to random effects). To this one might add other qualities such as simplifying, easily understood, representative and policy relevant (e.g. Gregory et al. 2005). Other authors have focused on desirable statistical qualities e.g. van Strien et al. (2012). It is very important that indicators be unbiased (or at least that sources of bias are known, understood, and factored into interpretations of indicator change), and ideally indicators should be precise and accompanied by estimates of precision (confidence intervals) and be amenable to quantitative reporting of changes (Sutherland 2006).

5. Whilst the characteristics of a successful indicator are well known, the ideal indicator is rarely possible, if ever, and it is recognised that indicators should be selected depending on the specific questions being asked and, perhaps more pertinently, the data available (Feest et al. 2010). In the case of biodiversity indicators, considerable challenges are posed by constraints in data availability; even with the UK's long-established biological recording and monitoring community, we have robust measures of change for only a minority of species. At the same time as developing an indicator that draws from existing data sources, can be updated regularly (preferably annually), is responsive to change, can be assessed with formal statistical assessments, but remains simple to understand and communicate, it also needs to account for the fact that the available data represents a biased sample of trends in Scottish biodiversity. Careful consideration will be required to overcome this issue.

6. In addition, careful thought has to be afforded to the relevance of an indicator to policy objectives, and the actions taken to achieve those objectives. If designed to measure progress towards narrowly defined objectives indicators can perform admirably, but their value will be limited to this function only, and they can easily become obsolete as objectives change or evolve. Conversely indicators designed to measure response to broader suites of objectives and actions can be less sensitive to change and harder to interpret in light of policy.

7. Heink & Kowarik (2010) identified a useful list of criteria that could be used to assess the suitability of biodiversity indicators, based on a review of 56 papers discussing indicator selection. They grouped these criteria into five groups:

• Feasibility (knowledge about species, portability, suitability for statistical analyses, existence of reference values);
• Efficiency (feasibility of data collection, universality, parsimony);
• Relationship between indicator and indicandum (precision of correlation between indicator and indicandum, construct validity, aggregation of substantial amount of ecological information);
• Information to be provided by the indicator (relevance, sensitivity to change, functional importance, distinction between natural and man-made change, rarity and threat);
• Perception of indicators (acceptance, comprehensibility and simplification of information, economic importance).

8. Whilst not all of these criteria are necessarily relevant to all biodiversity indicators, depending on the circumstances for which indicators are required, they do provide a useful suite of considerations against which to test candidate indicators. As Heink & Kowarik (2010) intended, these criteria, or a subset of them, could be used to enable the transparent selection of a biodiversity indicator, although they identified that different applications for indicators will require different 'patterns' of selection (i.e. the relative importance of criteria may vary).