Bottom‐up global biodiversity metrics needed for businesses to assess and manage their impact

Ensuring that companies can assess and manage their impacts on biodiversity will be crucial to solving the current biodiversity crisis, and regulatory and public pressure to disclose these impacts is increasing. Top‐down intactness metrics (e.g., Mean Species Abundance) can be valuable for generating high‐level or first‐tier assessments of impact risk but do not provide sufficient precision or guidance for companies, regulators, or third‐party assessors. New metrics based on bottom‐up assessments of biodiversity (e.g., the Species Threat Abatement and Restoration metric) can accommodate spatial variation of biodiversity and provide more specific guidance for actions to avoid, reduce, remediate, and compensate for impacts and to identify positive opportunities.


INTRODUCTION
Biodiversity is foundational to the global economy and to society at large.The WEF (2020) estimates that >50% of global Gross Domestic Product is highly or moderately dependent on living nature.However, more than 40,000 species are threatened with extinction (IUCN, 2023) and as many as a million may be (IPBES, 2019).Many vertebrate populations are in decline (Almond et al., 2022), and many of nature's contributions to people are being eroded (WEF, 2022).The loss of biodiversity is increasingly recognized as a material risk to private sector actors and as an opportunity for competitive advantage through avoiding risks and contributing to a nature-positive future (WWF, 2019).
Biodiversity-related risk can be dependency related, where the productive capacity of companies is reduced by the deterioration of ecosystem services.Or, it can be impact related, where a company's operations have negative impacts on biodiversity.Negative impacts on biodiversity may cause regulatory, reputational, or market risk (EFRAG, 2021).These 2 components of biodiversity risk are termed "double materiality," for instance, in the EU Sustainable Finance Disclosure Regulation (SFDR [EUROPA, 2022]).Double materiality requires that both impact materiality and financial materiality perspectives be applied in their own right without ignoring their interactions (EFRAG, 2021).
Increased awareness of these risks has sparked demand for action to protect, manage, and restore nature, from regulators, shareholders, and the general public (WWF, 2020).This has led to a renewed demand for tools and metrics companies can use to measure and manage biodiversity-related risk and opportunity.Dependency-related risk is closely linked to the provision of ecosystem services, and although ecosystem services are dependent on underlying biodiversity (Mace et al., 2012), ecosystem service provision varies widely among ecosystems and the quality and amount of the service may not be closely tied to the quality of the biodiversity at any particular site.Additionally, the change in ecosystem service provision with degradation of a site is neither consistent among ecosystems nor linear (Mace et al., 2012).Dependency-related risk assessment is not therefore ideally suited to the use of measurement metrics related to underlying biodiversity.Ecosystem service provision is well assessed through existing tools, such as ValuES ( 2022) and the Business and Biodiversity Interdependence Indicator of the Natural Capital Protocol (Capitals Coalition, 2023).We therefore do not consider dependency-related risk further.
Impact-related risks relate directly to the loss of biodiversity and are the subject of societal agreements, such as the Kunming-Montreal Global Biodiversity Framework and the United Nation's Sustainable Development Goals.To align their actions with societal goals, companies need to be able to assess their actions in relation to increased negative impacts on biodiversity, which may create risks to the company, and increased opportunity, where a company has the ability to contribute to societal goals by mitigating threats to biodiversity.
Increased awareness of biodiversity impact risks and opportunities has sparked demand for action from regulators, shareholders, and the general public to protect and restore nature (TNFD, 2022).However, there is little consensus on or standardization of the appropriate metrics to use for assessing impact-related risks or opportunities to mitigate them (EU, 2022).We therefore reviewed 2 types of metrics (topdown intactness and bottom-up global significance) that can be used by business for the assessment of biodiversity impact risks and opportunities (henceforth BIRO); considered how these metrics can be used to identify opportunities and design and implement interventions; and considered priorities for application and research.
Most BIRO assessments depend on the location of a company's direct operations (the equivalent of Scope 1 emissions in climate reporting) or on their use of commodities that affect biodiversity during their sourcing, production, or processing (purchased goods and services, part of Scope 3 in climate reporting).Because biodiversity varies at fine and coarse spatial scales (Crawley & Harral, 2001) and across its ecosystem, species, and genetic diversity components, the choice of BIRO metrics should capture this variability, and, because many companies do not have geographic commodity traceability, the metrics should accommodate a range of spatial scales.

TOP-DOWN INTACTNESS METRICS
Currently, BIRO are typically quantified based on modeled local losses of species or individuals relative to a reference level in a representative bioregion associated with threats and economic activities, policies, and finance flows that drive them.These modeled approaches depend on field research and expert opinion, which may be focused at the scales of populations and species.Bioregions in this context can refer to ecosystems, habitats, or other large-scale ecological units.Examples of these top-down intactness approaches include the GLOBIO model, for which the metric is Mean Species Abundance (MSA) (Schipper et al., 2020), and life cycle impact approaches (such as IMPACT World+), for which the metric is the Potentially Disappeared Fraction (PDF) of species (Bulle et al., 2019).
For a given impact, MSA expresses average percent population declines, whereas PDF expresses average percent species extirpated (exact definition varies among models).Both metrics convey the proportional changes in community intactness or ecosystem integrity relative to a bioregion-based reference level.These metrics are used in tools such as the WWF Biodiversity Risk Filter (WWF, 2023), which provides estimates of biodiversity risk of companies in a spatially explicit way.They are also used in footprinting tools, including the Global Biodiversity Score, Biodiversity Footprint for Financial Institutions, Corporate Biodiversity Footprint, and Biodiversity Footprint Calculator (review in Finance for Biodiversity [2022]), to calculate biodiversity footprints, expressed in, for example, MSA per square kilometer, which is equivalent to a reduction of ecosystem state (measured by modeled reduction in average population declines) of x km 2 of habitat.
The advantages of such approaches include that they can leverage existing global biodiversity models that are based on explicit pressure-impact relationships, thus enabling linkage to company activities through their contribution to impacts (e.g., BNP Paribas, 2022) and thereby allowing for rapid assessments of the likely scope and magnitude of BIRO.Another is that they allow companies to assess impacts from activities such as processing and transformation of commodities as well as production.
However, these approaches have significant limitations, 2 related to the metrics and 2 related to the underlying models.First, the metrics are "conservation agnostic" in that they give the same scores to impacts in bioregions with low and high global biodiversity significance (defined as containing species or ecosystems at risk of extinction or collapse) (Martin et al., 2019) (Figure 1a).Second, different manifestations of a particular impact cannot be separated.For instance, a uniform but limited loss of species across a bioregion gives the same value as a bioregion retaining intact sites surrounded by areas that are totally depleted (Figure 1b).In the first case, it does not much matter where impacts occur, whereas in the second case, it is important to make conservation interventions in the intact parts of the bioregion.Third, the underlying models rely on extrapolation from data that may be patchy or incomplete within or across bioregions, typically with little or no ground-truthing of the results (Schipper et al., 2020).This results in underrepresentation of some bioregions, taxa, and threats.Fourth, the threat drivers modeled do not include 2 of the most important threats to biodiversity: unsustainable resource use (other than hunting) and invasive species (Maxwell et al, 2016).
Owing to these limitations, BIRO values and corporate footprints generated by these metrics may underestimate company risk by omitting important impacts (false negatives) or may indicate a risk to a company, parts of companies' value chains, or even entire industry sectors that may not occur where the impact is actually happening (false positives).It is impossible to correct these errors without validation and ground-truthing within each of the bioregion-production method combinations, which is time-consuming and likely expensive.The low spatial precision of the models due to their use of global pressureimpact relationships means that even if precise sourcing information is available from a commodity producer, the footprint metric value may change very little or not at all (Figure 1), so companies cannot improve their BIRO estimation.
In Figure 1a, the top-down intactness metrics value is calculated from the fraction of the bioregion that is intact.In this hypothetical example, the species community is fully intact in 20% of the bioregion and all original species are extirpated in the remaining 80%, leading to a hypothetical bioregional intactness metric value of 20% of 10,000 (i.e., 2000) in both cases.The bottom-up global significance metric value is calculated from the number of threatened species in each, which is 10 times higher in the rainforest example than in the desert example.This means that the score of an intervention (which could relate either to an increase in risk, if company actions increase threats, or an opportunity, if company actions reduce them) based on top-down intactness metrics is the same whether placed in the desert or in the rainforest, whereas the score of an intervention in the rainforest bioregion based on the bottom-up global metric has a value 10 times greater than the score for the desert.
In Figure 1b, the potential risk or opportunity of an intervention is calculated from top-down intactness metrics from the proportion of the bioregion affected by the intervention (20% of 2000 = 400), meaning that it is possible to intervene in a place that will have no impact but still gives the same score as one where an intervention will carry significant risk or opportunity.The bottom-up global significance metric value in Figure 1b is derived from the number of threatened species present in the intervention areas in the bioregion, indicating that the risk or opportunity score for an intervention will be much greater in the area with the threatened species present.

BOTTOM-UP GLOBAL SIGNIFICANCE METRICS
A complementary approach to BIRO assessment is available through bottom-up global significance metrics, for which large data sets are compiled based on field research at fine geographical scales, which allows parameters to be observed, estimated, inferred, or suspected by experts.To achieve global coverage, bottom-up global significance metrics include a subset of all species for which data on area of habitat (Brooks et al., 2019) and threats are available.
Two examples are the Global Persistence Score (Durán et al., 2020) and the Species Threat Abatement and Restoration (STAR) metric (Mair et al., 2021).These bottom-up metrics assess the likelihood of persistence or, conversely, the extinction risk of species based on their global distribution and the threats they face.By incorporating information on species threats and parameters related to species persistence and extinction risk, these are conservation-aware metrics.The global persistence score estimates the likelihood of a species persisting within a study area based on the species' remaining area of habitat (Brooks et al., 2019).So far, this approach has been applied to assess the impact of production of soy in the Brazilian Cerrado on the persistence of 2009 species based on habitat loss due to land use.This approach could be extended to other regions, provided that the spatial distributions of the species and the impacts of different threats thereon can be estimated.For example, a recent study estimated the combined impacts of land use and hunting on the area of habitat of tropical mammals (Gallego-Zamorano et al., 2020), which could serve as a basis to quantify persistence scores.
The STAR metric is based on the International Union for the Conservation of Nature (IUCN) Red List of Threatened Species (IUCN, 2023) and estimates the potential for reduction of vertebrate extinction risk for every 5 × 5-km cell across the terrestrial realm.Extinction risk reduction can be generated by threat mitigation alone or through habitat restoration in combination with threat mitigation.STAR estimates the opportunity for reducing extinction risk in a particular pixel as the combination of the proportion of the threatened species' area of habitat contained in the pixel weighted by the extinction risk category of the species present (critically endangered, endangered, vulnerable, or near threatened) (complete methodology in Mair et al. [2021]).The STAR metric establishes a spatially explicit quantification of the relative importance of threats in driving species extinction across the globe, meaning that links can be made between components of biodiversity (species), threats to them, and impacts on these threats caused by company actions The STAR score for any area can be disaggregated by threat, so it is possible to evaluate the impacts of specific company activities linked to a particular threat by evaluating the extent and severity of the impacts of each threat separately.If the pathways of supply of specific commodities and industry activities are known, the impacts of corporate sourcing (for instance, for agricultural commodities) in particular places can be used to identify specific threat mitigation activities at the point of production, and thence set targets for threat reduction, allowing the reporting of impacts (Green et al., 2019).If there is little spatially explicit information on a company's activities, the STAR value, calculated, for instance, at national or subnational level, will have a larger error, as with top-down approaches.However, in situations where fine-scale impact data are available, STAR scores reflect more granular geographical BIRO (Figure 1).The nonnormalized STAR metric (nSTAR) (Irwin et al., 2022) may also be used to conduct finer scale impact assessment because it supports a more accurate allocation of the extinction risk impact to specific threats and therefore the measurement of company activity and related targets associated with those threats.
The main limitations of STAR are that it is currently calculated only from information on 3 comprehensively assessed taxonomic groups: mammals, birds, and amphibians.In addition, for some mammals and birds, and most amphibians, the scope and severity of threats have not been assessed, and thus median values for scope and severity are applied (Mair et al., 2021).This may mean that threats at specific sites may be overor underestimated.To improve taxonomic coverage, reptiles and freshwater and marine taxa will be included as of the end of 2023.
The STAR metric only relates to threatened species, thus ignoring threats to hitherto nonthreatened species and the ecosystems where they occur.Additionally, it does not provide information regarding impacts on genetic diversity, which remain poorly understood.Although threats to ecosystems vary widely across the globe, they are generally similar to threats to species at local scale, so local actions to conserve species seem likely to benefit ecosystems and genetic diversity (Bellard et al., 2022).
In tests conducted to refine calibration methodology (Schneck et al., 2023), STAR may include some threats or threatened species that do not occur at the site, meaning that the STAR value may be conservative (Schneck et al., 2023); the reverse may also be true.Ground-truthing of threatened species presence and threat intensity, to permit a calibrated STAR score, is one approach to resolve this problem and that of lack of threat scope and severity information (Schneck et al., 2023).

LINKING RISK TO ACTION
If BIRO assessment reveals high risk but does not clearly point to opportunities for risk reduction, companies may divest from high-risk assets or shift their supplier locations.This could have the unintended consequence of shifting or even increasing biodiversity impacts because the abandoned locations might be left to the large number of corporates and financial institutions currently not as concerned with their BIRO and potentially with weaker biodiversity safeguards, potentially resulting in a larger negative impact.By linking impact-related risk to identified threats that can be mitigated, bottom-up global significance metrics may be more able to encourage companies not to divest from high-risk locations but to improve practices-potentially through collective action.
A challenge for all BIRO approaches is that many companies do not yet have traceability to the production sites for the commodities they use (for instance, a company producing soybased cattle feed may not have any information about where the soy was produced).In these cases, top-down intactness approaches can give approximate, first-tier assessments of risk, albeit with caveats (Figure 1).By contrast, bottom-up global significance metrics can generate an estimate of BIRO by quantifying species extinction risk within the country or region of commodity production and then allocating a portion of this back to the company concerned.The allocation can be calculated, for instance, for agricultural commodities, as a proportion of the commodity produced in the country or region that the company buys.This approach will, however, become less necessary because supply chain transparency is rapidly increasing.There are new developments in remote sensing technology (that can reveal impacts attributable to particular suppliers) and blockchain traceability (e.g., zu Ermgassen et al., 2020) that allow analyses of extinction risk embedded in international trade that can be explored with nSTAR and economic input-output analyses (Irwin et al., 2022).

CONCLUSIONS
Top-down intactness metrics can be valuable for generating high-level or first-tier assessments of BIRO.However, the characteristics of the most commonly used top-down metrics and the nature of the data used to populate them mean they have limited value for evaluating risk to companies caused by biodiversity impacts because the outputs are conservation agnostic and thus not realistic measures of BIRO.Consequently, identifying, implementing, and measuring impact risk responses are impractical with top-down intactness metrics.Combined measures of ecosystem extent, integrity, connectivity, and conservation value might offer the best solution for the analysis of ecosystem impacts.Remote sensing provides a viable means to track some of these metrics, and linking to global databases, such as GLOBIO, can provide a global measure of intactness or integrity.
Bottom-up global significance metrics, such as STAR, now have global coverage and appropriate taxonomic span and can therefore provide companies with a quantifiable and comparable metric of impact risks to biodiversity within their physical footprint or associated with specific commodity sourcing, relevant to the scale of the impact.They also provide clear guidance for identification, implementation, and measurement of remedial actions.The 2 approaches in combination may provide a clear path for businesses to identify the most effective actions to reduce impacts.This will be crucial for compliance with emerging disclosure frameworks, such as the Taskforce on Nature-related Financial Disclosures (2022), and for setting science-based targets for nature (Science Based Targets, 2023).
In the future, it would be desirable to use metrics that encompass additional components of biodiversity-in particular, ecosystems and genetic diversity-in a similar bottom-up way while also incorporating measures of local values of biodiversity.
With the limited time available to address the nature crisis, societies need to make the most of the data and metrics available now.If top-down intactness metrics are used, then combining them with bottom-up global significance metrics, such as STAR, would help business understand risks, along with-cruciallyopportunities for potential mitigation actions.

FIGURE 1
FIGURE 1 Differences between a hypothetical biodiversity impact risks and opportunities (BIRO) assessment results based on top-down intactness metrics and bottom-up global significance approaches: (a) difference in conservation importance between 2 hypothetical bioregions and (b) importance of selecting the correct location within the bioregion for action (the higher the assigned scores, the higher the conservation value of the potential intervention).Metric score values are illustrative only.