Embedded Files
The SGBA and other metrics reviewed above belong to the "habitat-condition" class of proxy biodiversity metrics 1. As such, it is imperative to keep track of parallel research which utilizes standard scientific biodiversity metrics alongside such proxies. As of the time of publishing, no known parallel research has been reported for the SGBA due to its novelty in South East Asia. However, such parallel research has been progressively reported for England's SBM and its earlier iterations within about a decade before legislation of its current form in 2024. As such, those involving England's SBM are cited due to its similarities with the SGBA in order to provide a glimpse into possible implications for the latter.
This section reviews the strengths and limitations of the aforementioned proxy metrics collectively, with recommendations for improvement specifically for the SBM and, by implication, the SGBA which was modelled after the former. The metrics reviewed above are just some from among at least 8 countries which use some form of biodiversity metric 2. The variety of metrics available instead of one or a few standardised ones reflect the lack of consensus on how biodiversity ought to be quantified.
Strengths:
Strengths:
Although most of the metrics reviewed above differ widely on their methodology, calculations and data analyses, they share several broad commonalities. Overall, they present several practical advantages over ideal geographically wide-ranging field surveys: promoting repeatable, cost-effective assessments in place of resource-intensive surveys 3, with the exception of the US-CND metric insofar as it incorporates longitudinal sampling for greater scientific rigour. They were designed to exhibit attributes of a good biodiversity metric which include linguistic clarity and to demonstrate a clear relationship between a metric and actual biodiversity. The assumption underlying the latter is that the biodiversity values derived from the metric accurately represents any increases or decreases in either native floral and faunal species richness or abundance or both. Therefore, these metrics are designed to be surrogates or proxies of the capacity of a habitat to support native flora and fauna. It is noteworthy that the two Australian and North American metrics reviewed above are referred to as metrics for only “vegetation”. The SBM, however, is purported to be a proxy for general biodiversity, both vegetation and otherwise.
Limitations:
Limitations:
The extent to which proxy metrics accurately reflect biodiversity value remains open to scientific verification, especially if they have been designed and developed primarily by governments and industry, with less input from academia 2. In general, using habitats as proxy for biodiversity has been reportedly flawed: (1) the habitat extent, type and quality may not be the sole drivers of population viability for all species; (2) certain species (e.g., invertebrates) may require several habitats, some of which may be under-rated by a metric; (3) England's SBM shows no consistent relationship between biodiversity units and number of species of conservation significance. The aforementioned has given credence to studies examining the extent to which England's SBM preserves species.
Biodiversity, as measured by the SBM, has been reportedly found to be reflective of plant biodiversity only. In a study of the relationship between biodiversity units measured by the SBM and biodiversity indicator variables measuring three proxy taxa of biodiversity (birds, butterflies and vascular plants) across non-urban terrestrial broad habitat sites across England, significant correlations were obtained only for those of plant biodiversity variables (2 Marshall et al., 2024). For plants, significant correlations emerged on 3 out of 4 indicators (species richness, mean species range and change). Of these, only species richness correlated positively with HBU whereas this was negative for mean species range and change. Therefore, HBU was reflective of species richness. In contrast, mean species range and change declined as habitat rarity (thus, higher HBU) increased, as reflected by their inverse association with HBU. Nevertheless, the effect sizes for all correlations were small. The absence of detectable statistically significant relationships for other taxa was attributed to possible inadequate statistical power due to the far fewer sample sizes for birds and butterflies, narrower measurement ranges for the former (i.e., only two levels; e.g., no poor conditions recorded to represent lower range value levels) and calculation limitations imposed by methodological constraints (e.g., transects vs. plant quadrats).
In a similar vein, a critique on the SBM also reported species-related and other metric limitations 1. The SBM's oversimplified habitat assessment does not take account habitat heterogeneity vital to invertebrate survival. This is due to "scoring habitat components in isolation " instead of taking into account "ecotones (transitions between habitat types)" which provide favourable conditions and connectivity vital to different developmental stages and invertebrate migration, respectively. Ironically also, green space reduction occurs when large areas of low biodiversity value habitats are traded with smaller ones of higher value, since compensation is calculated in terms of biodiversity units instead of area. In addition, habitat condition criteria exhibit a one-size-fits-all approach. For instance, identical criteria are applied to 10 grassland types (e.g., bare ground where <5% of which might be considered reasonable on lowland meadow but not acid grasslands where 25–50% is expected).
Recommendations:
Recommendations:
The findings reported above have yielded the following recommendations for improvement to the SBM (2 Marshall et al., 2024), some of which may be directly applicable to the SGBA as the latter was modelled closely after the SBM.
Modify BNG targets:
Modify BNG targets:
The small effect sizes of significant relationships between HBU and selected biodiversity correlates reported above reinforces the perception that the target of achieving 10% BNG appears to be arbitrarily determined, lacking an explicit rationale and bringing little to no benefit to nature (2 Marshall et al., 2024). Instead, a higher target which would otherwise be more impactful could be adopted.
Separate taxa measurement indices:
Separate taxa measurement indices:
The findings above reinforce those of previous research which have reported inconsistent associations between the number of conservation priority species of any taxon with both habitat distinctiveness and condition 2. Reasons provided include: (1) the SBM not capturing important habitat features relevant to birds and butterflies; (2) the metric calculation scale does not adequately accommodate or is not applicable to (i) individuals with a wider range beyond the scale afforded by the SBM; (ii) invertebrates which occupy habitats at smaller spatial scales as the metric neither adequately captures nor measure habitat heterogeneity (e.g., habitat richness, edge density etc.), despite certain habitat condition criteria related to vegetation height and complexity. Therefore, the introduction of an index to represent mean change in species' range or population over time is warranted, based on periodically updated benchmarks for at least plants, birds and butterflies.
Trading rules:
Trading rules:
In another statistical analysis of the same study reported above, Non-Multidimensional Scaling (NMDS) displayed a clear separation of 3 taxa by broad habitat type (2 Marshall et al., 2024). Plants, birds and butterflies species composition formed respective clusters by 6 broad habitat types. This highlighted the differences among the 3 taxa across broad habitat types, providing further support for separate metrics for each taxa and their respective supporting habitats. While this justifies existing trading rules (i.e., compensation of medium or higher habitat distinctiveness by the same level of distinctiveness or higher), the "like-for-like" compensation requirements may actually limit the formation of habitat mosaics vital to biodiversity preservation in the face of climate change. More priority habitats should be explicitly specified in the SBM and designated as irreplaceable with very high distinctiveness status. Presently, irreplaceable habitats are removed from the baseline calculations of the SBM as they are either difficult or too time-consuming to recreate after destruction 4. Instead, for any losses or deterioration of such, compensation is required according to compliance with existing regulations.
Multipliers:
Multipliers:
Given that the attainment of BNG has a relatively low bar via the habitat creation of medium distinctiveness only, high value multipliers (i.e., > 1) should be applied to high distinctiveness and good condition habitats in baseline calculations instead of the 'negative' (i.e., < 1) reductionary penalty) multipliers associated those of habitat creation.
Quantitative considerations:
Quantitative considerations:
The SBM's linear measurement, calculations and assumptions of biodiversity should be replaced by non-linear ones to represent more realistically ecological relationships and distributions. For example, power relationships could typify the higher bands of habitat distinctiveness by converting the distinctiveness metric scale from a linear to a more logarithmic/ exponential one (e.g., doubling scale intervals) to reflect a 'power relationship scoring system'.
The SBM and SGBA could take a leaf (or two) from both the NSW VI and US-CND metrics with regards to their non-linear calculations and distributions accommodated. In the NSW VI, its vegetation attribute scores best assume a non-linear relationship between observed vegetation states and benchmarks as they mirror biological phenomena (e.g., species to area correlation). The NSW VI metric assumes that the relationship between attribute scores and both observed and benchmark data is continuous, non-linear and sigmoidal, as opposed to discrete categories of other metrics 3. Therefore, only very high attribute scores are allocated to a relatively broad range of observed vegetation attribute states which closely approximate benchmark values, thereby minimising underestimation bias. Conversely, only very low attribute scores are allocated to a relatively broad range of observed vegetation attribute states which fare poorly compared to benchmark values, thereby minimising overestimation bias.
The flexibility of the VQA (US/ CND) is yet again manifested in its accommodation of various data distributions which potentially yield higher or more favourable biodiversity estimates Boyle et al. (2024). For instance, the VQA (US/ CND) methodology advocates for the examination of not only overall vegetation Quality but also at the individual indicator level if different distribution patterns emerge. For instance, in a case study where indicators displayed bimodal distributions like tree cover which yielded lower Quality scores against the benchmark, it was proposed to divide them into separate classes if this was presumed to consist of natural and disturbed vegetation. However, taxonomic composition exhibited a normal-like (gamma) distribution, indicating no apparent differing subpopulations. Upon division into two indicators of tree cover, the same benchmark would then be used to assess tree cover, but assessed as separate focal plots/ samples. If, thereafter, the new indicator sub-categories were assessed with different Quality algorithms/ statistical methods, this might have resulted in more favourable Quality scores.
Habitat distinctiveness and condition:
Habitat distinctiveness and condition:
In light of the above, habitat distinctiveness and condition may need to be revised. Habitat distinctiveness would be better represented by mean range of species supported based on existing taxa data instead of habitat rarity and threat which are conceptually distinct and potentially contradictory (2 Marshall et al., 2024). Furthermore, habitat rarity should be applied uniformly to all habitats measured based on standardised reference sources (as opposed to the mixture of different places in England, the UK and EU). Therefore, this raises the call to collect, centralise, standardise and periodically update new data on English habitats and their conditions.
The revision of condition criteria was suggested for ecological reasons over agricultural ones, especially at apparent points of opposing priorities (2 Marshall et al., 2024). Certain attributes valuable to conservation might be undesirable agriculturally. This applies to where plants classified as 'undesirable' from an agricultural perspective supported both threatened native insects and pollinators. In addition, the assumption that a low distinctiveness habitat assigned with a fixed lowest condition score (a 'N/A' status, equivalent to 'poor' condition) has a negligible influence on overall biodiversity value has not been borne out by research (2 Marshall et al., 2024). For instance, certain habitats classified as low distinctiveness exhibited one or more of the following features not typically expected of such: (1) supporting threatened species; (2) particular biodiversity scores that were either comparable to those of higher distinctiveness (i.e., mean range and species richness) or were not necessarily among the lowest of scores (i.e., mean range change). Therefore, the combining of certain sub-habitats of the same broad habitat was proposed for those that exhibited comparable biodiversity equivalence (on individual biodiversity measurement variables).
Page updated
Report abuse