North America

Image: wikimedia

The Vegetation Quality Assessment, hereafter abbreviated “VQA (US-CND)” to distinguish it from Victoria state’s VQA above, was developed by a US-based private consultancy which published a case study on its use in British Columbia, Canada 1. While this section is labelled "North America" for simplicity, the VQA (US-CND) has been used in North and South America, Southern Europe across a variety of climes, including the tropics. While the VQA (US-CND) shares certain similarities with other metrics reviewed thus far in terms of the concept of restoration (also used interchangeably apparently with 'reclamation' or 're-vegetation' therein), it has several unique distinctive and adaptive features: 

Methodology: 

From the outset, perhaps what sets the VQA (US-CND) apart from other metrics reviewed above most is that it was designed to provide a statistical estimation of error to reduce observer bias via stratified, random and repeated sampling. Sampling stratification is according to vegetation class, disturbance, and reclamation treatment. Where excessive strata reduce sample sizes, vegetation classes are combined by ecological similarity, forming "ecosystem groups". Multiple assessment times are employed to collect data before, during and after the project. Each sampling assessment involves more than one data collection point as sufficient sampling is carried out to calculate a Confidence Interval (CI). 

Measurement indicators: 

While the measurement indicators (termed “ecological indicators”) of other metrics are necessarily fixed due to offset policies backed by legislation, the VQA (US-CND) remains flexible to the adoption of indicators so as to accommodate new data collection or retain compatibility with existing databases. Indicator selection is governed by assessment goals (e.g., benchmark comparisons, wildlife support, vegetation health, etc.) and resources (e.g., time, manpower, costs, etc.) available. The greater the number of indicators, the greater the constraints placed on resources. Comparisons with or reference to existing databases would require adoption of measurement indicators comparable to other metrics used therein based on taxonomic composition (e.g., species abundance) and vegetation structure (e.g., growth forms and size classes [height, diameter, etc.]). Besides the parameters directly measured by such indicators, data may also be reflective of other attributes (e.g., disturbance history and successional age). The different measurement indicators of the VQA (US-CND) are classified into at least two "functional groups", each comprising different measurement indicators: (1) "Composition" (e.g., "Species Richness") representing taxonomic composition; and (2) "Structure" (e.g., percent cover herbs/ trees, etc.) for relative abundance. A third optional functional group, termed "Environment" (e.g., percent cover surface water, decomposing wood, etc.), may be applied to reflect physical attributes. The "Environment" functional group may be deemed optional as it has not been reported as part of the VQA (US-CND) metric proper but demonstrated in a case study 1. For each functional group, the arithmetic mean of all relevant indicators is calculated. 

Calculations: 

The VQA (US-CND) employs a variety of statistical and calculation methods in the estimation of biodiversity value (“Quality”). Vegetation "Quality" is defined as the extent to which existing vegetation resembles native vegetation in the absence of human disturbance. Vegetation Quality is determined based on official provincial records of undisturbed native vegetation and used as a benchmark against which site samples were compared. 

The foremost method to determine quality is the "overlap quality" method whereby Quality is determined by calculating the area of the overlap between the probability distributions of sampling “focal” plots and benchmark plots: the greater the overlap, the higher the biodiversity value, and vice versa. The "overlap quality" method is calculated in two stages: (1) frequency plots for both benchmark types are fitted with probability distribution curves; (2) the probability distribution plots are superimposed onto each other and the overlapping area determined on a scale of 0-1, accompanied by Confidence Intervals (95%). 

The VQA (US-CND) has an arsenal of statistical tools to accommodate quality estimations under particular scenarios which would otherwise make assessment difficult. For example, while the "overlap quality" method is applicable to most biodiversity indicators used for assessment, there are specific scenarios where, due to the nature of a particular measurement indicator, both benchmark and focal probability distributions overlap less, thereby potentially understating biodiversity value. For instance, when a higher score for a specific indicator is desirable (e.g., species richness), another method of determining quality called the “one-tailed quality” approach is used. In this method, following specific criteria, certain non-overlapping areas are discounted so as not to unnecessarily decrease the overlapping area calculated, thus, resulting in a higher quality score. In other instances where benchmark data is lacking or missing altogether, the benchmark value of an indicator is fixed at '0'. This alternative method of determining quality is termed the “Fixed-benchmark quality” approach, whereby the difference between the focal mean of a specific indicator and the benchmark (fixed at '0') is calculated, with a smaller difference corresponding to greater quality (unit range: 0-1). This method is used when historical benchmark data is not available for comparison due to (1) human disturbance/ intervention (e.g., suppressing fire-induced forest regeneration and forest clearing accompanying human settlements), (2) difficulty in obtaining such (e.g., health indicators [disease, infestation, etc.] of old forests). Sample indicators include percent cover of invasive species and proportion of infected trees. 

As the “overlap quality” method is the primary means of determining quality (i.e.., by calculating the shared area by both focal and benchmark probability distributions), the adoption of alternative statistical methods (i.e., fixed-benchmark) for one or more indicators may be disadvantageous to Quality estimation yielding lower Quality estimation scores/ values. In order to circumvent this, a hybrid statistical adaptation is used, namely the "Overlap-means quality" method, whereby the weighted average of both methods is calculated instead as a Quality estimate. Therefore, the Overlap-means quality method is meant to maximise quality-overlap between two different distributions/ i.e., maximise biodiversity value/ assessment scores in order to minimise statistical disadvantages resulting from poor fitting/ superimposition of distributions.

For each measurement indicator, “Quality” is subsequently multiplied by “Area”, the product of which gives rise to “Quality Hectares” (QH). After estimating QH of individual indicators, overall quality is next determined by calculating the geometric mean of the three functional groups ("Composition", "Species Richness" and "Environment") as stated above. No weights are assigned the aforesaid functional groups as is done with the NSW VI metric "sub-indices", on the basis of reducing subjectivity. For each functional group, the arithmetic mean of all relevant indicators is calculated. 

The geometric mean is calculated by multiplying the arithmetic mean of each functional group followed by the root of the product according to the number of functional groups. This implies that additional functional groups may be considered in the determination of quality where the need arises, reflecting the versatility of the VQA (US-CND). Calculation of the geometric mean ensures that if one functional group approaches zero, this reduces the overall calculated quality value, thereby not allowing any functional group with high value to inflate or compensate for any single functional group with a low value. In comparison, the NSW VI metric avoids using the arithmetic mean altogether. However, biodiversity value estimation differs between the geometric and arithmetic means depending on the type of vegetation assessed 2. A CI (95%) of the net Quality Hectares (QH) exceeding Baseline Net QH (QHnet = 0) constitutes Net Positive Impact (NPI) attainment. 

Benchmarks: 

Similar to the Australian metrics reviewed above, benchmarks against which sampling (“focal”) plots of the VQA (US-CND) are compared may comprise pre-existing vegetation databases compiled by local authority; unlike the former, however,  benchmarks may also comprise onsite or adjacent undisturbed identical native vegetation type. Benchmarks and sampling (focal) plots are subject to identical random sampling methodology and measurement indicators used.

Baseline: 

Baseline Quality (Q) may be assumed to equal zero if offsite offset purchase and protection permanently guarantees the prevention of all native vegetation loss (e.g., agriculture conversion). This assumption poses a potential weakness in the lack of irreversible legislation. In contrast, Victoria State's VQA only confers up to a 20% additional gain in habitat score for the transfer of freehold to crown land status 3. Baseline assessment is ideally performed prior to development. If, however, development has already started, retrospective assessment is permitted according to a modelling method termed "backcasting" whereby pre-development vegetation is reconstructed. 

Gains: 

In the VQA (US-CND), gains may be derived from both offset and project sites, being optional for the latter. Site-dependent gains may be recognised based on location as follows:

restoration (offset & project sites), natural succession (project site) and averted losses (offset sites). Natural succession is not considered at offset sites as Quality Hectares (QH) are only attributable there to gains that are the result of active management (maintenance, etc.) of existing vegetation and which would not have otherwise occurred in the absence of such intervention.

Prediction: 

There is apparently no element of prediction demonstrated in a long-term mining case study which applies the VQA (US-CND) despite a cursory mention of “averted loss” which features in the Australian metrics reviewed above 1. 

Compensation: 

Any vegetation impacted by development ideally ought to be replaced, restored or compensated by the same vegetation type according to the "like-for-like or better" rule. However, exceptions allow for a different vegetation type for reasons of either cultural or conservation significance (e.g., replacement vegetation is of value to endangered species and wildlife, local or global vegetation rarity, etc.). The new vegetation substitute is subsequently evaluated using relevant vegetation benchmarks of the same type. 

From the preceding sections above, the VQA (US-CND) metric broadly resembles Victoria and NSW states’ metrics in terms of their vegetative attributes employed and the adoption of the geometric mean (NSW). What makes it stand out is its attempt to incorporate greater scientific rigour via its sampling and statistical methodologies. The general metric characteristics of the VQA (US-CND) metric and its ecological indicators comparable to the SGBA habitat condition criteria equivalents are presented below (see tables below). 

1: (Boyle et al., 2024)

2: (Dorrough and Oliver, 2024)

3: (O'Brien & Gordon, 2024)