Introduction
The following are some notes from a talk I gave for the BNG Virtual Symposium on 20th March 2024 about habitat condition assessment methods. The video from this talk will be available on the Biological Recording Company website, but initially only to those who bought a ticket for the event.
Habitat condition – what do we mean?
The statutory metric user guide says the following about condition:
‘Habitat condition is a measure of the state of a habitat and is used to >measure variation between parcels of the same habitat type. Condition is >often linked to past management, present management, and land use.’
So not much really! If you go back to the technical annexe from metric 4.0 there is a bit more detail provided:
’The approach used is based on the methodology used for Common Standards >Monitoring (CSM). Key indicators are used to make an overall assessment >of condition. However, the metric condition indicators are simpler than >under CSM and it is designed to be undertaken with a single visit to a >site, using visual indicators of likely wider habitat condition, whilst >still being objective and measurable.
The metric condition assessments look at a broader set of attributes than CSM, that cover both the best and poorest examples of each habitat. Thus, a high distinctiveness habitat could be assessed as being in poor condition because of the presence of invasive non-native species, signs of damage, or other impacts.’
So the condition assessment criteria are:
- based on CSM
- designed to be simpler and broader than CSM
- use visual indicators of likely wider habitat condition
- are designed to be objective and measurable
But that still doesn’t really help us understand what we mean by ‘condition’.
The Office for National Statistics uses the United Nations System of Environmental-Economic Accounting definition of condition, which is “the overall quality of an ecosystem asset in terms of its characteristics”. Again, not that clear.
Condition as a measure of function
What I think we mean by condition is ‘the ability of a habitat to fill all available ecological niches and fulfil all of its ecological functions, such that it is both resistant and resilient to disturbance.’
In other words, is all of the biodiversity present that we would expect, and can that habitat bounce back from disturbance?
I think therefore that habitat condition is a measure of function. How well is a habitat functioning? And how resilient is it to change? That is really important, because habitats that are not in good condition are more likely to be lost in the face of repeated disturbance or indeed other environmental change.
Condition in the metric
So we need to understand a habitat’s condition to understand its vulnerability to change, but also to identify opportunities to improve habitats.
In the metric, habitat condition is a key component of a BNG unit. Table 6 in the user guide for the statutory metric shows the scores assigned to each habitat condition level – and the condition of habitat can have a major impact on the number of units generated by a habitat.
So good condition assessments are important in getting the unit calculations right, and poor condition assessments could both over- or under-estimate the number of units provided by a habitat parcel.
The importance of data and evidence
All of our recommendations as ecologists should be underpinned by data and evidence. And this is of course also true for condition assessments. Officers at Local Planning Authorities are going to need confidence that your assessments of BNG and your metric calculations are based on data and supported by evidence.
So what data do you need to collect to provide evidence of habitat condition?
Criteria – but no method
DEFRA and Natural England have provided criteria for assessing the condition of habitats, but they haven’t provide a method for how to collect data to determine those criteria. And I would argue, that is not their job – it is not up to Natural England or DEFRA to teach ecologists how to do ecology.
However, in the absence of a method, everyone will do something slightly different and therefore the quality of the evidence will differ hugely. We will also end up in that situation where the same habitat will be assigned different condition scores by different people using different methods.
How not to do condition assessments
The worse possible thing you can do is wander around a habitat parcel and either tick or cross off each criteria and then count them up. I have seen this done, and all I know is that an ecologist thinks this many criteria have been passed or failed, but crucially, I do not know why. There are no data and no evidence to support that assertion. Later I will explain why I think that is so important when it comes to delivering net gains for biodiversity.
The condition assessments – which habitats are covered?
There are a total of 25 different habitats (or habitat groups) for which there are condition assessments. There are a number of cropland and urban habitats that do not require condition assessments. There are also certain habitats types (e.g. Bramble scrub) for which the condition is pre-set by the metric.
Each habitat has a number of criteria that need assessing to determine condition. For most there is no proposed method for doing this. The woodland assessment is built on some earlier work by the English Woodladn Biodiversity Group, and the criteria for lakes are based other work. CSM is the only other reference.
So I have drafted a proposed method for carrying out habitat condition assessments. I haven’t done so for every habitat; partly this is because I have very little experience of coastal and marine habitats, or lakes. So I have left those alone so I don’t propose anything inappropriate. But most other habitats have proposals. This also does not cover rivers, for which there is another method to follow.
Why is a method important? – standardised, reproducible, repeatable
I have proposed a method for carrying out condition assessments in England to support biodiversity net gain (BNG) assessments. The aim is to standardise the approach to carrying out condition assessments to aid repeatability and replicability. This will ensure that ecologists and regulators can be confident that condition assessments have been carried out to a consistent and appropriate standard.
It will also facilitate data sharing and scrutiny by providing a standard set of variables to record in the field and a standard way to store and share these data.
The philosophy here is: collect data first, assess condition second.
There is danger in trying to decide whether a criterion is passed or failed in real time in the absence of objective data (quantified where possible). It is better to collect data on which to base an assessment. Where possible, this could be automated based on the data, but it may require a more nuanced judgement based on experience.
An example – grassland
Lets look at an example for a grassland condition assessment. Here is the condition assessment for a Low distinctiveness grassland, something that turns up on a lot of sites we visit as ecologists. There are 7 criteria we need to look at to determine the condition of a grassland habitat parcel.
| Criterion | Method | |
|---|---|---|
| A | There are 6-8 vascular plants per square metre, including 2 forbs (which may include those listed in Footnote 1). Note - this criterion is essential for achieving Moderate or Good condition. | 10x 1 square metre quadrat: record species present in each. |
| B | Sward height is varied (at least 20% of the sward is less than 7 cm and at least 20% is more than 7 cm) | 10x 1 square metre quadrat: measure sward height 3 times at each quadrat. |
| C | Any scrub present accounts for less than 20% of the total grassland area. | 10x 6m diamter plots: record % cover scrub. |
| D | Physical damage is evident in less than 5% of total grassland area. | 10x 6m diamter plots:: record % cover of physical damage. |
| E | Cover of bare ground is between 1% and 10%, including localised areas (for example, a concentration of rabbit warrens). | 10x 1 square metre quadrat: record % cover bare gound. |
| F | Cover of bracken is less than 20%. | 10x 6m diamter plots: record % cover of bracken. |
| G | There is an absence of invasive non-native species (as listed on Schedule 9 of WCA, 1981). | 10x 1 square metre quadrat: record presence / absence of INNS species in each. |
Criterion A – There are 6-8 species per square metre, including at least 2 forbs (which can include the species previously referred to as ‘indicators of suboptimal condition’. Also, you must meet this to achieve moderate or good condition.
As we need to report the number of species per square metre, we need to count them. The criterion isn’t concerned about their identity, but as you will have identified them to work out the habitat type, you can report which species are present. So we need to mark out a square metre and count how many species. But once isn’t enough, because we know that the grassland we are in will vary, so we have to do this more than once to get an average across the whole parcel.
It is at this point we start to see the need for a method – how many squares, how to we locate them in the parcel, and so forth.
Criterion B – sward height is varied (at least 20% of the sward is less than 7 cm and at least 20% is more than 7 cm).
For this criterion we need to measure the height of the sward, or at least have a 7 cm ruler to work out the answer to this question. And again the need for a method is clear – how many times do I measure sward height, where, how? If we measure the actual height, we can report average sward height and also the % above and below 7 cm.
One of the issue with this criterion is that some grasslands at some time of year would pass this, and at others would fail. If you think about a Lowland meadow, just before it is cut for hay, it is likely that less than 20% of the sward will be below 7 cm. Shortly after it is cut for hay, less than 20% will be above 7 cm. So there are some issue to work out and while quantitative data are important, knowing and understanding your site is really important.
The remaining criteria have similar thresholds: • scrub less than 20% • physical damage less than 5% • bare ground between 1% and 10% • cover of bracken less than 20%
The final criterion is easier – if there are any INNS species present, you fail that criterion. So that is very straightforward to record.
A method for grassland
One of the things that is clear for grassland is that some of these criteria express themselves at different scales across a parcel. For example, while I can measure species present at the square metre, I can’t do this for scrub – if my entire square metre is scrub then I will skew my results. So we need to measure scrub cover in a different way. Also, how do I locate my sampling points? Fortunately there are a number of existing methods that we can use to help with this.
Therefore, my proposal is the random walk with stopping points. For each habitat parcel you walk a W or Z shape across the site, stopping ten times. At each stop you do a square metre quadrat to record plant species, bare ground, sward height and INNS, plus a 6 metre diameter circle plot to record the remaining criteria.
Proposed random walk method for grasslands
The aim is to provide unbiased stopping points across the parcel as well as collect data at appropriate scales for the different criteria. Of course other habitats require different methods and approaches to collect data that is appropriate for the condition assessment.
Target setting for condition – is it deliverable?
Earlier I showed how not to do condition assessments; now I will talk about why.
The aim of BNG is to make things better, and in our projects one of the things we can do is enhance existing habitats to make them better for biodiversity and deliver net gains. On some projects habitats can be retained and enhanced through management to deliver those gains.
To do this we set target conditions for habitats and our management and monitoring plans set out how this will be achieved. But if our condition assessment on has ticks and crosses or pass/fail, how do we know what management is required to deliver that target condition? This is where recording data and providing evidence makes it much easier to show how gains will be delivered for habitat parcels.
| Criterion | Current | Target |
|---|---|---|
| There are 6-8 vascular plants per square metre, including 2 forbs. | Average of 7 species per square metre | Increase species diversity thropugh seeding |
| Sward height is varied (at least 20% of the sward is less than 7 cm and at least 20% is more than 7 cm) | Average sward height 5cm; overgrazed. | Reduce grazing pressure to create sward heterogeneity |
| Any scrub present accounts for less than 20% of the total grassland area. | Scrub cover ~5% | Monitor grazing pressure to manage scrub. |
| Physical damage is evident in less than 5% of total grassland area. | Physical damage evident across ~15%; poaching. | Reduce grazing pressure to reduce poaching. |
| Cover of bare ground is between 1% and 10%. | Bare ground ~20%; overgrazed. | Reduce grazing pressure to increase grass coverage. |
| Cover of bracken is less than 20%. | Bracken absent. | Monitor bracken coverage. |
| There is an absence of invasive non-native species. | No INNS recorded. | Monitor INNS. |
Here we can see that the sward height criteria is fail, and crucially why. Therefore we can target management to create heterogeneity in sward height. Sometimes this shows us that a target condition is not deliverable because we cannot manage the site in a way that would mean all of the drivers of poor condition can be addressed. But in doing this exercise we can demonstrate that our design is deliverable. This is why a condition assessment, underpinned by data and evidence is so important.
Progress so far
So what progress has been made on developing this method?
I published the original method about a year ago, posted about it on LinkedIn and put a copy on Codeberg. I have since had a number of really useful conversations with ecologists about this and some helpful revisions.
I am grateful to Hannah Williams from WSP for her comments on the method which helped move it a long considerably. I have also been testing the approach with Adonis Blue Environmental Consultants on their BNG projects in Kent.
More recently the UKHab team and Digital Ecology are working together to refine and revise the method and tackle some of those habitats that I have not. Together we have support from the CIEEM Professional Standards Committee who will hopefully review the method in due course.
Accessing the methods
But we need your help. The wealth of knowledge and experience in the ecology community can only serve to make the proposed method stronger, so we need your input. What works? What doesn’t? What have we missed? Is the method practical? Is there a better way of doing things?
The method is now hosted on a wiki at conditionassessmentmethod.co.uk. I am sure you are familiar with wikipedia; this wiki is built on the same technology, which means you can sign up to an account and make your won contributions. This means you can do as much or as little as you like.
It is completely free and therefore accessible to anyone, to use to adapt and to improve. I hope that if you do make changes for the better that you will share them back via the wiki.
Collecting data – the QGIS project
And to support data collection, Digital Ecology has made available a QGIS project with condition assessment forms for a number of habitats. These help you collect quantified data that can be used in determining pass or fail against the criteria for each habitat.
This available free from our Codeberg site.
This project can then be used to create MerginMaps or QField projects for digital data collection in the field.
What next? - testing, refining, expanding
So what is next for the project?
Further testing of the method will be carried out this year by Digital Ecology, its partners and clients. No doubt issues will be identified and updates made to make sure it works for as many people as possible.
I am also hoping to run a field day to look at whether the proposed approach actually does result in more consistent condition assessments compared to existing approaches – so if you think that sounds interesting, please sign up to the Digital Ecology mailing list where further details will be shared.
And hopefully, you will give your feedback on the method, so that better condition assessments can be carried out.