From 2 November 2026, biodiversity net gain becomes part of what a development consent order application has to demonstrate. That means the habitat baseline stops being an internal working document and becomes evidence, submitted, published, and open to written questions from an examining authority and to challenge from interested parties.
Baselines that were adequate as a design input are not automatically adequate as evidence. The difference is not usually accuracy. It is traceability: whether you can explain how you arrived at a number, and why someone else should accept it.
Here are five questions worth rehearsing before submission, and what a defensible answer looks like.
1. "How did you decide what to exclude?"
This is the first question, and it is the one most likely to be asked, because the NSIP regime makes exclusion attractive.
BNG applies to habitats negatively impacted by the development, temporarily or permanently, and to habitats where you will create or enhance to deliver gains. Unimpacted habitat inside the order limits can be excluded from the baseline, but only where an ecologist confirms there will be no permanent loss, no temporary loss and no deterioration in condition.
On a long linear scheme, exclusion can materially reduce the assessed area. Which is precisely why it will be tested.
A weak answer points at the boundary plan and says the excluded land is not affected.
A defensible answer does three things. It shows what habitat is on the excluded land, because you cannot credibly say land is unaffected without having characterised it. It explains the mechanism by which it is unaffected, whether that is confirmed below-ground infrastructure, sufficient separation from works, or an above-ground structure with no negative effect. And it names the ecologist who made the judgement and describes their reasoning.
The guidance is direct that the robustness of the ecologist's reasoning matters here. That is an invitation to write the reasoning down at the time, not to reconstruct it eighteen months later.
2. "How current is this data?"
Ecological data ages, and NSIP programmes are long.
CIEEM's guidance on the lifespan of ecological reports gives the working framework. Under twelve months, likely valid in most cases. Twelve to eighteen months, likely valid with caveats for mobile species and for sites that have gained new features. Eighteen months to three years, a professional ecologist needs to revisit and review whether the data still holds. Over three years, unlikely to be valid, with most surveys needing to be redone.
A consent programme that runs five years from first survey to examination can very easily arrive with a baseline that has expired without anyone noticing.
A defensible answer states the survey dates plainly, sets out what has been reviewed and when, and explains what changed and what did not. Where data sits in the eighteen-month to three-year band, it records the ecologist's review rather than leaving the currency question open.
The version of this that goes badly is the application where survey dates are buried in an appendix and the first person to add them up is the examining authority.
3. "Was this surveyed in an appropriate season?"
Habitat condition assessment depends on being able to see and identify what is growing. Most habitat types are assessed in the growing season, broadly from mid to late April through August, with heathland and wetland often surveyed later. Protected species have their own windows.
Surveys outside the optimal window are not automatically invalid. But a negative result from a sub-optimal survey does not demonstrate absence, and a condition score derived from a February visit to a grassland invites a reasonable question.
A defensible answer does not pretend the constraint away. It states when each element was surveyed, identifies anything assessed outside the optimal window, explains what that means for confidence in the result, and sets out how the uncertainty was handled, whether through a precautionary assumption, a repeat visit or an explicit limitation.
Declaring a limitation is far stronger than hoping nobody checks the calendar. The seasonality problem, and how programmes work around it, is covered in the NSIP survey season problem.
4. "Does this baseline reflect the worst case, or the current design?"
NSIPs go to application with elements of the design unfixed. Guidance directs that where locations and construction areas remain uncertain, the baseline should reflect the realistic worst-case extent of habitat loss or deterioration, based on maximum design and construction parameters.
The failure mode is a baseline built on the design as it stood when the survey team went out, followed by a design that moved.
A defensible answer demonstrates that the assessed extent corresponds to the maximum parameters being applied for, not to a snapshot. It shows the assessed area against the envelope, explains how construction areas, compounds, haul routes and laydown were captured, and confirms that temporary impacts are included.
That last point catches people. Habitats of low or very low distinctiveness that are temporarily impacted do not have to be recorded as lost if restored to baseline condition within five years, which is a genuine concession for NSIPs. But they remain negatively impacted habitats, so they still belong in the baseline and on the boundary plan. The concession affects the scoring, not the scope.
5. "What is your confidence in this, and where is it weakest?"
This is the question that separates a baseline that survives examination from one that merely arrives at it.
Every baseline contains variation in certainty. Some parcels are unambiguous. Some habitats are hard to classify from any single method. Some areas were accessed easily and some were not. Some were surveyed in the right month and some were not.
A weak answer presents a single uniform map and implies even confidence across all of it.
A defensible answer says where the assessment is strong, where it is weaker, what was done about the weaker areas, and what residual uncertainty remains. It distinguishes accuracy from confidence. Accuracy tells you what the assessment concluded. Confidence tells you whether you should act on that conclusion without further checking.
Volunteering the weak points seems counterintuitive when you are trying to get an application through. In practice it does the opposite of what people fear. An applicant who has already identified their own uncertainties and addressed them looks like an applicant who understands their site. An applicant whose uncertainties are discovered by someone else looks like the reverse.
The pattern underneath all five
None of these questions is really about ecology. They are about whether the evidence can be traced, explained and defended by someone other than the person who produced it.
That is the shift the November change brings. Customers and consenting bodies do not want more data. They want evidence they can report, defend and compare. A baseline that is technically excellent but undocumented fails that test. A baseline that is honest about its limits and clear about its method passes it.
If you are assembling a baseline for a post-November application, the wider requirement is set out in BNG for NSIPs: what changes on 2 November 2026, and if you are working out what is achievable in the time you have, what is realistically deliverable before November is the practical companion.
How we think about this
Gentian maps habitat remotely at NSIP scale, and our parcel-level accuracy has been independently validated against chartered ecologists. The part of our work that matters most for the questions above is confidence: identifying where remote assessment is reliable, where field verification is needed, and where uncertainty genuinely remains.
We are clear about the limits of what we do. Remote assessment gives you a wide, consistent, repeatable view and a documented method. It does not replace the ecologist, the field verification, or the professional judgement behind the statutory submission. Resolution can be purchased. Trust has to be earned.
If you want a baseline you can hand to an examining authority without flinching, talk to us.