Points, Polygons and Plantlife: How QGIS & AI Can Transform Your Habitat Survey Workflows
Introduction
UKHab and NVC habitat surveys are a core part of an ecologist's role, and they are more technically demanding than they might appear. The classifications require ecologists to differentiate between habitats that are often very similar, and assess its condition. And to do that you need to be able to identify key indicator species in the field. And that means real botanical skills; building a fluency with plant ID that only comes with practice.
I cast my mind back to some memorable intertidal biotope surveys with JNCC on the shores of Anglesey and the West coast of Scotland. UKhab wasn’t around then, nor survey apps. But the experience taught me early on the importance of survey preparation, correct species ID, and efficient field data capture (tides!).
In this article, I want to explore how QGIS, alongside field apps like QField and Mergin Maps, can support the survey process; from pre-survey preparation, through to BNG reporting. I’ll also take an honest look at where AI plant ID tools like Pl@ntNet and Flora Incognita and Google lens fit in (and it isn’t replacing the botanical expertise at the heart of a habitat survey).
Phase 1, UKHab & NVC : Understanding the Habitat Survey Landscape
Before diving into the habitat survey workflow, it's worth a quick moment of context.
Phase 1
The Phase 1 habitat survey, developed by JNCC and first published in 1990, is still in use by some today.
It was the original framework for systematic habitat mapping in the UK. Its approach, and iconic symbology, became deeply embedded in ecological practice.
Many ecologists still have strong affection for it, and indeed, target notes remain popular in the field: a flexible, familiar way to capture what the classification codes can't quite express.
UKhab
Released in 2018, UKhab built on Phase 1's foundations and replaced it as the industry standard.
It offers a more granular classification, including 29 primary codes for grassland alone!
Although there is now some divergence between UKhab and the Statutory Metric habitats, if you're doing BNG work, you're using UKHab. That connection makes getting the habitat survey, and species ID, even more consequential.
NVC
Where a habitat is a bit special, or for a transitional habitat, or a detailed grassland assessment, the National Vegetation Classification (NVC) remains very much in active use.
Where UKHab operates at the level of habitat type, NVC goes deeper, characterising vegetation by its precise species composition and structure.
It is routinely required for statutory designations and UKhab and NVC are complementary. For some sites, you will need both.
Pre-Survey Preparation: Getting More Out of Your UKHab Survey Before You Arrive On Site
Good UKHab habitat surveys start long before you arrive on site, and QGIS is where that preparation happens.
Key datasets to load before heading out
OS National Geographic Database or OS Enhanced Landcover
Get accurate boundaries, habitat descriptions and site extents before you arrive, and save a lot of unnecessary digitising.
Aerial imagery (e.g. Bluesky)
Recent, dated, high-resolution imagery gives you additional insight into land cover ahead of the walkover.
Natural England Priority Habitat Inventory
Flag areas with potential for higher distinctiveness habitats early, so you know where to focus botanical effort on site.
FSC Biological Records Tool for QGIS
Import existing species records (e.g. from a LERC) directly into your project as mapped points or OS grid squares. Knowing a site has historic records of rare species changes how carefully you survey certain fields.
You can read more about choosing datasets to inform BNG assessments here.
In the Field: Capturing Botanical and Habitat Data with QField and Mergin Maps
Both QField and Mergin Maps let you take your QGIS project into the field on a tablet or phone, with pre-mapped parcels on screen overlaid on aerial imagery.
Key capabilities of these apps for UKHab surveys include
Digitise habitat boundaries once
Draw new parcels directly on top of aerial imagery as you walk the site, or cut the RLB into habitat areas, or simply re-classify your pre-survey data. Whichever is your approach, they are all so much more efficient than sketching on paper and redrawing back at the office.
Use full UKHab, and refer to habitat descriptions
Spatialsesh BNG includes the full UKhab classification set up as a structured dropdowns, with guidance for reference, so you can be confident selecting the correct code, rather than relying on memory or a separate crib sheet. You can set up similar for whatever habitat survey you are doing.
Accurate conditional assessments for full transparency
Record precise location of CA, and link to habitat polygons. Plus capture cover-abundance scores for NVC quadrats, or presence/absence of key indicator species for condition assessments.
Photo evidence
Attach images directly to individual habitat parcels or CA’s; invaluable when you need to evidence a pass or fail against a condition criterion, or if you want to check a species ID later.
Offline working
Data stores locally on your device and syncs back to QGIS when you're back on 5G/WiFi; no re-entry, no transcription errors, no illegible field notebooks.
For a deeper look at how this fits into a full BNG mapping workflow, see How can you streamline your BNG mapping workflow in QGIS and eliminate the usual friction?
Back in the Office: From Habitat Survey Data to BNG Report, Faster
Once your data is back in QGIS, the reporting stage is where the time savings really stack up:
Automated geometry fixing
GIS model builder allows you to develop custom QGIS tools that fix invalid polygons, overlaps, slivers, and coverage gaps within the red line boundary; what used to be a painstaking afternoon of manual fixes takes minutes.
Metric-ready export
Habitat parcels that are ready to export to the Metric, and condition scores that are ready to drop straight into the CA spreadsheets.
Professional map outputs
Consistent symbology and clean, well-attributed parcels produce polished deliverables for client reports that reflect the quality of the underlying survey work.
Can AI Plant ID Apps Support Botanical Survey Work? The Answer Is Yes.
Differentiating between habitat types, and assessing condition, requires real botanical skills. Building a fluency with plant ID takes practice. And today, AI plant identification apps like Pl@ntNet, Flora Incognita and Google Lens are widely used and very useful.
The Botanical Society of Britain and Ireland (BSBI) took a notably positive stance on these tools in their webinar Artificial Intelligence: the good, the bad, and how best to approach it as a botanical community and it's not hard to see why:
Accuracy is already impressive, and improving rapidly, making them a reliable first check for a wide range of species.
Real-time learning in the field; make an identification, check it against the app, and build visual recognition on the spot. The feedback loop accelerates botanical learning in a way that wasn't possible before.
Helping close the skills gap; in a profession where strong ID skills directly underpin UKHab classifications and condition assessments, anything that builds botanical confidence is a good thing.
Used thoughtfully, AI plant ID is not a shortcut around botanical knowledge, it's a tool for building it. Integrating these tools into QGIS workflows is an area of active interest to me (watch this space!).
A Smarter UKHab Survey Workflow, From First Prep to Final Report
Although AI may have more of an influence going forward, the ecology at the heart of habitat survey; knowing your plants, understanding the NVC communities or UKhab types, working methodically through a condition assessment; will always require human expertise.
But the GIS workflow around that ecology: the pre-survey preparation, the field data capture, the geometry fixes, the export to Defra metric and statutory CA spreadsheets; that's where spatial tools can make a big difference.
A UKHab habitat survey that flows from QGIS preparation, through QField or Mergin Maps in the field, and back into a clean project ready for metric export is faster, easier and more consistent than one pieced together from paper forms and manual data entry.
And in a professional landscape where BNG data is increasingly scrutinised, that matters.
Ready to improve your UKHab habitat survey workflow? Explore Spatialsesh BNG and QGIS for BNG training course (both built specifically for ecologists delivering Biodiversity Net Gain assessments in QGIS)