SFI 2026 agroforestry actions: tracking tree density on farm land

The Sustainable Farming Incentive took its agroforestry turn in the 2023 update and has consolidated that turn in each subsequent round. The 2026 version is expected to continue actions AGF1 and AGF2, covering in-field agroforestry at two density bands. These are relatively simple actions on paper: pay the farmer for keeping trees distributed through productive land at a defined density. The compliance details, however, are where application errors and audit failures bite.

This guide is a practical reference for farmers, agents, and consultants preparing SFI applications with agroforestry components. It covers the distinction between AGF1 and AGF2, what the RPA expects as evidence, how to calculate and document density, and the common pitfalls in agreements already running.

The two actions in summary

AGF1 — Maintain very low density in-field agroforestry. For eligible parcels carrying 15–50 trees per hectare, distributed across the productive area of the parcel rather than concentrated on boundaries. The agreement is annual, rolling, with payments per hectare of eligible parcel.

AGF2 — Maintain low density in-field agroforestry. For eligible parcels carrying 51–400 trees per hectare, with the same distribution and annual-payment structure. Payment rates are higher than AGF1, reflecting the larger area of productive land taken out of pure-crop production.

Payment rates vary by round and are published in the SFI handbook. In the 2025 round AGF1 was around £35/ha/year and AGF2 around £150/ha/year, with 2026 revisions expected to track input-cost inflation. Always reference the current handbook for the definitive figure.

Eligibility: what counts as in-field agroforestry

In-field, for the purposes of these actions, means trees distributed across the productive area rather than lined up on a boundary. A 50-hectare parcel with 100 trees scattered as single standards or in small clumps qualifies. A 50-hectare parcel with 200 trees in a linear belt on the southern boundary does not qualify for AGF1 or AGF2 (it may qualify for other actions, but not these).

Eligible parcels are permanent grassland, arable, temporary grassland, or mixed in-field productive use. Woodland parcels are not eligible (those are covered by separate woodland actions). Garden and amenity parcels are not eligible. Buffer zones along water bodies have their own actions and are not aggregated with AGF1/AGF2.

Eligible trees are defined by species (broadly, native or site-appropriate broadleaves and selected conifers), condition (alive, healthy, meeting the SFI handbook’s height and diameter thresholds), and protection (stock-proof if the parcel is grazed). Young trees below the height threshold can be counted only if they meet the establishment criteria; dead trees do not count.

Calculating density

Density is simply eligible trees divided by eligible parcel area. The parcel area comes from the Rural Payments mapping system record, which is the authoritative figure for SFI purposes. The tree count is your responsibility to produce and defend.

For small parcels with manageable tree numbers, a walking count with a clicker works. For larger parcels or irregular distributions, an aerial survey via drone is usually faster and more accurate. A nadir flight at 60–80 metres altitude produces imagery from which trees can be counted manually or via semi-automated tooling. Typical rate: 20–40 hectares per hour of flight, plus office time for the count.

Density can be uneven across a parcel. A 50-hectare parcel with dense tree cover on one half and bare on the other will average to a density figure that may mask local non-compliance. Some RPA inspections accept the whole-parcel figure; others ask for sub-parcel distribution. The defensible approach is to document both: report the whole-parcel density figure and flag the distribution uniformity (or lack of it) for the agreement.

Condition assessment

Density alone is not sufficient for AGF1/AGF2 compliance. The trees must be in “good condition”: alive, healthy canopy, stock-proof where applicable, and showing signs of being maintained. “Dead trees left standing” is an automatic non-compliance.

Condition evidence typically comprises dated photographs of a representative sample of trees, sized to the parcel. For a 10-hectare parcel, 20–30 photographs covering a range of trees is defensible. For a 100-hectare parcel with 200 trees, aim for 40–60 photographs including specific attention to any trees that look marginal. Photograph with GPS-tagging on so each image is locatable on the parcel map.

Drone capture works for condition assessment alongside density counting. A nadir flight for density gives you aerial imagery of every tree; side-on captures of a representative sample add the detail needed for condition judgement. This dual-capture approach is efficient for larger parcels and produces a comprehensive evidence pack.

Building the evidence pack

A compliant evidence pack for an AGF1 or AGF2 application, and for audit thereafter:

1. Parcel reference with RPA area.
2. Tree inventory: total count, species breakdown, approximate age distribution.
3. Density calculation: trees / hectare, plus sub-parcel distribution notes if relevant.
4. Aerial survey imagery with dated capture and capture parameters.
5. Ground-level condition photographs of representative trees, GPS-tagged.
6. Management plan: what maintenance is happening to keep the trees in compliant condition.
7. Any prior-year comparisons if the agreement is entering a renewal or multi-year phase.

The pack should be assembled once at application and refreshed annually, not scrambled together in response to an audit request. Audit prompts come unpredictably and are typically 14-day turnaround.

Combining with other actions

AGF1 and AGF2 cover in-field tree cover. Most working farms also have boundary features — hedges, shelterbelts, earth banks — that qualify for other actions. The typical combined agreement:

• AGF1 or AGF2 for in-field trees, annual SFI payment.
• HRW1 / HRW2 / HRW3 for hedgerow management, annual SFI payment.
• Capital Grants BN5, BN7 for boundary restoration or new planting, one-off capital payment.
• Capital Grants TE1–TE6 for tree planting, one-off capital payment.

Each action has its own evidence requirements, but the capture infrastructure overlaps. A single drone flight can produce imagery that supports in-field density counting (AGF1/AGF2), boundary condition assessment (HRW), and structural condition documentation (Capital Grants). The analysis tooling is different for each — tree counting is not the same as porosity measurement — but the raw imagery is shared.

See 2026 Capital Grants boundary evidence for the boundary side of this, and tree line density with drone photography for the drone capture workflow.

Common audit findings

Patterns from recent SFI audit cycles:

• Inflated tree counts. Counts that include trees on boundaries (not in-field), dead trees, or trees below the height threshold. The density appears to satisfy the action but the audit count lands lower.
• Clustered distribution. The whole-parcel average meets the density threshold but the trees are clustered in a small area, with most of the parcel effectively treeless. Some audits accept this, others don’t.
• Poor condition trees. Canopy dieback, uncontrolled grazing damage, missing stock protection. The agreement is in principle met by density but breaches the “good condition” requirement.
• Species drift. Agreements with a species specification that is not reflected in the actual planting. More common on newer plantings where the original design list was not fully realised in ground.
• Stale evidence. Photographs undated, or copied forward from prior-year evidence without fresh capture. A 2026 audit wants 2026 photographs.

Frequently asked questions