What the Bracknell trial is telling us: pavement ecological succession, the Kick Test, and a 49% reduction in active substance use

In February 2026, Kersten UK, Complete Weed Control, and Bracknell Town Council began a trial that we think will change how local authorities approach hard surface weed management. Six months in, with the first growing season underway and the first pesticide application data now in, the results are confirming what the baseline research suggested - and adding something more concrete than we had before.

This post summarises where the trial stands, what it has found, and what it means for councils managing urban highway networks under NAP 2025.

The starting point: weed presence is the wrong problem to solve

The conventional approach to hard surface weed management treats weed presence as the primary problem. Our research argues that weed presence is a symptom, and that the real problem is the accumulated detritus at the kerb edge that provides the biological growth medium in which weeds establish.

This distinction matters because it changes what you measure, when you intervene, and how much it costs. A council that manages weeds by appearance is making decisions without the information it needs. A council that manages the substrate - measuring depth, testing root anchorage, identifying indicator species - is making decisions that can prevent capital expenditure rather than simply defer it.

The Bracknell trial was designed to generate the empirical data that turns that argument from a principle into a validated, reproducible field framework.

Pavement ecological succession: what the baseline data shows

The pre-season baseline survey collected 502 species-depth measurement pairs across six sites - two surface car parks, two park-edge kerb environments, an enclosed paved area, and a residential footway. The statistical analysis confirmed something that had not previously been demonstrated in formally published literature: plant species assemblages on hard urban surfaces are significantly stratified by detritus accumulation depth.

The Kruskal-Wallis result (H = 125.35, p = 6.03 x 10⁻²⁸) is not a marginal finding. The effect size (epsilon-squared = 0.247) is large by conventional thresholds, meaning depth alone accounts for approximately a quarter of all variance in species composition across the dataset. For a single field variable measured with a Vernier gauge, that is a striking result.

Three indicator species emerged with statistically significant stage concentration:

• Moss (p < 0.001) - concentrated in Stages 1 and 2. Its presence reliably rules out Stage 4 structural damage.
• Oakleaf fleabane (p = 0.021) - 85% of observations in the 11-25mm band. The strongest single-species signal that the intervention window is open.
• Bramble (p < 0.001) - 70% of observations at Stage 4 depth, never recorded at Stage 1. Its presence is the strongest single-species indicator that the window has closed and structural damage is underway.

Dandelion and meadow grass - the two species most commonly cited in practitioner guidance as weed indicators - showed no statistically significant stage stratification whatsoever (p = 0.754 and p = 0.715 respectively). Both are distributed across the full accumulation gradient. Relying on either to trigger or characterise an intervention decision has no statistical justification.

The four-stage framework and the Kick Test

The baseline data established a four-stage diagnostic framework - Pavement Ecological Succession - based on detritus depth and root anchorage state:

The most important diagnostic decision in the entire framework is the Stage 2 / Stage 3 distinction. Both stages present at 11-25mm depth and can look identical above ground. The cost differential between them is 10:1 to 15:1. Depth measurement alone cannot resolve it.

The Kick Test - applying lateral foot pressure to the base of the weed - was developed from empirical observation during extraction at Bracknell, where root masses at identical measured depths behaved differently. Some lifted freely with the substrate as a complete unit. Others broke cleanly at substrate level, leaving root material embedded in pavement cracks. That observation is now formalised as a reproducible pre-extraction field diagnostic that any operative can apply before deploying any equipment. One additional observation worth noting from the baseline extraction: at Birch Hill Car Park 1, extraction revealed a previous resurfacing course laid directly over established bramble root systems - confirming that resurfacing over unextracted root mass leaves the underlying problem in place.

Try the Kick Test diagnostic below:

June 2026: first pesticide application data

The trial completed its first pesticide application in June 2026, generating the initial treatment efficacy data from the growing season. The headline result is a 49% reduction in active substance used on brush-and-spray plots compared to spray-only plots across five comparable sites.

The reduction is attributable to the removal of the detritus matrix before spraying. Without the accumulated growth medium, the target surface area and weed biomass requiring treatment is significantly smaller. The mechanism is straightforward: mechanical extraction is not just a weed control intervention, it is a pesticide reduction intervention. The two objectives are the same action.

The Elms was excluded from this comparison because a surface condition mismatch between brushed and unbrushed plots at baseline made a like-for-like comparison unreliable at this stage. That is the correct handling of the data - reporting the result you can stand behind, not the one that looks best.

What this means for NAP 2025

Local authorities are under regulatory pressure from NAP 2025 to demonstrate progressive pesticide reduction with data-driven operational plans and measurable detritus thresholds. The Bracknell data provides exactly the kind of evidence that makes that case internally - not a policy commitment to reduce herbicide use, but a measured, quantified reduction achieved through a specific operational change at a specific set of sites.

A 49% reduction in active substance use is not a modelled projection. It is a measured outcome from the first application of the first growing season. The end-of-season report, anticipated October 2026, will add recolonisation rates, re-accumulation data, and full treatment efficacy across all plots and all treatment types.

Mid-season monitoring: what we are tracking now

Mid-season monitoring is currently underway across all six Bracknell sites. Each visit records detritus depth at permanently marked measurement points, species composition, weediness scale scores, and recolonisation state following treatment. This is the first post-treatment monitoring round and will generate the initial recolonisation sequence data - which species return first, at what depth, and at what rate following different treatment combinations.

The full picture so far

The Bracknell trial is six months into a programme that runs to October 2026. What the baseline and first-season data have established so far:

• Species assemblages on hard urban surfaces are statistically stratified by detritus depth - a large-effect result confirmed across 502 measurement pairs.
• Three indicator species provide reliable field proxies for cascade stage. Dandelion and meadow grass do not.
• The Stage 2 / Stage 3 distinction carries a 10:1 to 15:1 cost differential and cannot be resolved by depth measurement alone. The Kick Test resolves it.
• Mechanical extraction before herbicide application produced a 49% reduction in active substance used across five comparable sites in the first growing season application.

The full baseline paper is available to download from the trial page. End-of-season data will be published as the trial concludes.

The Bracknell Integrated Weed Management Trial is a collaboration between Bracknell Town Council, Complete Weed Control Ltd, and Kersten UK. For the full trial overview, Kick Test simulator, and baseline paper download, visit kerstenuk.com/bracknell-trial.