Each droplet leaving the sprayed field carries a fixed pesticide mass and slides downhill with runoff. Inside the vegetated buffer strip, dense stems and thatch intercept and infiltrate the flow, removing pesticide mass roughly exponentially with the distance travelled through the strip — the standard vegetative filter strip (VFS) trapping model used in agricultural runoff studies:
C_out / C_in = e^(−k·W)
η = 1 − e^(−k·W)
where W is the strip width (m) and k is a vegetation-density extinction coefficient (1/m) — denser grass, higher stem density and more surface roughness raise k. The simulation applies this per unit time travelled inside the strip (trap probability per step = 1 − e^(−k·v·dt)); telescoping that per-step survival over the full crossing time W/v reduces exactly to e^(−k·W), so the measured retention η converges to the closed-form curve above as more droplets cross the strip (verified numerically: 200,000-trial Monte-Carlo of the per-step rule gives η≈0.658 against the closed form's 0.660 at k=0.18, W=6).
- Rainfall / runoff intensity — scales downhill flow speed and how many droplets are mobilised per second.
- Spray application dose — scales how much pesticide mass each droplet represents, feeding directly into the stream concentration readout.
- Strip width / vegetation density — the two buffer-strip parameters in the formula above; widen the strip or thicken the vegetation and the fraction reaching the stream falls exponentially, not linearly.
Real-world relevance: this is exactly the trade-off regulators lean on for setback/buffer-strip rules under IPM guidance — a narrow, sparse strip barely helps, but width and density compound because they sit in the same exponent. The top-down field view (drag to pan, scroll/pinch to zoom) shows the geometry directly; the strip chart below traces your measured η against the theoretical curve.