Fire: the Two Jobs and the Legal Checks

#Two farmers, one month, opposite results

Both light a block of Highveld grass in the second week of September. One is stripping out three years of moribund, greyed-off grass so the spring flush comes back palatable. The other is trying to knock back a dwarf shrub that has been thickening for a decade. Same month, same match, same veld type — and at least one of them has wasted the fire.

Fire here is not one tool. It is two tools that look identical from the bakkie, and the timing that serves one defeats the other. Get it wrong and you have paid for a fire, carried the risk of a fire, and bought nothing.

#Job one: remove moribund grass. Job two: hurt the shrubs.

The distinction is set out plainly in practitioner guidance, and it is routinely taught as one rule when it is two (SA Grain).

The job	When	Conditions on the day	Fuel needed
Remove moribund grass, restore grazing quality	After the first spring rains, as soon as possible	Cooler — under about 20 °C, humidity above about 50%	No specific load stated
Kill or check shrubs and encroaching bush	Before the first spring rains	Hotter — above about 25 °C, humidity below about 30%	3–4 t DM/ha of flammable grass

The physiology is the whole argument. For the shrub burn you wait for the moment when the woody plants have pushed reserves into new growth and the grass has not yet moved; burn then and the fire costs the shrub far more than it costs the sward. For the grassland-maintenance burn the concern runs the other way: light it too early and the stimulated grass regrows on reserves with no soil water behind it, while the bared surface loses ash nutrients to late-winter wind.

For the maintenance job, the long-standing recommendation in Highland Sourveld is biennial burning in spring, as soon as possible after the first spring rains (J.D. Scott; KZN DARD). Biennial — every second year. And if you have grazed the block to the ground you do not have a fire at all: below roughly 3–4 t DM/ha a burn does not carry hot enough to touch a shrub whenever you light it. Fuel load is a grazing decision taken months earlier. A block you intend to burn for bush control must be rested into the fire, which makes the burn plan and the grazing plan one plan.

Whole-block uniform burns are also not the recommended pattern. SANBI's biodiversity guidance is patch-mosaic burning rather than uniform whole-block burns (SANBI). That document could not be read for this course beyond the recommendation itself, so no fire-return interval and no figure for how much of a farm to burn per year is given here — get both from the guideline or your extension office before you plan a rotation of burns.

#The paperwork takes longer than the weather window

This is where most farmers lose. The legal lead time on a burn runs in weeks; the weather window runs in hours.

CARA governs whether you may burn. It says nothing about firebreaks, readiness or what happens when a fire gets away. That is a different Act.

The legal position

National Veld and Forest Fire Act 101 of 1998 (Act text):

s10(2) — the hard stop. Once a high-fire-danger warning is published for your region, no person may light, use or maintain a fire in the open air there. It is a first category offence (s25(1)) carrying up to two years' imprisonment on a first conviction (s24(1)). No plan, permission or agreed date survives it.
s5(1)(c) and s5(4)(c) — your FPA's rules bind you. An FPA must make rules binding its members, and those rules must provide for controlled burning. This, not the firebreak section, is where the permission and notification requirements for a management burn live.
s12(1) — the firebreak duty. Every owner on whose land a veldfire may start or burn, or from whose land it may spread, must prepare and maintain a firebreak on their side of the boundary — whether or not they ever intend to burn anything.
s12(2)–(4) — burning that firebreak. Agree a date with each adjoining owner and inform the FPA; failing agreement, give adjoining owners and the FPA at least 14 days' written notice. The adjoining owner must burn the same boundary the same day or attend, with enough people to stop a spread. Even after complying you may not burn if the FPA objects, a high-fire-danger warning is out, or conditions are not conducive.
s13 — firebreak standards. Wide and long enough to have a reasonable chance of stopping a spread, must not cause soil erosion, and reasonably free of flammable material.
s17(1) — readiness. Prescribed or reasonably required equipment, protective clothing and trained personnel, and responsible people present in your absence who will fight the fire and alert neighbours and the FPA.
ss24–25 — penalties. Three offence categories: maximum first-conviction penalties of two years (first), one year (second), and a fine or up to six months' community service (third). A second conviction in the second or third category is sentenced one category up (s24(4)).

Two traps. First, s12 governs firebreaks, not your management burn — a course that teaches s12(2) neighbour notification as the check before a grazing burn has given you the wrong section. Second, the prescribed s17 equipment and personnel standards could not be obtained for this course, so no kit list appears here. Ask your FPA; it is their rules that bind you.

Do this on the farm

Count backwards from the day you want black ground, every time:

Name the job. Moribund grass, or shrubs? That fixes the season and the fuel load.
Day −60 or earlier: for a shrub burn, rest the block into the fire so it carries 3–4 t DM/ha.
Day −30 at the latest: lodge the CARA Regulation 12 application, asking for permission to graze the burn in the same letter, and check your date falls inside the burning periods on record at the extension office.
Day −14 at the latest: firebreak dates agreed in writing with every adjoining owner, or 14 days' written notice served on them and the FPA.
Day −7: firebreak cut to the s13 standard — wide enough, clear of fuel, and cut so it will not start a donga.
Day −1: FPA membership current, no high-fire-danger warning, no FPA objection.
On the day: temperature, humidity and wind against the job in the table. If they do not match, you do not burn. The paperwork keeps.

#What a runaway fire is actually worth

Without membership you walk into court already losing: the neighbour proves the fire came off your land, and the burden of disproving negligence is yours. With membership, the neighbour builds the case. That change in who does the work is the cheapest legal position available in South African veld management.

What it is worth in rands depends on what burnt. No South African range of damages actually awarded in veldfire claims could be sourced for this course, so no figure is offered here — distrust anyone who offers one without a case reference. What sits in such a claim is clear enough: the neighbour's grazing for the rest of the season, fencing, sheds, stock, and your own losses on top. Price your boundary fence per kilometre off a live quote and you have a floor for one line of it.

#Do not burn for the carbon

There is a real temptation to bolt a soil-carbon justification onto a burn plan. Do not. The South African evidence points in two directions, and both camps have long-term data from the same experiment.

Contested — here is both sides

Against frequent fire. Sixty-four years of annual burning at Ukulinga lessened soil organic carbon and nitrogen in humid subtropical grassland (Global Change Biology, 2021). Long-term South African burning plots showed 70% lower organic carbon in the top 0–10 mm — 0.8% against 2.7% — but only 12% lower organic carbon and 13% lower total nitrogen over the 0–100 mm layer (du Preez et al. 2011). Quote both or neither: the alarming figure is the shallowest one, and sampling depth is doing most of the rhetorical work here.

For frequent fire. A 2025 study on the same Ukulinga experiment reports that early-season frequent prescribed burns produced greater soil carbon concentrations and higher sequestration rates than longer burn intervals, with accumulation continuing over a twenty-year window (African Journal of Ecology, 2025).

The likely reconciling variables are sampling depth, burn season and whether grazing was excluded. Nobody has settled it. Treat it as an open question, not as a side to pick.

So the burn decision is made on the grazing objective — moribund grass, or shrubs — and on the law. Not on a carbon claim in either direction. If someone tells you burning destroys your soil carbon, ask at what depth. If someone tells you it builds it, ask in what season and under what grazing.

#What to do on Monday

Put two dates in the diary: the day your CARA Regulation 12 application must be lodged — thirty days before the earliest day you would light — and the day your firebreak notices go out. Then confirm in writing that your FPA membership is current. If you cannot name which of the two jobs your spring burn does, you are not planning a burn; you are planning a fire.

#Check yourself

3 questions — answers explained as you go

1A stand of dwarf shrub has been thickening on a Highveld block for years and you want to check it with fire. When do you burn, and what has to be standing?
Why: The shrub burn works because the woody plants have pushed reserves into new growth while the grass is still dormant, so the fire costs the shrub more than the sward. Cool post-rain burns are the maintenance job and will barely mark a shrub. Without roughly 3–4 t DM/ha of fuel the fire does not carry hot enough to matter whenever you light it — which makes fuel load a grazing decision taken months earlier.
2Your neighbour's fire ran onto your block in August. It has greened up beautifully and your weaners need it. What does the law require?
Why: Regulation 12 has two limbs: no land user may burn veld, or use as grazing any veld that has burned, except on written permission. It does not care who lit it. The 30-day lead time is the trap — by the time the flush is worth grazing, an application not yet lodged is already too late.
3You are a paid-up member of your area's fire protection association and a fire off your land damages a neighbour. What has membership done for you?
Why: s34(1) reverses the presumption of negligence for FPA members; without it you start the case already losing. But s34(2) leaves the plaintiff's onus to prove wrongfulness intact. Membership shifts one element of a claim — it is not cover, and it does not replace the firebreak, the crew or the s17 readiness duties.

Fire: the Two Jobs and the Legal Checks

By the end of this lesson you can

#Two farmers, one month, opposite results

#Job one: remove moribund grass. Job two: hurt the shrubs.

#The paperwork takes longer than the weather window

#What a runaway fire is actually worth

#Do not burn for the carbon

#What to do on Monday

#Check yourself

Sources for this lesson