The Capacity Map and the Legal LSU Table

#There is a number attached to your farm that you did not choose

In a provincial office there is a topo-cadastral map with your farm on it and a figure printed over the polygon: hectares per large stock unit. You did not survey it. Nobody asked you. And under Regulation 10 of the Conservation of Agricultural Resources Act 43 of 1983, that figure is the legal default for how many animals you may keep.

Read "default", not "permit". Nobody issues you a stocking licence, nobody inspects a smallholding annually, and Regulation 11 carries a proviso that expressly allows the number to be exceeded where the veld is protected. What the map does is decide who carries the burden of proof. Run at the mapped figure and nobody asks you anything. Run above it and the number is what you must argue against — and the only currency that argument accepts is a dated record of your own veld condition.

Most farmers react in one of two wrong ways. Either the map becomes gospel — "the department says 6 hectares, so 6 hectares it is" — or it becomes a joke, a number drawn by a committee that has never walked their kloof. Both reactions skip the same thing: the department already told you what the map can and cannot do, in its own background document. Almost nobody reads it.

#How the number was made

The original 1993 map was drawn by expert opinion. Pasture scientists, extension officers and farmers marked capacity onto 1:250 000 topo-cadastral sheets, and the former homeland areas were excluded entirely.

The current map was rebuilt with GIS. Median and modal zonal statistics were computed using the Mucina & Rutherford vegetation map as the zones and the 1993 map as the reference dataset, then workshopped province by province. Additional inputs were the national Land Type Survey and, in KwaZulu-Natal, Camp's 1999 Bioresource classification (DALRRD background document).

One consequence matters before you look up your own value: the polygons are not drawn on a consistent basis across the country. Northern Cape polygons follow farm and district boundaries; Eastern Cape polygons follow vegetation boundaries. A polygon edge may be an ecological fact or an administrative one, and the map does not tell you which. The 2017 revision was gazetted for public comment as GN 971 in Government Gazette 41100 of 8 September 2017 (notice; gov.za record).

#The four limits the department states about its own map

This is the part to memorise, because it is what converts the map from a verdict into an input.

1. The values assume veld in relatively good condition, and are described as a guideline for planning. If your camps are below benchmark, the polygon figure is optimistic for you specifically.
2. A utilisation factor of 40–50% is already built in. The map does not tell you how much grass grows; it tells you what you can take while leaving the rest as cover, litter and plant reserves. Which is why "but there is still grass standing" is not an argument against the number — the standing grass is part of the calculation.
3. Tree crown cover was not incorporated, and browsing capacity is not included at all. On encroached bushveld the map overstates what you can run on grass, and it is silent on a browse resource that may be substantial.
4. The map was built for use at 1:250 000. Sub-polygon variation exists on every farm — the vlei, the north slope, the old lands — and none of it is visible at that scale.

Read together: the map is a ceiling drawn for an average property in good condition, with no trees and no browsers, at a scale that cannot see your farm. It is still the number you work with. It is not the number you stock to.

#How CARA turns a map into a duty

The regulations under CARA 43 of 1983 (GN R1048 in GG 9238 of 25 May 1984, as amended) do the converting.

Regulation 11 also carries a proviso, quoted here in full because people quote it loosely:

"Provided that such number may on occasion be exceeded on condition that the veld shall under all circumstances effectively be protected against deterioration and destruction."

Two things about that sentence are textually solid. Regulation 11 caps the total number of animals on the veld of the farm unit — a stocking-rate constraint, not a limit on how tightly you pack one camp for a day. And the proviso genuinely does permit exceeding the number.

#The legal LSU table is not the ARC table

Here is the trap that catches people who did the last lesson properly. That lesson used the ARC's frame-size factors — the feed-demand science. Regulation 11(2) and its Table 2 give different numbers, because they do a different job: they are the legal accounting factors that test compliance with the ceiling.

The most useful line in that table is the first one. A medium-frame cow is 1,21 LSU in law — not 1,0. Almost every farmer counting his herd in his head assumes a cow is one unit, and lands about a fifth under the truth. The ARC makes the same point from the science side: grazing capacity and stocking rate "sometimes differ by as much as 300%, leading to total degradation" (Mokolobate et al. 2017).

Use the ARC factors to plan feed. Use CARA Table 2 whenever you are answering to CARA. Write down which table produced which number, because they will not agree and someone will eventually ask.

#Work it on a real farm

A 300 ha Highveld property, all of it veld, sitting in a Grassland polygon at 6 ha/LSU.

The legal ceiling: 300 ÷ 6 = 50 LSU.

The stock: 30 medium-frame cows, 20 medium-frame weaners over 7 months, 12 mutton-breed ewes, 2 light-horse mares.

Counting the way most people do — a cow is one unit, a weaner half, sheep and horses barely register: 30 + 10 + 2 + 2 = 44 LSU. Comfortably inside 50. Nothing to see.

Counting on CARA Table 2:

• 30 cows × 1,21 = 36,30
• 20 weaners × 0,53 = 10,60
• 12 ewes × 0,15 = 1,80
• 2 mares × 0,94 = 1,88

Total: 50,58 LSU.

Same animals, same farm. One count is 12% under the ceiling and the other is over it. Nobody bought a cow; the arithmetic just stopped being wrong.

And that is before the four limits bite. If your veld is below benchmark, if 15% of the property is encroached, if the vlei is grazeable four months a year — the honest working figure is below 50, not at it. Elsenburg puts the recommendation plainly: keep fewer animals than the recommended grazing capacity allows. The pay-off it lists is reserves that build, a drought survived without buying fodder, and long-run capacity that rises.

#When the permit is the right route

If your own assessment genuinely differs from the polygon, arguing at the bakkie window achieves nothing. There is a lawful route.

Regulation 10(3) is that route: the executive officer may set a different capacity for your farm unit where actual capacity differs appreciably. Individual-farm exceptions are handled by Resource Auditors, and the department states these permits are valid for one year. A revision of the mapped figure itself requires a veld condition survey by a registered pasture scientist or an equivalently qualified person (DALRRD).

A one-year permit is a number agreed for a season, not a permanent reclassification. It is the right route when you have an evidenced case — a farm-specific survey, several years of records, a genuinely mismatched polygon boundary. It is the wrong route for legitimising a herd you have already bought.

The evidence has to exist before you need it, and almost nobody starts it in time. If you intend to run above the mapped figure, start the file in your first season, not in the season somebody asks.

#What to do on Monday

Get your polygon figure in writing, with the date, from the executive officer or the extension office. Convert your herd and flock on the Table 2 factors. Put the two numbers side by side on one page.

If you are over, you know it before an inspector does, and Regulation 9(1)(e) gives you three lawful ways down. If you are under, you have the start of the record that makes the Regulation 11 proviso arguable rather than hopeful.