Grazing guidance for communal land run by Communal Property Associations and village committees in South Africa
Give every camp time to grow back, and let the village decide how.
Soil moisture probes, grass cameras and herd counts give a simple picture of how hard each camp is being grazed. Veldwise suggests a rotation, drafts the management plan and prints it for the meeting. The committee and herders decide. Nobody's right to graze changes.
Grazing pressure by campillustrative
The black line is what the camp can carry. Past it, the grass starts to lose.
Illustrative example, moving every second. Not live telemetry.
Original illustration: the herder walks the herd through the gate to a camp that has had time to rest.Original illustration: the plan on the phone is a starting point. People in the room make the decision.
Who it is for
CPA and village committees
A clear picture of each camp, a rotation to discuss, and a plan document ready for the meeting.
Herders and livestock owners
Simple words, big buttons and a phone view. Herders help check the camps on foot.
Extension officers and funders
A shared record of what was suggested, who agreed, and what happened to the grass.
What Veldwise does
Shows how hard each camp is being grazed, and how the grass is doing.
Suggests a rotation and drafts a plan for people to discuss.
Keeps a record of what the committee agreed.
What it does not do
It does not change or allocate anyone's right to graze.
It does not name or rank individual owners or herds.
It does not replace the committee, traditional leadership or the extension officer.
It does not claim to know one perfect number of animals.
Questions people ask
Will this tell me how many animals I may keep?
No. It shows pressure on a camp, not limits on a person. Any change to who grazes where is agreed by the community.
Who owns the information about our camps?
The community. Camp data is recorded by camp, not by owner, and the committee decides who else sees it. The full principles are on the Pilot and practicalities page.
What if the phone or sensors stop working?
Herders can walk the camps and record what they see. The Field kit in the demo shows this routine. The system says plainly when a reading is missing.
Is rotational grazing always the answer?
No. Researchers disagree about how much rest, fencing and destocking help on communal land. Veldwise presents options and evidence so people can judge for themselves.
Which languages does it speak?
Plain words first, in the languages people use at the meeting, written and tested with local speakers. The languages will be confirmed with each pilot community. See the Pilot and practicalities page.
What if we have no smartphone?
Committee members and herder representatives can get short SMS messages, and the plan and camp sheet can be printed for the meeting. The demo shows both.
The problem
When the herd can't move on, the grass can't grow back.
On shared land, grass that is grazed again and again without rest gets thinner each season. Bare ground, erosion and poorer animals follow. The cause is rarely one person. It is usually that no one has a shared picture to decide from.
Dry spellRain is late or short.
Herds stay putNo rested camp is in reach.
Grass thinsPalatable species lose out.
Soil showsBare patches spread.
Rain runs offWater and soil wash away.
Animals lose conditionLess milk, weight and income.
Arguments and lossesHard to agree when grass is short.
Three situations the system tells apart
Short dry spell
Grass cover dips but recovers once it rains and the camp rests. The suggestion is to keep watching.
Chronic overgrazing
Cover keeps falling season after season while pressure stays high. This needs a shared decision about rotation.
Not enough data
A camera lens is dirty or a probe is offline. The system says so and asks herders to check instead of guessing.
Leads, not verdicts. Veldwise gives the committee leads to discuss. It never judges a person, a herd or a right to graze, and carrying capacity is an estimate that researchers still debate.
How it works
From a soil probe to a plan the village signs off.
Ingest
Soil moisture probes, grass cameras and herd counts from herders' phones or tag readers arrive each day or week.
Score
A pressure model compares animals with what each camp can carry, given recent rain and grass cover.
Explain
A language model writes a short plain note: what changed, what the evidence is, and what could be done.
Pattern-check
Years of history help tell a short dry spell from a slow decline.
Verify
A vision model estimates grass cover from camera photos. Herders confirm on foot when the two disagree.
Visualise
A camp map, a phone view and a printable plan show the same facts.
The committee decides
The committee and herder representatives agree, change, decline, ask herders to check, or ask the extension officer. Nothing becomes a plan before that.
The six roles
Five helpers, one committee that decides.
Each helper has one separate job, so each can be checked on its own. None of them moves a herd, changes a right to graze, or signs a plan. Here is what each one reads, what it produces, how it helps, and what a person still checks.
1. Grazing pressure model
Forecasting model
Reads
Soil moisture readings, rain over the last 30 days, grass cover history, herd counts for each camp, days since the camp last rested, and the season.
Produces
A pressure score for each camp (animals compared with what the camp can carry right now) and a four-week outlook for grass cover, with a confidence value.
How it helps
Turns scattered readings into one comparable number per camp, so the committee sees which camp needs attention first, weeks before the grass is gone. Because it uses recent rain, the camp's capacity moves with the season instead of staying a fixed guess.
A person checks
That the herd count is right, and whether local knowledge disagrees, for example after a burn, a disease outbreak or a new water point.
2. Rotation advisor
Language model
Reads
The pressure scores, each camp's rest history, where water points are, and the rules the committee has already agreed, such as minimum rest periods.
Produces
Two or three rotation options in plain words, each with its trade-offs: walking distance for herders, days of rest gained, and which camps are affected.
How it helps
The committee starts from options instead of a blank page, and can compare them side by side. Explanations are written in short, simple sentences, and can be translated into the languages people speak at the meeting.
A person checks
Fairness between households, water access, agreements between neighbouring villages, and the rules of traditional leadership, none of which the model can see.
3. Management plan writer
Long-context language model
Reads
The record of decisions, camp maps, meeting notes and Field kit entries for the year.
Produces
A draft CPA grazing management plan: the camps, the rotation calendar, the rules agreed, how the grass will be monitored, and a signature page.
How it helps
Saves the committee secretary days of writing, and gives extension officers and funders a consistent document in the same format every time. Meeting notes become a plan the village can read back and correct.
A person checks
That every line matches what was actually agreed. The committee edits the draft and signs it. Nothing is final without signatures.
4. Pattern reader
Long-context language model
Reads
Several seasons of rain, grass cover, herd counts and past committee decisions.
Produces
A label for what is happening in each camp, such as dry-spell dip, slow decline or recovering, with the evidence behind it and any other explanation it cannot rule out.
How it helps
Stops a panic after one dry month and stops a slow decline from being ignored. After a rotation starts, it shows whether the grass is actually coming back.
A person checks
Whether the explanation fits what people saw, including fire, disease, old fields, bush encroachment and wood harvesting.
5. Grass cover checker
Vision model
Reads
Photos from fixed camera posts, and photos herders take along a short walking line with their phones.
Produces
An estimate of green cover, bare ground and bush, with a confidence value. It also flags a dirty lens, a blocked view or photos taken at a different angle.
How it helps
Gives regular grass readings for camps that are far from the village, using the same viewpoint every time so changes are easy to see. It builds a photo record that can be shown to the meeting and to funders.
A person checks
When the camera and a walk on foot disagree, herders confirm. Shade, dust and the colour of winter grass can all fool a camera.
6. The committee
Human reviewer (mandatory)
Reads
Everything above, plus what no sensor can see: who lives where, who depends on which camp, past agreements and local rules.
Decides
Agree, change, decline, ask herders to check, or ask the extension officer. Herder representatives are part of the decision.
How it helps
Keeps the decision with the people who live on the land. The models give leads to discuss. The committee turns a lead into a plan, or does not.
Advised by
The agricultural extension officer, who can explain options and what research says.
A worked example (simulated)
An illustration of how the roles fit together. Not real data.
Readings arrive
Camp A's camera and soil probe report, and herders update the herd count. Grass cover has fallen from 71% to 38% over six weeks.
Pressure model scores Camp A
About 1.4 times what the camp can carry, with low confidence on the herd count.
Pattern reader labels it
Slow decline, not a dry-spell dip. It notes that rain was near normal, so weather alone does not explain it.
Rotation advisor offers three options
Move a third of the herds to rested Camp C for three weeks, move all herds for two weeks, or ask herders to walk Camp A first.
The committee decides
It picks the first option and asks herders to confirm the cover on foot. The plan writer adds it to the draft management plan.
The loop closes
Three weeks later the grass cover checker shows whether Camp A is recovering, and the committee reviews the plan.
What the models never do
Move a herd or send anyone anywhere.
Change, allocate or remove a right to graze.
Name, rank or report on individual owners.
Present a draft as final without signatures.
Hide uncertainty. Every output shows confidence and what the model cannot see.
How we keep them honest
Outputs are probabilities and estimates, never facts about a person.
Herders' walks and photos are the check on every camera estimate.
Model and provider names are left out on purpose: choose current offerings when building and check provider documentation.
Deployment options under consideration: Azure IoT Hub with Azure OpenAI, AWS IoT with Amazon Bedrock, or Google Cloud with Vertex AI, with optional edge processing at the camp.
3D model
Crowd one camp and watch the grass thin.
Drag to look around, scroll or use the buttons to zoom, and move the slider to put more animals in Camp A.
Soil sensor
Grass camera
Phone with grazing plan
Camp A: grazed now
Camp B: resting
The 3D model needs WebGL and an internet connection to load its library. Try another browser or check your connection.
Grazing pressure0.6 × capacity
Grass cover in Camp A80%
StatusSTABLE
Proposed herder actionHerd stays in Camp A
Committee boardNo action needed
Left field: Camp A, where the herd grazes.
Right field: Camp B, which has been resting.
Brown and white blocks: cattle.
Person with a stick: the herder.
Short post with a glowing top: soil sensor.
Tall pole with a box: grass camera.
Phone on the stand: the grazing plan.
Spinning wireframe ball: the pressure model.
Dark board at the back: the committee, who decide.
Simplified for teaching. Real carrying capacity depends on rainfall, soil, grass type and the season. In real use the committee agrees before any herd moves.
Research and evidence
The need is real, and the science is still debated.
A 2021 review cites estimates that nearly 60% of South Africa's land is degraded, and that communal land makes up about 14% of the country.
From a People and Nature review of land degradation and justice. Other researchers argue that blaming communal tenure or overstocking alone is too simple, so Veldwise treats its numbers as guidance for discussion, not proof.
How each part of Veldwise helps
Probes, cameras and herd counts give the shared picture. The pressure model and pattern reader help separate dry spells from slow decline. The committee step keeps decisions with the people who live on the land, in line with participatory approaches already tried in South African communal areas.
A counterpoint: argues that common assumptions about overstocking and privatisation are an inappropriate basis for policy on rangeland commons.
Coverage
Rain decides how much grass there is.
Current temperature and rain over the last 30 days for six candidate communal areas. This is the only live data on the site. It comes from Open-Meteo and refreshes every 10 minutes while this page is open. If your browser or host blocks outside requests, the page says so instead of showing numbers.
Loading live weather…
Rainfall is gridded model data for each town, not a rain gauge on a camp. The "Prospective" stage is the company's own placeholder designation, not derived from the weather feed.
Pilot and practicalities
How we would work with your village.
The technology is the small part. What matters is agreement, language, access, and who looks after the equipment. This page sets out the proposed approach, and every point is open to change in discussion with each community.
Proposals, not contracts. Nothing here is binding until a community has agreed it in its own meeting. Highlighted terms are for the company to complete.
Whose data is it?
The community owns its camp data.Readings, photos and plans belong to the CPA or village, not to the company.
Counts are by camp, not by owner.No individual owner, herd or household is named, ranked or reported.
The committee decides who sees what.Extension officers, funders and researchers see only what the committee chooses to share.
The community can leave.If the pilot ends, the data is handed back or deleted at the community's request.
Personal information is handled under POPIA.The system is designed to hold very little of it. [Legal review to be completed]
Final termsWritten into the pilot agreement. [Data agreement reference and final terms]
In your language
Plain words first
Short sentences, no jargon, and pictures and colours alongside words. Status never relies on colour alone, and the text size buttons make everything larger.
Written with local speakers
Translations are written and tested with people who live in the area, including herders. Machine translation is a first draft only. The committee can correct terms such as camp and grass names.
Languages to confirm
[Languages to be confirmed with each pilot community, for example isiZulu, isiXhosa, Sesotho, Setswana, Sepedi, Xitsonga or Afrikaans]
Without a smartphone
Short SMS messages
Committee members and herder representatives can receive alerts of under 160 characters, and reply with a simple code to confirm a herd count or say a camp was checked. Reply codes are a proposal to be tested with herders.
Paper for the meeting
A one-page camp sheet and the draft plan can be printed in large type for the meeting and pinned up afterwards. A herder representative can update the notice each week.
Shared phones
One phone can serve a herder group, so nobody needs to own a smartphone to take part.
When the signal drops
Readings are saved at the camp and sent later. No reading is lost because the network was down.
Who is in the room
Veldwise works inside the CPA constitution and alongside traditional leadership. It does not replace either. The proposed steps to start in a village:
Ask first
Meet the traditional authority and the CPA executive, explain the idea, and be clear that they can say no.
Open meeting
Present to the whole community in their language. Invite questions from everyone, including women, youth, herders and people with few animals.
Agree the rules
Decide who makes decisions, what data is kept and who sees it, and what the tool will never do.
Install with permission
Choose sensor sites together with herders. Nothing is installed at a site the community has asked to leave alone, such as a sacred or sensitive place.
Meet on a schedule
The committee meets regularly with herder representatives. Disagreements go to the extension officer for advice, not to the system.
The right to pause or leave
The community can pause or end the pilot at any time and take its data with it.
What a pilot involves
1. Agreement
Meetings, rules and a signed pilot agreement.
[duration]
2. Set up and baseline
Install equipment, take first grass photos, and walk the camps with herders.
[duration]
3. A grazing season
Regular check-ins, committee decisions and plans through the season.
[duration]
4. Review
Did the grass change, and what did people think? Decide together whether to continue.
[duration]
Item
Who provides it
Who looks after it
Sensors and cameras
[to be agreed]
[to be agreed]
Solar power and field gateway
[to be agreed]
[to be agreed]
Phones, SMS credit or a shared device
[to be agreed]
[to be agreed]
Training for herder representatives
[to be agreed]
[to be agreed]
Data storage and backups
[to be agreed]
[to be agreed]
How we would judge success (proposed, to be agreed with each community): the grass cover trend in rested camps, how many agreed rotations were actually carried out, how committee members and herders rate the process, and how long it takes to produce a plan. Pilot fee or funding source: [to be confirmed].
What we would install
Device
What it measures
Power
Connection
Design notes
Soil moisture probe
Moisture in the root zone at a fixed depth
Small battery, optional mini solar panel
Short-range radio to the gateway
Buried, with cable protected from animals
Grass camera
Photos of the same patch at set times, for grass cover
Solar panel with battery
Short-range radio or cellular
Mounted high on a pole out of reach, fixed viewpoint, cattle-proof post
Herd count
Estimated number of animals in a camp
Herder's phone, or an optional tag reader at a gate or water point
App, SMS or radio
Recorded by camp, never by owner. Tags are optional
Gate sensor
Whether a gate is open or closed
Battery
Short-range radio
Simple magnet switch
Field gateway
Collects readings, runs simple checks at the camp, stores them if signal drops
Solar panel with battery
Cellular, with satellite as an option where there is no coverage
Weatherproof, locked and raised off the ground
Rural coverage is patchy, so every device is designed to keep working offline and send later. [Hardware models, suppliers and maintenance plan to be confirmed]
I started this company because too many decisions about land are made without the people who live on it. I want communities to have clear, honest information about their own camps, so that the people who graze the land are the ones who decide how it is cared for. If we get this right, the veld recovers and nobody loses their right to graze.
Veldwise
DEMO MODE — ALL TELEMETRY, PREDICTIONS AND AI OUTPUTS ON THIS SCREEN ARE SIMULATED
Sign in to the demo
Demo access only. No account is created or verified.