The season ahead is a decision problem.

From 25 to 28 August, the region's meteorological services convene in Swakopmund for the thirty-third Southern Africa Regional Climate Outlook Forum. What emerges from that room is the consensus seasonal outlook: rainfall probabilities by zone and sub-season, the closest this region has to an official statement about the summer ahead. Within days, it will be quoted in boardrooms, extension offices, credit committees and cabinet memos from Windhoek to Maputo.

And in almost every one of those rooms, the same thing will happen. The outlook will be read, nodded at, and filed, because a probability is not a decision. Even a forty per cent chance of below-normal rainfall in a given zone does not tell a dryland grain farmer whether to plant the full area or half of it. It does not tell an agro-dealer how much working capital to carry into October, a processor whose supply contracts are now exposed, or a lender what happens to a client's debt service if the dry tercile arrives. The outlook is the start of the season's hardest analytical work, and in most institutions, that work has no owner.

The stakes this season are not ordinary

The World Food Programme's current outlook puts the probability of a very strong El Niño at 81 per cent and projects 49 million additional people at risk of acute food insecurity across 45 vulnerable countries by the end of 2027; in Southern Africa alone, the numbers are expected to rise by nearly three-quarters. Its findings document something less noticed and, for anyone who works with models, more alarming: a shrinking monitoring base, what WFP calls a growing drought of food security data. When data thin out, the weight carried by every model that stands in for them grows, and so does the cost of a bad one.

Now put the money next to the risk. South Africa tracked an annual average of R188.3 billion in climate finance across 2022 and 2023, against an estimated need of up to R499 billion a year, according to the South African Climate Finance Landscape 2025. Of that tracked finance, 11.3 per cent went to adaptation, well below the African average of 33.7 per cent. Read those two numbers together, and the shape of the problem becomes clear. The problem is not only the volume of capital. Adaptation also lacks investable pipelines, credible benefit cases and, where commercial finance is expected, defensible cash-flow logic. Strip away the vocabulary, and the unresolved bottleneck is not another climate product; it is the decision-modelling work that must follow it.

The missing layer

Between the forecast and finance, a layer is missing, and it belongs to nobody. Extension services are usually organised around practices, while banks are organised around risk and repayment. The translation between them is rarely owned end to end: turning tercile probabilities into yield, price, cost and cash-flow scenarios for real enterprises; setting the decision dates and stop rules; pricing the impact of the downside tercile on working capital before it happens, rather than after. That translation is a model, built to a standard that a lender could audit. We have argued elsewhere in these notes that the model, not the report, is the deliverable that keeps working after the consultant leaves. A seasonal outlook is the purest test of that argument: a document that is all probabilities and no decisions, arriving at the exact moment every decision in the system comes due.

What this practice will do about it, in the open

Within seventy-two hours of the final official outlook becoming publicly available, Auctus Agri will publish the Southern Africa 2026/27 Climate-to-Cashflow Briefing, free, on this site. It will not contain a forecast from us; we do not make weather predictions, and we treat the official probabilities with the discipline they demand. What it will contain is the translation: exposure read through crop, livestock and irrigated systems; what the outlook means for input and planting decisions; the working-capital and debt-service arithmetic under each tercile; contract, credit and insurance consequences; and a dated decision calendar from September to February with named triggers drawn from official monitoring products, so that early action is a schedule rather than a slogan.

If you run an institution that has to turn this outlook into decisions, an extension service, a lender, a commodity organisation, a processor, the briefing is where to start, and we are reserving one founding-client slot for an institution that wants the full translation built around its own data, portfolio and decision calendar. The forecast is coming either way. The only question any of us controls is whether it lands on a decision framework or on a filing tray.