What is going wrong

Sand is extracted from the bed and banks of the Rwizi at scale, because Mbarara is building and construction sand has to come from somewhere. The extraction is often informal, frequently unlicensed, and in places has altered the course of the river outright.

The damage is not simply the removal of material. Taking sand from a channel lowers the bed, which steepens the banks either side and undercuts them. Vegetation that was stable slides in. The river then carries that soil downstream and drops it somewhere else, which is why you see fresh sand and silt bars in stretches nobody has mined.

It is also a livelihood. The people digging are not the people profiting most, and enforcement that targets the diggers while leaving the trade intact has a poor record here and elsewhere. We say this because conservation writing that treats sand miners as villains tends to produce policy that does not work.

Riverside soil is the best soil available: flat, deep, moist through the dry season. So it gets farmed, and in many stretches it is farmed right up to the water.

With nothing rooted at the edge there is no structure holding the bank together. Every heavy rain moves topsoil off the field and into the channel. The farmer loses their field gradually and the river silts up quickly. Both parties lose, which is what makes this the pressure most amenable to negotiation.

Statutory buffer zones exist. Enforcement across a catchment this size, on land that in many cases was cultivated before the rules arrived, is another matter. Our own work concentrates here because it is the pressure where a conversation with one household can produce a measurable change within a season.

Reporting and assessment on the Rwizi has documented untreated waste entering the river from dairies, soft drink manufacturers, hospitals, hotels, markets, a brewery, universities, a research institution, abattoirs and residential buildings.

A 2021 study of trace metals in the Rwizi ecosystem found heavy metals in both the abiotic and biotic components of the system — that is, in the water and sediment, and in the organisms living in them — and assessed the risk to human health. Fertiliser residue and insecticides have also been recorded.

This is a regulatory problem more than a community one. We monitor and we publish, and we say plainly that stopping industrial discharge is the job of the National Environment Management Authority and the Ministry of Water and Environment, not of a organisation our size. Claiming otherwise would be dishonest.

Eucalyptus grows fast, makes straight poles, and sells. For a household with a wet corner of land and school fees to pay, it is a rational choice, and for decades it was actively encouraged.

Near water it also draws the table down around it. Planted in or beside a wetland, a eucalyptus stand quietly does the opposite of what a wetland does.

We planted it ourselves in our first season, on a site where we should have known better. We stopped in 2023. We mention this on our own website because an organisation that only publishes its good decisions is not telling you anything useful.

Papyrus swamp is the catchment's filter and its sponge. It slows water down, drops sediment out, takes up nutrients, and releases what it holds gradually.

Drained for cultivation, brick clay or building, it stops doing all of that at once. Floods arrive faster and drought bites harder. Wetland is reported at roughly 3.5% of catchment land cover, forest at about 2.5%, open water at 1% — thin margins for a system carrying this much demand.

Wetland restoration has begun in places. Kakigaani in Rwampara district is one documented example, and bamboo planting in demarcated buffer zones is now established practice. It works. It is slower and less visible than the conversion that preceded it.