Most of the damage to the Rwizi is at the water's edge, in a strip a few metres wide. That is also where the cheapest and most effective repair is available.
The problem, precisely
Where cultivation runs to the waterline there is nothing rooted holding the bank. Rain moves topsoil off the field into the channel; the channel silts; the field shrinks. Both the farmer and the river lose, which is unusual among environmental problems and is the reason this one is negotiable.
Restoring the strip means agreeing that a few metres come out of production, planting them with species that root deep and tolerate seasonal flooding, and protecting them long enough to establish. Three years is roughly the point at which a buffer holds without help.
What we plant
On the bank itself: bamboo, which roots aggressively and binds soil faster than anything else available, and indigenous species including Markhamia lutea and Ficus natalensis. Where papyrus has been cut but the hydrology is intact, often the best intervention is simply to stop cutting and let it return.
Behind the buffer, on the household's own land, we supply fruit and timber species so the family gets something back for the ground they have given up. This is not generosity; it is what makes the agreement stick.
We do not plant eucalyptus anywhere near water. We did in our first season, on a site where we should have known better, and we stopped in 2023.
- Bank replanted
- 6.4 kmMeasured: Surviving at twelve months. Field count, June 2026.
- Twelve-month survival
- 61%Measured: Average across all sites. We replace the losses.
- To establishment
- 3 yearsIndicative: Roughly when a buffer holds without further intervention.
- Measured: Measuredcounted by us, in the field
- Cited: Citedpublished elsewhere, cited here
- Indicative: Indicativean approximation, not a measurement
61% is not a good number and we publish it anyway
Nearly two in five seedlings do not reach a year. Dry-season losses, livestock, and sites we misjudged all contribute. Quoting planting totals without this figure would make our work look roughly 60% more effective than it is.
Wetland work
Papyrus recovers on its own if the hydrology is intact and the cutting stops. Much of our wetland work is therefore negotiation rather than planting: agreeing a rest period with the households using a stretch, and finding them something else for that period.
Where drainage channels have been cut we sometimes block them, which is cheap and immediately effective. Where the wetland has been converted to permanent cultivation, honest assessment usually says recovery is a decade-scale proposition and our effort is better spent on the margin that has not gone yet.
Thirty metres of buffer costs about 350,000 shillings
Negotiated, planted, and monitored for a full year including replacement of losses. That figure is what it actually costs us, not the cost of the seedlings alone.