Skip to main content
River Rwizi Charity Conservation
Donate
Wide view of green vegetation surrounding the River Rwizi in Mbarara

Protecting the water six million people depend on.

We work alongside the households, schools and districts of the catchment to restore the river’s banks and protect the wetlands that feed it.

Support the workWhat we have achieved

Mbarara City · Photograph BalukuBrian, CC BY-SA 4.0

Field-based conservation · Founded 2021

Based at Biharwe, Mbarara. Working along the River Rwizi.

10
catchment districts
6
monitoring points
6.4 km
bank replanted

Counted in the field, not estimated. Every figure on this site carries the date it was taken and the method behind it.

How we measure

What have we achieved so far?

Numbers, with the caveats attached.

Every figure below is a field count on a marked transect. Tap one to see how it was measured and when.

6.4 kmbank replanted

Cumulative length of riverbank with buffer planting established and surviving at twelve months.

Method
Physical count on marked transects
Measured
June 2026
11,200seedlings planted

Total planted since 2022. Survival at twelve months averages 61%, which is why we replace.

Method
Physical count on marked transects
Measured
June 2026
1,840people trained

Attendance at community and school sessions. Counts people, not sessions.

Method
Physical count on marked transects
Measured
June 2026
6monitoring points

Fixed sampling locations, visited quarterly for water quality and bank condition.

Method
Physical count on marked transects
Measured
June 2026
See all results

How much bank is standing?

Six point four kilometres, one stretch at a time.

  • 20220.4 km
  • 20231.5 km
  • 20243.4 km
  • 20255.2 km
  • 20266.4 km
Cumulative kilometres of riverbank with buffer planting established and still surviving at twelve months. Field counts on marked transects, to June 2026.

Which seasons survived?

Planting is the number everyone quotes. This is the one that counts.

Every cohort we have planted since 2022, counted at twelve months. The failures stay on the chart. A season we lost entirely tells a reader more than the ones that went well.

  • 2022 wet60%

    720 of 1,200 · First cohort, Biharwe upper. Planted in numbers we could report, before follow-up was costed in.

  • 2023 dry0%

    0 of 900 · Nyakayojo — land use not agreed. Total loss.

  • 2023 wet58%

    1,392 of 2,400 · Planted before the rains settled

  • 2024 wet70%

    2,170 of 3,100 · Onset confirmed on the ground first

  • 2025 wet76%

    1,824 of 2,400 · Species mix revised for drought

  • 2026 wet79%at 6 months

    948 of 1,200 · Six-month figure. The twelve-month count is not due until the next dry season.

61% average across every cohort that has completed a full year. The newest cohort is a six-month count and is excluded until its year is up.
Why one season shows zero

The 2023 dry-season cohort was 900 seedlings at Nyakayojo, planted after agreeing with the landowner but without speaking to the household actually farming the land. We lost the lot within a season. It is the reason stage two of the method exists, and it stays on this chart rather than being quietly dropped from the average.

What we changed after it

What is the problem, and what are we doing about it?

Soil is going into the river. We are putting roots back into the bank.

Four sentences, in the order they matter. If you read nothing else on this page, this is the organisation.

  1. 01The problem

    Nothing is holding the bank, so the bank goes into the river

    The soil beside a river is the best soil available, so it gets farmed to the waterline. With nothing rooted at the edge, every heavy rain moves topsoil off the field and into the channel. Sand extraction lowers the bed at the same time, steepening what is left.

    ~6 millionpeople depend on the Rwizi system

  2. 02The goal

    A rooted strip along the bank, kept by the households who farm beside it

    Not a fence and not a reserve. Bamboo, Markhamia and indigenous species between the crop and the water, agreed with the person actually farming the land and worth more to them standing than cleared.

    3 mthe usual buffer width we negotiate

  3. 03The method

    Survey it, agree it, plant it in the rains, then go back and count

    Five stages, published in full. The one that decides everything is the last: returning at six and twelve months, counting what survived, and publishing the rate whether or not it flatters us.

    61%alive at twelve months since 2022

  4. 04How far we have got

    6.4 kilometres of bank, on a river that runs 90

    6.4 km standing90 km to Lake Victoria
    We have not turned that into a percentage, because a percentage needs a denominator: how much of this bank actually needs replanting. Nobody has surveyed that, ourselves included.

And four things this work cannot fix

  • What the city discharges into the river below Mbarara
  • Sand extraction, which is licensed and policed nationally
  • Deforestation in the Buhweju headwaters, where we hold no sites
  • Rainfall that arrives later and harder each year
The pressures in detail

Who are we, and where do we work?

This is not just a line on a map.

The Rwizi rises in the Buhweju hills and carries water through farms, wetlands and towns across south-western Uganda before reaching Lake Victoria. About six million people draw on the system for water, farming or grazing.

  1. Buhweju hills0 kmWhere it rises · 1,800 m
  2. Rwentondo22 kmBuffer planting · 1,469 m
  3. Mbarara City28 kmAbout 200,000 people · 1,420 m
  4. Lake Mburo45 kmFirst of three lakes · 1,370 m
  5. Lake Victoria90 kmOutflow · 1,134 m

What happens upstream does not stay upstream.

Soil stripped from a hillside, waste from a town and a drained wetland all travel. The catchment, not the district, is the unit that matters.

Understand the Rwizi

What is damaging the river?

The river is changing for ordinary reasons.

Not one dramatic event. Pressure builds through many small decisions along the catchment, most of them made for good reasons at the time. That is what makes it fixable, and what makes it hard.

Sand being extracted from a river channel, with excavated banks visible
Onitsha, Nigeria · Sand extraction from a river channel. Lowering the bed steepens the banks either side and undercuts vegetation that was stable. Photographed on the River Niger, not the Rwizi. Ngostary2k, CC0

We hold no publishable photograph of extraction on the Rwizi. This one shows the same practice on another African river, and we would rather say so.

01

Sand mining moves the channel

Extraction from the bed and banks does not just remove sand. It changes where the river drops its load, undercuts banks that were stable, and in places has shifted the course outright. It is also a livelihood for people with few alternatives, which is why enforcement alone has never worked.

Source: Ministry of Water and Environment; district reports

02Cultivation runs to the waterline

The soil beside a river is the best soil available, so it gets farmed. With nothing rooted at the edge there is no structure holding the bank, and every heavy rain moves topsoil into the channel. The farmer loses the field slowly and the river silts up quickly.

Source: Our own bank surveys

03Untreated waste goes straight in

Discharge into the Rwizi has been documented from dairies, abattoirs, hospitals, hotels, markets, a brewery and residential buildings. Studies of the system have found heavy metals, fertiliser residue and insecticides in the water and in the organisms living in it.

Source: Peer-reviewed trace-metal study of the Rwizi ecosystem, 2021

04Eucalyptus drinks the wetland dry

Planted for poles and firewood because it grows fast and sells, eucalyptus near water draws down the table around it. We planted it ourselves in our first season. We stopped in 2023 and we would not do it again.

Source: Internal decision record

05The wetlands are being converted

Papyrus swamp is the catchment's filter and its sponge. Drained for cultivation and building, it stops doing both jobs at once — so floods arrive faster and droughts bite harder. Reported figures put wetland at around 3.5% of the catchment's land cover.

Source: Catchment management assessments

The full picture

Where does the river run?

Follow the Rwizi from its headwaters to Lake Victoria.

Three of the ten districts in the catchment carry programmes of ours. The rest are part of the same watershed and carry none, and we draw that difference rather than shading the whole basin one colour.

BuhwejuMbararaLake MburoLake Victoria
  • Sites we work on
  • Places on the course
~90 km
source to lake
10
catchment districts
6
monitoring points

Drawn from published coordinates. Site positions are marked approximate where GPS-surveyed coordinates are not available, and the course is traced at catchment scale.

Open the full river explorer

Where along the river does the water get worse?

Clean off the hills. Four times dirtier crossing the city.

Every other chart here asks whether the river is improving over time. This one asks where, and the answer names a stretch our work does not reach.

  1. RW-0588NTU

    Upper catchment · control · Rwampara

  2. RW-0462NTU

    Upper catchment · restored · Rwampara

  3. RW-0134NTU

    Above the city · Mbarara

  4. RW-02131NTU

    In the city · Mbarara

  5. RW-03174NTU

    Below the discharge · Mbarara

  6. RW-0651NTU

    Below the lakes · Isingiro

Read top to bottom, the river runs downstream. It arrives at Mbarara at 34 and leaves the discharge reach at 174; the lakes and wetlands below bring it back to 51. The one comparison that is ours is the first pair — 62 on a restored reach against 88 on the same reach left alone.

Our own quarterly sampling, 30 June 2026. Structured sample data until the field records are loaded. Points are shown in downstream order rather than at surveyed kilometre marks.

Open the catchment map

Which districts does it cross?

Ten districts. Programmes in three.

The rest are part of the same watershed and carry none of our work. We separate the two because a map that shades every district the same colour implies work we are not doing.

  • Buhweju
  • Bushenyi
  • Sheema
  • Ntungamo
  • Mbarara
  • Rwampara
  • Isingiro
  • Kiruhura
  • Lyantonde
  • Rakai and Kyotera

Filled: active programmes (Mbarara, Rwampara, Isingiro). Outlined: catchment only.

What do we do, day to day?

Six programmes. One river system.

Different pressures need different work, run through district structures rather than around them.

Close view of riverside plants growing along the edge of the River Rwizi

River and wetland restoration

Replanting buffer strips and letting papyrus come back where it was cut out.

6.4 kmbank replanted · field count, June 2026
Explore restoration

How do we restore a riverbank?

Planting is one day. Restoration takes longer.

  1. 01Walk the stretch first.
  2. 02Find out who owns and who uses the land.
  3. 03Agree the buffer together.
  4. 04Plant for the site, not for the photograph.
  5. 05Come back.

6 months.12 months.

That is when we go back and count what survived. A seedling in the ground is not a restored riverbank, and the difference between the two is the only thing worth reporting.

5 stages, each one recorded. The full protocol names what gets written down at every step.

Read the methodology

What do we plant, and why?

Species chosen for the site, not the photograph.

Bank species where the water rises, farm species behind the buffer, and fruit on the household’s own ground so the exchange is visible from their doorway.

BambooRiverbank, front line
Roots aggressively. Binds soil faster than anything else available here.
Markhamia luteaBank and farm boundary
Indigenous, deep-rooted, useful timber, tolerates seasonal wet.
Ficus natalensisBank and homestead
Indigenous, culturally valued, an excellent soil stabiliser.
Fruit speciesHousehold plots behind the buffer
A direct return to the family giving up ground.
EucalyptusNowhere near waterDiscontinued 2023
Draws the water table down. We planted it in our first season, stopped in 2023, and will not resume.
The agroforestry programme

Why does the work take a year?

Everything here is timed against the rains.

Two wet seasons, two dry ones. Planting only works inside the wet windows, so a season missed is four months lost — which is most of what looks slow about this work from outside.

JFMAMJJASOND
  • Historic average
  • Recent years
Millimetres per month, averaged across the catchment. Indicative pattern, not a single station series.
1068 mm
historic annual total
1033 mm
recent years

The total has barely moved. The distribution has: the wet months are wetter, the dry months drier, and the gap between them longer. An average that holds steady while the pattern underneath it breaks is exactly the kind of figure that hides a problem — and it is why we now confirm the onset of rains on the ground before anything leaves the nursery.

How a shifting rainfall pattern changes the work

What does one site take, start to finish?

It took three months to plant four hundred metres

Close view of riverside plants growing along the edge of the River Rwizi
Mbarara City · Bank plants doing the job we are trying to restore elsewhere: holding soil in place. NUWAHWERA PHIONAH, CC BY-SA 4.0

The difficult part was not the planting. It was walking the bend with a notebook, working out who farmed which piece, and asking eleven households whether three metres of their best ground could come out of production.

Two agreed on the spot. Six wanted to know what they were getting back. Three said no — and one of them was right to. Then the part that decides everything: coming back at six months, counting, and replacing the fifth that had died.

The land near the water is the best land we have. Nobody leaves it idle because a stranger asks. But if you show me what to plant behind it, that is a different conversation.

Said to a field officer at Rwentondo, recorded with permission

11
households on the bend
400 m
planted across two days
63%
alive at twelve months

Site record, Rwentondo · June 2026

Read the Rwentondo field note

How do we know the work lasted?

We return to the same places and measure again.

Six fixed points, sampled quarterly since 2023. One of them is a stretch we have deliberately never restored. Without it, an improvement at a planted site could just be a wetter year.

Turbidity at two monitoring pointsNTU, lower is better
0591192023 Q12026 Q2
  • RW-04 · downstream of buffer planting
  • RW-05 · control, never restored
How to read this chart

Turbidity is how much soil is suspended in the water, in nephelometric units. Lower is better. RW-04 sits downstream of our earliest buffer planting; RW-05 is an unrestored stretch, so the gap between the two lines is the part that is plausibly ours. Two of the six points sit below the city’s discharge and got worse — bank planting cannot influence that, and we publish it anyway.

Downstream of our earliest buffer planting

Latest sampling
30 June 2026
Locations
6 fixed monitoring points
Explore the monitoring data

How did the organisation get here?

What changed, and what changed our minds.

Six moments from the record. Two of them are things we got wrong, which is roughly the proportion an honest list has.

  1. 2021

    Registered and based at Biharwe

    On the eastern edge of Mbarara City, about a kilometre from a stretch of the Rwizi we have been replanting ever since.

  2. 2022

    First cohort in the ground

    1,200 seedlings at Biharwe upper. Planted in numbers we could report, before anyone had costed the visits that would tell us whether they lived.

  3. 2023Got it wrong

    Nyakayojo: 900 seedlings, all of them lost

    Agreed with the landowner, never discussed with the household actually farming the land. Gone within a season. It is why finding out who uses the land is now stage two of the method rather than an assumption.

  4. 2023Got it wrong

    Stopped planting eucalyptus near water

    It grows fast, it sells, and every nursery in the district had it. It also draws the water table down, which is precisely what we exist to stop. The stands already in are being left to mature rather than felled.

  5. 2023

    Six fixed monitoring points, sampled quarterly

    One of them a stretch deliberately never restored, so an improvement at a planted site can be told apart from a wetter year.

  6. 2026

    6.4 km standing, and a survival rate we publish

    11,200 seedlings planted since 2022, 61% alive at twelve months, 1,840 people trained. Field counts on marked transects, June 2026.

More about the organisation

What have we got wrong?

Three things we would do differently.

Conservation work does not become more credible by pretending every decision worked. Any organisation four years in has a list like this; most do not publish it.

  1. 012022–23

    We planted eucalyptus near the water

    It grows fast, it sells, and every nursery in the district had it. It also draws down the water table around it, which is precisely the thing we exist to stop.

    What changed. We stopped planting it near water in 2023 and it is not coming back. The stands we put in are being left to mature rather than felled, because pulling them now would open the bank we were trying to hold.

  2. 022023

    We counted seedlings planted, not seedlings alive

    We planted in numbers we could report and did not budget for the follow-up. Two-fifths of the first seedlings were dead within a year, mostly from planting too late in the season and from goats.

    What changed. Six- and twelve-month counts are now costed into every project before it starts, and we publish the survival rate beside the planting total instead of the planting total alone.

  3. 032025

    Turbidity below the city got worse, not better

    Our monitoring point below Mbarara read higher last dry season than the year before, despite the work upstream. Bank planting does not offset what a city discharges.

    What changed. We publish it, and we stopped implying that riverbank restoration alone fixes water quality. Effluent is a regulatory matter and we say so rather than taking credit we have not earned.

See what changed

How can any of this be verified?

You should not have to take our word for it.

The method, the monitoring records, the governing documents and the accounts are published for anyone who wants to check the work behind the claims.

NGO Bureau and URSB registration numbers are not yet published — they go up on the governance page as each certificate is issued. We leave the space empty rather than filling it with a claim.

How to check us

What happens next?

Five things due, and what each one is waiting on.

Nothing here is a promise we have not already made elsewhere on this site, and nothing carries a date we cannot keep.

  1. 01 · Next dry seasonCommitted

    The 2026 cohort reaches its twelve-month count

    1,200 seedlings planted last wet season are currently a six-month figure, which is why they sit outside the published average. The twelve-month count goes on the site whichever way it falls.

  2. 02 · As each is issuedWaiting on others

    Registration numbers go on the governance page

    The NGO Bureau and URSB certificates are with the registrars. Each number is published the day it arrives.

  3. 03 · First audited yearNeeds funding

    An external auditor replaces our own expenditure figures

    Every spending figure here is our own working split and is labelled indicative. Audited figures replace them, and the superseded numbers stay in the archive.

  4. 04 · Once field records loadCommitted

    Real monitoring records replace the sample series

    The quarterly series behind the charts is structured sample data and says so on every chart. The field notebooks are being entered now.

  5. 05 · Not yet scheduledNeeds funding

    Monitoring extended into the Buhweju headwaters

    What happens at 1,800 metres decides how much water is still moving ninety kilometres downstream in August, and we hold no sites up there. We will not claim a date until somebody funds the travel.

Committed means budgeted. Needs funding means we know how and cannot yet pay. Waiting on others means the decision is not ours.

Plans and reporting in full

Can I have the underlying data?

Everything behind the numbers, in a form you can open.

Not a summary of the evidence. The evidence itself. Each file is generated from the same records the charts above draw from, so a figure in a spreadsheet cannot quietly differ from the figure printed beside it.

Published under CC BY 4.0 — use them, including to contradict us, with attribution. Last built 30 June 2026.

Request the raw records

Who actually does the work?

Conservation here is mostly conversation.

Field officers who walk the bank and write down what they find. Farmers who agree to take three metres out of production. Teachers who run the school sessions themselves. District officers who know the land tenure. Nursery groups who raise what gets planted.

A woman leading a practical demonstration to a group of community members
Uganda · A community trainer demonstrating a technique. Sessions run early morning or late afternoon, around farming hours. Vmusika, CC BY-SA 4.0

We do not publish staff photographs or names without asking first, in the person’s own language. The team page carries the profiles as consent is recorded.

Meet the team

What have we written down lately?

Reports from the field.

We have not published a field note yet. Rather than filling this space with headlines nobody wrote, here is what does exist and carries a date.

What do people ask us most?

Answered the way we would answer them on the telephone.

How do I get involved?

Four ways, and none of them needs a long commitment. You can give once or monthly; you can join a planting day or a bank survey, most of which take a morning; schools in the catchment districts can host a session run with their own teachers; and districts, institutions and companies with land near the water can work with us on a buffer. Each has its own page under Get involved, with what is required and who to contact.

What has the organisation achieved so far?

Since 2021: 6.4 kilometres of riverbank replanted and still standing at twelve months, 11,200 seedlings planted, 1,840 people trained through school and community sessions, and six fixed points sampled quarterly since 2023. Every one of those is our own field count on a marked transect, dated June 2026, and the method behind them is published in full.

Where does my donation go?

Ten dollars , about 38,000 shillings, raises roughly a dozen seedlings at a nursery run by a local group, covers the transport to a stretch we have already surveyed, and pays for the two return visits at six and twelve months. On our current working figures 78% of income goes to programmes, 14% to staff and field costs and 8% to administration. Those proportions are indicative until the first external audit, and they are labelled that way wherever they appear.

How do you decide where to work?

By where the bank is failing and where a household will agree to a buffer. We survey the stretch on foot first, establish who owns the land and who farms it, rarely the same person, and only then negotiate a width and a term. A site nobody has agreed to is a site that will be cleared within a season, which is why the agreement comes before the seedlings.

Ask us a question
A man watering and tending rows of young tree seedlings in a plant nursery

How can I help?

Help us keep coming back.

Planting is the cheap part. Restoration also needs the transport, the follow-up visits, the replacement planting and the time spent with the households using the bank.

A seedling costs about UGX 3,000 to raise, plant and check twice.

An approximate working cost to June 2026, including both follow-up visits. It is not a guaranteed allocation of any particular gift.

DonateBecome a member
See where the money goes

Questions, press or research requests reach a person, not a queue. info@riverrwizicharityconservation.org · +256 781 013 376