This is a small dataset from a small programme, described honestly. It is not a national monitoring network and we will not present it as one.
What we can share
| Dataset | Coverage | Format |
|---|---|---|
| Water quality readings | 6 points, quarterly, 2023 to present | CSV |
| Planting and survival records | All sites, 2022 to present | CSV |
| Fixed-point photography | 6 points, quarterly, with coordinates | JPEG |
| Site boundaries | Active sites | GeoJSON |
| Buffer agreements | Anonymised only | CSV |
What we ask in return
- Tell us what you find, including if it contradicts something we have published.
- Cite the data so others can trace it.
- Do not attempt to re-identify individuals from the anonymised agreement records.
- If you publish, send us a copy. We will link to it from the research library even if it is critical of us.
Students
We take three to six month research placements, typically from environmental science, hydrology or development studies programmes. A stipend covering transport and lunch is provided.
Undergraduate and masters students are welcome to use our data for dissertations without a placement. Just ask. There is no form.
What is in the data, and what is wrong with it
The water quality series covers six points sampled quarterly since 2023, one of which is a control on an unrestored stretch. Quarterly sampling is too coarse to separate a restoration effect from ordinary seasonal variation, and with a series this short the confidence intervals on any trend are wide. We publish it anyway, with that stated, because a small honest series is more useful than none.
The planting records run from 2022 and carry the count made at the site on the day, not seedlings dispatched. The 2022 cohort is the weakest part of the set: there was no monitoring protocol worth the name in that first year and the six-month returns from it are incomplete. Sites that failed remain in the records and in the denominator of every survival figure we publish.
The fixed-point photography is the most straightforwardly useful thing we hold — the same frame, the same coordinates, quarterly. It answers questions about bank position and vegetation cover that the numeric series cannot.
Buffer agreement records are shared only in anonymised form. They describe household-level land use decisions, and the households agreed to a restoration project rather than to publication.
Questions we would like somebody else to answer
We are not well placed to analyse our own data and we know which questions we have been ducking.
- Does the difference between the restored points and the control survive proper statistical treatment, or are we reading noise?
- What actually predicts survival at twelve months: species, bank aspect, distance from a household, or the quality of the negotiation?
- Is bamboo at the waterline holding the bank as well as we assume, or is the assumption doing the work?
- What happens to a buffer strip after the livelihood support ends?
- How much of the turbidity at any point originates upstream of anything we have touched?
Practical questions
What licence does the data come under?
Our site text and data are published under CC BY 4.0, which permits reuse including commercially, with attribution. Photographs by other people carry their own licences and those travel with the image.
How long does a request take?
Usually within a week. Tell us what you need and what you intend to do with it, in a sentence or two.
Can I use this for a dissertation?
Yes, without a placement and without a form. Just ask. Send us the finished piece and we will link to it from the research library even if it is critical of us.
Will you co-author?
Only where we have genuinely contributed to the analysis. Adding our name to a paper we did not do the work on would misrepresent both of us.
Can I sample at your monitoring points?
Yes, and we would welcome an independent set of readings alongside ours. Coordinates come with the fixed-point photography.
Request data
Tell us what you need and what you are doing with it. Usually turned around within a week.