Skip to main content
lab-notes.aiPublic-service AI playbooks
Opportunity 17/17InfrastructureEnvironment, land and water

Water Resource Management

Explore how published rainfall, river-level, and abstraction figures could be compared, and where their different scales limit a water-management question.

Inspect starter dataSet up the build partner
Published sources
3
Starter data
20 synthetic records
Domain build partner
build-water-management
Last reviewed

From the draft strategy

Opportunity

The draft strategy names AI for water management as a potential public-service application, covering flooding, pollution control, and how water is shared between users, without saying which of those decisions it means or how far ahead it is meant to look.

Where the draft says it

Table 2 — potential public-service applications

Read the draft strategy source (opens in a new tab)

Source register

Research already done

Real published source

We reviewed these published sources so you do not have to start from zero. Each source shows what it covers, how it can be accessed, and why it matters.

3 sources

  1. Department for Infrastructure

    Open

    DfI Rivers water level network

    What it covers
    A network of around 130 hydrometric stations recording water levels in Northern Ireland's rivers and loughs, shown on a map viewer that also lets the station records be downloaded.
    Why it matters here
    It is the published level series this playbook's synthetic records stand in for, and it shows what is missing as clearly as what is there: levels are recorded here, rainfall is published by somebody else.

    Open the source (opens in a new tab)

  2. Met Office

    Open

    UK and regional climate series

    What it covers
    Monthly, seasonal, and annual rainfall, temperature, and sunshine series for the UK and its regions, with the Northern Ireland rainfall series running from 1931 and downloadable as plain tables.
    Why it matters here
    It is where the rainfall side of this example would come from, at a regional monthly grain far coarser than any catchment decision needs, which is the gap the synthetic file makes visible.

    Open the source (opens in a new tab)

  3. Department of Agriculture, Environment and Rural Affairs

    Open

    Abstraction and impoundment licensing requirements

    What it covers
    When taking water from a river, lough, or borehole has to be notified or licensed in Northern Ireland, the daily volume thresholds that trigger each step, and the sectors that abstract the most.
    Why it matters here
    Its daily volume thresholds are the bands the synthetic records use, so the competing-demand part of this example can be discussed in the units the rules already work in.

    Open the source (opens in a new tab)

Starter data

Starter dataset

Synthetic working data

Use 20 synthetic monthly rows across four catchments to explore rainfall, river-level, and abstraction questions together.

AI-assisted research helped identify and interpret the published sources. We then created a small, non-sensitive synthetic dataset shaped by the information those sources expose. It is for exploration—not evidence, training, or operational decisions.

Records
20
Fields
6

Field preview

table view
  • id

    SYN-WM-01 · SYN-WM-02 · SYN-WM-03

  • catchment

    Catchment A · Catchment B · Catchment C

  • month

    2025-03 · 2025-04 · 2025-05

  • rainfallMm

    96 · 54 · 38

  • riverLevelBand

    above normal · normal · below normal

  • abstractionBand

    20-100 m3/day · over 100 m3/day · 10-20 m3/day

Inspect all 20 recordsView repository file (opens in a new tab)Download JSON (opens in a new tab)

How it was prepared

AI authored fictional rainfall totals, river-level bands, and abstraction bands based on published licensing thresholds.

Limitations

  • The catchments are letters and the figures are invented, so nothing here describes rainfall, a river, or an abstraction anywhere in Northern Ireland.
  • Monthly rows cannot show a flood, which happens in hours, and a level band against normal is not a flow.
  • Flooding, pollution, and sharing water out are three different decisions, and one small table appearing to serve all three is part of what this example is meant to expose.

Clone and build

Domain build partner

Build with a domain-aware coding agent

Clone the repository, then ask your coding agent to load the skill that ships with it. The skill brings the opportunity, sources, starter data, known unknowns and constraints into the conversation before anything is proposed. It follows the open Agent Skills standard, so it works in Claude Code, Codex and any other agent that reads skills.

01 / Clone

git clone https://github.com/Hypership-Software/lab-notes.ai.git

02 / Invoke

Claude Code
/build-water-management
Codex
$build-water-management

Conditions, not a checklist

Before you build

These are not footnotes. They are conditions that any responsible prototype would need to address.

  • Signal errors distribute costs unevenly

    A missed or false signal here reaches emergency response, pollution control, and whether there is water to go round, and those costs do not fall on the same people.

  • Asset details can be sensitive

    Detail about assets and networks can be sensitive in itself, which is why the catchments in this file are letters.

  • Ordinary months can hide sensor failure

    A sensor that has failed or drifted can hide the event that matters most, and anything judged on average performance is being judged mostly on ordinary months.