India Water Canvas · Methodology · v1.0 · 2026-05-08

How the Health × Availability × Impact scores are derived

This document is the canonical reference for how each state's H/A/I score is computed. The artefact's credibility rests on this being open, auditable, and contestable. Reviewers comment here. Anyone can fork the spreadsheet, propose new weightings, dispute the math, and replicate or contest the result.

v1.0 status: Methodology is published as author-judgment pending external review. Named methodology reviewers (Mihir Shah, Aditi Mukherji, Veena Srinivasan, Himanshu Thakkar) have been invited; sign-offs pending. Any reviewer dissent is recorded publicly in review-log.md. The dashboard's footer + caution flag this state explicitly.

1 · The three axes — definitions

The artefact treats Health, Availability, and Impact as conceptually distinct axes. Each is composed only of parameters belonging to its definition. This separation matters: a state can have stable Availability but collapsing Health (Bengaluru) or vice versa.

Health · the state of the water itself

Water quality parameters and ecosystem state. Not human-health consequences (those sit under Impact).

Geogenic contaminants — arsenic (As), fluoride (F), uranium (U), nitrate (NO₃), pesticide residues
Microbiological + organic load — BOD (biochemical oxygen demand), DO (dissolved oxygen), faecal coliform
Industrial pollution — heavy metals (Cr, Pb, Hg), dye effluent, pharmaceutical residues
Physical state — salinity (TDS, EC), turbidity, pH, temperature
Emerging contaminants — microplastics, PFAS
Ecosystem health — lake + wetland area lost since 1990 baseline, springshed perennial vs seasonal rate, river ecosystem health

Availability · the quantity + reliability of water reaching its intended use

Per-capita freshwater (m³/yr)
% blocks safe / over-exploited (CGWB GEC classification)
Surface storage stress (reservoir level vs design capacity)
Springshed perennial rate
Tank cascade functional rate
JJM tap functionality (does the infrastructure actually deliver?)
NRW (non-revenue water) loss in urban systems
Seasonal reliability + climate-stationarity of supply

Impact · the human, economic, and climate consequence

Population at risk in critical-classified zones
Climate vulnerability composite (cyclone + GLOF + sea-level rise + drought exposure)
Economic exposure (% state GDP at water-risk · agricultural output exposure)
Marginalised groups disproportionately affected — caste-disaggregated water access · women's water-fetching cost · urban informal · displaced communities · tribal commons
Sewage-borne disease load
Geogenic-exposure-driven health burden — As-induced cancer, F-induced fluorosis, U-induced kidney damage

2 · The formulas

Health score (0–100, lower = worse) — state of the water itself

Health = 100
  − geogenic_penalty · As + F + U + NO₃ + pesticides exposure
    weighted by % population on contaminated source water · max 25
  − microbiological_penalty · river BOD > limit · faecal coliform load
    weighted by % population dependent on basin · max 20
  − industrial_pollution · named cluster cities × discharge severity
    Cr / Pb / Hg / pharma / dye · max 15
  − salinity_penalty · % land area + coastal districts affected
    by salinity intrusion · max 15
  − ecosystem_loss_penalty · % wetland + lake area lost since 1990
    springshed perennial → seasonal rate · max 15
  − emerging_contam · microplastics + PFAS data presence /
    monitoring rate (proxy for unmonitored exposure) · max 10

Note: drinking-water-at-tap functionality is moved out of Health into Availability — functionality is delivery, not the water's intrinsic state.

Availability score (0–100, lower = worse) — quantity + delivery

Availability = % blocks safe (CGWB GEC, direct contribution) · max 35
  + per-capita freshwater (m³/yr / 5,177 baseline-1951 normalised) · max 25
  + surface storage stress (reservoir level vs design) · max 10
  + springshed perennial rate · max 10
  + tank cascade functional rate · max 5
  + JJM tap functionality (% functional, audited not claimed) · max 10
  + climate-stationarity of supply (rainfall variability low?) · max 5

Impact score (0–100, lower = worse) — human + economic + climate consequence

Impact = 100
  − population_at_risk · % pop in critical-classified zones · max 25
  − climate_vulnerability · cyclone + GLOF + SLR + drought composite · max 20
  − economic_exposure · % state GDP at water-risk · agri output · max 15
  − marginalised_affected · caste-disaggregated access ·
    gender · urban informal · tribal · displaced · max 20
  − health_burden · As cancer · F fluorosis · U kidney ·
    cholera/typhoid/JE/dengue load · max 10
  − subsistence_burden · women's water-fetching hours/day ·
    rural drinking-water tariff burden · max 10

3 · Open spreadsheet

The full computation per state is open as a CSV. Anyone can fork it, propose new weightings, submit changes via PR.

h-a-i-derivation.csv — per-state H/A/I scores with input parameters and source links
funder-ladder.csv — 22+ funders × annual ₹ × layer
intersections.csv — energy/waste/markets/health KPIs per state
place-kpis.csv — state-specific intersection numbers
bottlenecks.csv — 8 bottlenecks × severity × layer × loop
unit-econ.csv — 9 unit-economics assets × cost × axis
sheds.csv — 14 SHEDS × signals × needs × build phase
hidden-dynamics.csv — 8 dynamics × severity × axis affected

4 · Source-data trust grades

Per-input-parameter trust grade per data-and-groundtruth.md:

Parameter	Source	Trust	Lag
% blocks over-exploited	CGWB GEC 2023	silver (direction trustworthy)	12-18 mo
Per-capita freshwater	CWC + Census denominators	silver	1-3 yr
River BOD	CPCB monitoring stations	bronze (lab quality varies)	30-90 days
JJM functionality	CAG audit + field surveys	red (definition slippage)	2-3 yr post-audit
Geogenic contamination (As/F/U)	Academic studies + state PHED	silver-bronze (state-specific)	per study
Salinity intrusion	INCOIS + ground research	bronze (sparse coverage)	per study
Glacier mass loss	GSI + ICIMOD	gold (satellite-confirmed)	annual
Cyclone vulnerability	IMD + INCOIS	gold	real-time
OCEMS industrial discharge	State PCBs (legally restricted)	red (politically captured)	real-time within agency

5 · Limitations

Author judgment: weightings within each axis are author-judgment-pending-review. Reviewer feedback may shift these substantially.
Aggregation: state-level scoring papers over basin-level + district-level + ward-level variation. A state-level Health of 38 hides ward-level scores ranging from 18 to 70.
Confidence interval: ±15 points per axis-state combination is the author's honest uncertainty band.
Source heterogeneity: some parameters draw from gold-trust sources (IMD, GSI), others from red-trust sources (OCEMS, JJM functionality). The composite mixes these.
Temporal mismatch: parameters update on different cycles (CGWB GEC 2-3 yr, CPCB monthly, IMD daily). The H/A/I composite represents a recent snapshot, not a single calendar year.
Missing data: some states have sparse parameter coverage; the formula falls back to category-mean defaults, marked in the spreadsheet.

6 · Reviewer feedback log

Reviewer feedback received and incorporated. See review-log.md for the full review history.

v1.0: invitations open, no sign-offs received yet. Methodology is published with this status flagged on the dashboard.

7 · How to contribute / dispute

Open an issue at parent repository, title: [Methodology] [your concern]
State which axis (Health / Availability / Impact), which parameter, what change you propose
If proposing a new weighting, attach computation showing the difference in 1-2 example states
Author response within 14 days; recorded in review-log.md
If accepted, change incorporated in next minor version (v1.x → v1.x+1) with credit + erratum logged