How the IVH v4.1 is built and checked.
The Water Vulnerability Index (IVH v4.1) condenses five territorial dimensions into a single, transparent score for every municipality. Here is how it is weighted, what has been checked (it has not been validated), and which open data feeds it.
Five components, one score.
The IVH v4.1 is a weighted sum of five normalised components, each on a 0–100 scale. A higher score means higher water vulnerability. The weights are an expert choice by the project; they have not been estimated from observed water outcomes.
IVH v4.1 = 0.30·Climate/Drought + 0.30·SafeWater + 0.20·RecordedEvents + 0.15·SocialVulnerability + 0.05·WaterStorage
Drought & climate
SPI-12 (50 %), SPEI-12 (35 %) and SPI-48 (15 %) of the latest month with data. SPI from CHIRPS with a gamma fit per calendar month; SPEI from CHIRPS minus FAO-56 Penman-Monteith reference evapotranspiration computed from ERA5-Land, with a 3-parameter log-logistic fitted by L-moments; both on the 1981–2010 normal. Each index is scored on the WMO drought categories (−2 → 100, 0 → 50, +2 → 0). Computed at department level. Since IVH v4.1 (October 2026), after an independent technical audit.
Safe water access (municipal)
Share of households without piped or improved water, from the 2024 INE Census at municipal resolution. The only component computed per municipality.
Recorded events
Water-related events recorded in DesInventar (1970–2015: floods, flash floods, overflows, mudflows, rain, hail, landslides and droughts) per 100,000 inhabitants (2024 Census), per department, with double weight for events from 2000 onward.
Social vulnerability
Unmet basic needs (NBI) from the 2024 Census, as a proxy for the population's capacity to cope with water stress.
Water storage
GRACE-FO terrestrial water storage anomaly (TWSA, 60 %) and JRC surface-water occurrence (40 %, excluding Lake Titicaca from La Paz), describing the physical state of stored water. GRACE measures total storage, not aquifers.
Compared with past events.
The index was compared with the historical record of events. The correlation is weak and not statistically significant, and the comparison is circular in part, because event records make up 20 % of the index. (This page used to show ρ ≈ +0.87; that figure came from a calculation error that counted each department's events once per municipality, and has been corrected.) An independent test that uses only the rainfall anomaly found no statistically significant relation with recorded events (ρ = +0.02, p = 0.75, 9 departments × 25 years). The index is a structural comparison tool; it does not forecast events and has no external validation yet. The checks and their limits are public at terranava.org/algoritmo/alerta/backtest/.
Open, traceable inputs.
Every input is an open, citable dataset published by an official provider. Results can be traced from the final score back to the original data and the transformation applied.
| Source | Provides | Provider |
|---|---|---|
| ERA5-Land | Climate reanalysis, temperature & precipitation anomalies | Copernicus / ECMWF |
| CHIRPS | Satellite-gauge precipitation (SPI / SPEI inputs) | UCSB Climate Hazards Center |
| GRACE-FO | Terrestrial water storage anomaly (TWSA) | NASA JPL |
| SEAS5 | Probabilistic seasonal precipitation forecast | ECMWF |
| 2024 Census | Safe water, sanitation & unmet basic needs (municipal) | INE Bolivia |
| DesInventar | Historical event record (1970–2015) | UNDRR DesInventar |
| HydroSHEDS | Watershed & river-network delineation | WWF HydroSHEDS |
Where the water goes.
P = AET + Q + ΔS
Alongside the vulnerability index, TerraNava closes the watershed water balance: precipitation (P) is partitioned into actual evapotranspiration (AET), runoff (Q) and the change in water storage (ΔS). The partition between evaporation and runoff follows the Budyko–Fu framework, driven by an ensemble of precipitation products and remote-sensing evapotranspiration, with the storage term checked against GRACE-FO.
This gives, for each watershed and sub-basin, a physically consistent estimate of how much water enters, leaves and is stored — with uncertainty bands — rather than a single unverified number. Explore the closed balance for Bolivia's three drainage systems and 90 sub-basins in the live water balance.
From method to municipality.
Open a real municipal profile to see the IVH v4 and its five components computed on live data, with no login required.