Why AERSI?

AQI tells you the moment.
AERSI tells you the story.

Standard air quality indices answer one question: how bad is the air right now? But chronic health damage from air pollution is shaped by something far more complex — persistence, volatility, multi-pollutant burden, and data reliability. AERSI is built to measure all of them.

"The burden of air pollution on human health is not determined by the worst single day — it is determined by the cumulative weight of every day, every spike, every season of exposure."

The Problem

What AERSI adds to AQI

AQI (what it measures):

  • ✓ How bad is the air right now?
  • ✗ How often has it been unsafe over 30 days?
  • ✗ Is this pollution stable or wildly swinging?
  • ✗ Which pollutants are compounding the danger?
  • ✗ How complete and trustworthy is the data?
  • ✗ How does this station compare to others over time?

What AERSI adds:

  • ✓ Combined pollution burden, WHO-normalized across all pollutants
  • ✓ How many of the last 30 days exceeded safe thresholds
  • ✓ How unpredictably the air quality swings day to day
  • ✓ How complete the station's data actually is
  • ✓ A confidence label so you know how much to trust the score
  • ✓ A single number that compounds all dimensions honestly
AQI and AERSI answer different questions. Both are useful. AQI tells you how bad the air is today. AERSI tells you how severe, persistent, and volatile the exposure has been over the past 30 days.
Four Dimensions

What AERSI measures

Each component captures something AQI cannot. Together they compound multiplicatively.

PL

Pollution Load

WHO-normalized weighted sum of all pollutants present, with a soft-saturation transform that reflects the sublinear dose-response relationship for fine particles. Missing pollutants are handled via weight renormalization — not silently excused.

EPF

Exposure Persistence

How often has AQI exceeded 100 over 30 days? A station unsafe 28 out of 30 days is fundamentally more dangerous than one that spikes once. EPF captures this, dampened proportionally when data is sparse.

VSF

Variability Severity

Unpredictable swings cause acute health events. VSF uses median absolute day-to-day AQI change — robust to sensor spikes — to capture instability that consistently bad air does not produce.

CF

Data Confidence

A station with two pollutants and 10 days of data should not be presented with the same confidence as a fully observed one. CF is computed and shown as a confidence label — making data quality visible, not hidden.

Real Scenarios

Where AQI misleads

Scenario 1

The Consistent Polluter

AQI of 180 every single day for 30 days. Dangerous, but predictable. The damage accumulates slowly and steadily.

✓ AERSI captures the high EPF — persistent exposure penalized correctly
Scenario 2

The Unpredictable Spiker

AQI swings between 40 and 400 with no pattern. Average is also 180. AQI treats these identically. AERSI does not — the high VSF correctly distinguishes the spiker.

✓ AERSI captures the high VSF — acute spike risk penalized correctly
Scenario 3

The Multi-Pollutant Station

AQI reports only the worst single pollutant. A station with PM2.5 at 3× and NO2 at 2× looks identical to one with just PM2.5 at 3×. AERSI combines them — the multi-pollutant burden is visible.

✓ AERSI captures combined burden — PL sums all pollutants with renormalized weights
Scenario 4

The Sparse Data Station

A station only reports PM2.5 with 15 days of data. AQI treats this identically to a fully observed station. AERSI shows "Low Confidence" — the uncertainty is honest, not hidden.

✓ AERSI flags data gaps honestly — sparse stations scored but confidence labelled
See the Full Formula → Explore the Map