A Practical Guide to SPI and SPEI for Drought Monitoring

Ask ten hydrologists which drought index to use and you'll get at least three different answers. In practice, SPI and SPEI cover most operational drought-monitoring needs — the real skill is knowing which one, at which timescale, answers your specific question.

SPI: Precipitation Only

The Standardized Precipitation Index normalizes precipitation totals over a chosen accumulation period (commonly 1, 3, 6, or 12 months) against the long-term climatological distribution for that period and location. A SPI value of -1.5 means conditions are 1.5 standard deviations drier than the historical average for that timescale — nothing more, nothing less. It's simple, well-understood, and requires only a long, reliable precipitation record.

SPEI: Adding the Demand Side

SPEI extends SPI by incorporating potential evapotranspiration, effectively asking not just "how much rain fell" but "how much of a water deficit resulted, once atmospheric demand is accounted for." In a warming climate, this matters — the same rainfall total can represent a very different drought severity depending on temperature-driven evapotranspiration demand, and SPEI captures that where SPI cannot.

Choosing a Timescale

Short accumulation periods (1-3 months) track agricultural drought and near-term soil moisture stress well. Longer periods (6-12 months) better reflect hydrological drought — reservoir levels, groundwater, streamflow — which respond to cumulative deficits over many months, not a single dry spell.

A Common Mistake

We regularly see SPI computed over too short a baseline period, producing misleadingly extreme values because the climatological distribution itself is poorly estimated. We recommend a minimum 30-year baseline wherever the record allows, and explicit documentation of the baseline period used — a detail that matters enormously for anyone trying to compare your results against a different study.

In Practice

For most of our drought monitoring engagements, we compute both SPI and SPEI across multiple timescales and present them alongside satellite-derived Vegetation Condition Index — because no single index tells the whole story, and triangulating across meteorological, agricultural, and vegetation-response indicators produces a far more defensible drought assessment than any one number in isolation.