Terrain (DEM)

TWI: Predicting Where the Landscape Stays Wet

Concept article · Updated · by Dr. Anant Kumar Pathak

Water does not distribute evenly across terrain — it accumulates where large upslope areas funnel into gentle ground. The Topographic Wetness Index predicts those persistently wet zones from the shape of the land alone.

How it works

TWI takes the log of the upslope contributing area divided by the tangent of the local slope. Large catchments feeding flat ground yield high TWI (wet); small catchments on steep ground yield low TWI (dry). It is a steady-state proxy for soil-moisture potential.

The formula

TWI = ln( a / tan(β) ) — a = upslope contributing area, β = slope

Reading the values

Typical range: Typically ≈ 0–30 (higher = wetter)

Low on well-drained ridges · high in valley bottoms and convergence zones prone to saturation.

Advertisement

The data it needs

Derived fromDigital Elevation Model (SRTM / Copernicus, 30 m) — not an optical band index.

In practice

It informs soil mapping, wetland and hydric-soil prediction, digital soil-carbon modelling, and siting for agriculture and drainage.

Where it struggles

Assumes steady-state hydrology and breaks down on flat terrain and coarse DEMs; the flow-routing algorithm changes results.

TWI assumes steady-state hydrology and breaks down on flat terrain (where slope approaches zero and the log blows up) and on coarse DEMs. Your flow-accumulation algorithm materially changes the result, so document it.

Compute TWI on your own study area

Skip the code. Draw or upload a boundary and Spatial Research Suite runs TWI on live data — with cloud masking, exports and citations built in.

Run this analysis in GISforus →

Frequently asked

What does a high TWI value indicate?

A location likely to stay wet — valley bottoms and convergence zones with large upslope area and gentle slope; low TWI marks well-drained ridges.

Why does TWI misbehave on flat terrain?

As slope approaches zero the tangent in the denominator does too, sending TWI toward infinity, so nearly flat cells need special handling.

Primary reference: Beven, K.J. & Kirkby, M.J. (1979). A physically based, variable contributing area model of basin hydrology. Hydrological Sciences Bulletin 24(1), 43–69.