Soil Moisture

Soil moisture is the water held in the pore space of the soil column. Measured as a volumetric fraction or as a depth of water, it is a modest reservoir by volume and a disproportionately important one by influence, because it sits at the junction of the water balance and the surface energy balance and controls the exchange rate between them.

The quantity that matters is rarely the total. Most of the water in a soil is held tightly enough that plants cannot extract it. The useful measure is the water available between the wilting point and field capacity, and that range differs by a factor of several between a sand and a clay. A volumetric moisture content of 0.20 can mean comfortably wet or approaching wilting depending entirely on soil texture — which makes absolute soil moisture values close to meaningless without the soil properties alongside them.

The two regimes

Soil moisture controls how the land surface partitions available energy between latent and sensible heat, and it does so differently in two distinct regimes:

• Energy-limited. With ample water, evaporation is set by available energy and atmospheric demand. Adding water changes little. Most humid temperate regions sit here most of the time.
• Water-limited. Below a critical soil moisture, evaporation is limited by supply. Now the moisture state directly sets the flux partitioning, and drying soil pushes energy into sensible heat.

The interesting behaviour is in the transition between them, and it is where land-atmosphere coupling is strongest. Regions that spend time near the transition — the Mediterranean, the Sahel, the Great Plains, parts of central Europe — show the strongest coupling, and models disagree most about exactly where the threshold sits.

Memory, and why it matters for prediction

Soil moisture anomalies persist for weeks to months, far longer than any atmospheric anomaly. That memory is the main source of land-driven predictability at subseasonal range: knowing the soil is dry in May carries real information about July temperature in a water-limited region.

It is also the main route by which errors accumulate quietly. The drought feedback runs dry soil to more sensible heat to a deeper, drier boundary layer to suppressed convection to less rain to drier soil. This amplifies heatwaves and droughts, and it is why a regional model with a soil moisture bias produces a temperature bias that grows rather than staying constant.

The practical consequence for modelling is initialisation. Soil moisture is poorly observed, so it comes from reanalysis or spin-up — and a value taken from one land-surface scheme is not physically equivalent to the same number in another, because the schemes define field capacity and wilting point differently. Initialising from a mismatched source produces weeks of drift that is easily mistaken for model error. Spinning up the land surface properly is among the highest-return steps in a downscaling setup.

Observing it

Point measurements are straightforward and unrepresentative — soil moisture varies sharply over metres with texture and topography. Satellite microwave retrievals give broad coverage but sense only the top few centimetres, while the root zone that matters for evaporation extends a metre or more, and relating the two requires a model. This gap between what is observed and what is dynamically relevant is the central difficulty in constraining soil moisture, and it is why reanalysis soil moisture should be treated as a model product rather than an observation.

See also: land-surface models for how it is represented, evapotranspiration for the flux it controls, water balance for the hydrological context, and precipitation for its main input.