Abstract. In this study, the frequency and intensity of soil-cooling rains is assessed
using in situ observations of atmospheric and soil profile variables in
southern France. Rainfall, soil temperature, and topsoil volumetric soil
moisture (VSM) observations, measured every 12 min at 21 stations of the
SMOSMANIA (Soil Moisture Observing System – Meteorological Automatic Network
Integrated Application) network, are analyzed over a time period of 9 years,
from 2008 to 2016. The spatial and temporal statistical distribution of the
observed rainfall events presenting a marked soil-cooling effect is
investigated. It is observed that the soil temperature at a depth of 5 cm
can decrease by as much as 6.5 ∘C in only 12 min during a
soil-cooling rain. We define marked soil-cooling rains as rainfall events
triggering a drop in soil temperature at a depth of 5 cm larger than
1.5 ∘C in 12 min. Under Mediterranean and
Mediterranean–mountain climates, it is shown that such events occur up to nearly
3 times a year, and about once a year on average. This frequency decreases to
about once every 3.5 years under semi-oceanic climate. Under oceanic climate,
such pronounced soil-cooling rains are not observed over the considered
period of time. Rainwater temperature is estimated for 13 cases of marked
soil-cooling rains using observed changes within 12 min in soil temperature
at a depth of 5 cm, together with soil thermal properties and changes in
VSM. On average, the estimated rainwater temperature is generally lower than
the observed ambient air temperature, wet-bulb temperature, and topsoil
temperature at a depth of 5 cm, with mean differences of −5.1, −3.8, and
−11.1 ∘C, respectively. The most pronounced differences are
attributed to hailstorms or to hailstones melting before getting to the soil
surface. Ignoring this cooling effect can introduce biases in land surface
energy budget simulations.