Combining observations with convection permitting simulations to better predict extreme rainfall in ungauged locations

 Can we better (more accurately) estimate subdaily extreme rainfall in ungauged locations?

To answer to this question we combined rainfall observations together with high-resolution (~2.5km) convective permitting climate models, and geostatistical tecniques. We then compared the proposed approach with traditional state of the art methods, i.e geostatistical spatialization from raingauge observations.

The proposed, climate model aided, approach equalises traditional interpolation methods in the data rich case (1 station in ~200km2), with the exception of low duration extreme rainfall (lower than 6 hours).
When moving towards data scarce areas cases (1 station in 400km2 and 1/800km2) the proposed approach outperforms traditional methods, providing more accurate estimate of subdaily extreme rainfall.

Please have look to the paper by (me) Giuseppe Formetta, Eleonora Dallan, marco borga, and Francesco Marra.

----> The link to the paper is here