TY - GEN
AU - Karan Ruparell
AU - Kieran Hunt
AU - Hannah Cloke
AU - Christel Prudhomme
AU - Florian Pappenberger
AU - Matthew Chantry
AB - The Continuous Ranked Probability Score (CRPS) is a vital metric in earth system forecasting, how-ever it is still unintuitive for many users. In hydrology, we are increasingly transitioning to proba-bilistic and ensemble decisions making, and so informed use of metrics such as the CRPS is crucial. In this technical memorandum, we provide an overview of the CRPS. We highlight its growing im-portance, particularly in the context of machine learning, and then explain its relationship to various other metrics and scores commonly used in earth system forecasting, such as the Brier score and Mean Absolute Error. We then go on to apply the CRPS in the context of hydrological forecasting.
BT - Technical Memoranda
CY - Reading
DA - 12/2025
DO - 10.21957/16101872b5
M1 - 935
N2 - The Continuous Ranked Probability Score (CRPS) is a vital metric in earth system forecasting, how-ever it is still unintuitive for many users. In hydrology, we are increasingly transitioning to proba-bilistic and ensemble decisions making, and so informed use of metrics such as the CRPS is crucial. In this technical memorandum, we provide an overview of the CRPS. We highlight its growing im-portance, particularly in the context of machine learning, and then explain its relationship to various other metrics and scores commonly used in earth system forecasting, such as the Brier score and Mean Absolute Error. We then go on to apply the CRPS in the context of hydrological forecasting.
PB - ECMWF
PP - Reading
PY - 2025
T2 - Technical Memoranda
TI - A Hydrologist’s guide to the CRPS
UR -
ER -