ARNO-RAIN: a comprehensive dataset of observed (20 years) and simulated (2000 years) sub-hourly rainfall series for 70 rain gauges
摘要
ARNO-RAIN, a benchmark dataset of observed and simulated rainfall time series for 70 rain gauges in the Arno River basin (central Italy), is here presented. For each station, the dataset includes: (i) 20 years of continuous 15-minute and daily rainfall observations; (ii) a 2000-year synthetic daily rainfall series generated using the CoSMoS-2s stochastic model; and (iii) a corresponding 2000-year synthetic 15-minute series obtained by multifractal canonical disaggregation calibrated against observed intensity–duration–frequency curves. The integrated modelling framework enables the generation of sub-daily rainfall series from daily data. While ARNO-RAIN is specific to the Arno River basin, the approach can be transferred to other regions with similar data availability. ARNO-RAIN is a reproducible, well-documented, and openly accessible dataset designed to support benchmarking, evaluation, and intercomparison of stochastic rainfall generators and disaggregation methods. The dataset is complemented by station metadata and validation notebooks to enhance transparency and support reproducible research.