A quality-controlled hourly precipitation dataset for the analysis of intense precipitation over Italy

Despite the growing abundance of precipitation datasets, the availability of high temporal and spatial resolution observations from rain gauges is still limited and fragmented. However, these data are essential especially when the focus is on intense precipitation, since other products (e.g., satellite, radar, and reanalysis) may be affected by important biases. In Italy, hourly precipitation measurements are managed independently by regional or sub-regional institutions, resulting in the absence of a unified national-scale dataset. To address this gap, we present the first comprehensive hourly precipitation database for Italy, obtained by integrating observations from ~ 3,000 continuously monitoring rain gauges. The database spans several decades, with some time series beginning in the early 1980s, while the highest spatial coverage is achieved from the early 2000s up to 2024. An extensive pre-processing phase was carried out to standardize and organize the dataset, e.g., by removing duplicate stations and standardizing the coordinates and the timing to a common reference system. To ensure data reliability and consistency, a comprehensive quality control procedure was also applied, by adapting to the specific characteristics of the Italian climate a set of well-established methodologies from the literature (e.g., Blenkinsop et al., 2017, Lewis et al, 2021). Quality control was designed to identify and correct common issues such as the erroneous aggregation of daily totals into single hourly records, outliers (detected using statistical thresholds based on observed data extremes), and unrealistically high values occurring after prolonged data gaps, usually indicative of sensor malfunction. The resulting dataset represents a robust basis for a wide range of applications. For instance, it allowed us to characterize how intense precipitation is distributed across the Italian territory in terms of magnitude and seasonality, and to further investigate the diurnal cycle of extreme rainfall. Another key outcome concerns the probabilistic analysis of extreme precipitation. Although the temporal extent of the dataset is not adequate to support analyses based on classical extreme value theory, it can be analyzed with more effective recent approaches, such as the MEV (Marani & Ignaccolo, 2015) and the SMEV (Marra et al., 2019) frameworks. Finally, beyond research applications, the dataset offers a valuable support for risk management, adaptation planning, and infrastructure design under changing climate conditions.