Accurate rainfall observations with high spatial and temporal resolutions are key for hydrological applications, in particular for reliable flood forecasts. However, rain gauge networks operated by regional or national environmental agencies are often sparse, and weather radars tend to underestimate rainfall. As a complementary source of information, rain gauges from personal weather stations (PWSs), which have a network density 100 times higher than dedicated rain gauge networks in the Netherlands, can be used. However, PWSs are prone to additional sources of error compared to dedicated gauges, because they are generally not installed and maintained according to international guidelines. A systematic long-term analysis involving PWS rainfall observations across different seasons, accumulation intervals, and rainfall intensity classes has been missing so far. Here, we quantitatively compare rainfall estimates obtained from PWSs with rainfall recorded by automatic weather stations (AWSs) from the Royal Netherlands Meteorological Institute (KNMI) over the 2018–2023 period, including a sample of 1760 individual rainfall events in the Netherlands. This sample consists of the 10 highest rainfall accumulations per season and accumulation intervals (1, 3, 6, and 24 h) over a 6-year period. It was found that the average of a cluster of PWSs severely underestimates rainfall (around 36 % and 19 % for 1 h and 24 h intervals, respectively). By adjusting the data with areal reduction factors to account for the spatial variability of rainfall extremes and applying a bias correction factor of 1.22 to compensate for instrumental bias, the average relative bias decreases to −5 % for 1 h intervals or almost zero for intervals of 3 h and longer. The highest correlations (0.85 and 0.86) and lowest coefficients of variation (0.14 and 0.18) were found for 24 h intervals during winter and autumn, respectively. We show that most PWSs are able to capture high rainfall intensities up to around 30 mm h−1, indicating that these can be utilized for applications that require rainfall data with a spatial resolution of the order of kilometres, such as for flood forecasting in small, fast-responding catchments. PWSs did not observe the most intense rainfall events, which were associated with return periods exceeding 10 or 50 years (above approximately 30 mm h−1) and occurred in spring and summer. However, the spatial distribution of rainfall likely played a large role in the observed differences rather than instrumental limitations. This emphasizes the importance of having a dense rain gauge network. In addition, the variation in undercatch is likely partly due to the disproportional underestimation of tipping bucket rain gauges with increasing intensities. Outliers during winter were likely caused by solid precipitation and can potentially be removed using a temperature module from the PWS. We recommend additional research on dynamic calibration of the tipping volumes to improve this further.

On 29 October 2024, torrential rainfall locally exceeding 300 mm within less than 24 h, caused devastating floods in the province of Valencia in Spain. In this study we quantify and describe the spatial and temporal structure of the rainfall event on this day using rainfall observations from approximately 225 personal weather stations (PWSs), low-cost commercial devices primarily operated by citizens. The network density of PWSs is ∼7 times higher compared to the dedicated rain gauge network operated by the Spanish Meteorological Agency (AEMET) in the province of Valencia, allowing a more detailed analysis of the spatial and temporal rainfall dynamics. In addition, PWS observations are available in near real-time to the public with a temporal resolution of 5 min, whereas the data from AEMET are not available in real time for the public and at a lower publicly available temporal resolution (1 h). Daily rainfall sums recorded by the PWSs showed a high correlation (r=0.94) and low bias (underestimation of 4 %) compared to rainfall reported by AEMET. In the upstream parts of the Magro catchment (1661 km2), a first burst of extreme rainfall, reaching up to 180 mm of rainfall in a few hours, started in the morning, leading to the generation of a first flood wave in the upstream parts of the catchment. While the resulting flood wave was propagating downstream through the channel network, a second rainfall peak occurred, which moved downstream along with the flood wave. This spatial and temporal coincidence has likely exacerbated the devastating power of this event. Based on the PWS data, it could have been anticipated that the extreme rainfall already occurring early in the morning would likely result in flooding in the Magro catchment. Areal rainfall maps based on interpolating PWS data indicated catchment average rainfall exceeding 150 mm d−1 across an area of more than 2500 km2. However, the total accumulated rainfall remains uncertain due to interrupted measurements likely caused by power outage and inherent uncertainty associated with interpolating point measurements. For the Rambla de Poyo catchment, the resulting average discharge was around 900 m3 s−1. The estimated return period of the catchment-average rainfall and resulting discharge from this event exhibits large uncertainties, with on average exceeding 10 000 and 900 years, respectively. This study shows the potential of PWSs for real-time rainfall monitoring and potentially flood early warning systems, by complementing dedicated rain gauge networks in order to reduce the uncertainty from areal rainfall estimates and to localize potential flooding more accurately.