The Royal Dutch Meteorological Institute (Dutch: Koninklijk Nederlands Meteorologisch Instituut, pronounced [ˈkoːnɪŋklək ˈneːdərlɑnts ˌmeːteːoːroːˈloːɣis ˌɪnstiˈtyt]; KNMI) is the Dutch national weather forecasting service, which has its headquarters in De Bilt, in the province of Utrecht, central Netherlands.The primary tasks of KNMI are weather forecasting, monitoring of climate changes and monitoring seismic activity. KNMI is also the national research and information centre for climate, climate change and seismology.
Tropical cyclones (TCs) pose significant risks due to their associated hazards, including powerful winds, inland and coastal flooding, and wind waves. However, more reliable TC records are required to ensure a robust statistical analysis for risk assessment. To overcome this limitation, researchers have developed methods to generate synthetic tropical cyclones (STCs) that provide a larger sample size of occurrences at specific locations. This study compares STC databases from different sources such as Massachusetts Institute of Technology (MIT), Columbia HAZard model (CHAZ), Synthetic Tropical cyclOne geneRation Model (STORM), and Deltares with historical TCs from the International Best Track Archive for Climate Stewardship (IBTrACS) on a basin-wide scale in the North Atlantic Basin. The aim is to assess the effectiveness of STCs in replicating crucial historical tropical cyclones parameters for risk analysis and to identify potential biases in the STC generation models. The comparison uses a hexagonal mesh to evaluate characteristics such as maximum winds, translation speed, and residence time. The study acknowledges the validation paradox arising from the limited IBTrACS data at specific locations that make it difficult to rigorously validate the accuracy of STCs in those areas and from systematic differences across the STC datasets. Despite the historical TCs database limitation, comparing STC with IBTrACS characteristics remains the only viable method for assessing biases in STC generation models. The evaluated STCs reveal spatial bias patterns, which may indicate deficiencies in the underlying hazard models. Identifying and describing these biases aim to guide the use of these events and highlight key aspects for further development in STC generation methods.
Between the 25th and the 31st of January 2023 an eruption occurred several kilometres East of Epi Island, Vanuatu, an area known to host a submarine volcanic zone including several cones. The largest cone (Epi B) entered regularly in activity in the last century, and images of daily monitoring satellites confirmed its responsibility for the January 2023 eruption. The eruption evolved to Surtseyan on the 31st, producing ash and gas columns reaching more than 100 m above sea level and small pyroclastic surge moving across the water, with subsequent pumice rafts and water discolouration. The eruptive activity is visible on satellite imagery and was well recorded on seismic, hydroacoustic and infrasound stations that are part of the Comprehensive Nuclear-Test-Ban Treaty Organization (CTBTO) International Monitoring System (IMS). Although this had no apparent consequences on the nearby communities, a small tsunami was however reported by locals and recorded on the coastal tide gauges of Port Vila 125 km to the south and Luganville 175 km to the north showing maximum amplitude of 5 cm. This data is used to validate the location and timing of the eruption and tentatively propose a tsunami mechanism. As a large part of the Vanuatu population is living in coastal areas, understanding tsunami mechanisms and assessing tsunami hazards associated with submarine eruptions is of main concern. Ultimately, this study aims to bring scientists and risk managers’ attention to a potentially hazardous volcano, through satellite and seismo-acoustic data analysis, numerical simulations of tsunami, and GIS mapping. • Submarine eruption in Vanuatu triggered a tsunami in January 2023 • Space-time origin is determined using satellite images and seismo-acoustic and sea-level data • Tsunami modelling via column collapse mechanism better matches gauge records
In July 2021 record-breaking extreme rainfall occurred in Western Europe causing huge damage and loss of life. The event was associated with a persistent cut-off low pressure system. We assess the influence of the large-scale dynamics on the extreme rainfall, evaluating the sensitivity of the rainfall to the specific circulation pattern of this event. Using multiple lines of evidence we show small changes in the position and magnitude of the cut-off low act to reduce the total rainfall over the region. We use reanalysis data, ensemble forecasts of the event, a boosted ensemble of a dynamically similar event, and pseudo global warming simulations of the event. We show that, for an event with analogous large-scale dynamics, future dynamical changes may outweigh thermodynamical effects, reducing the rainfall should such an event occur in a warmer climate.
Regional machine learning weather prediction (MLWP) models based on graph neural networks have recently demonstrated remarkable predictive accuracy, outperforming numerical weather prediction models at lower computational costs. In particular, limited-area model (LAM) and stretched-grid model (SGM) approaches have emerged for generating high-resolution regional forecasts. While LAM uses lateral boundaries from an external global model, SGM incorporates a global domain at lower resolution. This study aims to understand how differences in model design impact relative performance and potential applications. Specifically, in a near-identical setup, the strengths and weaknesses of these two approaches are identified for generating deterministic regional forecasts over Europe. Results show that both LAM and SGM are competitive deterministic MLWP models with generally accurate and comparable forecasting performance over the regional domain. Various differences were identified in the performance of the models across applications. LAM is able to successfully exploit high-quality boundary forcings to make predictions within the regional domain and is more suitable when training data is only available in a limited region. SGM is fully self-contained for easier operationalisation, can take advantage of more training data and shows signs of increased (temporal) generalisability. Our paper can serve as a starting point for meteorological institutes to guide their choice between LAM and SGM in developing an operational data-driven forecasting system.
Clouds and aerosol–cloud interactions remain major sources of uncertainty in climate projections. Here, we improve the representation of mixed-phase clouds (MPCs) in the EC-Earth3-AerChem Earth System Model by replacing the default temperature-dependent nucleation scheme with a physically based aerosol-sensitive heterogeneous ice nucleation parameterization. This scheme accounts for immersion freezing by K-feldspar, quartz, and marine organic aerosols, and is combined with a machine-learning-based parameterization of secondary ice production (SIP) to represent ice crystal multiplication processes. The new configuration improves agreement with global in situ ice nucleating particle (INP) observations and reveals realistic spatial patterns of ice crystal number concentrations (ICNC) across diverse environments. While these improvements do not eliminate the persistent structural cloud biases in EC-Earth3-AerChem, the aerosol-sensitive primary ice production scheme increases supercooled liquid water and cloud cover, particularly in the extratropics. Critically, the addition of SIP rebalances the cloud phase by enhancing ICNC in regions with low primary ice formation. Compared to the default scheme, the aerosol-sensitive primary ice production configuration with SIP reduces cloud radiative effect biases at mid- and high latitudes, while increasing them in the lower latitudes, leading to comparable global biases across configurations. Our results highlight the importance of explicitly representing both aerosol-sensitive nucleation and SIP for realistic simulations of MPCs and their radiative impacts. Unlike previous schemes, in which ice concentrations depend directly on INPs, the presence of effective SIP enhances ice formation in all MPCs and reduces the sensitivity of ICNC to aerosols, especially at low INP levels.
