Hindcast Mode Research Articles

An efficient metamodel to downscale total water level in open beaches

A hybrid methodology is described to obtain the wave set-up and infragravity wave level components for estimating total water level (TWL) in coastal areas, taking into account the surf zone hydrodynamics involved. By combining statistical tools and numerical methods, high-resolution spatial distributions of wave height and water level components are calculated efficiently and seamlessly. The proposed methodology is tested and numerically validated on La Salvé beach (Cantabria, Spain), showing a robust performance for diverse scenarios. The resulting TWL can efficiently evaluate coastal hazards in forecast and hindcast mode, having strong potential as a guideline for coastal engineers, managers and emergency planners in the elaboration of adaptation and mitigation strategies. As an application of the metamodel to these coastal management tasks, significant past storm events have been reconstructed, and their associated TWL-driven coastal flooding hazard has been analyzed using a impact rank scale.

Three-dimensional hydrodynamic modelling of saltwater ingression and circulation in a Brackish Lake Shinji, Japan

Lake Shinji is a mesohaline lake that has intermittent saline water intrusion from its east coast connection of Ohashi River. This intrusion is often in the shape of a very thin layer that moves along the Lake bottom and stayed at the center due to bowl shape bathymetry of the Lake having maximum depth at the center. To elucidate the intrusion mechanism a 3-D hydrodynamic model (EcoPARI) was employed in a hindcast mode covering the year 2012 from the beginning of January to the end of September. Initially, the core emphasis was given to the vertical turbulence diffusion model and optimum model parameters were determined to accurately simulate the hydrodynamics. It was highlighted by the numerical simulations that alteration of one model constant, which is the multiplying coefficient of Richardson number, “α”, is influential instead of altering all the constants to tune the turbulence model. The possibility of using differential model constants for temperature and salinity was also tested and it helped in the improvement of simulations. Finally, the performance of the model was also assessed under extreme weather conditions and a historical upwelling event was simulated. The model was able to reproduce the intrusion in the summer months as well as the upwelling event in September due to strong westerly winds caused by a nearby passing typhoon.

Using a physics-based hydrological model and storm transposition to investigate machine-learning algorithms for streamflow prediction

Machine learning (ML) algorithms have produced remarkable advances in streamflow prediction, exceeding the performance of calibrated conceptual and physics-based hydrological models that have been developed over many decades. ML algorithms seem to overcome the issue of errors known to be present in rainfall and streamflow estimates that have hindered the performance of hydrological models for decades. In this paper, we propose a methodology for testing and benchmarking ML algorithms using artificial data generated by physically-based hydrological models. Our approach makes it possible to design controlled numerical experiments that can improve our understanding of this new generation of black-box models. We conducted a diagnostics study to demonstrate our methodology in which we attempted to determine if ML algorithms can identify a function relating streamflow and rainfall. This exercise combined the implementation of the distributed hillslope-link hydrological model (HLM) on a 4,385 km2 basin driven by precipitation fields created using the stochastic storm transposition (SST) framework, and an advanced deep learning algorithm based on gated recurrent unit (GRU)-Attention neural networks. The data generated allowed us to create prediction scenarios that are equivalent to the hindcast and real-time forecast problems. We proposed a set of scale-independent performance metrics to evaluate the results of our experiment and found that the GRU can correctly identify a predictive function for all analyzed locations in the river network. We concluded that under the circumstances tested in this study, deep learning can identify the transformation function when trained in Hindcast Mode, making it a powerful tool to determine the streamflow response of a basin to predetermined rainfall scenarios. However, it fails to significantly outperform the predictions of temporal persistence when tested in Forecast Mode.

Aeolus Data Validation for an Extreme Precipitation Event in Greece with the COSMO NWP Model

The study of atmospheric models and climate systems has been hampered by the restricted availability of wind profile measurements. By recording wind profiles in near real time and giving useful information for Numerical Weather Prediction (NWP) models, the Doppler wind lidar technology used in the European Space Agency’s Aeolus mission is expected to unravel this problem. By analyzing the precipitation from IANOS medicane, which occurred over the Thessaly plain in September 2020, a case study utilizing the COSMO NWP model illustrates the potential usefulness of Aeolus data in strengthening NWP models. Run in hindcast mode and forced by analyses with and without Aeolus, the model assimilated data that were produced at the European Centre for Medium-Range Forecasts (ECMWF). Evaluation against observations from Greek synoptic stations showed that the model precipitation using the Aeolus-assimilated data preponderated over the model results without their inclusion. This work aims to demonstrate the additional value of the Aeolus project towards the expansion of our knowledge of Earth’s atmosphere, particularly the improvement of our capacity to estimate severe weather events via the use of Aeolus with NWP models.

In the wind: Invasive species travel along predictable atmospheric pathways.

Invasive species such as insects, pathogens, and weeds reaching new environments by traveling with the wind, represent unquantified and difficult-to-manage biosecurity threats to human, animal, and plant health in managed and natural ecosystems. Despite the importance of these invasion events, their complexity is reflected by the lack of tools to predict them. Here, we provide the first known evidence showing that the long-distance aerial dispersal of invasive insects and wildfire smoke, a potential carrier of invasive species, is driven by atmospheric pathways known as Lagrangian coherent structures (LCS). An aerobiological modeling system combining LCS modeling with species biology and atmospheric survival has the potential to transform the understanding and prediction of atmospheric invasions. The proposed modeling system run in forecast or hindcast modes can inform high-risk invasion events and invasion source locations, making it possible to locate them early, improving the chances of eradication success.

A three-dimensional mixotrophic model of Karlodinium veneficum blooms for a eutrophic estuary

Blooms of dinoflagellate Karlodinium veneficum are widely distributed in estuarine and coastal waters and have been found to cause fish kills worldwide. K. veneficum has a mixed nutritional mode and relies on both photosynthesis and phagotrophy for growth; it is a mixotroph. Here, a model of mixotrophic growth of K. veneficum (MIXO) was developed, calibrated with previously-reported laboratory physiological data, and subsequently embedded in a 3D-coupled hydrodynamic (ROMS)-biogeochemical (RCA) model of eutrophic Chesapeake Bay, USA. The resulting ROMS-RCA-MIXO model was applied in hindcast mode to investigate seasonal and spatial distributions. Simulations showed that K. veneficum blooms occurred during June-August and were confined to the upper and middle Bay, consistent with long-term field observations. Autotrophic growth dominated in spring but heterotrophic growth dominated during the summer. The number of prey ingested by K. veneficum varied from 0.1 to 0.6 day−1 and the food vacuole content reached up to 50% of the core mixotroph biomass. The ingestion rate increased with prey density and also when P:N ratio fell below ∼0.03 (N:P ∼ 33), indicating that K. veneficum only switched to mixotrophic feeding in P-deficient waters when sufficient prey were available; this occurred during the summer months. The digestion rate increased with both the food vacuole content and temperature. The modeling analysis affirms K. veneficum as a phagotrophic ‘alga’ which is primarily photosynthetic but switches to mixotrophic feeding under nutrient deficient conditions.

A climate-dependent spatial epidemiological model for the transmission risk of West Nile virus at local scale

In this study, initial elements of a modelling framework aimed to become a spatial forecasting model for the transmission risk of West Nile virus (WNV) are presented. The model describes the dynamics of a WNV epidemic in population health states of mosquitoes, birds and humans and was applied to the case of Greece for the period 2010–2019. Calibration was performed with the available epidemiological data from the Hellenic Centre for Disease Control and Prevention and the environmental data from the European Union's earth observation program, Copernicus. Numerical results of the model for each municipality were evaluated against observations. Specifically, the occurrence of WNV, the number of infected humans and the week of incidence predicted from the model were compared to the corresponding numbers from observations. The results suggest that dynamic downscaling of a climate-dependent epidemiological model is feasible down-to roughly 80km2. This below nomenclature of territorial units for statistics (NUTS) 3 level represents the municipalities being the lowest level of administrative units, able to cope with WNV and take actions. The average detection probability in hindcast mode was 72%, improving strongly as the number of infected humans increased. Using the developed model, we were also able to show the fundamental importance of the May temperatures in shaping the WNV dynamics. The modeling framework couples epidemiological and environmental dynamical variables with surveillance data producing risk maps downscaled at a local level. The approach can be expanded to provide targeted early warning probabilistic forecasts that can be used to inform public health policy decision making.

Predicting Interannual Variability in Sea Surface Height Along the West Coast of Australia Using a Simple Ocean Model

AbstractSea surface height (SSH) along the west coast of Australia is key to local climate and is strongly forced by remote surface wind variability related to El Niño Southern Oscillation (ENSO) in the tropical Pacific Ocean. This study provides a method to predict interannual variability in SSH along the west coast of Australia using a simple 1.5‐layer dynamical ocean model forced by a statistical atmospheric model for ENSO‐related winds. The model has realistic coastlines and is driven by reanalysis surface winds regressed onto an ENSO index. The model when run in hindcast mode to predict past variability can simulate tide gauge observations at Fremantle along the west coast of Australia up to 13 months in advance, which outperforms persistence. We conclude that this methodology can be useful as a baseline for gauging the performance of more sophisticated forecast models for predicting SSH variations along the west coast of Australia.

Evaluating the accuracy and uncertainty of atmospheric and wave model hindcasts during severe events using model ensembles

Various uncertainties exist in a hindcast due to the inabilities of numerical models to resolve all the complicated atmosphere-sea interactions, and the lack of certain ground truth observations. Here, a comprehensive analysis of an atmospheric model performance in hindcast mode (Hurricane Weather and Research Forecasting model—HWRF) and its 40 ensembles during severe events is conducted, evaluating the model accuracy and uncertainty for hurricane track parameters, and wind speed collected along satellite altimeter tracks and at stationary source point observations. Subsequently, the downstream spectral wave model WAVEWATCH III is forced by two sets of wind field data, each includes 40 members. The first ones are randomly extracted from original HWRF simulations and the second ones are based on spread of best track parameters. The atmospheric model spread and wave model error along satellite altimeters tracks and at stationary source point observations are estimated. The study on Hurricane Irma reveals that wind and wave observations during this extreme event are within ensemble spreads. While both Models have wide spreads over areas with landmass, maximum uncertainty in the atmospheric model is at hurricane eye in contrast to the wave model.

A Web-Based Observatory for Biogeochemical Assessment in Coastal Regions

The concept of water observatories is extended to create a highly versatile tool for both the daily and the long-term management of estuarine ecosystems. Coastal observatories are evolving from simple data repositories to include forecasts, scenarios’ analyses and indicators, integrated in web platforms that provide multiple products and services. In a context of climate change (CC) and growing anthropogenic pressures, the assessment of the ecological health implies that the biogeochemical status is adequately quantified and incorporated in the coastal management decision-making procedure. This quantification requires accurate models for hydrodynamics and ecology that account for all relevant processes at the right scales. These models must be applied in forecast mode for emergency purposes and in hindcast mode to explore multiple scenarios as part of the CC adaptation strategy, creating a complex, vast amount of information to be shared with the coastal managers. A web-based portal supported by a comprehensive modeling and forecasting framework and materialized along the main water quality/biogeochemistry themes, from data to indicators, is developed and demonstrated in two distinct yet complex coastal systems: the Tagus estuary and the Ria Formosa lagoon. The paper starts with the requirements analysis from both ecological and computer science perspectives and then presents the overall multi-layered architecture of the framework and its key software components. The observatory portal implementation and demonstration explore its usefulness for coastal management.

Coupling large-scale climate indices with a stochastic weather generator to improve long-term streamflow forecasts in a Canadian watershed

This paper aims at improving long-term streamflow forecasts by implementing a novel technique based on conditioning the parameters of a stochastic weather generator on large-scale climate indices, with varying lengths of training periods during the establishment of correlations. The most important climate indices are identified by looking at yearly correlations between a set of 40 indices and meteorological data (precipitation and temperature) at the watershed scale. A linear model is then constructed to identify precipitation and temperature anomalies to induce perturbations in the stochastic weather generator. Time windows of 5, 10, 15, 20 and 30 years are used in determining the optimal linear model. The performance of the proposed approach is assessed against that of a resampling of past climatology and using the same stochastic weather generator unconditioned on climate indices. Each member of the ensemble weather forecast is then fed to a hydrological model to create the Ensemble Streamflow Forecasts (ESF) with a one-year forecasting horizon. The three approaches are tested in hindcast mode over a 30-year period at 12 forecast dates. Results show that temperatures are significantly correlated with large-scale climate indices, whereas precipitation is only weakly related to the same indices. The length of the time window has a considerable impact on the prediction ability of the linear models. The precipitation models based on short duration time windows performed better than those based on longer windows, while the reverse was found for the temperature models. A comparison between all three Ensemble Streamflow Forecast approaches is assessed using the Continuous Ranked Probability Score (CRPS) metric. Results show that the proposed method improves long-term streamflow forecasting, particularly the volumetric bias and the peak flows during the spring flood.

Inverse modeling of fire emissions constrained by smoke plume transport using HYSPLIT dispersion model and geostationary satellite observations

Abstract. Smoke forecasts have been challenged by high uncertainty in fire emission estimates. We develop an inverse modeling system, the HYSPLIT-based Emissions Inverse Modeling System for wildfires (or HEIMS-fire), that estimates wildfire emissions from the transport and dispersion of smoke plumes as measured by satellite observations. A cost function quantifies the differences between model predictions and satellite measurements, weighted by their uncertainties. The system then minimizes this cost function by adjusting smoke sources until wildfire smoke emission estimates agree well with satellite observations. Based on HYSPLIT and Geostationary Operational Environmental Satellite (GOES) Aerosol/Smoke Product (GASP), the system resolves smoke source strength as a function of time and vertical level. Using a wildfire event that took place in the southeastern United States during November 2016, we tested the system's performance and its sensitivity to varying configurations of modeling options, including vertical allocation of emissions and spatial and temporal coverage of constraining satellite observations. Compared with currently operational BlueSky emission predictions, emission estimates from this inverse modeling system outperform in both reanalysis (21 out of 21 d; −27 % average root-mean-square-error change) and hindcast modes (29 out of 38 d; −6 % average root-mean-square-error change) compared with satellite observed smoke mass loadings.

Effect of Fluvial Discharges and Remote Non-Tidal Residuals on Compound Flood Forecasting in San Francisco Bay

Accurate and timely flood forecasts are critical for making emergency-response decisions regarding public safety, infrastructure operations, and resource allocation. One of the main challenges for coastal flood forecasting systems is a lack of reliable forecast data of large-scale oceanic and watershed processes and the combined effects of multiple hazards, such as compound flooding at river mouths. Offshore water level anomalies, known as remote Non-Tidal Residuals (NTRs), are caused by processes such as downwelling, offshore wind setup, and also driven by ocean-basin salinity and temperature changes, common along the west coast during El Niño events. Similarly, fluvial discharges can contribute to extreme water levels in the coastal area, while they are dominated by large-scale watershed hydraulics. However, with the recent emergence of reliable large-scale forecast systems, coastal models now import the essential input data to forecast extreme water levels in the nearshore. Accordingly, we have developed Hydro-CoSMoS, a new coastal forecast model based on the USGS Coastal Storm Modeling System (CoSMoS) powered by the Delft3D San Francisco Bay and Delta community model. In this work, we studied the role of fluvial discharges and remote NTRs on extreme water levels during a February 2019 storm by using Hydro-CoSMoS in hindcast mode. We simulated the storm with and without real-time fluvial discharge data to study their effect on coastal water levels and flooding extent, and highlight the importance of watershed forecast systems such as NOAA’s National Water Model (NWM). We also studied the effect of remote NTRs on coastal water levels in San Francisco Bay during the 2019 February storm by utilizing the data from a global ocean model (HYCOM). Our results showed that accurate forecasts of remote NTRs and fluvial discharges can play a significant role in predicting extreme water levels in San Francisco Bay. This pilot application in San Francisco Bay can serve as a basis for integrated coastal flood modeling systems in complex coastal settings worldwide.

Fast Ice Prediction System (FIPS) for land-fast sea ice at Prydz Bay, East Antarctica: an operational service for CHINARE

AbstractA Fast Ice Prediction System (FIPS) was constructed and is the first regional land-fast sea-ice forecasting system for the Antarctic. FIPS had two components: (1) near-real-time information on the ice-covered area from MODIS and SAR imagery that revealed, tidal cracks, ridged and rafted ice regions; (2) a high-resolution 1-D thermodynamic snow and ice model (HIGHTSI) that was extended to perform a 2-D simulation on snow and ice evolution using atmospheric forcing from ECMWF: either using ERA-Interim reanalysis (in hindcast mode) or HERS operational 10-day predictions (in forecast mode). A hindcast experiment for the 2015 season was in good agreement with field observations, with a mean bias of 0.14 ± 0.07 m and a correlation coefficient of 0.98 for modeled ice thickness. The errors are largely caused by a cold bias in the atmospheric forcing. The thick snow cover during the 2015 season led to modeled formation of extensive snow ice and superimposed ice. The first FIPS operational service was performed during the 2017/18 season. The system predicted a realistic ice thickness and onset of snow surface melt as well as the area of internal ice melt. The model results on the snow and ice properties were considered by the captain of R/VXuelongwhen optimizing a low-risk route for on-ice transportation through fast ice to the coastal Zhongshan Station.

Development and validation of a storm surge forecasting/hindcasting modelling system for the extensive Río de la Plata Estuary and its adjacent Continental Shelf

The Rio de la Plata (RdP) Estuary is affected by significant surges several times per year. This phenomenon has historically caused catastrophic water-level enlargements of up to 4.44 m, threatening and claiming human lives and producing major economic and material damages. The negative surges are less frequent, but when they do occur, inhibit the access to the principal harbors and waterways and disable the drinking water intakes of the Metropolitan Area of Buenos Aires (the Capital City of Argentina) with a population of more and 16 million people. Recent works suggest that the number and strength of the surge events have been increasing with time. Nevertheless, a state-of-the-art system for the forecast of those events is not available yet. In this work, the implementation of a numerical modelling system for the forecast/hindcast of storm surges and the associated currents in the RdP and the adjacent continental shelf are presented and validated. This pre-operational system is based on an adaptation of the CROCO community ocean model to solve the dynamics associated with the surge. The model was implemented using a set of routines written in open-source programming language (Linux and Python) to be cheap and efficient and to ensure an easy future transfer to the services responsible for the alerts. For a better representation of the regional atmospheric dynamics, wind speed and sea-level pressure used to force the simulations were corrected using direct observations collected at an oceanographic buoy anchored at the estuary. The model system performance in hindcast mode was quantified by comparison with observations from tidal gauges and current meters at several locations of the estuary and the adjacent shelf. Percent errors for water level over the whole estuary and currents in the intermediate and exterior estuary drove to average results of 8 and 13%, respectively. The skill scores resulted, on average, of 0.90 and 0.80, respectively. The model performance in both hindcast and forecast modes was evaluated during historical extreme storm surges. Results support the good performance of the model to simulate even extreme events with average skill scores of 0.97 and 0.92, respectively. Results are encouraging, particularly taking into account the limitations in the atmospheric forcing for the region, where only a relatively small number of direct observations are assimilated by the reanalysis and forecast models.

Stochastic ensemble climate forecast with an analogue model

Abstract. This paper presents a system to perform large-ensemble climate stochastic forecasts. The system is based on random analogue sampling of sea-level pressure data from the NCEP reanalysis. It is tested to forecast a North Atlantic Oscillation (NAO) index and the daily average temperature in five European stations. We simulated 100-member ensembles of averages over lead times from 5 days to 80 days in a hindcast mode, i.e., from a meteorological to a seasonal forecast. We tested the hindcast simulations with the usual forecast skill scores (CRPS or correlation) against persistence and climatology. We find significantly positive skill scores for all timescales. Although this model cannot outperform numerical weather prediction, it presents an interesting benchmark that could complement climatology or persistence forecast.

Connectivity modelling of areas closed to protect vulnerable marine ecosystems in the northwest Atlantic

Over the course of the past decade, in response to United Nations General Assembly resolutions calling for the protection of vulnerable marine ecosystems (VMEs), the Northwest Atlantic Fisheries Organization has closed 14 areas around the high-seas portion of Grand Bank and Flemish Cap to protect deep-sea coral and sponge habitats from impacts by bottom-contact fishing gears. Structural and functional connectivity for those areas were not explicitly considered in the area-selection process. We applied a particle-tracking model in each of four seasons to produce dispersal trajectories at the surface and 100 m from start points within the closed areas. These were run in forecast and hindcast modes to identify dispersal kernels. Currents at the surface, 100 m, 1000 m and “on bottom” were examined under an independent model (NEMO) to infer structural connectivity among the areas at relevant depths not available in the particle-tracking model. Spawning times and planktonic larval duration of the dominant sponges, sea pens and gorgonian corals were then considered to evaluate the trajectories as bio-physical models, while species distribution models identified potential source populations from hindcast projections. Five of the 14 areas, including the three largest closures, showed particle retention, with three others showing retention within 10 km of their boundaries. The regional pattern of currents and their topographic forcing emerged as a strong structuring agent. A system of weakly-connected closed areas to protect sea pen VMEs on Flemish Cap was identified. The conducted approach illustrates the added value of assessing/modelling networking properties when designing MPAs.

Seasonal drought prediction for semiarid northeastern Brazil: verification of six hydro-meteorological forecast products

Abstract. A set of seasonal drought forecast models was assessed and verified for the Jaguaribe River in semiarid northeastern Brazil. Meteorological seasonal forecasts were provided by the operational forecasting system used at FUNCEME (Ceará's research foundation for meteorology) and by the European Centre for Medium-Range Weather Forecasts (ECMWF). Three downscaling approaches (empirical quantile mapping, extended downscaling and weather pattern classification) were tested and combined with the models in hindcast mode for the period 1981 to 2014. The forecast issue time was January and the forecast period was January to June. Hydrological drought indices were obtained by fitting a multivariate linear regression to observations. In short, it was possible to obtain forecasts for (a) monthly precipitation, (b) meteorological drought indices, and (c) hydrological drought indices. The skill of the forecasting systems was evaluated with regard to root mean square error (RMSE), the Brier skill score (BSS) and the relative operating characteristic skill score (ROCSS). The tested forecasting products showed similar performance in the analyzed metrics. Forecasts of monthly precipitation had little or no skill considering RMSE and mostly no skill with BSS. A similar picture was seen when forecasting meteorological drought indices: low skill regarding RMSE and BSS and significant skill when discriminating hit rate and false alarm rate given by the ROCSS (forecasting drought events of, e.g., SPEI1 showed a ROCSS of around 0.5). Regarding the temporal variation of the forecast skill of the meteorological indices, it was greatest for April, when compared to the remaining months of the rainy season, while the skill of reservoir volume forecasts decreased with lead time. This work showed that a multi-model ensemble can forecast drought events of timescales relevant to water managers in northeastern Brazil with skill. But no or little skill could be found in the forecasts of monthly precipitation or drought indices of lower scales, like SPI1. Both this work and those here revisited showed that major steps forward are needed in forecasting the rainy season in northeastern Brazil.

Ash Metrics for European and Trans‐Atlantic Air Routes During the Eyjafjallajökull Eruption 14 April to 23 May 2010

AbstractMetrics for the risk associated with the threat that airborne volcanic ash particles pose to commercial jet aircraft are presented using simulations based on a Lagrangian particle transport and dispersion model driven by satellite measurements for the Eyjafjallajökull volcanic eruption, Iceland, for the period 14 April to 23 May 2010. The study utilizes a four‐dimensional data set of simulated ash concentrations together with European and trans‐Atlantic air routes to determine metrics corresponding to the total mass intercepted (defined as the dose), the mass interception rate (the dose rate), and the concentration (the exposure) over time (the dosage) that a jet aircraft encounters along the air route. The methodology can be used as a logistical and flight planning tool in a forecast mode and also in hindcast mode to assess the extent of airline fleet exposure to ash following an eruption, thereby providing operators with information useful for flight safety.

Impact of rheology on probabilistic forecasts of sea ice trajectories: application for search and rescue operations in the Arctic

Abstract. We present a sensitivity analysis and discuss the probabilistic forecast capabilities of the novel sea ice model neXtSIM used in hindcast mode. The study pertains to the response of the model to the uncertainty on winds using probabilistic forecasts of ice trajectories. neXtSIM is a continuous Lagrangian numerical model that uses an elasto-brittle rheology to simulate the ice response to external forces. The sensitivity analysis is based on a Monte Carlo sampling of 12 members. The response of the model to the uncertainties is evaluated in terms of simulated ice drift distances from their initial positions, and from the mean position of the ensemble, over the mid-term forecast horizon of 10 days. The simulated ice drift is decomposed into advective and diffusive parts that are characterised separately both spatially and temporally and compared to what is obtained with a free-drift model, that is, when the ice rheology does not play any role in the modelled physics of the ice. The seasonal variability of the model sensitivity is presented and shows the role of the ice compactness and rheology in the ice drift response at both local and regional scales in the Arctic. Indeed, the ice drift simulated by neXtSIM in summer is close to the one obtained with the free-drift model, while the more compact and solid ice pack shows a significantly different mechanical and drift behaviour in winter. For the winter period analysed in this study, we also show that, in contrast to the free-drift model, neXtSIM reproduces the sea ice Lagrangian diffusion regimes as found from observed trajectories. The forecast capability of neXtSIM is also evaluated using a large set of real buoy's trajectories and compared to the capability of the free-drift model. We found that neXtSIM performs significantly better in simulating sea ice drift, both in terms of forecast error and as a tool to assist search and rescue operations, although the sources of uncertainties assumed for the present experiment are not sufficient for complete coverage of the observed IABP positions.