Search

This study is anchored in the H2020 SEAMLESS project (https://www.seamlessproject.org , last access: 29 January 2024), which aims to develop ensemble assimilation methods to be implemented in Copernicus Marine Service monitoring and forecasting systems, in order to operationally estimate a set of targeted ecosystem indicators in various regions, including uncertainty estimates. In this paper, a simplified approach is introduced to perform a 4D (space–time) ensemble analysis describing the evolution of the ocean ecosystem. An example application is provided, which covers a limited time period in a limited subregion of the North Atlantic (between 31 and 21 ∘ W, between 44 and 50.5 ∘ N, between 15 March and 15 June 2019, at a 1/4 ∘ and a 1 d resolution). The ensemble analysis is based on prior ensemble statistics from a stochastic NEMO (Nucleus for European Modelling of the Ocean)–PISCES simulator. Ocean colour observations are used as constraints to condition the 4D prior probability distribution. As compared to classic data assimilation, the simplification comes from the decoupling between the forward simulation using the complex modelling system and the update of the 4D ensemble to account for the observation constraint. The shortcomings and possible advantages of this approach for biogeochemical applications are discussed in the paper. The results show that it is possible to produce a multivariate ensemble analysis continuous in time and consistent with the observations. Furthermore, we study how the method can be used to extrapolate analyses calculated from past observations into the future. The resulting 4D ensemble statistical forecast is shown to contain valuable information about the evolution of the ecosystem for a few days after the last observation. However, as a result of the short decorrelation timescale in the prior ensemble, the spread of the ensemble forecast increases quickly with time. Throughout the paper, a special emphasis is given to discussing the statistical reliability of the solution. Two different methods have been applied to perform this 4D statistical analysis and forecast: the analysis step of the ensemble transform Kalman filter (with domain localization) and a Monte Carlo Markov chain (MCMC) sampler (with covariance localization), both enhanced by the application of anamorphosis to the original variables. Despite being very different, the two algorithms produce very similar results, thus providing support to each other's estimates. As shown in the paper, the decoupling of the statistical analysis from the dynamical model allows us to restrict the analysis to a few selected variables and, at the same time, to produce estimates of additional ecological indicators (in our example: phenology, trophic efficiency, downward flux of particulate organic matter). This approach can easily be appended to existing operational systems to focus on dedicated users' requirements, at a small additional cost, as long as a reliable prior ensemble simulation is available. It can also serve as a baseline to compare with the dynamical ensemble forecast and as a possible substitute whenever useful. [ABSTRACT FROM AUTHOR]

This paper describes an attempt to use data obtained from sodar (sound detection and ranging) for short-term forecasting of PM 10 concentration levels in Krakow. Krakow is one of the most polluted cities in central Europe (CE) in terms of PM 10 concentration. This is due to the high municipal emissions. Thanks to intensive corrective actions taken by the city authorities, these are being effectively eliminated, but the unfavourable topographic location of the city limits natural ventilation. The article describes all these conditions, focusing on presenting the method of short-term correction of air quality for the time needed to take quick corrective actions by the city authorities in the event of anticipated exceedances of the permissible values. Based on several years of measurements of the physical properties of the atmosphere with sodar, the authors of the paper suggest that sodar data could be considered for operational use to generate short-term predictions. The proposed method is based on the use of the spectrum, i.e. the set of amplitudes of signals returning to the sodar receiver from the reflection of a single-frequency sound transmission and its characteristic properties depending on the physical state of the atmosphere. Similar spectra were parameterised with a single numerical value using statistical methods. It was found that, in some cases preceding high concentrations of PM 10 , the spectral parameters had similar values. This made it possible to develop a forecasting method for such concentrations by using data mining to search for conditions in historical data closest to the state of the atmosphere at the time of forecasting. In this part of the study, data from 2017–2018 were used. In the next step, three methods of using the sodar data developed for PM 10 prediction were proposed, comparing them with the method without sodar use. The study results were tested on independent material using data concerning the episodes of high concentrations of pollutants from October 2021 to March 2022 in Krakow. The findings were considered to be encouraging, also taking into account the speed and low cost of preparing the forecast. [ABSTRACT FROM AUTHOR]

This study is anchored in the H2020 SEAMLESS project (www.seamlessproject.org), which aims to develop ensemble assimilation methods to be implemented in Copernicus Marine Service monitoring and forecasting systems, in order to operationally estimate a set of targeted ecosystem indicators in various regions, including uncertainty estimates. In this paper, a simplified approach is introduced to perform a 4D (space-time) ensemble analysis describing the evolution of the ocean ecosystem. An example application is provided, which covers a limited time period in a limited subregion of the North Atlantic (between 31° W and 21° W, between 44° N and 50.5° N, between March 15 and June 15, 2019, at a 1/4° and a 1 day resolution). The ensemble analysis is based on prior ensemble statistics from a stochastic NEMO/PISCES simulator. Ocean colour observations are used as constraints to condition the 4D prior probability distribution. As compared to classic data assimilation, the simplification comes from the decoupling between the forward simulation using the complex modelling system and the update of the 4D ensemble to account for the observation constraint. The shortcomings and possible advantages of this approach for biogeochemical applications are discussed in the paper. The results show that it is possible to produce a multivariate ensemble analysis continuous in time and consistent with the observations. Furthermore, we study how the method can be used to extrapolate analyses calculated from past observations into the future. The resulting 4D ensemble statistical forecast is shown to contain valuable information about the evolution of the ecosystem for a few days after the last observation. However, as a result of the short decorrelation time scale in the prior ensemble, the spread of the ensemble forecast is increasing quickly with time. Throughout the paper, a special emphasis is given to discussing the statistical reliability of the solution. Two different methods have been applied to perform this 4D statistical analysis and forecast: the analysis step of the Ensemble Transform Kalman Filter (with domain localization) and a Monte Carlo Markov Chain (MCMC) sampler (with covariance localization), both enhanced by the application of anamorphosis to the original variables. Despite being very different, the two algorithms produce very similar results, thus somehow validating each other. As shown in the paper, the decoupling of the statistical analysis from the dynamical model allows us to restrict the analysis to a few selected variables and, at the same time, to produce estimates of additional ecological indicators (in our example: phenology, trophic efficiency, downward flux of particulate organic matter). This approach can easily be appended to existing operational systems to focus on dedicated users' requirements, at small additional cost, as long as a reliable prior ensemble simulation is available. It can also serve as a baseline to compare with the dynamical ensemble forecast, and as a possible substitute whenever useful. [ABSTRACT FROM AUTHOR]

The MyOcean marine core service, implementing the Global Monitoring for Environment and Security (GMES) objectives, targets the provision of ocean state data from various platforms to assist, among other downscaled activities, the needs of the operational response to marine safety, particularly concerning oil spills. The MEDSLIK oil spill and trajectory prediction system makes use of the MyOcean regional and Cyprus Coastal Ocean Forecasting and Observing System (CYCOFOS) downscaled forecasting products for operational application in the Mediterranean and pre-operational use in the Black Sea. Advanced Synthetic Aperture Radar (ASAR) satellite remote-sensing images from European Space Agency (ESA) and European Maritime Safety Agency - CleanSeaNet (EMSA-CSN) provide the means for routine monitoring of the southern European seas for the detection of illegal oil discharges. MEDSLIK offers various ways, to be described in this paper, of coupling the Pan-European capacity for Ocean Monitoring and Forecasting (MyOcean) forecasting data with ASAR imageries to provide both forecasts and hindcasts for such remotely-observed oil slicks. The main concern will be the drift of the oil slick and also, in the case of the forecast mode, its diffusive spreading, although some attempt is also made to estimate the changes in the state of the oil. The successful link of the satellite-detected oil slicks with their operational predictions using the MyOcean products contributes to the operational response chain and the strengthening of maritime safety for accidental or illegal spills, in implementation of a Mediterranean decision support system for marine safety regarding oil spills. [ABSTRACT FROM AUTHOR]

First results of a coupled modeling and forecasting system for the pelagic fisheries are being presented. The system consists currently of three mathematically fundamentally different model subsystems: POLCOMS-ERSEM providing the physical- biogeochemical environment implemented in the domain of the North-West European shelf and the SPAM model which describes sandeel stocks in the North Sea. The third component, the SLAM model, connects POLCOMS-ERSEM and SPAM by computing the physical-biological interaction. Our major experience by the coupling model subsystems is that well-defined and generic model interfaces are very important for a successful and extendable coupled model framework. The integrated approach, simulating ecosystem dynamics from physics to fish, allows for analysis of the pathways in the ecosystem to investigate the propagation of changes in the ocean climate and lower trophic levels to quantify the impacts on the higher trophic level, in this case the sandeel population, demonstrated here on the base of hindcast data. The coupled forecasting system is tested for some typical scientific questions appearing in spatial fish stock management and marine spatial planning, including determination of local and basin scale maximum sustainable yield, stock connectivity and source/sink structure. Our presented simulations indicate that sandeels stocks are currently exploited close to the maximum sustainable yield, but large uncertainty is associated with de20 termining stock maximum sustainable yield due to stock eigen dynamics and climatic variability. Our statistical ensemble simulations indicates that the predictive horizon set by climate interannual variability is 2-6 yr, after which only an asymptotic probability distribution of stock properties, like biomass, are predictable. [ABSTRACT FROM AUTHOR]

Since December 2010, the global analysis and forecast of the MyOcean system consists in the Mercator Oc'ean NEMO global 1/4° configuration with a 1/12° "zoom" over the Atlantic and Mediterranean Sea. The zoom open boundaries come from the global 1/4° at 20° S and 80° 5 N. The data assimilation uses a reduced order Kalman filter with a 3-D multivariate modal decomposition of the forecast error. It includes an adaptative error and a localization algorithm. A 3D-Var scheme corrects for the slowly evolving large-scale biases in temperature and salinity. Altimeter data, satellite temperature and in situ temperature and salinity vertical profiles are jointly assimilated to estimate the initial conditions for the numerical ocean forecasting. This paper gives a description of the recent systems. The validation procedure is introduced and applied to the current and future systems. This paper shows how the validation impacts on the quality of the systems. It is shown how quality check (in situ, drifters) and data source (satellite temperature) impacts as much as the systems design (model physics and assimilation parameters). The validation demonstrates the accuracy of the MyOcean global products. Their quality is stable in time. The future systems under development still suffer from a drift. This could only be detected with a 5 yr hindcast of the systems. This emphasizes the need for continuous research efforts in the process of building future versions of MyOcean2 forecasting capacities. [ABSTRACT FROM AUTHOR]

The Black Sea coastal nowcasting and forecasting system was built within the framework of EU FP6 ECOOP project for five regions: the south-western basin along the coasts of Bulgaria and Turkey, the North-Western shelf along the Romanian and Ukrainian coasts, coastal zone around of the Crimea peninsula, the north-eastern Russian coastal zone and the coastal zone of Georgia. The system operates in the real-time mode during the ECOOP project and afterwards. The forecasts include temperature, salinity and current velocity fields. Ecosystem model operates in the off-line mode near the Crimea coast. [ABSTRACT FROM AUTHOR]

Machine tools are subject to multiple sources of error during machining, resulting in deviations in the dimensions of the part and a reduction in contour accuracy. This paper proposes a contour error prediction model based on a long short-term memory (LSTM) neural network, taking hexagonal recess machining as an example and considering the power, vibration, and temperature signals that affect the contour error. The experimental data show that the model can accurately predict the contour error of the machined part. A more accurate and robust contour error prediction model can provide data support for online compensation of contour errors. [ABSTRACT FROM AUTHOR]

It is highly likely that the Mediterranean region will experience increased aridity and hydrological droughts. Therefore, seasonal forecasts of soil moisture can be a valuable resource for groundwater management. However, their accuracy in this region has not been evaluated against observations. This paper presents an evaluation of soil moisture in the Central Mediterranean region during the period 2001–2021 using the seasonal forecast system SEAS5. Standardized anomalies of soil moisture are compared with observed values in ERA5 reanalysis. In terms of the average magnitude of the forecast error and the anomaly correlation coefficient, the forecasts demonstrate good performance only in certain regions of the domain for the deepest soil layer at 289 cm, the most interesting for groundwater management. No clear overlap with specific land features such as orography, land cover, or distance from the coast has been observed with respect to the forecast performance. Accordingly, seasonal forecasts can be used to detect wet and dry events for the deepest soil layer in certain regions, with lead-times of up to 6 months. In these regions, the area under the Relative Operating Characteristic (ROC) curve can reach values larger than 0.8. Dry events are generally better captured than wet events for all soil layers. We also analyzed the effectiveness of seasonal forecasts in predicting wet and dry events in Northern and Central Italy for the 2012–2013 period, with a lead-time of 6 months. We found that seasonal forecasting has great potential for groundwater management in certain areas of the Central Mediterranean. However, improvements are needed in observations, data assimilation methods, and the seasonal forecasting system to ensure reliable forecasts for upper soil layers and other regions. [ABSTRACT FROM AUTHOR]

ECOOP project allowed the improvement of the regional and coastal operational forecasting systems for the different European Seas. In the Iberia-Biscay-Ireland area (IBI) a regional system has been developed and improved for the project in order to provide IBI partners the best initial and boundary conditions to their embedded coastal systems. End users could also get access to the regional hindcasts and forecasts through the ECOOP website. This system has been upgraded as follow: a first existing version V0, a second base-line version V1 ready for the ECOOP Target Operational Period and a third one V2 which consists in a new generation regional system. This paper especially pays attention to the improvements from the V1 system, whose physics are close to a large scale basin system, to the V2 one which physics are more adapted to shelf and coastal issues. Strong developments which allow further regional physics resolution in the NEMO OGCM such as tide, non linear free surface and adapted vertical mixing scheme among other have been carried out from V1 to V2 versions for the project. Thus, regional thermal fronts due to tidal mixing appear in V2 solution and are quite well placed. Moreover, simulation of the stratification in shelf areas is also improved in V2. [ABSTRACT FROM AUTHOR]

Wind-wave forecasts play a crucial role in the North American Great Lakes region towards ensuring the safety of communities, enhancement of the economy, and protection of property. Modeling wind waves in closed and relatively shallow basins with complex bathymetry like the Great Lakes is a challenge that is successfully tackled in part by using variable-resolution triangular unstructured meshes with no limits in terms of computational scalability and maximum resolution in the coastal areas. In this paper, we discuss recent advances in developing unstructured mesh capabilities as part of the spectral wave model WAVEWATCH III, in the context of National Oceanic and Atmospheric Administration (NOAA) operational requirements such as model robustness, efficiency, and accuracy. We revisit the history of developments leading to the transition from rectilinear to curvilinear grids and finally to an unstructured mesh version of NOAA's operational Great Lakes wave modeling system (GLWUv2.0). The article describes the development of the operational GLWUv2.0, from mesh design and scalability analysis to validation and verification for hindcast of storm cases and re-forecast using 4 months of retrospective simulations. In closed Great Lakes basins untouched by swell from distant sources, the atmospheric model's direct impact on wave behavior stands apart, showing reduced forecast accuracy over time, while maintaining consistent precision in accurately wind-hindcasted stormy conditions. [ABSTRACT FROM AUTHOR]

As air pollution is regarded as the single largest environmental health risk in Europe it is important that communication to the public is up to date and accurate and provides means to avoid exposure to high air pollution levels. Long- and short-term exposure to outdoor air pollution is associated with increased risks of mortality and morbidity. Up-to-date information on present and coming days' air quality helps people avoid exposure during episodes with high levels of air pollution. Air quality forecasts can be based on deterministic dispersion modelling, but to be accurate this requires detailed information on future emissions, meteorological conditions and process-oriented dispersion modelling. In this paper, we apply different machine learning (ML) algorithms – random forest (RF), extreme gradient boosting (XGB), and long short-term memory (LSTM) – to improve 1, 2, and 3 d deterministic forecasts of PM 10 , NOx , and O 3 at different sites in Greater Stockholm, Sweden. It is shown that the deterministic forecasts can be significantly improved using the ML models but that the degree of improvement of the deterministic forecasts depends more on pollutant and site than on what ML algorithm is applied. Also, four feature importance methods, namely the mean decrease in impurity (MDI) method, permutation method, gradient-based method, and Shapley additive explanations (SHAP) method, are utilized to identify significant features that are common and robust across all models and methods for a pollutant. Deterministic forecasts of PM 10 are improved by the ML models through the input of lagged measurements and Julian day partly reflecting seasonal variations not properly parameterized in the deterministic forecasts. A systematic discrepancy by the deterministic forecasts in the diurnal cycle of NOx is removed by the ML models considering lagged measurements and calendar data like hour and weekday, reflecting the influence of local traffic emissions. For O 3 at the urban background site, the local photochemistry is not properly accounted for by the relatively coarse Copernicus Atmosphere Monitoring Service ensemble model (CAMS) used here for forecasting O 3 but is compensated for using the ML models by taking lagged measurements into account. Through multiple repetitions of the training process, the resulting ML models achieved improvements for all sites and pollutants. For NOx at street canyon sites, mean squared error (MSE) decreased by up to 60 %, and seven metrics, such as R2 and mean absolute percentage error (MAPE), exhibited consistent results. The prediction of PM 10 is improved significantly at the urban background site, whereas the ML models at street sites have difficulty capturing more information. The prediction accuracy of O 3 also modestly increased, with differences between metrics. Further work is needed to reduce deviations between model results and measurements for short periods with relatively high concentrations (peaks) at the street canyon sites. Such peaks can be due to a combination of non-typical emissions and unfavourable meteorological conditions, which are rather difficult to forecast. Furthermore, we show that general models trained using data from selected street sites can improve the deterministic forecasts of NOx at the station not involved in model training. For PM 10 this was only possible using more complex LSTM models. An important aspect to consider when choosing ML algorithms is the computational requirements for training the models in the deployment of the system. Tree-based models (RF and XGB) require fewer computational resources and yield comparable performance in comparison to LSTM. Therefore, tree-based models are now implemented operationally in the forecasts of air pollution and health risks in Stockholm. Nevertheless, there is big potential to develop generic models using advanced ML to take into account not only local temporal variation but also spatial variation at different stations. [ABSTRACT FROM AUTHOR]

While deep learning (DL) models are effective in rainfall-runoff modelling, their dependence on data and lack of physical mechanisms can limit their use in hydrology. As there is yet no consensus on the consideration of the fundamental water balance for DL models, this paper presents an in-depth investigation of the effects of water balance constraint on the long-short term memory (LSTM) network. Specifically, based on the Catchment Attributes and Meteorology for Large-sample Studies (CAMELS) dataset, the LSTM and its architecturally mass-conserving variant (MC-LSTM) are trained basin-wise to provide rainfall-runoff prediction and then the robustness of the LSTM and MC-LSTM against data sparsity, random parameters initialization and contrasting climate conditions are assessed across the contiguous United States. Through large-sample tests, the results show that the water balance constraint evidently improves the robustness of the basin-wise trained LSTM. On the one hand, as the amount of training data increases from 1 year to 15 years, the incorporation of the water balance constraint into the LSTM network decreases the sensitivity from 95.0 % to 32.7 %. On the other hand, the water balance constraint contributes to the stability of the LSTM for 450 (85 %) basins when there are 3 years' training data. In the meantime, the water balance constraint improves the transferability of the LSTM from the driest years to the wettest years for 318 (67 %) basins. Overall, the in-depth investigations of this paper facilitate insights into the use of DL models for rainfall-runoff modelling. [ABSTRACT FROM AUTHOR]

Marine heatwaves (MHWs), defined as prolonged period of extremely warm sea surface temperature (SST), have been receiving a lot of attention in the past decade as their frequency and intensity increase in a warming climate. This paper investigates the extent to which the seasonal occurrence and duration of MHWs can be predicted with the European Centre for Medium-Range Weather Forecast (ECMWF) operational seasonal forecast system. The prediction of the occurrence of MHW events, the number of MHW days per season, and their intensity and spatial extent are derived from seasonal SST forecasts and evaluated against an observation-based SST analysis using both deterministic and probabilistic metrics over the 1982–2021 period. Forecast scores show useful skill in predicting the occurrence of MHWs globally for the two seasons following the starting date. The skill is the highest in the El Niño region, the Caribbean, the wider tropics, the north-eastern extra-tropical Pacific, and southwest of the extra-tropical basins. The skill is not as good for other midlatitude eastern basins nor for the Mediterranean, with the forecast system being able to represent the low-frequency modulation of MHWs but showing poor skill in predicting the interannual variability of the MHW characteristics. Linear trend analysis shows an increase in MHW occurrence at a global scale, which the forecasts capture well. [ABSTRACT FROM AUTHOR]

There remains a gap between the production of scientifically robust forecasts and the translation of these forecasts into useful information such as daily "bulletins" for decision-makers in early warning systems. There is significant published literature on best practice in communicating risk information but very little to guide and provide advice on the process of how these bulletins have been, or should be, developed. This paper reviews two case studies where bulletins were developed for national and district-level government agencies and humanitarian responders: daily reports in response to cyclones Idai and Kenneth in Mozambique and prototype landslide forecast bulletins in the Nilgiris and Darjeeling districts of India. Primary data were collected from producers and intermediaries of the bulletins via interview, and secondary data were analysed on iterative changes in the bulletin development, minutes from internal discussions, and feedback from users to extract learning on both the content and process of developing the bulletins. There were significant similarities in the type of content included in the bulletins, such as the layout, choice of words, and use of visualisation that was consistent with published best practices. Both case studies experienced challenges dealing with uncertainty, complexity, and whether to include advice. There were also similarities in the processes and approaches taken to develop the bulletins. Both case studies took an iterative approach, developed feedback mechanisms, benefitted from experienced multidisciplinary teams, and emphasised the need for strong inter-relationships and the importance and value of preparedness and protocols. A major challenge was the difficulty in balancing science capabilities, including issues related to data scarcity, with user needs, which did not become significantly easier to deal with given more time availability. In particular, there were tensions between developing new forecast products that were urgently needed by users against the limited time for testing and refinement of those forecasts and the risk of misinforming decisions due to uncertainty in the information based on limited data. The findings indicate that whilst more research is needed into existing or best practice processes to develop content for forecast bulletins, there is an existing body of experiential and intuitive knowledge and learning that already exists but that is not yet captured in an appropriate format that could be of significant interest and value to those developing forecast information. This paper goes some way to capturing some of the learning from translating scientific forecasts into useful information, in particular on both the content and the process of developing forecast bulletins for decision-making. [ABSTRACT FROM AUTHOR]

Operational global-scale hydrological forecasting systems are used to help manage hydrological extremes such as floods and droughts. The vast amounts of raw data that underpin forecast systems and the ability to generate information on forecast skill have, until now, not been publicly available. As part of the Global Flood Awareness System (GloFAS; https://www.globalfloods.eu/ , last access: 3 December 2022) service evolution, in this paper daily ensemble river discharge reforecasts and real-time forecast datasets are made free and openly available through the Copernicus Climate Change Service (C3S) Climate Data Store (CDS). They include real-time forecast data starting on 1 January 2020 updated operationally every day and a 20-year set of reforecasts and associated metadata. This paper describes the model components and configuration used to generate the real-time river discharge forecasts and the reforecasts. An evaluation of ensemble forecast skill using the continuous ranked probability skill score (CRPSS) was also undertaken for river points around the globe. Results show that GloFAS is skilful in over 93 % of catchments in the short (1 to 3 d) and medium range (5 to 15 d) against a persistence benchmark forecast and skilful in over 80 % of catchments out to the extended range (16 to 30 d) against a climatological benchmark forecast. However, the strength of skill varies considerably by location with GloFAS found to have no or negative skill at longer lead times in broad hydroclimatic regions in tropical Africa, western coast of South America, and catchments dominated by snow and ice in high northern latitudes. Forecast skill is summarised as a new headline skill score available as a new layer on the GloFAS forecast Web Map Viewer to aid user interpretation and understanding of forecast quality. [ABSTRACT FROM AUTHOR]

A novel method for three-dimensional variational assimilation of Lagrangian data with a primitive-equation ocean model is proposed. The assimilation scheme was implemented in the Mediterranean ocean Forecasting System and evaluated for a 4-month 5 period. Four experiments were designed to assess the impact of trajectory assimilation on the model output, i.e. the sea-surface height, velocity, temperature and salinity fields. It was found from the drifter and Argo trajectory assimilation experiment that the forecast skill of surface-drifter trajectories improved by 15 %, that of intermediate-depth float trajectories by 20 %, and moreover, the forecasted sea-surface height fields im10 proved locally by 5% compared to satellite data, while the quality of the temperature and salinity fields remained at previous levels. In conclusion, the addition of Lagrangian trajectory assimilation proved to reduce the uncertainties in the model fields, thus yielding a higher accuracy of the ocean forecasts. [ABSTRACT FROM AUTHOR]

Ensemble predictions of atmospheric dispersion that account for the meteorological uncertainties in a weather forecast are constructed by propagating the individual members of an ensemble numerical weather prediction forecast through an atmospheric dispersion model. Two event scenarios involving hypothetical atmospheric releases are considered: a near-surface radiological release from a nuclear power plant accident and a large eruption of an Icelandic volcano releasing volcanic ash into the upper air. Simulations were run twice-daily in real time over a 4-month period to create a large dataset of cases for this study. The performance of the ensemble predictions is measured against retrospective simulations using a sequence of meteorological fields analysed against observations. The focus of this paper is on comparing the spread of the ensemble members against forecast errors and on the calibration of probabilistic forecasts derived from the ensemble distribution. Results show good overall performance by the dispersion ensembles in both studies but with simulations for the upper-air ash release generally performing better than those for the near-surface release of radiological material. The near-surface results demonstrate a sensitivity to the release location, with good performance in areas dominated by the synoptic-scale meteorology and generally poorer performance at some other sites where, we speculate, the global-scale meteorological ensemble used in this study has difficulty in adequately capturing the uncertainty from local- and regional-scale influences on the boundary layer. The ensemble tends to be under-spread, or over-confident, for the radiological case in general, especially at earlier forecast steps. The limited ensemble size of 18 members may also affect its ability to fully resolve peak values or adequately sample outlier regions. Probability forecasts of threshold exceedances show a reasonable degree of calibration, though the over-confident nature of the ensemble means that it tends to be too keen on using the extreme forecast probabilities. Ensemble forecasts for the volcanic ash study demonstrate an appropriate degree of spread and are generally well-calibrated, particularly for ash concentration forecasts in the troposphere. The ensemble is slightly over-spread, or under-confident, within the troposphere at the first output time step T + 6, thought to be attributable to a known deficiency in the ensemble perturbation scheme in use at the time of this study, but improves with probability forecasts becoming well-calibrated here by the end of the period. Conversely, an increasing tendency towards over-confident forecasts is seen in the stratosphere, which again mirrors an expectation for ensemble spread to fall away at higher altitudes in the meteorological ensemble. Results in the volcanic ash case are also broadly similar between the three different eruption scenarios considered in the study, suggesting that good ensemble performance might apply to a wide range of eruptions with different heights and mass eruption rates. [ABSTRACT FROM AUTHOR]

Improving the prediction of clouds in shallow cumulus regimes via turbulence parameterization in the planetary boundary layer (PBL) will likely increase the global skill of global climate models (GCMs) because this cloud regime is common over tropical oceans where low cloud fraction has a large impact on Earth's radiative budget. This study attempts to improve the prediction of PBL structure in tropical trade-wind regimes in the Community Atmosphere Model (CAM) by updating its formulation of momentum flux in CLUBB (Cloud Layers Unified by Binormals), which currently does not by default allow for upgradient momentum fluxes. Hindcast CAM output from custom CLUBB configurations which permit countergradient momentum fluxes are compared to in-situ observations from weather balloons collected during the ElUcidating the RolE of Cloud–Circulation Coupling in ClimAte and Atlantic Tradewind Ocean–Atmosphere Mesoscale Interaction Campaign (EUREC4A/ATOMIC) field campaign in the Tropical Atlantic in early 2020. Comparing a version with CAM-CLUBB with a prognostic treatment of momentum fluxes results in vertical profiles that better match previously published LES results. Countergradient fluxes are frequently simulated between 950 hPa and 850 hPa over the EUREC4A/ATOMIC period in CAM-CLUBB. Further modification to the PBL parameterization by implementing a more generalized calculation of the turbulent length scale reduces model bias and RMSE relative to sounding data. Benefits are also seen in the diurnal cycle, although more systematic model errors persist. A cursory budget analysis suggests the buoyant production of momentum fluxes, both above and below the jet maximum, significantly contributes to the frequency and depth of countergradient vertical momentum fluxes in the study region. This paper provides evidence that higher-order turbulence parameterizations may offer pathways for improving the simulation of trade-wind regimes in global models. [ABSTRACT FROM AUTHOR]

This paper investigates how to refine the ground meteorological observation network to greatly improve the PM 2.5 concentration forecasts by identifying sensitive areas for targeted observations that are associated with a total of 48 forecasts in eight heavy haze events during the years of 2016–2018 over the Beijing–Tianjin–Hebei (BTH) region. The conditional nonlinear optimal perturbation (CNOP) method is adopted to determine the sensitive area of the surface meteorological fields for each forecast, and a total of 48 CNOP-type errors are obtained including wind, temperature, and water vapor mixing ratio components. It is found that, although all the sensitive areas tend to locate within and/or around the BTH region, their specific distributions are dependent on the events and the start times of the forecasts. Based on these sensitive areas, the current ground meteorological stations within and around the BTH region are refined to form a cost-effective observation network, which makes the relevant PM 2.5 forecasts starting from different initial times for varying events assimilate fewer observations, but overall, it achieve the forecasting skill comparable to and even higher than that obtained by assimilating all ground station observations. This network sheds light on the idea that some of the current ground stations within and around the BTH region are very useless for improving the PM 2.5 forecasts in the BTH region and can be greatly scattered to avoid unnecessary work. [ABSTRACT FROM AUTHOR]

The 100th anniversary of the Liverpool Tidal Institute (LTI) was celebrated during 2019. One aspect of tidal science for which the LTI acquired a worldwide reputation was the development and use of tide prediction machines (TPMs). The TPM was invented in the late 19th century, but most of them were made in the first half of the 20th century, up until the time that the advent of digital computers consigned them to museums. This paper describes the basic principles of a TPM, reviews how many were constructed around the world and discusses the method devised by Arthur Doodson at the LTI for the determination of harmonic tidal constants from tide gauge data. These constants were required in order to set up the TPMs for predicting the heights and times of the tides. Although only 3 of the 30-odd TPMs constructed were employed in operational tidal prediction at the LTI, Doodson was responsible for the design and oversight of the manufacture of several others. The paper demonstrates how the UK, and the LTI and Doodson in particular, played a central role in this area of tidal science. [ABSTRACT FROM AUTHOR]

The South China Sea Operational Oceanography Forecasting System (SCSOFS), constructed and operated by the National Marine Environmental Forecasting Center of China, has been providing daily updated hydrodynamic forecasting in the South China Sea (SCS) for the next 5 d since 2013. This paper presents recent comprehensive updates to the configurations of the physical model and data assimilation scheme in order to improve the forecasting skill of the SCSOFS. This paper highlights three of the most sensitive updates: the sea surface atmospheric forcing method, the discrete tracer advection scheme, and a modification of the data assimilation scheme. Intercomparison and accuracy assessment among the five sub-versions were performed during the entire upgrading process using the OceanPredict Intercomparison and Validation Task Team Class 4 metrics. The results indicate that remarkable improvements have been made to the SCSOFSv2 with respect to the original version (known as SCSOFSv1). The domain-averaged monthly mean root-mean-square errors of the sea surface temperature and sea level anomaly have decreased from 1.21 to 0.52 ∘ C and from 21.6 to 8.5 cm, respectively. [ABSTRACT FROM AUTHOR]

Container technology provides a pathway to facilitate easy access to unified modeling systems and opens opportunities for collaborative model development and interactive learning. In this paper, we present the implementation of software containers for the System for High-resolution prediction on Earth-to-Local Domains (SHiELD), a unified atmospheric model for weather-to-seasonal prediction. The containerized SHiELD is cross-platform and easy to install. Flexibility of the containerized SHiELD is demonstrated as it can be configured as a global, a global–nest, and a regional model. Bitwise reproducibility is achieved on various x86 systems tested in this study. Performance and scalability of the containerized SHiELD are evaluated and discussed. [ABSTRACT FROM AUTHOR]

Accurately predicting the seasonal streamflow supply (SSS), i.e., the inflow into a reservoir accumulated during the snowmelt season (April to August), is critical to operating hydroelectric dams and avoiding hydrology-related hazard. Such forecasts generally involve numerical models that simulate the hydrological evolution of a basin. The operational department of the French electric company Electricité de France (EDF) implements a semi-distributed model and has carried out such forecasts for several decades on about 50 basins. However, both scarce observation data and oversimplified physics representation may lead to significant forecast errors. Data assimilation has been shown to be beneficial for improving predictions in various hydrological applications, yet very few have addressed the seasonal streamflow supply prediction problem. More specifically, the assimilation of snow observations, though available in various forms, has been rarely studied, despite the possible sensitivity of the streamflow supply to snow stock. This is the goal of the present paper. In three mountainous basins, a series of four ensemble data assimilation experiments – assimilating (i) the streamflow (Q) alone, (ii) Q and fractional snow cover (FSC) data, (iii) Q and local cosmic ray snow sensor (CRS) data and (iv) all the data combined – is compared to the climatologic ensemble and an ensemble of free simulations. The experiments compare the accuracy of the estimated streamflows during the reanalysis (or assimilation) period September to March, during the forecast period April to August, and the SSS estimation. The results show that Q assimilation notably improves streamflow estimations during both reanalysis and the forecast period. Also, the additional combination of CRS and FSC data to the assimilation further ameliorates the SSS prediction in two of the three basins. In the last basin, the experiments highlight a poor representativity of the CRS observations during some years and reveal the need for an enhanced observation system. [ABSTRACT FROM AUTHOR]

This paper presents a methodological framework designed for the event-based evaluation of short-range hydrometeorological ensemble forecasts, in the specific context of an intense flash-flood event characterized by high spatiotemporal variability. The proposed evaluation adopts the point of view of end users in charge of the organization of evacuations and rescue operations at a regional scale. Therefore, the local exceedance of discharge thresholds should be anticipated in time and accurately localized. A step-by-step approach is proposed, including first an evaluation of the rainfall forecasts. This first step helps us to define appropriate spatial and temporal scales for the evaluation of flood forecasts. The anticipation of the flood rising limb (discharge thresholds) is then analyzed at a large number of ungauged sub-catchments using simulated flows and zero-future rainfall forecasts as references. Based on this second step, several gauged sub-catchments are selected, at which a detailed evaluation of the forecast hydrographs is finally achieved. This methodology is tested and illustrated for the October 2018 flash flood which affected part of the Aude River basin (southeastern France). Three ensemble rainfall nowcasting research products recently proposed by Météo-France are evaluated and compared. The results show that, provided that the larger ensemble percentiles are considered (75th percentile for instance), these products correctly retrieve the area where the larger rainfall accumulations were observed but have a tendency to overestimate its spatial extent. The hydrological evaluation indicates that the discharge threshold exceedances are better localized and anticipated if compared to a naive zero-future rainfall scenario but at the price of a significant increase in false alarms. Some differences in the performances between the three ensemble rainfall forecast products are also identified. Finally, even if the evaluation of ensemble hydrometeorological forecasts based on a low number of documented flood events remains challenging due to the limited statistical representation of the available data, the evaluation framework proposed herein should contribute to draw first conclusions about the usefulness of newly developed rainfall forecast ensembles for flash-flood forecasting purpose and about their limits and possible improvements. [ABSTRACT FROM AUTHOR]

There is a scientific consensus that the Mediterranean region (MEDR) is warming and as the temperature continues to rise, extreme events such as droughts and heat waves are becoming more frequent, severe, and widespread. Given the detrimental effects of droughts, it is crucial to accelerate the development of forecasting and early warning systems to minimize their negative impact. This paper examines the current state of knowledge in drought modeling and prediction using statistical, dynamical, and hybrid statistical-dynamical models, and suggests some research prospects to further improve drought prediction in this region. The review finds that while all methods have their strengths and shortcomings, hybrid statistical-dynamical methods can perform the most skillful prediction with a long lead time. However, the application of these methods is still challenging due to the lack of high-quality observational data and the limited computational resources. Finally, the paper concludes by discussing the importance of using a combination of sophisticated methods such as data assimilation techniques, machine learning models, and copula models and integrating data from different sources (e.g., remote sensing data, in-situ measurements, and reanalysis) to improve the accuracy and efficiency of drought forecasting. [ABSTRACT FROM AUTHOR]

Ensemble predictions of atmospheric dispersion that account for the meteorological uncertainties in a weather forecast are constructed by propagating the individual members of an ensemble numerical weather prediction forecast through an atmospheric dispersion model. Two event scenarios involving hypothetical atmospheric releases are considered: a near-surface radiological release from a nuclear power plant accident, and a large eruption of an Icelandic volcano releasing volcanic ash into the upper air. Simulations were run twice-daily in real time over a four month period to create a large data set of cases for this study. Performance of the ensemble predictions is measured against retrospective simulations using analysed meteorological fields. The focus of this paper is on comparing the spread of the ensemble members against forecast errors and on the calibration of probabilistic forecasts derived from the ensemble distribution. Results show good overall performance by the dispersion ensembles in both studies, but with simulations for the upper air ash release generally performing better than those for the near-surface release of radiological material. The near-surface results demonstrate a sensitivity to the release location, with good performance in areas dominated by the synoptic-scale meteorology and generally poorer performance at some other sites where, we speculate, the global-scale meteorological ensemble used in this study has difficulty in adequately capturing the uncertainty from local and regional scale influences on the boundary layer. The ensemble tends to be under-spread, or over-confident, for the radiological case in general, especially at earlier forecast steps. The limited ensemble size of 18 members may also affect its ability to fully resolve peak values or adequately sample outlier regions. Probability forecasts of threshold exceedances show a reasonable degree of calibration, though the over-confident nature of the ensemble means that it tends to be too keen on using the extreme forecast probabilities. Ensemble forecasts for the volcanic ash study demonstrate an appropriate degree of spread and are generally well-calibrated, particularly for ash concentration forecasts in the troposphere. The ensemble is slightly over-spread, or under-confident, within the troposphere at the first output time step T+6, thought to be attributable to a known deficiency in the ensemble perturbation scheme in use at the time of this study, but improves with probability forecasts becoming well-calibrated here by the end of the period. Conversely, an increasing tendency towards over-confident forecasts is seen in the stratosphere, which again mirrors an expectation for ensemble spread to fall away at higher altitudes in the met ensemble. Results in the volcanic ash case are also broadly similar between the three different eruption scenarios considered in the study, suggesting that good ensemble performance might apply to a wide range of eruptions with different heights and mass eruption rates. [ABSTRACT FROM AUTHOR]

For modeling flow and transport processes in the soil–plant–atmosphere system, knowledge of the unsaturated hydraulic properties in functional form is mandatory. While much data are available for the water retention function, the hydraulic conductivity function often needs to be predicted. The classical approach is to predict the relative conductivity from the retention function and scale it with the measured saturated conductivity, Ks. In this paper we highlight the shortcomings of this approach, namely, that measured Ks values are often highly uncertain and biased, resulting in poor predictions of the unsaturated conductivity function. We propose to reformulate the unsaturated hydraulic conductivity function by replacing the soil-specific Ks as a scaling factor with a generally applicable effective saturated tortuosity parameter τs and predicting total conductivity using only the water retention curve. Using four different unimodal expressions for the water retention curve, a soil-independent general value for τs was derived by fitting the new formulation to 12 data sets containing the relevant information. τs was found to be approximately 0.1. Testing of the new prediction scheme with independent data showed a mean error between the fully predicted conductivity functions and measured data of less than half an order of magnitude. The new scheme can be used when insufficient or no conductivity data are available. The model also helps to predict the saturated conductivity of the soil matrix alone and thus to distinguish between the macropore conductivity and the soil matrix conductivity. [ABSTRACT FROM AUTHOR]

On average every 2 years, the stratospheric polar vortex exhibits extreme perturbations known as sudden stratospheric warmings (SSWs). The impact of these events is not limited to the stratosphere: but they can also influence the weather at the surface of the Earth for up to 3 months after their occurrence. This downward effect is observed in particular for SSW events with extended recovery timescales. This long-lasting stratospheric impact on surface weather can be leveraged to significantly improve the performance of weather forecasts on timescales of weeks to months. In this paper, we present a fully data-driven procedure to improve the performance of long-range forecasts of the stratosphere around SSW events with an extended recovery. We first use unsupervised machine learning algorithms to capture the spatio-temporal dynamics of SSWs and to create a continuous scale index measuring both the frequency and the strength of persistent stratospheric perturbations. We then uncover three-dimensional spatial patterns maximizing the correlation with positive index values, allowing us to assess when and where statistically significant early signals of SSW occurrence can be found. Finally, we propose two machine learning (ML) forecasting models as competitors for the state-of-the-art sub-seasonal European Centre for Medium-Range Weather Forecasts (ECMWF) numerical prediction model S2S (sub-seasonal to seasonal): while the numerical model performs better for lead times of up to 25 d, the ML models offer better predictive performance for greater lead times. We leverage our best-performing ML forecasting model to successfully post-process numerical ensemble forecasts and increase their performance by up to 20%. [ABSTRACT FROM AUTHOR]

This paper reviews the state of the art in storm surge forecasting and its particular application in the northern Adriatic Sea. The city of Venice already depends on operational storm surge forecasting systems to warn the population and economy of imminent flood threats, as well as help to protect the extensive cultural heritage. This will be more important in the future, with the new mobile barriers called MOSE (MOdulo Sperimentale Elettromeccanico , Experimental Electromechanical Module) that will be completed by 2021. The barriers will depend on accurate storm surge forecasting to control their operation. In this paper, the physics behind the flooding of Venice is discussed, and the state of the art of storm surge forecasting in Europe is reviewed. The challenges for the surge forecasting systems are analyzed, especially in view of uncertainty. This includes consideration of selected historic extreme events that were particularly difficult to forecast. Four potential improvements are identified: (1) improve meteorological forecasts, (2) develop ensemble forecasting, (3) assimilation of water level measurements and (4) develop a multimodel approach. [ABSTRACT FROM AUTHOR]

Springback is an inevitable problem in the local bending process of hull plates, which leads to low processing efficiency and affects the assembly accuracy. Therefore, the prediction of the springback effect, as a result of the local bending of hull plates, bears great significance. This paper proposes a springback prediction model based on a backpropagation neural network (BPNN), considering geometric and process parameters. Genetic algorithm (GA) and improved particle swarm optimization (PSO) algorithms are used to improve the global search capability of BPNN, which tends to fall into local optimal solutions, in order to find the global optimal solution. The result shows that the proposed springback prediction model, based on the BPNN optimized by genetic algorithm, is faster and offers smaller prediction error on the springback due to local bending. [ABSTRACT FROM AUTHOR]

During the last decades, the process of explaining life-threatening natural hazards to the public has become a major public issue from the point of view of effective prevention policies. The avalanche risk and the communication methods aimed at its forecasting and prevention constitute the focus of this paper. Among the strategies for an effective communication of environmental risks, cartography plays a pivotal role. It has proved to be essential not only for communication purposes, but also for the planning of prompt and efficient preventive interventions; in so doing, it contributes to the reduction of avalanche-caused damages and deaths. The paper investigates prevention and forecasting activities of the Meteomont Service of the Alpine Troops Command (COMTA) of Bolzano (capital city of the province of South Tyrol - North Italy), resulting in the daily publication of avalanche bulletins (Bollettini valanghe), which also include hazard maps. Specifically, the phases that contribute to the production of the avalanche bulletin and the embedded avalanche risk maps will be firstly examined; secondly, such maps will be analysed in order to assess their communicative potential for the purpose of a correct interpretation aimed at the effective prevention of snow-related risks in mountain areas. Possible improvement will be proposed on the basis of the experience of several avalanche warning services worldwide. [ABSTRACT FROM AUTHOR]

In this paper we define RAL2 – the second Regional Atmosphere and Land (RAL) science configuration for regional modelling. RAL2 uses the Unified Model (UM) as the basis for the atmosphere and the Joint UK Land Environment Simulator (JULES) for the land. RAL2 defines the science configuration of the dynamics and physics schemes of the atmosphere and land and builds on the baseline of RAL1. There are two RAL2 sub-releases, one for mid-latitudes (RAL2-M) and one for tropical regions (RAL2-T). We document the differences between them and where appropriate discuss how RAL2 relates to RAL1 and the corresponding configuration of the global forecasting model. Our results show an increase in medium and low cloud amounts in the mid-latitudes leading to improved cloud forecasts. The increase in cloud amount leads to a reduced diurnal cycle of screen temperature. There is also a reduction in the frequency of heavier precipitation rates. RAL2 is expected to be the last RAL science configuration with two sub-releases as research effort is focused on producing a single defined configuration of the model that performs effectively in all regions of the world. [ABSTRACT FROM AUTHOR]

This paper presents an initial analysis of predicting time series derived from long-term interferometric Synthetic Aperture Radar (InSAR) data. Time series analysis provides insights into the temporal evolution, variation, and dynamic nature of events. In this study, we focus on the Istanbul region, which is the most populous city in Turkey and spans both Europe and Asia. While the area is prone to seismic risks caused by active tectonic faults, it is also susceptible to other risks due to various phenomena. Therefore, this study investigates landslides triggered by geological structure and human-induced activities, particularly in Tepekent, a landslide-prone area in the town of Buyukcekmece located on the European side. We utilized the StaMPS persistent scatterer InSAR (PSI) method to detect slow movements over time. A total of 157 archive Copernicus Sentinel-1 data, acquired over the region between June 2017 and August 2022, were processed, primarily observing human structures with a maximum displacement amount of approximately 1 cm/year. About 500 persistent scatterer points were identified in the region, and the time series was formed by taking the average of these points. The Long Short-Term Memory (LSTM) neural network method was used to estimate motion. We trained the model with data from the first three years of the time series and used the data from the remaining two years for estimation, while the accuracy analysis was performed with the 5-year time series data. The RMSE values for the training and test data were determined to be 0.725 mm/year and 0.656 mm/year, respectively. Additionally, we estimated the time series for the period from August 2022 to August 2024. The observation and prediction results could be beneficial in developing efficient mitigation risk actions and sustainable urban management strategies. [ABSTRACT FROM AUTHOR]

This paper investigates the influence of multi-decadal climatic variability on the temporal evolution of root zone storage capacities (Sr,max) and its implications for streamflow predictions at the catchment scale. Through a comprehensive analysis of 286 catchments across Europe and the US, we analyse the deviations in evaporative ratios (IE) from expected values based on catchment aridity (IA) and their subsequent impact on Sr,max predictions. Our findings reveal that while catchments do not strictly adhere to their specific parametric Budyko curves over time, the deviations in IE are generally very minor, with an average ΔIE = 0.01 and an interquartile range IQR= -0.01 to 0.03. Consequently, these minor deviations lead to limited changes in predictions of Sr,max, mostly ranging between -10.5 and +21.5 mm (-5.1 % to +9.9 %). When these uncertainties in Sr,max are incorporated into hydrological models, the impact on streamflow predictions is found to be marginal, with the most significant shifts in monthly evaporation and streamflow not exceeding 4 % and 12 %, respectively. Our study underscores the utility of parametric Budyko-style equations for first order estimates of future Sr,max in hydrological models, even in the face of climate change and variability. This research contributes to a more nuanced understanding of hydrological responses to changing climatic conditions and offers valuable insights for future climate impact studies in hydrology. [ABSTRACT FROM AUTHOR]

Despite the wealth of existing climate forecast data, only a small part is effectively exploited for sectoral applications. A major cause of this is the lack of integrated tools that allow the translation of data into useful and skillful climate information. This barrier is addressed through the development of an R package. Climate Services Toolbox (CSTools) is an easy-to-use toolbox designed and built to assess and improve the quality of climate forecasts for seasonal to multi-annual scales. The package contains process-based, state-of-the-art methods for forecast calibration, bias correction, statistical and stochastic downscaling, optimal forecast combination, and multivariate verification, as well as basic and advanced tools to obtain tailored products. Due to the modular design of the toolbox in individual functions, the users can develop their own post-processing chain of functions, as shown in the use cases presented in this paper, including the analysis of an extreme wind speed event, the generation of seasonal forecasts of snow depth based on the SNOWPACK model, and the post-processing of temperature and precipitation data to be used as input in impact models. [ABSTRACT FROM AUTHOR]

This paper analyses the variability of the sea level in the Mediterranean Sea and its reproduction with a barotropic model, with and without applying data assimilation. The impact of data assimilation is considered in hindcast and forecast simulations, considering its usefulness for both reanalysis studies and short-term forecasts. We used a two-dimensional finite element barotropic model with an ensemble Kalman filter, assimilating hourly sea-level observations from 50 stations along all the Mediterranean coasts. The results show a great improvement given by data assimilation in hindcast simulations for the reproduction of astronomical tide, surge and total sea level, even in coastal areas far from the assimilated stations (e.g. the Eastern Mediterranean Sea). The improvement is consistent also in forecasts, especially for the first day (~37 % average error reduction), and in case of storm surge events with a strong presence of seiche oscillations. Since these oscillations depend only on the initial state and not on the boundary conditions, they are corrected very effectively by data assimilation. Finally, based on observations, this article estimates the periods of the normal barotropic modes (seiches) in the Adriatic Sea, where they have been extensively studied, and in the Mediterranean Sea, where the present documentation is scarce. [ABSTRACT FROM AUTHOR]

Accurate crime prediction plays an important role in public safety, providing technical guidance and decision support for the police and government departments. Due to the dynamics and imbalance of crime distribution, it is difficult to build predictive models for it. Specifically, the fine-grained and non-linear spatiotemporal dependencies of crime data cannot be captured accurately. In this paper, a neural network model ST-ACLCrime based on ConvLSTM and SE block was proposed to predict the number of theft crimes in hotspot areas. By overlaying ConvLSTM layers, fine-grained spatiotemporal dependencies are captured while preserving spatial location information. To further enhance the global channel feature representation, SE block is used to recalibrate the channel features and enhance the channel inter-dependencies. In addition, the closeness and the period components are set to dynamically capture the dependence of different time trends. We choose the city of Chicago as the study case, and use a multi-level spatial grid to divide the whole city area. The experimental results show that the proposed model exceeds all baseline model, such as HA, CNN, LSTM, CNN-LSTM and ConvLSTM. It was effectively capturing spatiotemporal dependence and improving prediction accuracy. [ABSTRACT FROM AUTHOR]

Flash floods are among the most devastating and lethal natural hazards. In 2018, three flash-flood episodes resulted in 46 casualties in the deserts of Israel and Jordan alone. This paper presents the hydrometeorological analysis and forecasting of a substantial storm (25–27 April 2018) that hit an arid desert basin (Zin, ∼1400km2 , southern Israel) claiming 12 human lives. This paper aims to (a) spatially assess the severity of the storm, (b) quantify the timescale of the hydrological response, and (c) evaluate the available operational precipitation forecasting. Return periods of the storm's maximal rain intensities were derived locally at 1 km2 resolution using weather radar data and a novel statistical methodology. A high-resolution grid-based hydrological model was used to study the intra-basin flash-flood magnitudes which were consistent with direct information from witnesses. The model was further used to examine the hydrological response to different forecast scenarios. A small portion of the basin (1 %–20 %) experienced extreme precipitation intensities (75- to 100-year return period), resulting in a local hydrological response of a high magnitude (10- to 50-year return period). Hillslope runoff, initiated minutes after the intense rainfall occurred, reached the streams and resulted in peak discharge within tens of minutes. Available deterministic operational precipitation forecasts poorly predicted the hydrological response in the studied basins (tens to hundreds of square kilometers) mostly due to location inaccuracy. There was no gain from assimilating radar estimates in the numerical weather prediction model. Therefore, we suggest using deterministic forecasts with caution as it might lead to fatal decision making. To cope with such errors, a novel cost-effective methodology is applied by spatially shifting the forecasted precipitation fields. In this way, flash-flood occurrences were captured in most of the subbasins, resulting in few false alarms. [ABSTRACT FROM AUTHOR]

Agricultural drought, which occurs due to a significant reduction in the moisture required for vegetation growth, is the most complex amongst all drought categories. The onset of agriculture drought is slow and can occur over vast areas with varying spatial effects, differing in areas with a particular vegetation land cover or specific agro-ecological sub-regions. These spatial variations imply that monitoring and forecasting agricultural drought require complex models that consider the spatial variations in a given region of interest. Hierarchical Bayesian models are suited for modelling such complex systems. Using partially pooled data with sub-groups that characterise spatial differences, these models can capture the sub-group variation while allowing flexibility and information sharing between these sub-groups. This paper's objective is to improve the accuracy and precision of agricultural drought forecasting in spatially diverse regions with a hierarchical Bayesian model. Results showed that the hierarchical Bayesian model was better at capturing the variability for the different agro-ecological zones and vegetation land covers compared to a regular Bayesian auto-regression distributed lags model. The forecasted vegetation condition and associated drought probabilities were more accurate and precise with the hierarchical Bayesian model at 4- to 10-week lead times. Forecasts from the hierarchical model exhibited higher hit rates with a low probability of false alarms for drought events in semi-arid and arid zones. The hierarchical Bayesian model also showed good transferable forecast skills over counties not included in the training data. [ABSTRACT FROM AUTHOR]

As a challenge in the construction of a "seamless forecast" system, improving the prediction skills of subseasonal forecasts is a key issue for meteorologists. In view of the evolution characteristics of numerical models and recent deep learning models for subseasonal forecasts, as forecast times increase, forecast results tend to become intraseasonal low-frequency components, which are essential to the change in general circulation on the subseasonal timescale as well as persistent extreme weather. In this paper, the Global Subseasonal Forecast Model (GSFM v1.0) first extracted the intraseasonal oscillation (ISO) components of atmospheric signals and used an improved deep learning model (SE-ResNet) to train and predict the ISO components of geopotential height at 500 hPa (Z500) and temperature at 850 hPa (T850). The results show that the 10-30 day prediction performance of the model used in this paper is better than that of the model trained directly with original data. Compared with other models/methods, the SE-ResNet model has a good ability to depict the subseasonal evolution of the ISO components of Z500 and T850. In particular, although the CFSv2 results have a better prediction performance through 10 days, the SE-ResNet model is substantially superior to CFSv2 through 10-30 day, especially in the middle and high latitudes. The SE-ResNet model also has a better effect in predicting 3-8 planetary waves, which leads to the difference in model prediction performance in extratropical areas. A case study shows that the SE-ResNet model depicted the phase change and propagation characteristics of planetary waves well. Thus, the application of data-driven subseasonal forecasts of atmospheric ISO components may shed light on improving the skill of seasonal forecasts. [ABSTRACT FROM AUTHOR]

Modern machine learning, especially deep learning, which is used in a variety of applications, requires a lot of labelled data for model training. Having an insufficient amount of training examples leads to models which do not generalize well to new input instances. This is a particular significant problem for tasks involving aerial images: often training data is only available for a limited geographical area and a narrow time window, thus leading to models which perform poorly in different regions, at different times of day, or during different seasons. Domain adaptation can mitigate this issue by using labelled source domain training examples and unlabeled target domain images to train a model which performs well on both domains. Modern adversarial domain adaptation approaches use unpaired data. We propose using pairs of semantically similar images, i.e., whose segmentations are accurate predictions of each other, for improved model performance. In this paper we show that, as an upper limit based on ground truth, using semantically paired aerial images during training almost always increases model performance with an average improvement of 4.2% accuracy and.036 mean intersection-over-union (mIoU). Using a practical estimate of semantic similarity, we still achieve improvements in more than half of all cases, with average improvements of 2.5% accuracy and.017 mIoU in those cases. [ABSTRACT FROM AUTHOR]

Dust storms pose significant threats to human health and property. Accurate forecasting is crucial for taking precautionary measures. Dust models have suffered from uncertainties from emission and transport factors. Data assimilation can help refine biased models by incorporating available observations, leading to improved analyses and forecasts. The Ensemble Kalman Filter (EnKF) is a widely-used assimilation algorithm that effectively tunes models, particularly in terms of intensity adjustment. However, when the position of the simulation does not align consistently with the observations which is referred to as position error, the EnKF algorithm struggles. This is because its background covariance normally represents intensity uncertainty, while the positional errors in the long distance transport are difficult to be quantified and were usually neglected. In this paper, we propose a novel Neighboring Time Ensemble Kalman Filter (NTEnKF). In addition to the original ensembles quantifying dust loading variation, this methodology introduces extra ensembles from neighboring time for describing the potential spread of dust position. The enlarged ensemble captures both intensity and positional errors, allowing observations to be thoroughly resolved into the assimilation calculations. We tested this method on three major dust storm events that occurred in spring 2021. The results show that position error significantly degraded dust forecasting in terms of RMSE and NMB, and hindered the EnKF from assimilating valid observations. In contrast, the NTEnKF yielded substantial improvements in both dust analysis fields and forecasts compared to the EnKF. [ABSTRACT FROM AUTHOR]

The impact of assimilating GNSS-ZTD (global navigation satellite system–zenith total delay) on the precipitable water vapor and precipitation forecast over Italy is studied for the month of October 2019, which was characterized by several moderate to intense precipitation events, especially over northwestern Italy. The WRF (Weather Research and Forecasting) model, version 4.1.3, is used with its 3D-Var data assimilation system to assimilate ZTD observations from 388 GNSS receivers distributed over the country. The dataset was built collecting data from all the major national and regional GNSS permanent networks, achieving dense coverage over the whole area. The water vapor forecast is verified for the forecast hours of 1–6 h after the last data assimilation time. Results show that WRF underestimates the atmospheric water vapor content for the period, and GNSS-ZTD data assimilation improves this underestimation. The precipitation forecast is verified in the phases of 0–3 and 3–6 h after the last data assimilation time using more than 3000 rain gauges spread over Italy. The application of GNSS-ZTD data assimilation to a case study improved the precipitation forecast by increasing the rainfall maximum and by better focusing the precipitation pattern over northeastern Italy, with the main drawback being the prediction of false alarms. Considering the study over the whole period, GNSS-ZTD data assimilation had a positive impact on rainfall forecast, with an improvement in the performance up to 6 h and with statistically significant results for moderate to intense rainfall thresholds (25–30 mm (3 h) -1). [ABSTRACT FROM AUTHOR]

Marine Heatwaves (MHWs), defined as prolonged period of extremely warm Sea Surface Temperature (SST), have been receiving a lot of attention in the past decade as their frequency and intensity increase in a warming climate. This paper investigates the extent to which the seasonal occurrence and duration of MHWs can be predicted with the European Centre for Medium-range Weather Forecast (ECMWF) operational seasonal forecast system. The prediction of the occurrence of MHW events, the number of MHW days per season, their intensity and spatial extent is derived from SST seasonal forecasts and evaluated against an observation-based SST analysis using both deterministic and probabilistic metrics over the 1982–2021 period. Forecast scores show useful skill in predicting the occurrence of MHWs globally for the two seasons following the starting date. The skill is the highest in the El-Niño region, the Caribbean, the wider Tropics, the Northeast Extra-tropical Pacific and Southwest of the Extra-tropical basins. The skill is not as good for other midlatitudes eastern basins, nor for the Mediterranean, the forecast system being able to represent the low frequency modulation of MHWs but showing poor skill in predicting the interannual variability of the MHW characteristics. Linear trend analysis shows an increase of MHW occurrence at a global scale, which the forecasts capture well. [ABSTRACT FROM AUTHOR]

Marine Heatwaves (MHWs), defined as prolonged period of extremely warm Sea Surface Temperature (SST), have been receiving a lot of attention in the past decade as their frequency and intensity increase in a warming climate. This paper investigates the extent to which the seasonal occurrence and duration of MHWs can be predicted with the European Centre for Medium-range Weather Forecast (ECMWF) operational seasonal forecast system. The prediction of the occurrence of MHW events, the number of MHW days per season, their intensity and spatial extent is derived from SST seasonal forecasts and evaluated against an observation-based SST analysis using both deterministic and probabilistic metrics over the 1982-2021 period. Forecast scores show useful skill in predicting the occurrence of MHWs globally for the two seasons following the starting date. The skill is the highest in the El-Niño region, the Caribbean, the wider Tropics, the Northeast Extra-tropical Pacific and Southwest of the Extra-tropical basins. The skill is not as good for other midlatitudes eastern basins, nor for the Mediterranean, the forecast system being able to represent the low frequency modulation of MHWs but showing poor skill in predicting the interannual variability of the MHW characteristics. Linear trend analysis shows an increase of MHW occurrence at a global scale, which the forecasts capture well. [ABSTRACT FROM AUTHOR]

Land subsidence is a geological disaster. It will lead to the decline of land elevation, resulting in the potential safety hazards of urban facilities. Thus, the prediction of land subsidence displacement is significant. Among the existing prediction methods, the methods based on the time-series prediction model only analyze the settlement series data without considering the settlement mechanism, so they are easy to apply. However, they less consider the influence of other factors on land subsidence. Besides, they independently input displacement time-series data from different monitoring points without considering their relationship. To solve these problems, we take the monitoring point as the entity and take the DEM, soil type, building height, and land subsidence displacement sequence at the corresponding position of the monitoring point as the attributes to construct the knowledge graph. And then, we propose a framework Graph-TAP for modeling temporal attribute prediction of the knowledge graph's entity. This framework learns the representation of events with the specific entity at first. Then it captures the temporal dependency between historical events using GRU. Finally, it predicts the entity's displacement attribute. We randomly selected 61 subsidence monitoring points in Shenzhen, China. We used the land subsidence displacement InSAR time-series data (12-day time interval) and other attribute data from June 22, 2015, to April 5, 2016, for model training, validation, and testing. The experimental results show that our method is better than the time-series prediction based on the LSTM model and the DARNET (a knowledge graph temporal attribute prediction framework). [ABSTRACT FROM AUTHOR]

Volcanic ash advisories are produced by specialised forecasters who combine several sources of observational data and volcanic ash dispersion model outputs based on their subjective expertise. These advisories are used by the aviation industry to make decisions about where it is safe to fly. However, both observations and dispersion model simulations are subject to various sources of uncertainties that are not represented in operational forecasts. Quantification and communication of these uncertainties are fundamental for making more informed decisions. Here, we develop a data assimilation method that combines satellite retrievals and volcanic ash transport and dispersion model (VATDM) output, considering uncertainties in both data sources. The methodology is applied to a case study of the 2019 Raikoke eruption. To represent uncertainty in the VATDM output, 1000 simulations are performed by simultaneously perturbing the eruption source parameters, meteorology, and internal model parameters (known as the prior ensemble). The ensemble members are filtered, based on their level of agreement with the ash column loading, and their uncertainty, of the Himawari–8 satellite retrievals, to produce a constrained posterior ensemble. For the Raikoke eruption, filtering the ensemble skews the values of mass eruption rate towards the lower values within the wider parameters ranges initially used in the prior ensemble (mean reduces from 1 to 0.1 Tg h -1). Furthermore, including satellite observations from subsequent times increasingly constrains the posterior ensemble. These results suggest that the prior ensemble leads to an overestimate of both the magnitude and uncertainty in ash column loadings. Based on the prior ensemble, flight operations would have been severely disrupted over the Pacific Ocean. Using the constrained posterior ensemble, the regions where the risk is overestimated are reduced, potentially resulting in fewer flight disruptions. The data assimilation methodology developed in this paper is easily generalisable to other short duration eruptions and to other VATDMs and retrievals of ash from other satellites. [ABSTRACT FROM AUTHOR]

This paper presents a phenomenological framework for forecasting the area-integrated fire radiative power from wildfires. In the method, a region of interest is covered with a regular grid, whose cells are uniquely and independently parameterized with regard to the fire intensity according to (i) the fire incidence history, (ii) the retrospective meteorological information, and (iii) remotely sensed high-temporal-resolution fire radiative power taken together with (iv) consistent cloud mask data. The parameterization is realized by fitting the predetermined functions for diurnal and annual profiles of fire radiative power to the remote-sensing observations. After the parametrization, the input for the fire radiative power forecast is the meteorological data alone, i.e. the weather forecast. The method is tested retrospectively for south-central African savannah areas with the grid cell size of 1.5∘×1.5∘. The input data included ECMWF ERA5 meteorological reanalysis and SEVIRI/MSG (Spinning Enhanced Visible and Infra-Red Imager on board Meteosat Second Generation) fire radiative power and cloud mask data. It has been found that in the areas with a large number of wildfires regularly ignited on a daily basis during dry seasons from year to year, the temporal fire radiative power evolution is quite predictable, whereas the areas with irregular fire behaviour, predictability was low. The predictive power of the method is demonstrated by comparing the predicted fire radiative power patterns and fire radiative energy values against the corresponding remote-sensing observations. The current method showed good skills for the considered African regions and was useful in understanding the challenges in predicting the wildfires in a more general case. [ABSTRACT FROM AUTHOR]