Search

Ensemble-based data assimilation methods often suffer sampling errors due to limited ensemble sizes and model errors, which can result in filter divergence. One method to avoid filter divergence is inflation, which inflates ensemble perturbations to increase ensemble spread and account for model errors. The commonly applied inflation methods, including the multiplicative inflation, relaxation to prior spread (RTPS), and additive inflation, often use a constant inflation parameter. To capture different error growths at different scales, a scale-dependent inflation is proposed here, which applies different inflation magnitudes for variables associated with different scales. Results from the two-scale Lorenz05 model III show that for the ensemble square root filter (EnSRF) and integrated hybrid ensemble Kalman filter with ensemble mean updated by hybrid background error covariances (IHEnKF-Mean), scale-dependent inflation is superior to constant inflation. Constant inflation overinflates small-scale variables and results in increased small-scale errors, which then propagate to large-scale variables through the coupling between large- and small-scale variables and lead to increased large-scale errors. Scale-dependent inflation applies larger inflation for large-scale variables and imposes no inflation for small-scale variables, since large-scale errors have larger magnitudes than small-scale ones and small-scale errors grow faster than large-scale ones. But IHEnKF-Ensemble that updates both the ensemble mean and perturbations with hybrid background error covariances is much less sensitive to scale-dependent inflation, compared to EnSRF and IHEnKF-Mean, because updating ensemble perturbations with hybrid background error covariances can play a role similar to the scale-dependent inflation. Significance Statement: Ensemble-based data assimilation and ensemble forecasts often have smaller ensemble spread than errors. Strategies used by ensemble-based data assimilation to combat insufficient ensemble spread usually focus on the short-term ensemble forecasts, rather than considering the whole ensemble forecasts over different lead times. Thus, there are obvious gaps between the ensemble-based data assimilation and ensemble forecasts. A scale-dependent inflation that can capture different error growths at different scales is proposed, which obtains improved consistency between ensemble spread and errors at different lead times and effectively links the ensemble-based data assimilation and ensemble forecasts. [ABSTRACT FROM AUTHOR]

This letter raises the possibility that ergodicity concerns might have some bearing on the signal‐to‐noise paradox. This is explored by applying the ergodic theorem to the theory behind ensemble weather forecasting and the ensemble mean. Using the ensemble mean as our best forecast of observations amounts to interpreting it as the most likely phase‐space trajectory, which relies on the ergodic theorem. This can fail for ensemble forecasting systems if members are not perfectly exchangeable with each other, the averaging window is too short and/or there are too few members. We argue these failures can occur in cases such as the winter North Atlantic Oscillation (NAO) forecasts due to intransitivity or regime behaviour for regions such as the North Atlantic and Arctic. This behaviour, where different ensemble members may become stuck in different relatively persistent flow states (intransitivity) or multi‐modality (regime behaviour), can in certain situations break the ergodic theorem. The problem of non‐ergodic systems and models in the case of weather forecasting is discussed, as are potential mitigation methods and metrics for ergodicity in ensemble systems. [ABSTRACT FROM AUTHOR]

Deep learning is revolutionizing various scientific fields, with medical applications at the forefront. One key focus is automating image segmentation, a process crucial in many clinical services. However, medical images are often ambiguous and challenging even for experts. To address this, reliable models need to quantify their uncertainty, allowing physicians to understand the model's confidence in its segmentation. This paper explores how the performance and uncertainty of a model are influenced by the number of annotations per input sample. We examine the effects of both single and multiple manual annotations on various deep learning architectures. To tackle this question, we employ three widely recognized deep learning architectures and evaluate them across four publicly available datasets. Furthermore, we explore the effects of dropout rates on Monte Carlo models by examining uncertainty models with dropout rates of 20%, 40%, 60%, and 80%. Subsequently, we evaluate the models using various measurement metrics. The findings reveal that the influence of multiple annotations varies significantly depending on the datasets. Additionally, we observe that the dropout rate has minimal or no impact on the model's performance unless there is a substantial loss of training data, primarily evident in the 80% dropout rate scenario. [ABSTRACT FROM AUTHOR]

Deep Learning (DL) has the potential to enhance patient outcomes in healthcare by implementing proficient systems for disease detection and diagnosis. However, the complexity and lack of interpretability impede their widespread adoption in critical high-stakes predictions in healthcare. Incorporating uncertainty estimations in DL systems can increase trustworthiness, providing valuable insights into the model's confidence and improving the explanation of predictions. Additionally, introducing explainability measures, recognized and embraced by healthcare experts, can help address this challenge. In this study, we investigate DL models' ability to predict sex directly from electroencephalography (EEG) data. While sex prediction have limited direct clinical application, its binary nature makes it a valuable benchmark for optimizing deep learning techniques in EEG data analysis. Furthermore, we explore the use of DL ensembles to improve performance over single models and as an approach to increase interpretability and performance through uncertainty estimation. Lastly, we use a data-driven approach to evaluate the relationship between frequency bands and sex prediction, offering insights into their relative importance. InceptionNetwork, a single DL model, achieved 90.7% accuracy and an AUC of 0.947, and the best-performing ensemble, combining variations of InceptionNetwork and EEGNet, achieved 91.1% accuracy in predicting sex from EEG data using five-fold cross-validation. Uncertainty estimation through deep ensembles led to increased prediction performance, and the models were able to classify sex in all frequency bands, indicating sex-specific features across all bands. [ABSTRACT FROM AUTHOR]

The East African "short rains" from October–December (OND) are crucial for the region's cultural and agricultural landscape. Traditional climate studies have often treated these rains as a single mode, representing the average rainfall across the region. This approach, however, fails to capture the complex geographical variations in seasonal rainfall. In our study, we analyse 4200 reforecasts from a seasonal prediction system spanning 1981–2022, identifying distinct clusters that represent different geographical patterns of the short rains. We explore the influence of tropical sea‐surface temperature patterns, upper‐level tropospheric flow, and low‐level moisture fluxes on these clusters. A key revelation of our research is the limited predictability of certain geographical rainfall structures based on large‐scale climatic drivers. This finding highlights a gap in current forecasting methodologies, emphasising the necessity for further research to understand and predict these intricate patterns. Our study illuminates the complexities of regional rainfall variability in East Africa, underlining the importance of continued investigation to improve climate resilience strategies in the region. [ABSTRACT FROM AUTHOR]

Model ensembles may provide estimates of uncertainties arising from unknown initial conditions and model deficiencies. Often, the ensemble mean is taken as the best estimate, and quantities such as the mean‐squared error between model mean and observations decrease with the number of ensemble members. But the ensemble size is often limited by available resources, and so some idea of how many ensemble members that are needed before the error has saturated would be advantageous. The behaviour with ensemble size is often estimated by producing subsamples from a large ensemble. But this strategy requires that this large ensemble is already available. Fortunately, in many situations, the dependence on ensemble size follows simple analytical relations when the quantity under interest (such as the mean‐squared error between ensemble mean and observations) is calculated over many grid points or time points. This holds both for ensemble means and the related sampling variance. Here, we present such relations and demonstrate how they can be used to estimate the gain of increasing the ensemble. Whereas previous work has mainly focused on the size of the model ensemble, we recognize that uncertainties in observations play a role. We therefore also study the effect of using the mean of an ensemble of reanalyses. We show how the analytical relations can be used to estimate the point where the gain of increasing the size of the model ensemble is dwarfed by the gain of increasing the number of reanalyses. We demonstrate these points using two climate model ensembles: a large multimodel ensemble and a large single‐model initial‐condition ensemble. [ABSTRACT FROM AUTHOR]

Ensemble Kalman Filters (EnKFs), which assimilate observations based on statistics derived from an ensemble of samples of ocean states, have become the norm for ocean data assimilation (DA) and forecasting. These schemes are commonly implemented with inflation and localization techniques to increase their ensemble spread and to filter out spurious long‐range correlations resulting from the limited‐size ensembles imposed by computational burden constraints. Such ad‐hoc methods were found to be not necessary in ensemble DA experiments with simplified ocean/atmospheric models and large ensembles. Here, we conduct a series of one‐year‐long ensemble experiments with a fully realistic EnKF‐DA system in the Red Sea using tens ‐to thousands of ensemble members. The system assimilates satellite and in‐situ observations and accounts for model uncertainties by integrating a 4‐km‐resolution ocean model with European Center for Medium Range Weather Forecast (ECMWF) atmospheric ensemble fields, perturbed internal physics and initial conditions for forecasting. OceanOur results indicate that accounting for model uncertainties is more beneficial than simply increasing the ensemble size, with the improvements due to large ensembles leveling off at about 250 members. Besides, and in contrast to what is commonly observed with simplified models, the investigated ensemble DA system still required localization even when implemented with thousands of members. These findings are explained by: (i) amplified spurious long‐range correlations produced by the low‐rank nature of the ECMWF atmospheric forcing ensemble; and (ii) non‐Gaussianity generated by the perturbed internal physical parameterization schemes. Large‐ensemble forcing fields and non‐Gaussian DA methods might be needed to get full benefits from large ensembles in ocean DA. [ABSTRACT FROM AUTHOR]

Abstract Deep Learning (DL) has the potential to enhance patient outcomes in healthcare by implementing proficient systems for disease detection and diagnosis. However, the complexity and lack of interpretability impede their widespread adoption in critical high-stakes predictions in healthcare. Incorporating uncertainty estimations in DL systems can increase trustworthiness, providing valuable insights into the model’s confidence and improving the explanation of predictions. Additionally, introducing explainability measures, recognized and embraced by healthcare experts, can help address this challenge. In this study, we investigate DL models’ ability to predict sex directly from electroencephalography (EEG) data. While sex prediction have limited direct clinical application, its binary nature makes it a valuable benchmark for optimizing deep learning techniques in EEG data analysis. Furthermore, we explore the use of DL ensembles to improve performance over single models and as an approach to increase interpretability and performance through uncertainty estimation. Lastly, we use a data-driven approach to evaluate the relationship between frequency bands and sex prediction, offering insights into their relative importance. InceptionNetwork, a single DL model, achieved 90.7% accuracy and an AUC of 0.947, and the best-performing ensemble, combining variations of InceptionNetwork and EEGNet, achieved 91.1% accuracy in predicting sex from EEG data using five-fold cross-validation. Uncertainty estimation through deep ensembles led to increased prediction performance, and the models were able to classify sex in all frequency bands, indicating sex-specific features across all bands.

The December of 2019, marked the start of one of the biggest pandemics that the human race had seen for some centuries. COVID-19 after originating from China was in full force and was spreading quickly. This, however, was different from the previous pandemics as this is the age of technology and social circles on the internet. Thus, a sinister form of situation arose where fake news and misinformation flooded social media. The situation got to the point that WHO termed it as an "infodemic". Thus, NLP was again implored to find a solution and massive research was conducted for the detection of fake news on these platforms. The success of fake news detection improved and by today i.e. in 2023 the techniques have matured quite a bit. Keeping both of these aspects in mind, we have conducted a detailed review on fake news detection techniques for COVID-19. We have discussed the collection of data by providing a deep analysis of 7 COVID-19 Fake News datasets. Moreover, during the analysis of different methodologies, domination of deep learning and hybrid models was observed - specifically ensemble of transformer based models. Additionally, we explored the practical implications of COVID-19 Fake News detectors as components in generative AI models and as browser extensions to keep the common people safe. Finally, we discussed the limitations in existing research and how it can be improved in the future by exploring multi-modal, feature rich and cross-lingual approaches. [ABSTRACT FROM AUTHOR]

This paper investigated the use of linear models to forecast crude oil futures prices (WTI) on a monthly basis, emphasizing their importance for financial markets and the global economy. The main objective was to develop predictive models using time series analysis techniques, such as autoregressive (AR), autoregressive moving average (ARMA), autoregressive integrated moving average (ARIMA), as well as ARMA variants adjusted by genetic algorithms (ARMA-GA) and particle swarm optimization (ARMA-PSO). Exponential smoothing techniques, including SES, Holt, and Holt-Winters, in additive and multiplicative forms, were also covered. The models were integrated using ensemble techniques, by the mean, median, Moore-Penrose pseudo-inverse, and weighted averages with GA and PSO. The methodology adopted included pre-processing that applied techniques to ensure the stationarity of the data, which is essential for reliable modeling. The results indicated that for one-step-ahead forecasts, the weighted average ensemble with PSO outperformed traditional models in terms of error metrics. For multi-step forecasts (3, 6, 9 and 12), the ensemble with the Moore-Penrose pseudo-inverse showed better results. This study has shown the effectiveness of combining predictive models to forecast future values in WTI oil prices, offering a useful tool for analysis and applications. However, it is possible to expand the idea of applying linear models to non-linear models. [ABSTRACT FROM AUTHOR]

Polar lows (PLs) are intense mesoscale cyclones that form over polar oceans during colder months. Characterized by high wind speeds and heavy precipitation, they profoundly impact coastal communities, shipping, and offshore activities. Amid the substantial environmental changes in polar regions due to global warming, PLs are expected to undergo noteworthy transformations. In this study, we investigate the response of PL development in the Barents Sea to climate warming based on two representative PLs. Sensitivity experiments were conducted including the PLs in the present climate and the PLs in a pseudo–global warming scenario projected by the late twenty-first century for Shared Socioeconomic Pathway (SSP) 2-4.5 and SSP 3-7.0 scenarios from phase 6 of the Coupled Model Intercomparison Project (CMIP6). In both warming climate scenarios, there is an anticipated decrease in PL intensity, in terms of the maximum surface wind speed and minimum sea level pressure. Despite the foreseen increase in latent heat release in the future climate, contributing to the enhancement of PL intensity, other primary factors such as decreased baroclinic instability, heightened atmospheric static stability, and reduced overall surface heat fluxes play pivotal roles in the overall decrease in PL intensity in the Barents Sea under warming conditions. The augmentation of surface latent heat flux, however, results in increased precipitation associated with PLs by enhancing the latent heat release. Furthermore, the regional steering flow shifts in the warming climate can influence the trajectory of PLs during their development. Significance Statement: Global warming is anticipated to impact cyclone systems worldwide. Polar lows (PLs), intense mesocyclones in polar regions with potential socioeconomic and human life implications, pose uncertainties regarding intensity changes in a warming climate. In this study, we aimed to better understand how PLs over the Barents Sea will respond to the environmental changes in future climate conditions [Shared Socioeconomic Pathway (SSP) 2-4.5 and SSP 3-7.0] by the end of the twenty-first century. Our results find that the intensity of PLs is expected to decrease in the future while there is an expected increase in precipitation associated with PLs in the warming climate. These findings aim to contribute valuable insights for disaster management strategies in the face of evolving climate scenarios. [ABSTRACT FROM AUTHOR]

An accurate understanding of the current and future water cycle over the Third Pole is of great societal importance, given the role this region plays as a water tower for densely populated areas downstream. An emerging and promising approach for skillful climate assessments over regions of complex terrain is kilometer-scale climate modeling. As a foundational step towards such simulations over the Third Pole, we present a multi-model and multi-physics ensemble of kilometer-scale regional simulations for the hydrological year of October 2019 to September 2020. The ensemble consists of 13 simulations performed by an international consortium of 10 research groups, configured with a horizontal grid spacing ranging from 2.2 to 4 km covering all of the Third Pole region. These simulations are driven by ERA5 and are part of a Coordinated Regional Climate Downscaling EXperiment Flagship Pilot Study on Convection-Permitting Third Pole. The simulations are compared against available gridded and in-situ observations and remote-sensing data, to assess the performance and spread of the model ensemble compared to the driving reanalysis during the cold and warm seasons. Although ensemble evaluation is hindered by large differences between the gridded precipitation datasets used as a reference over this region, we show that the ensemble improves on many warm-season precipitation metrics compared with ERA5, including most wet-day and hour statistics, and also adds value in the representation of wet spells in both seasons. As such, the ensemble will provide an invaluable resource for future improvements in the process understanding of the hydroclimate of this remote but important region. [ABSTRACT FROM AUTHOR]

To improve the representation of all relevant scales in initial conditions for large-domain convection-allowing models, a new multiscale ensemble Kalman filter (MEnKF) algorithm is developed and implemented within the Gridpoint Statistical Interpolation analysis system (GSI) data assimilation framework coupled with the Finite-Volume Cubed-Sphere Dynamical Core (FV3) limited-area model. The algorithm utilizes ensemble background error covariances filtered to match the observations assimilated. This is realized in a sequential manner. 1) When assimilating coarse-resolution observations such as radiosondes, ensemble background perturbations are filtered to remove scales smaller than those the observations can represent, along with relatively large horizontal localization radii to ensure low-noise and balanced analysis increments. 2) The resulting ensemble analyses from the first step then serve as the background to assimilate denser observations such as radar data with smaller localization radii. Several passes can be taken to assimilate all observations. In this paper, vertically increasing horizontal filter scales are used when assimilating rawinsonde and surface observations together, while radar data are assimilated in the second step. The algorithm is evaluated through six convective storm cases during May 2021, with cycled assimilation of either conventional data only or with additional radar reflectivity followed by 24-h ensemble forecasts. Overall, positive impacts of the MEnKF on forecasts are obtained regardless of reflectivity data; its advantage over the single-scale EnKF is most significant in surface humidity and temperature forecasts up to at least 12 h. More accurate hourly precipitation forecasts with MEnKF can last up to 24 h for light rain. Furthermore, MEnKF forecasts higher ensemble probabilities for the observed hazardous events. [ABSTRACT FROM AUTHOR]

Univariate regression trees have representation problems for non-orthogonal regression functions. Ensembles of univariate regression trees have better representational power. In some cases, weighted ensembles have shown better performance than unweighted ensembles. In this paper, we study the properties of ensembles of regression trees by using regression classification models. We propose a theoretical framework to study the representational power of infinite-sized weighted ensembles, consisting of randomised finite-sized regression trees. We show for some datasets that the weighted ensembles may have better representational power than unweighted ensembles, but the performance is highly dependent on the weighting scheme and the properties of datasets. Our model cannot be used for all the datasets. However, for some datasets, we can accurately predict the experimental results of ensembles of regression trees. [ABSTRACT FROM AUTHOR]

When using sub‐km models to forecast convection, it is important to have a large enough domain to allow convection to fully spin‐up from the lateral boundaries. However, running large domains is computationally expensive and while it may be feasible for research purposes it is not yet feasible for routinely run models, such as the Met Office 300‐m London model. To try and mitigate the spin‐up issues in the London model, a variable‐resolution 300‐m London Model (the 'LMV') has been developed, which allows the boundaries of the London model to be further away from areas of interest (e.g., London Heathrow) at lower computational cost. Results from several cases of summertime convection show that the convective storms in the variable‐resolution model are more like those in a large fixed‐resolution 300‐m model than those in the much smaller London model. This implies variable resolution is a viable option for increasing the size of the London model domain without increasing the computational costs too much. Extended evaluation of the LMV was conducted during summer 2022, running as an ensemble nested inside the Met Office's operational UK ensemble (MOGREPS‐UK). Overall, the LMV looks promising for high‐impact convective events as it is better able to represent the organisation of convection into lines or larger storms whereas MOGREPS‐UK tends to simulate isolated, circular storms. This often leads to more reliable probabilities of heavy rainfall in the LMV ensemble compared to MOGREPS‐UK. However, there is an issue with the LMV producing too many small precipitating showers in situations where there should only be shallow clouds. This is thought to be a result of shallow clouds getting too deep in the model and precipitating erroneously. [ABSTRACT FROM AUTHOR]

Thunderstorms pose a major hazard to society and the economy, which calls for reliable thunderstorm forecasts. In this work, we introduce SALAMA, a feedforward neural network model for identifying thunderstorm occurrence in numerical weather prediction (NWP) data. The model is trained on convection‐resolving ensemble forecasts over central Europe and lightning observations. Given only a set of pixel‐wise input parameters that are extracted from NWP data and related to thunderstorm development, SALAMA infers the probability of thunderstorm occurrence in a reliably calibrated manner. For lead times up to 11 h, we find a forecast skill superior to classification based only on NWP reflectivity. Varying the spatiotemporal criteria by which we associate lightning observations with NWP data, we show that the time‐scale for skillful thunderstorm predictions increases linearly with the spatial scale of the forecast. [ABSTRACT FROM AUTHOR]

The constraint of computational power and the huge number of degrees of freedom of the atmosphere means a sampling uncertainty exists in probabilistic ensemble forecasts. In our previous study, the uncertainty could be quantified, creating a convergence measure which converges proportional to n−1/2 in the limit of large ensemble size n. This power law can then be extrapolated to determine how sampling uncertainty would decrease with larger ensemble sizes and hence find the necessary ensemble size. It is unknown, however, how the sampling uncertainty depends on different weather regimes. This study extends the previous idealised ensemble developed, by including weak and strong forcing convective weather regimes, to look at how sampling uncertainty convergence differs in each. Two 5,000‐member ensembles were run, with weak and strong forcing respectively. Comparisons with a kilometre‐scale weather prediction model ensured realistic weak and strong forcing regimes by comparing the rain, convective available potential energy (CAPE), convective adjustment timescale, and distribution shapes throughout the diurnal cycle. Differences in distribution shape between the regimes led to differences in the convergence measure. Large differences in spread between weak and strong forcing runs throughout the 24 hr period led to large differences in sampling uncertainty of the mean and standard deviation, which could be quantified according to well‐known equations. The timing of these differences was case‐dependent. For extreme statistics such as the 0.95 quantile and for cases where there was precipitation, the moisture variables for the weak forcing case had the largest sampling uncertainty and required the most members for convergence proportional to n−1/2. This was due to the tails of the weak forcing moisture variables containing the least amount of density. Different ensemble sizes will hence be required depending on whether one is in the weak or strong forcing convective weather regime. [ABSTRACT FROM AUTHOR]

As well as having an impact on the background state of the climate, global warming due to human activities could affect its natural oscillations and internal variability. In this study, we use four initial-condition ensembles from the CMIP6 framework to investigate the potential evolution of internal climate variability under different warming pathways for the twenty-first century. Our results suggest significant changes in natural climate variability and point to two distinct regimes driving these changes. The first is a decrease in internal variability of surface air temperature at high latitudes and all frequencies, associated with a poleward shift and the gradual disappearance of sea ice edges, which we show to be an important component of internal variability. The second is an intensification of the interannual variability of surface air temperature and precipitation at low latitudes, which appears to be associated with El Niño–Southern Oscillation (ENSO). This second regime is particularly alarming because it may contribute to making the climate more unstable and less predictable, with a significant impact on human societies and ecosystems. [ABSTRACT FROM AUTHOR]

Vehicular flow rate is an essential measure commonly collected by inductive-loop detectors for transportation agencies to evaluate freeways and highways. Loop detectors are typically located in urban areas due to installation and maintenance costs, and do not provide large spatial coverage. Crowdsourced data provide large spatial coverage, but typically do not capture vehicular flow rates. Therefore, a dynamically weighted ensemble (DWE) comprised of XGBoost and neural network models is proposed to expand the spatial coverage of vehicular flow rates by estimating flow rates for the Phoenix, AZ, metropolitan area using crowdsourced data. The model is evaluated using K-fold cross-validation methods, achieving a cross-validated mean absolute percent error of 21.74%, outperforming all other comparison models. The trained model is then used to estimate vehicular flow rates along highways and freeways throughout the state of Arizona. The proposed method provides transportation professionals with a transferable, cost-effective solution for large-scale flow rate estimation. [ABSTRACT FROM AUTHOR]

The skill of subseasonal (multiweek) forecasts of tropical‐cyclone (TC) occurrence over the Northern Hemisphere is examined in the Australian Bureau of Meteorology's (BoM) multiweek to seasonal prediction system, ACCESS‐S2. ACCESS‐S2 shows a good representation of the spatial distribution of TCs in the Northern Hemisphere; however, TC track frequency is generally underpredicted in the western North Pacific to the east of the Philippines and in the eastern North Pacific. The reduced activity relative to observations could be due to a significant positive bias in 850–200‐hPa wind shear in both of these regions, as well as a significant negative sea‐surface temperature (SST) bias in the eastern North Pacific. Despite biases in climatological TC frequency, the observed change in TC track frequency across the Northern Hemisphere with the phase of the Madden–Julian Oscillation (MJO) is well captured by ACCESS‐S2. Changes in the large‐scale environment (e.g., precipitation, 600‐hPa relative humidity, 850‐hPa absolute vorticity and 850–200‐hPa wind shear) are also well represented, with the location and size of the anomalies comparable to ERA‐Interim, apart from SST which shows a different response during some phases. ACCESS‐S2 shows skill relative to climatology for multiweek predictions of TC occurrence out to week 5 in the western North Pacific, eastern North Pacific and North Atlantic; and out to week 2 for the North Indian Ocean. Assessment of real‐time forecasts for Typhoon Rai (December 2021) showed that ACCESS‐S2 provided good guidance of the development and potential landfall of a TC in the Philippines at four weeks lead time. [ABSTRACT FROM AUTHOR]

The extratropical stage of Hurricane Ida (2021) brought extreme subdaily rainfall and devastating flooding to parts of eastern Pennsylvania, New Jersey, and New York. We investigate the predictability and character of this event using 31-member ensembles of perturbed initial condition hindcasts with the Tropical Atlantic version of GFDL's System for High-resolution prediction on Earth-to-Local Domains (T-SHiELD), a ∼13-km global weather forecast model with a ∼3-km nested grid. At lead times of up to 4 days, the ensembles are able to capture the most extreme observed hourly and daily rainfall accumulations but are negatively biased in the spatial extent of heavy precipitation. Large intraensemble differences in the magnitudes and locations of simulated extremes suggest that although impacts were highly localized, risks were widespread. In Ida's tropical stage, interensemble spread in extreme hourly rainfall is well predicted by large-scale moisture convergence; by contrast, in Ida's extratropical stage, the most extreme rainfall is governed by mesoscale processes that exhibit chaotic and diverse forms across the ensembles. Our results are relevant to forecasting and communication in advance of extratropical transition and imply that flood preparedness efforts should account for the widespread possibility of severe localized impacts. Significance Statement: After making landfall in Louisiana, Hurricane Ida (2021) transitioned to an extratropical storm which brought extreme rainfall and unprecedented flooding to parts of the northeastern United States. To what extent were these impacts knowable in advance? We use a numerical weather model with very high resolution to produce ensemble hindcasts—simulations of a past weather event initialized with tiny perturbations to the initial conditions, representing dozens of equally plausible versions of Ida's extratropical stage. We find that the observed hourly and daily rainfall maxima fall within the simulated outcomes of ensembles initialized with lead times of about 4 days or less. The location and intensity of the heaviest rainfall vary widely across these ensembles, suggesting that many locations across the Northeast were exposed to some likelihood of extreme rainfall. [ABSTRACT FROM AUTHOR]

The purpose of this descriptive study was to examine school string teachers' beliefs about and practices of teaching nontraditional music courses. We surveyed middle and high school string teachers (N = 42) about what nontraditional music courses and ensembles (NMCEs) they currently offer and would like to offer to their students. In addition, we examined the educational rationales that string teachers have for offering NMCEs. Our findings indicated that the most frequently offered NMCEs were guitar class, pit orchestra, and music technology. Participants expressed strongest interest in adding fiddle club, popular music, and mariachi ensemble. The participants explained that NMCEs provide opportunities to teach content that closely may more reflect students' cultures and possibly attract a wider range of students. The participants used NMCEs to cultivate musical skills that are often overlooked in large-group setting, including improvisation, playing by ear, and collaborative arranging and composition. [ABSTRACT FROM AUTHOR]

This paper examines the tools used to mediate intersubjectivity as a central element in Jean-Paul Sartre's phenomenological theory of ensembles. It first presents a brief account of ordinary individuals acting in and through violent groups from the viewpoints of psychology and phenomenology. Next, using Sartre's ontology of consciousness, the paper establishes the phenomenological structure of consciousness and intersubjectivity to explain, with recent psychological findings, how individual agents in violent groups come to deny their moral responsibility for the group's ideology and action. Finally, through Sartre's theory of ensembles, the paper shows how collectives of individuals evolve into groups. In violent groups, instruments of terror are used to mediate intersubjective relationships, and this explains how individual perpetrators and those complicit may sense a lack of control and perceive that they have diminished moral responsibility even though they are responsible for collective violence. [ABSTRACT FROM AUTHOR]

In reinforcement learning (RL), training a policy from scratch with online experiences can be inefficient because of the difficulties in exploration. Recently, offline RL provides a promising solution by giving an initialized offline policy, which can be refined through online interactions. However, existing approaches primarily perform offline and online learning in the same task, without considering the task generalization problem in offline-to-online adaptation. In real-world applications, it is common that we only have an offline dataset from a specific task while aiming for fast online-adaptation for several tasks. To address this problem, our work builds upon the investigation of successor representations for task generalization in online RL and extends the framework to incorporate offline-to-online learning. We demonstrate that the conventional paradigm using successor features cannot effectively utilize offline data and improve the performance for the new task by online fine-tuning. To mitigate this, we introduce a novel methodology that leverages offline data to acquire an ensemble of successor representations and subsequently constructs ensemble Q functions. This approach enables robust representation learning from datasets with different coverage and facilitates fast adaption of Q functions towards new tasks during the online fine-tuning phase. Extensive empirical evaluations provide compelling evidence showcasing the superior performance of our method in generalizing to diverse or even unseen tasks. [ABSTRACT FROM AUTHOR]

Information regarding the uncertainty associated with weather forecasts, particularly when they are related to a localized area at convective scales, can certainly play a crucial role in enhancing decision-making. In this study, we discuss and evaluate a short-range forecast (0–75 h) from of a regional ensemble prediction system (NEPS-R) running operationally at the National Centre for Medium Range Weather Forecasting (NCMRWF). NEPS-R operates at a convective scale (~ 4 km) with 11 perturbed ensemble members and a control run. We assess the performance of the NEPS-R in comparison to its coarser-resolution global counterpart (NEPS-G), which is also operational. NEPS-R relies on initial and boundary conditions provided by NEPS-G. The NEPS-G produces valuable forecast products and is capable of predicting weather patterns and events at a spatial resolution of 12 km. The objective of this study is to investigate areas where NEPS-R forecasts could add value to the short-range forecasts of NEPS-G. Verification is conducted for the period from 1st August to 30th September 2019, covering the summer monsoon over a domain encompassing India and its neighboring regions, using the same ensemble size (11 members). In addition to standard verification metrics, fraction skill scores, and potential economic values are used as the evaluation measures for the ensemble prediction systems (EPSs). Near-surface variables such as precipitation and zonal wind at 850 hPa (U850) are considered in this study. The results suggest that, in some cases, such as extreme precipitation, there is a benefit in using regional EPS forecast. State-of-the-art probabilistic measures indicate that the regional EPS has reduced under-dispersion in the case of precipitation compared to the global EPS. The global EPS tends to provide higher skill scores for U850 forecasts, whereas the regional EPS outperforms the global EPS for heavy precipitation events (> 65 mm/day). There are instances when the regional EPS can provide a useful forecast for cases, including moderate rainfall, and can add more value to the global EPS forecast products. The investigation of diurnal variations in precipitation forecasts reveals that although both models struggle to predict the correct timing, the time phase and peaks in precipitation in the convection-permitting regional model are closer to the observations. [ABSTRACT FROM AUTHOR]

Abstract: The increasing complexity and impacts of fire seasons in the United States have prompted efforts to improve early warning systems for wildland fire management. Outlooks of potential fire activity at lead‐times of several weeks can help in wildland fire resource allocation as well as complement short‐term meteorological forecasts for ongoing fire events. Here, we describe an experimental system for developing downscaled ensemble‐based subseasonal forecasts for the contiguous US using NCEP's operational Climate Forecast System version 2 model. These forecasts are used to calculate forecasted fire danger indices from the United States (US) National Fire Danger Rating System in addition to forecasts of evaporative demand. We further illustrate the skill of subseasonal forecasts on weekly timescales using hindcasts from 2011 to 2021. Results show that while forecast skill degrades with time, statistically significant week 3 correlative skill was found for 76% and 30% of the contiguous US for Energy Release Component and evaporative demand, respectively. These results highlight the potential value of experimental subseasonal forecasts in complementing existing information streams in weekly‐to‐monthly fire business decision making for suppression‐based decisions and geographic reallocation of resources during the fire season, as well for proactive fire management actions outside of the core fire season.

In situ approaches can accelerate the pace of scientific discoveries by allowing scientists to perform data analysis at simulation time. Current in situ workflow systems, however, face challenges in handling the growing complexity and diverse computational requirements of scientific tasks. In this work, we present Wilkins, an in situ workflow system that is designed for ease-of-use while providing scalable and efficient execution of workflow tasks. Wilkins provides a flexible workflow description interface, employs a high-performance data transport layer based on HDF5, and supports tasks with disparate data rates by providing a flow control mechanism. Wilkins seamlessly couples scientific tasks that already use HDF5, without requiring task code modifications. We demonstrate the above features using both synthetic benchmarks and two science use cases in materials science and cosmology.

Abstract This letter raises the possibility that ergodicity concerns might have some bearing on the signal‐to‐noise paradox. This is explored by applying the ergodic theorem to the theory behind ensemble weather forecasting and the ensemble mean. Using the ensemble mean as our best forecast of observations amounts to interpreting it as the most likely phase‐space trajectory, which relies on the ergodic theorem. This can fail for ensemble forecasting systems if members are not perfectly exchangeable with each other, the averaging window is too short and/or there are too few members. We argue these failures can occur in cases such as the winter North Atlantic Oscillation (NAO) forecasts due to intransitivity or regime behaviour for regions such as the North Atlantic and Arctic. This behaviour, where different ensemble members may become stuck in different relatively persistent flow states (intransitivity) or multi‐modality (regime behaviour), can in certain situations break the ergodic theorem. The problem of non‐ergodic systems and models in the case of weather forecasting is discussed, as are potential mitigation methods and metrics for ergodicity in ensemble systems.

Cotton is a major economic crop predominantly cultivated under rainfed situations. The accurate prediction of cotton yield invariably helps farmers, industries, and policy makers. The final cotton yield is mostly determined by the weather patterns that prevail during the crop growing phase. Crop yield prediction with greater accuracy is possible due to the development of innovative technologies which analyses the bigdata with its high-performance computing abilities. Machine learning technologies can make yield prediction reasonable and faster and with greater flexibility than process based complex crop simulation models. The present study demonstrates the usability of ML algorithms for yield forecasting and facilitates the comparison of different models. The cotton yield was simulated by employing the weekly weather indices as inputs and the model performance was assessed by nRMSE, MAPE and EF values. Results show that stacked generalised ensemble model and artificial neural networks predicted the cotton yield with lower nRMSE, MAPE and higher efficiency compared to other models. Variable importance studies in LASSO and ENET model found minimum temperature and relative humidity as the main determinates of cotton yield in all districts. The models were ranked based these performance metrics in the order of Stacked generalised ensemble > ANN > PCA ANN > SMLR ANN > LASSO> ENET > SVM > PCA SMLR > SMLR SVM > SMLR. This study shows that stacked generalised ensembling and ANN method can be used for reliable yield forecasting at district or county level and helps stakeholders in timely decision-making. [ABSTRACT FROM AUTHOR]

The stratosphere has previously been shown to be a significant source of subseasonal tropospheric predictability. The ability of ensemble prediction systems to appropriately exploit this depends on their ability to reproduce the statistical properties of the real atmosphere. In this study, we investigate predictability properties of the coupled stratosphere‐troposphere system in the sub‐seasonal to seasonal prediction project hindcasts by fitting a simple, minimal model. We diagnose the signal and noise components of each system in the stratosphere and troposphere and their coupling. We find that while the correlation skill scores are similar in most systems, the signal to noise properties can be substantially different. In the stratosphere, some systems are significantly overconfident, with a quantifiable impact on the tropospheric confidence. We link the method and details of the design of a prediction system to these predictive properties. Plain Language Summary: Subseasonal weather prediction systems make use of ensemble forecasting approaches, in which each forecast is made up of a number of complimentary predictions (ensemble members), that produce more skillful predictions by averaging and allows forecasters to anticipate the uncertainty in any particular forecast. There is a "battle" between the useful signal, hidden in the initial conditions of a simulation run, against the noise that chaotically builds up over the run from inaccuracies in the representation of the initial conditions. We use a simple statistical model with a minimal number of parameters to investigate how well the prediction systems capture both the signal and noise properties of the real atmosphere. We find that while the skill in the troposphere and stratosphere are in line with expectations from the real‐world, the ratio between the signal and the noise is too large, particularly in the stratosphere. We find this has an impact on the troposphere below, increasing the signal‐to‐noise ratio there too—artificially inflating it—giving it an overconfidence which is not wholly statistically or physically justified. We link the method and details of the design of a prediction system to these predictive properties. Key Points: Fitting a detailed statistical model to subseasonal prediction systems reveals some have significant overconfidence in the stratosphereLarge diversity in stratospheric error growth properties between different systemsSystems that are overconfident in the stratosphere have an overconfidence bias in the troposphere [ABSTRACT FROM AUTHOR]

The first version of the Taiwan Central Weather Bureau one-tier (TCWB1T) fully coupled global atmospheric and oceanic modeling forecast system had been developed and implemented as a routine operation for seasonal prediction at Central Weather Bureau (CWB) in 2017, with a minor revision in 2020. Based on NCEP CFSv1, the global atmospheric model in NCEP CFSv1 was replaced by CWB's atmospheric global spectral model (GSM) and coupled with the GFDL MOM3. Several parameters have been tested and tuned in the CWB atmospheric GSM, achieving an optimal configuration with better sea surface temperature (SST) predictions for integration more than one year. Using NCEP CFSR as the initial condition, TCWB1T conducted hindcasts from 1982 to 2011 and forecasts from 2012 to 2019 to analyze its performance. The results of these hindcasts and forecasts show that the TCWB1T can make useful predictions as verified against the observations of OISST, ERSST, CFSR, and GPCP based on the methods of EOF, RMSE, anomaly correlation, ranked probability skill score (RPSS), reliability diagram (RD), and relative operating characteristics (ROCs). TCWB1T also has the same level of skill scores as NCEP CFSv2 and/or the ECMWF fifth-generation seasonal forecast system (SEAS5), based on EOF, anomaly pattern correlation, climatological bias, RMSE, temporal correlation, and anomaly correlation percentage of forecast skill. TCWB1T shows forecast skill that is better in winter than in summer. Overall, it indicates that TCWB1T can be used for seasonal ENSO predictions. [ABSTRACT FROM AUTHOR]

Hurricane Ian made landfall along Florida's west coast at 1905 UTC 28 September 2022 near the Fort Myers area as a high-impact storm. Here, we examine the potential link between track forecast errors near the time of landfall and errors in both the synoptic-scale upper-level flow and a shortwave moving within that flow. Five days before the actual landfall (0000 UTC 23 September), most model guidance indicated landfall would occur close to where Ian eventually came ashore. But by 0000 UTC 25 September, model forecasts were all forecasting landfall in the Florida Panhandle. One day later, the models again agreed with each other but for a landfall 100–200 km north of Tampa, Florida. By 0000 UTC 27 September, forecast models indicated landfall would occur near Tampa. Model forecasts continued shifting to the right and finally converged on Punta Gorda, Florida, as the landfall location, less than 24 h before landfall. In this short article, we hypothesize that the track of Ian depended on subtle interactions with an extratropical wave in the middle and upper atmosphere. Deterministic and ensemble model forecasts reveal that the interactions were very sensitive to the characteristics of this wave and the synoptic-scale flow in which the wave was embedded. A 1–2-dam difference in the geopotential heights played a major role in whether Ian moved north into the Panhandle or toward the east, making landfall in central Florida. Significance Statement: The purpose of this work is to look at possible causes of error in forecasts of Hurricane Ian's landfall with the intent of saving lives and reducing damage, which can run to the billions of dollars. The path that hurricanes take is strongly influenced by the larger-scale weather patterns around them. We find that small errors in the upper-level flow patterns north of the hurricane appear to have played an important role in the forecast. These errors were present in both the general flow and a wave moving through that flow, resulting in small but significant changes to the forecast path. Understanding the cause of forecast errors is a key step both in improving forecasts and in identifying types of events that are more difficult to forecast. [ABSTRACT FROM AUTHOR]

The forecast uncertainty, particularly for precipitation, serves as a crucial indicator of the reliability of deterministic forecasts. Traditionally, forecast uncertainty is estimated by ensemble forecasting, which is computationally expensive since the forecast model is run multiple times with perturbations. Recently, deep learning methods have been explored to learn the statistical properties of ensemble prediction systems due to their low computational costs. However, accurately and effectively capturing the uncertainty information in precipitation forecasts remains challenging. In this study, we present a novel spatiotemporal transformer network (ST-TransNet) as an alternative approach to estimate uncertainty with ensemble spread and probabilistic forecasts, by learning from historical ensemble forecasts. ST-TransNet features a hierarchical structure for extracting multiscale features and incorporates a spatiotemporal transformer module with window-based attention to capture correlations in both spatial and temporal dimensions. Additionally, window-based attention can not only extract local precipitation patterns but also reduce computational costs. The proposed ST-TransNet is evaluated on the TIGGE ensemble forecast dataset and Global Precipitation Measurement (GPM) precipitation products. Results show that ST-TransNet outperforms both traditional and deep learning methods across various metrics. Case studies further demonstrate its ability to generate reasonable and accurate spread and probability forecasts from a single deterministic precipitation forecast. It demonstrates the capacity and efficiency of neural networks in estimating precipitation forecast uncertainty. [ABSTRACT FROM AUTHOR]

Large explosive volcanic eruptions cause short-term climatic impacts on both regional and global scales. Their impact on tropical climate variability, in particular El Niño–Southern Oscillation (ENSO), is still uncertain, as is their combined and separate effect on tropical and global precipitation. Here, we investigate the relationship between large-scale temperature and precipitation and tropical volcanic eruption strength, using 100-member MPI-ESM ensembles for idealized equatorial symmetric Northern Hemisphere summer eruptions of different sulfur emission strengths. Our results show that for idealized tropical eruptions, global and hemispheric mean near-surface temperature and precipitation anomalies are negative and linearly scalable for sulfur emissions between 10 and 40 Tg S. We identify 20 Tg S emission as a threshold where the global ensemble-mean near-surface temperature and precipitation signals exceed the range of internal variability, even though some ensemble members emerge from variability for lower eruption strengths. Seasonal and ensemble mean patterns of near-surface temperature and precipitation anomalies are highly correlated across eruption strengths, in particular for larger emission strengths in the tropics, and strongly modulated by ENSO. There is a tendency to shift toward a warm ENSO phase for the first postvolcanic year as the emission strength increases. Volcanic cooling emerges on a hemisphere-wide scale, while the precipitation response is more localized, and emergence is mainly confined to the tropics and subtropics. Significance Statement: The purpose of this study is to investigate at which strength the climate responses of volcanic forcing can be distinguished from the internal climate variability and whether the responses will linearly increase as the emission strengths become stronger. We ran 100-member MPI-ESM ensembles of idealized equatorial volcanic eruptions of different sulfur emission strengths and find that seasonal and ensemble mean patterns of near-surface temperature and precipitation anomalies are distinguishable and linearly scalable for sulfur emissions from 10 to 40 Tg S if their forcing patterns are similar. The identification of volcanic fingerprints is important for seasonal to decadal forecasts in the case of potential future eruptions and could help to prepare society for the regional climatic consequences of such an event. [ABSTRACT FROM AUTHOR]

Considering a whole summer season in central Europe, we find that the operational, convection‐permitting ICON‐D2 ensemble prediction system is spatially underdispersive in convective precipitation forecasts. The spatial spread of hourly precipitation is insufficient to capture the inherent error adequately across all scales (up to 300 km) and forecast times (up to 24 h). This lack of spread becomes more pronounced in the weak convective forcing regime. Using physically based stochastic perturbations in the planetary boundary layer is beneficial and leads to a reduction in spatial error at scales larger than 20 km and increases the spread at scales less than 50 km during weak forcing of convection, whereas the effect is almost neutral during strong forcing. Complementing the stochastic perturbations by perturbed parameters in the microphysics scheme shows an additive effect on spatial error and spread for a characteristic case study. Assessing the practical predictability of convective precipitation in a flow‐dependent manner is crucial, and our approach of combining multiple sources of uncertainty proves beneficial in mitigating the spatial underdispersion across scales, particularly during weak convective forcing. [ABSTRACT FROM AUTHOR]

Rank and probability integral transform histograms are established tools to assess the calibration of probabilistic forecasts. They not only check whether a forecast is calibrated, but they also reveal what systematic biases (if any) are present in the forecasts. Several extensions of rank histograms have been proposed to evaluate the calibration of probabilistic forecasts for multivariate outcomes. These extensions introduce a so‐called pre‐rank function that condenses the multivariate forecasts and observations into univariate objects, from which a standard rank histogram can be produced. Existing pre‐rank functions typically aim to preserve as much information as possible when condensing the multivariate forecasts and observations into univariate objects. Although this is sensible when conducting statistical tests for multivariate calibration, it can hinder the interpretation of the resulting histograms. In this article, we demonstrate that there are few restrictions on the choice of pre‐rank function, meaning forecasters can choose a pre‐rank function depending on what information they want to extract concerning forecast performance. We introduce the concept of simple pre‐rank functions and provide examples that can be used to assess the mean, spread, and dependence structure of multivariate probabilistic forecasts, as well as pre‐rank functions that could be useful when evaluating probabilistic spatial field forecasts. The simple pre‐rank functions that we introduce are easy to interpret, easy to implement, and they deliberately provide complementary information, meaning several pre‐rank functions can be employed to achieve a more complete understanding of multivariate forecast performance. We then discuss how e‐values can be employed to formally test for multivariate calibration over time. This is demonstrated in an application to wind‐speed forecasting using the EUPPBench post‐processing benchmark dataset. [ABSTRACT FROM AUTHOR]