Current Prediction Systems Research Articles

A Simple Statistical Postprocessing Scheme for Enhancing the Skill of Seasonal SST Predictions in the Tropics

Abstract Seasonal prediction systems are subject to systematic errors, including those introduced during the initialization procedure, that may degrade the forecast skill. Here we use a novel statistical postprocessing correction scheme that is based on canonical correlation analysis (CCA) to relate errors in ocean temperature arising during initialization with errors in the predicted sea surface temperature fields at 1–12-month lead time. In addition, the scheme uses CCA of simultaneous SST fields from the prediction and corresponding observations to correct pattern errors. Finally, simple scaling is used to mitigate systematic location and phasing errors as a function of lead time and calendar month. Applying this scheme to an ensemble of seven seasonal prediction models suggests that moderate improvement of prediction skill is achievable in the tropical Atlantic and, to a lesser extent, in the tropical Pacific and Indian Ocean. The scheme possesses several adjustable parameters, including the number of CCA modes retained, and the regions of the left and right CCA patterns. These parameters are selected using a simple tuning procedure based on the average of four skill metrics. The results of the present study indicate that errors in ocean temperature fields due to imperfect initialization and SST variability errors can have a sizable negative impact on SST prediction skill. Further development of prediction systems may be able to remedy these impacts to some extent. Significance Statement The prediction of year-to-year climate variability patterns, such as El Niño, offers potential benefits to society by aiding mitigation and adaptation efforts. Current prediction systems, however, may still have substantial room for improvement due to systematic model errors and due to imperfect initialization of the oceanic state at the start of predictions. Here we develop a statistical correction scheme to improve prediction skill after forecasts have been completed. The scheme shows some moderate success in improving the skill for predicting El Niño and similar climate patterns in seven prediction systems. Our results not only indicate a potential for improving prediction skill after the fact but also point to the importance of improving the way prediction systems are initialized.

STGNets: A spatial–temporal graph neural network for energy consumption prediction in cement industrial manufacturing processes

Energy consumption is an essential indicator for conserving energy and reducing emissions in industrial manufacturing processes. However, the industrial production chain is a multi-variable, time-varying, nonlinear system, which poses significant challenges for energy consumption prediction, particularly under complex physical–chemical processes, even when the system is in a steady state. Furthermore, the poor accuracy cannot meet practical production requirements due to current prediction systems cannot integrate spatial–temporal information effectively. To solve this issue, this research provides a novel technique for final online goods quality prediction based on deep spatial–temporal graph neural networks (GNN). Our approach can capture hidden spatial information relationships and manage long-time sequences in the processing data by using a learnable dependency matrix and a stacked dilated convolution component. Moreover, these two primary components are organically merged into a single end-to-end optimization framework. The experimental results on the real-world test dataset demonstrate that our technique outperforms rival machine learning techniques. Our research exemplifies how graph deep learning architecture may be utilized to solve real-world problems in industries.

Monthly Arctic sea ice prediction based on a data-driven deep learning model

There is growing interest in sub-seasonal to seasonal predictions of Arctic sea ice due to its potential effects on midlatitude weather and climate extremes. Current prediction systems are largely dependent on physics-based climate models. While climate models can provide good forecasts for Arctic sea ice at different timescales, they are susceptible to initial states and high computational costs. Here we present a purely data-driven deep learning model, UNet-F/M, to predict monthly sea ice concentration (SIC) one month ahead. We train the model using monthly satellite-observed SIC for the melting and freezing seasons, respectively. Results show that UNet-F/M has a good predictive skill of Arctic SIC at monthly time scales, generally outperforming several recently proposed deep learning models, particularly for September sea-ice minimum. Our study offers a perspective on sub-seasonal prediction of future Arctic sea ice and may have implications for forecasting weather and climate in northern midlatitudes.

Skillful U.S. Soy Yield Forecasts at Presowing Lead Times

Abstract Soy harvest failure events can severely impact farmers, insurance companies, and raise global prices. Reliable seasonal forecasts of misharvests would allow stakeholders to prepare and take appropriate early action. However, especially for farmers, the reliability and lead time of current prediction systems provide insufficient information to justify within-season adaptation measures. Recent innovations increased our ability to generate reliable statistical seasonal forecasts. Here, we combine these innovations to predict the 1–3 poor soy harvest years in the eastern United States. We first use a clustering algorithm to spatially aggregate crop producing regions within the eastern United States that are particularly sensitive to hot–dry weather conditions. Next, we use observational climate variables [sea surface temperature (SST) and soil moisture] to extract precursor time series at multiple lags. This allows the machine learning model to learn the low-frequency evolution, which carries important information for predictability. A selection based on causal inference allows for physically interpretable precursors. We show that the robust selected predictors are associated with the evolution of the horseshoe Pacific SST pattern, in line with previous research. We use the state of the horseshoe Pacific to identify years with enhanced predictability. We achieve high forecast skill of poor harvests events, even 3 months prior to sowing, using a strict one-step-ahead train-test splitting. Over the last 25 years, when the horseshoe Pacific SST pattern was anomalously strong, 67% of the poor harvests predicted in February were correct. When operational, this forecast would enable farmers to make informed decisions on adaption measures, for example, selecting more drought-resistant cultivars or change planting management. Significance Statement If soy farmers would know that the upcoming growing season will be hot and dry, they could decide to take anticipatory action to reduce losses, that is, buy more drought resistant soy cultivars or change planting management. To make such decisions, farmers would need information even prior to sowing. On these very long lead times, a predictable signal can emerge from low-frequency processes of the climate system that can affect surface weather via teleconnections. However, traditional forecast systems are unable to make reliable predictions at these lead times. In this work, we used machine learning techniques to train a forecast model based on these low-frequency components. This allowed us to make reliable predictions of poor harvest years even 3 months prior to sowing.

Ocean data assimilation for the initialization of seasonal prediction with the Community Earth System Model

Seasonal prediction highly depends on its initialization scheme. Currently, the operational utilization of advanced data assimilation methods in the oceanic initialization is still relatively less and its benefit is far from being fully unveiled. In this study, the Community Earth System Model (CESM) is configured with an Ensemble Adjustment Kalman Filter (EAKF) to initialize the seasonal prediction. The performance of the established prediction system is assessed by comparing it with a benchmark prediction system, which is based on the same model but with a nudging scheme. Results show that the assimilation of only oceanic observations can not only effectively constrain the oceanic variables, but also improves the atmospheric variables through coupling processes, which helps to produce accurate and compatible initial conditions. With the initial conditions, the current prediction system is better than the benchmark prediction system and can produce skillful predictions for sea surface temperature, ocean heat content, air temperature at 2 m, precipitation and major climate variabilities including El Niño–Southern Oscillation (ENSO) and Indian Ocean Dipole (IOD). The current prediction system also shows better ensemble consistency and less prediction drifts, which could contribute to the improvement of prediction skill.

Framework for Bidirectional Knowledge-Based Maintenance of Wind Turbines.

Artificial intelligence (AI) techniques, such as machine learning (ML), are being developed and applied for the monitoring, tracking, and fault diagnosis of wind turbines. Current prediction systems are largely limited by their inherent disadvantages for wind turbines. For example, frequency or vibration analysis simulations at a part scale require a great deal of computational power and take considerable time, an aspect that can be essential and expensive in the case of a breakdown, especially if it is offshore. An integrated digital framework for wind turbine maintenance is proposed in this study. With this framework, predictions can be made both forward and backward, breaking down barriers between process variables and key attributes. Prediction accuracy in both directions is enhanced by process knowledge. An analysis of the complicated relationships between process parameters and process attributes is demonstrated in a case study based on a wind turbine prototype. Due to the harsh environments in which wind turbines operate, the proposed method should be very useful for supervising and diagnosing faults.

A wavelet artificial neural network method for medium‐term rainfall prediction in Queensland (Australia) and the comparisons with conventional methods

AbstractThis study aims to develop a medium‐term rainfall forecast model to predict monthly rainfall 1, 3, 6, and 12 months in advance using hybrid wavelet‐artificial neural networks (WANN) for Queensland, Australia. To assess the performance of the models, seven input sets comprised of either past rainfall magnitudes, selected climate anomalies, or combinations of rainfall with different climate anomalies as independent variables were defined. The data of the years 1908–1999 and 2000–2016 from 10 weather stations in Queensland were used for training and verification of the models, respectively. The results showed that the WANN enhances the average prediction accuracy in terms of root‐mean‐square‐error (RMSE) with 90, 52, 32, and 15% at 1, 3, 6, and 12 months lead time, respectively compared to of the artificial neural networks (ANN). Also, the current prediction system of the Australian Bureau of Meteorology (BOM), Australian Community Climate Earth‐System Simulator–Seasonal (ACCESS‐S) provided the predictions with the average RMSE values of 91.0, 102.7, and 100.3 mm for 1, 3, and 6 months lead times, respectively. In contrast, the corresponding RMSE values of 8.2, 37.6, and 52.4 mm are obtained through the WANN. Moreover, performance of the WANN method was compared with the autoregressive integrated moving average (ARIMA) and multiple linear regression (MLR) models. ARIMA and MLR produced similar performances at same lead times; corresponding average RMSE values of 78.8, 82.72, 80.9, and 79.5 mm were obtained through ARIMA compared to the 75.9, 79.5, 79.0, and 78.9 mm through the MLR. The results of the current study indicate than the performance of the WANN is more accurate than the ANN, ARIMA, MLR, and ACCESS‐S forecasts.

Predicting Daily Mean Wind Speed in Europe Weeks ahead from MJO Status

ABSTRACTThe Madden–Julian oscillation (MJO), a prominent feature of the tropical atmospheric circulation at subseasonal time scales, is known to modulate atmospheric variability in the Euro-Atlantic region. However, current subseasonal prediction systems fail to accurately reproduce the physical processes involved in these teleconnection mechanisms. This paper explores the observed impact of strong MJO events on surface wind speed over Europe. It is found that some MJO phases are accompanied by strong wind anomalies in Europe. After showing that this teleconnective mechanism is not present in the predictions of the ECMWF monthly forecasting system, a methodology to reconstruct forecasts of daily mean wind speed in the continent weeks ahead is proposed. This method combines MJO forecasts from the S2S project database and the observed teleconnection impacts in the historical records. Although it is found that strong MJO events cannot be skillfully predicted more than 10 days ahead with current prediction systems, a theoretical experiment shows that this method can effectively transform a dynamical MJO forecast into a probabilistic wind speed prediction in Europe.

External validation of current prediction systems of improvement after decompression surgery in Chiari malformation type I patients: can we do better?

The Chiari Severity Index (CSI) and points-based algorithm of Thakar et al. are two prognostic tools that have been developed to predict the likelihood of improvement after suboccipital decompression in adult patients with Chiari malformation type I (CM-I). This study aimed to externally validate and critically evaluate these algorithms in the interest of guiding the development of improved prediction systems. A consecutive cohort of CM-I patients undergoing suboccipital decompression between September 2006 and September 2018 were included. The CSI and Thakar point score were computed for all patients, and associations with improvement were analyzed. The ability of both prediction systems to predict improvement as measured by different Chicago Chiari Outcome Scale (CCOS) cutoffs was assessed using receiver operating curve analysis. Post hoc correlations between the algorithms and different CCOS subcomponents were also assessed. The surgical cohort was composed of 149 adult CM-I patients, of whom 39 (26%) had a syrinx. Most patients experienced improvement after surgery (80% CCOS ≥ 13; 96% CCOS ≥ 11). The proportion of patients improving decreased with increasing CSI, but the results were not statistically significant (p = 0.246). No statistically significant difference in the mean Thakar point score was identified between improved and nonimproved patients using both CCOS cutoffs (p = 0.246 for a cutoff of 13 and p = 0.480 for a cutoff of 11). The CSI had a poor ability in identifying improved patients at a CCOS cutoff of 13 (area under the curve [AUC] 0.582) and 11 (AUC 0.646). The Thakar point score similarly had poor discrimination at a cutoff of 13 (AUC 0.467) and 11 (AUC 0.646). Neither algorithm had significant correlation with any of the CCOS subcomponents except for CSI and nonpain symptom improvement (coefficient = -0.273, p = 0.001). Previously published algorithms failed to provide prediction value with regard to clinically meaningful improvement following suboccipital decompression in adult CM-I patients. Future models and practical scoring systems are still required to improve the decision-making process.

The digitizing algorithm for precipitation in the atmosphere on the base of radar measurements

There is an increasing demand for automated high-quality very-short-range forecasts and nowcasts of precipitation on small scales and at high update frequencies. Current prediction systems use different methods of determining precipitation such as area tracking, individual cell tracking and numerical models. All approaches are based on radar measurements. World-leading manufactories of meteorological radars and attendant visualization software are introduced in the paper. Advantages of the numerical modelling against inertial schemes designed on statistical characteristics of convective processes are outlined. On this way, radar data assimilation systems as a necessary part of numerical models are intensively developed. In response to it, the use of digital formats for processing of radar measurements in numerical algorithms became important. In the focus of this work is the developing of a unified code for digital processing of radar signals at the preprocessing, filtration, assimilation and numerical integration steps. The proposed code also includes thinning, screening or superobbing radar data before exploring them for the assimilation procedures. The informational model manages radar data flows in the metadata and binary array forms. The model constitutes an official second-generation European standard exchange format for weather radar datasets from different manufactories. Results of radar measurement processing are presented for both, the single radar and radar overlying network.

Assessing probabilistic predictions of ENSO phase and intensity from the North American Multimodel Ensemble

Here we examine the skill of three, five, and seven-category monthly ENSO probability forecasts (1982–2015) from single and multi-model ensemble integrations of the North American Multimodel Ensemble (NMME) project. Three-category forecasts are typical and provide probabilities for the ENSO phase (El Niño, La Niña or neutral). Additional forecast categories indicate the likelihood of ENSO conditions being weak, moderate or strong. The level of skill observed for differing numbers of forecast categories can help to determine the appropriate degree of forecast precision. However, the dependence of the skill score itself on the number of forecast categories must be taken into account. For reliable forecasts with same quality, the ranked probability skill score (RPSS) is fairly insensitive to the number of categories, while the logarithmic skill score (LSS) is an information measure and increases as categories are added. The ignorance skill score decreases to zero as forecast categories are added, regardless of skill level. For all models, forecast formats and skill scores, the northern spring predictability barrier explains much of the dependence of skill on target month and forecast lead. RPSS values for monthly ENSO forecasts show little dependence on the number of categories. However, the LSS of multimodel ensemble forecasts with five and seven categories show statistically significant advantages over the three-category forecasts for the targets and leads that are least affected by the spring predictability barrier. These findings indicate that current prediction systems are capable of providing more detailed probabilistic forecasts of ENSO phase and amplitude than are typically provided.

A new approach to surface currents monitoring

There is growing recognition in the offshore industry of the need for high resolution surface currents. To this end, a new current prediction system has been developed. This system is highly portable and massively based on satellite (altimetry, sea surface temperature, ocean colour) and in-situ observations (surface drifters). It provides daily near real-time maps of the surface currents at high resolution (4km, hourly data) and has a 48h forecast ability. Results presented for a configuration deployed offshore Brazil show good agreement with observations and highlight the impact of the ocean colour data.

A generic approach to evaluate how B-cell epitopes are surface-exposed on protein structures

Methods that predict antibody epitopes could help to promote the development of diagnostic tools, vaccines or immunotherapies by affecting the epitope binding of antibodies during an immunological response to antigens.It is generally assumed that there is a direct relationship between antibody accessibility to antigens and accessible surface of proteins. Based on this assumption, prediction systems often includes solvent accessibility values calculated from the primary sequence of proteins or from their three dimensional structures as a predictive criterion. However, the current prediction systems seem weakly efficient in view of benchmark tests.We were interested in evaluating how amino acids that have been experimentally identified as epitopic elements could differ from the rest of the antigenic molecule at the level of surface exposure, hence we assessed the average accessibility of epitopes. The approach used here utilises published epitopes deduced from numerous identification techniques, including sequence scanning and structure visualisation after crystallography, and it involves many types of antigens from toxins to allergens. Our results show that epitopic residues are not distributed among any specific Relative Surface Accessibility and Protrusion Index values and that, in some cases, epitopes cover the entire antigenic sequence.These results led to the conclusion that the classification of known epitopes with respect to the experimental conditions used to identify them should be introduced before attempting to characterise epitopic areas in a generic way.

Multiphysics superensemble forecast applied to Mediterranean heavy precipitation situations

Abstract. The high-impact precipitation events that regularly affect the western Mediterranean coastal regions are still difficult to predict with the current prediction systems. Bearing this in mind, this paper focuses on the superensemble technique applied to the precipitation field. Encouraged by the skill shown by a previous multiphysics ensemble prediction system applied to western Mediterranean precipitation events, the superensemble is fed with this ensemble. The training phase of the superensemble contributes to the actual forecast with weights obtained by comparing the past performance of the ensemble members and the corresponding observed states. The non-hydrostatic MM5 mesoscale model is used to run the multiphysics ensemble. Simulations are performed with a 22.5 km resolution domain (Domain 1 in http://mm5forecasts.uib.es) nested in the ECMWF forecast fields. The period between September and December 2001 is used to train the superensemble and a collection of 19~MEDEX cyclones is used to test it. The verification procedure involves testing the superensemble performance and comparing it with that of the poor-man and bias-corrected ensemble mean and the multiphysic EPS control member. The results emphasize the need of a well-behaved training phase to obtain good results with the superensemble technique. A strategy to obtain this improved training phase is already outlined.

A real-time nearshore wave and current prediction system

A nearshore wave, tide and current prediction system was demonstrated during the MREA04 Trial in the Portuguese coastal waters near Pinheiro da Cruz during the early spring of 2004. Daily forecasts of regional scale wave and tidal information and nearshore waves and currents were generated in DIOPS utilizing a suite of regional and nearshore models forced with data from meteorological and oceanographic production centers. A limited beach experiment was conducted with three Nortek current meters deployed in the surf zone and a video imagery system to generate 10-min time exposures used to identify the locations of wave breaking. In this study, Delft3D, a coastal hydrodynamic modeling system, capable of simulating hydrodynamic processes due to waves, tides, rivers, winds and coastal currents, is used to predict the nearshore wave and longshore currents near Pinheiro da Cruz. The nearshore bathymetry used in this study was based on LIDAR data collected in February 2000. Delft3D shows mixed results when compared with the measured wave height and nearshore currents. Improved Delft3D results can be achieved in the future if migrating sand bars can be measured and included in the modeling.

Prediction of consensus structural motifs in a family of coregulated RNA sequences

Given a set of homologous or functionally related RNA sequences, the consensus motifs may represent the binding sites of RNA regulatory proteins. Unlike DNA motifs, RNA motifs are more conserved in structures than in sequences. Knowing the structural motifs can help us gain a deeper insight of the regulation activities. There have been various studies of RNA secondary structure prediction, but most of them are not focused on finding motifs from sets of functionally related sequences. Although recent research shows some new approaches to RNA motif finding, they are limited to finding relatively simple structures, e.g. stem-loops. In this paper, we propose a novel genetic programming approach to RNA secondary structure prediction. It is capable of finding more complex structures than stem-loops. To demonstrate the performance of our new approach as well as to keep the consistency of our comparative study, we first tested it on the same data sets previously used to verify the current prediction systems. To show the flexibility of our new approach, we also tested it on a data set that contains pseudoknot motifs which most current systems cannot identify. A web-based user interface of the prediction system is set up at http://bioinfo. cis.nctu.edu.tw/service/gprm/.