Time Series Of Events Research Articles

Investigating the Performance of the Informer Model for Streamflow Forecasting

Recent studies have shown the potential of transformer-based neural networks in increasing prediction capacity. However, classical transformers present several problems such as computational time complexity and high memory requirements, which make Long Sequence Time-Series Forecasting (LSTF) challenging. The contribution to the prediction of time series of flood events using deep learning techniques is examined, with a particular focus on evaluating the performance of the Informer model (a particular implementation of transformer architecture), which attempts to address the previous issues. The predictive capabilities of the Informer model are explored and compared to statistical methods, stochastic models and traditional deep neural networks. The accuracy, efficiency as well as the limits of the approaches are demonstrated via numerical benchmarks relating to real river streamflow applications. Using daily flow data from the River Test in England as the main case study, we conduct a rigorous evaluation of the Informer efficacy in capturing the complex temporal dependencies inherent in streamflow time series. The analysis is extended to encompass diverse time series datasets from various locations (>100) in the United Kingdom, providing insights into the generalizability of the Informer. The results highlight the superiority of the Informer model over established forecasting methods, especially regarding the LSTF problem. For a forecast horizon of 168 days, the Informer model achieves an NSE of 0.8 and maintains a MAPE below 10%, while the second-best model (LSTM) only achieves −0.63 and 25%, respectively. Furthermore, it is observed that the dependence structure of time series, as expressed by the climacogram, affects the performance of the Informer network.

Characterisation of Dansgaard–Oeschger events in palaeoclimate time series using the matrix profile method

Abstract. Palaeoclimate time series, reflecting the state of Earth's climate in the distant past, occasionally display very large and rapid shifts showing abrupt climate variability. The identification and characterisation of these abrupt transitions in palaeoclimate records is of particular interest as this allows for understanding of millennial climate variability and the identification of potential tipping points in the context of current climate change. Methods that are able to characterise these events in an objective and automatic way, in a single time series, or across two proxy records are therefore of particular interest. In our study the matrix profile approach is used to describe Dansgaard–Oeschger (DO) events, abrupt warmings detected in the Greenland ice core, and Northern Hemisphere marine and continental records. The results indicate that canonical events DO-19 and DO-20, occurring at around 72 and 76 ka, are the most similar events over the past 110 000 years. These transitions are characterised by matching transitions corresponding to events DO-1, DO-8, and DO-12. They are abrupt, resulting in a rapid shift to warmer conditions, followed by a gradual return to cold conditions. The joint analysis of the δ18O and Ca2+ time series indicates that the transition corresponding to the DO-19 event is the most similar event across the two time series.

Modeling the chaotic dynamics of bubble growth under hydrodynamic interactions in pool boiling

This study seeks to understand the origins of intermittency in quantities of interest in pool boiling, such as bubble departure diameter and growth time. The intermittency of nucleation site activity due to hydrodynamic and thermal interactions with nearby nucleation sites is well-established; however, few mechanistic models have been developed that predict such intermittency. Here we assume a fluctuating pressure field at a nucleation site due to an adjacent bubble which alters bubble growth depending on the phase of the pressure oscillation with respect to the instant of bubble nucleation. The results suggest that even when a single departure frequency is assumed for nearby bubbles, its effect on the nucleation site being considered causes aperiodicity and intermittency in bubble departure quantities. The effects of pressure field phase angle, degree of superheat, and choice of force balance model on the bubble departure quantities are examined. The phase angle is shown to play a major role in determining bubble departure diameter and growth time. The departure diameter is observed to have a broad distribution, belying the assumption of a unique value. Period doubling of the ebullition cycle is observed for some conditions, a phenomenon documented by other investigators. A multifractal analysis of the time series of departure events indicates the presence of long term temporal correlations, which become weaker with increasing superheat. The second order Hurst exponent of the structure functions of departure events appears to have a universal behavior with Jakob number, independent of the choice of liquid. The Hurst exponent rises from zero to a value close to 0.5 at large superheats, suggesting a continuous transition from an ordered state to a disordered state along the boiling curve.

Koreksi Bias Data Hujan Satelit Persiann Sebagai Alternatif Pos Stasiun Hujan dengan Pendekatan Quantile Mapping di Sub Das Brantas Hulu

The rainfall data is usually obtained from meteorological stations, which many sill observed manually in Indonesia and can lead to an observation error. Remote sensing rainfall data observed by satellite offer interesting alternatives to overcome the limitation of existing meteorological stations in Indonesia. It offers accuracy, spatial coverage, timeliness and cost effectiveness. However it contains bias if compared to land based station. This research will evaluate rainfall data observed in PERSIANN (Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks) in Brantas Hulu Watershed. The rainfall data of 13 years from 12 stations is evaluated to obtain area averaged rainfall using Thiessen Polygon method. Three datasets from PERSIANN, PERSIANN CCS and PERSIANN CDR are compared statistically with the area averaged area for daily, monthly and yearly rainfall. Bias correction is performed using Quantile Mapping to match the cumulative distribution curve from the area averaged rainfall. Verification of bias correction is performed using 2 years data from 2019 to 2020. Finally, the extreme rainfall analysis is performed to test the corrected data for design rainfall calculations. The correlation of daily data shows poor agreement for all three datasets with correlation coefficient below 0,40. The monthly and yearly data gives better correlation with coefficient above 0,80. It shows that for time series event, the satellite data gives poor correlation with the area averaged rainfall. However, quantile mapping correction gives consistent cumulative distribution agreement for correction and model verification. The verification analysis shows an increase in coefficient correlation from 0.387 to 0.43 (daily) and 0.887 to 0.915 (monthly). In addition, there is a decrease in normalized root mean square error (NRMSE) from 31.03% to 16.91% (daily) and 46.03% to 14.03% (monthly). A decrease in normalized mean absolute error (NMAE) occurred from 16.12% to 9.74% (daily) and 31.67% to 10.07% (monthly). Beside that, the results of extreme rainfall analysis produce values that are very close to the rain station. So it is found that the PERSIANN CCS quantile mapping regression model is the best in improving data quality.

Negative storm surges in the Río de la Plata Estuary: mechanisms, variability, trends and linkage with the Continental Shelf dynamics

Negative storm surge (NSS) events frequently affect the Río de la Plata Estuary (RdP), limiting the access to major ports and waterways and emptying the freshwater intakes of one of the most socially developed basins of the Southwestern Atlantic Continental Shelf (SWACS). It is important to re-examine this issue, as the available studies are outdated, there are some contradictions in the literature and a regional framework of the adjacent shelf dynamics is not yet available. In this work (i) nonlinear statistical analyses were applied to review and update the climatological description of NSS in the RdP; and, then, (ii) an ocean numerical model was employed to describe and understand the surge dynamics framed into the larger scale dynamics of the atmospherically forced SWACS. An analysis of 87 years of hourly residual sea level observations (RSL) reveals that the number of yearly NSS events time series presents (i) a statistically significant multidecadal non-linear trend or envelope and (ii) two significant pseudo-cycles centred at 3.5 and 11.5 years that explain a large fraction of the interannual variability embedded in the multidecadal modulation.Numerical simulations results show that the surges (both negative and positive) that are generated in the SWACS and reach the RdP result from the passage of atmospheric planetary waves over Patagonia which produce winds with a significant alongshore (N-S) component. These winds force an Ekman transport either offshore or onshore at the coast, generating a negative or positive sea level anomaly, which will be larger as the atmospheric system is slower and more persistent. When synoptic atmospheric systems migrate away from the coast on their northeastward propagation, the forced sea level anomalies propagate along the coast to the north as free Kelvin waves, eventually reaching the estuary; thus the forcing for surges at the RdP can be remote. In this sense, despite the large gulfs and bays, the response of the Patagonian coast is similar to that of a linear coastline oriented in a north–south direction. In addition to remotely forced surges, locally forced events occur in the RdP as a result of the estuarine geometry and its resonant response to winds with a component longitudinal to the estuary axis (NW-SE), which could be forced by other atmospheric configurations besides Rossby waves, such as cyclogenesis of the Argentine Littoral.

Comprehensive assessment of five near-real-time satellite precipitation products in the Lower Yangtze River Basin and the Lixiahe region, China: Dual perspectives from time series and extreme events

Near real-time satellite precipitation products (NSPPs) have significant potential in meteorological hydrological monitoring and forecasting due to their timeliness. Previous evaluation-oriented studies have either focused on the uncertainty of time series or on the monitoring capability of extreme events. There is limited research that comprehensively considers both aspects to provide a systematic knowledge. Moreover, there is still insufficient understanding of the accuracy of hourly precipitation estimates. This study conducts a comprehensive evaluation between five NSPPs (IMERG Early, IMERG Late, GSMaP NRT, GSMaP NRT Gauge and PERSIANN CCS) from two main perspectives at hourly scale: an assessment of time series spanning from 2018 to 2022 over the Lower Yangtze River Basin and the Lixiahe region and an examination of the four extreme events over the Tai Lake Basin and the Chao Lake Basin. HP-metrics are proposed to evaluate bias in the location of extreme rainfall, rainy area, and precipitation amount over time, comprehensively analyzing performance in terms of extreme events. The results show that, for general and seasonal assessment, IMERG Late generally performed best with good statistical metrics, with the highest CSI (0.343), CC (0.445), KGE ‘(0.295) in annual results, and the best agreement in capturing convective rain diurnal curves (summer CC = 0.820). GSMaP showed a better detection capability especially in spring and summer, and also performed better in capturing frequency curves, but always showed large estimation errors. IMERG and GSMaP detected worse CC region close to the water body. And the effect of high terrain on GSMaP NRT is more obvious in winter. For the evaluation in detecting the extreme events, IMERG and GSMaP overestimated the area and intensity and had a certain ability to discriminate the location of extreme precipitation, based on the temporal variation of HP-metrics. The rapid expansion or contraction of the extent of the extreme precipitation region could degrade the detection capability of HP-π. The comprehensive understanding of NSPPs provides valuable insights for upgrading satellite retrieval algorithms and making informed selection for precipitation products in meteorological hydrological applications.

Identification of sinkhole‐prone zones by successive coincidence deficit index analysis

AbstractHydrological data‐driven models require the time series of several hydrological events with different time resolutions. The interpretation of any time series event is generally difficult without some sort of filtering or converting it to a single index value. The simultaneous analysis of two or more hydrological events over a definite time span may be more informative about the region of interest. For this purpose, a new index, referred to as the successive coincidence deficit index (SCDI), was introduced to identify sinkhole‐prone regions using the persistent water deficit concept. In this study, monthly integrated multi‐satellite retrievals for GPM based precipitation (P) and gravity recovery and climate experiment‐based groundwater storage (GWS) datasets over Konya Closed Basin (KCB) in Türkiye were used to analyse the sinkhole occurrence. The main finding of this study is that SCDI distribution with high index values, concentrated on the southwestern part of KCB, is in line with the sinkholes occurred mainly after 2010. The proposed SCDI could also serve as a kind of drought index, which enables practitioners to quantify the relationship between drought and sinkhole occurrence. Moreover, the event coincidence analysis was utilized to detect deficiency in GWS over the KCB, which was associated with a rate of 0.8 for P deficiency, and this rate reaches up to 0.9 in the sinkhole region analysed in this study. As a conclusion, the proposed methodology can detect sinkhole‐prone regions to construct risk maps for stakeholders, policymakers, and end users.

Analysis of Fluid-Injection-Induced Seismicity Using a Dynamic Sliding Model Incorporating the Rate- and State-Dependent Friction Law

Summary Earthquakes can be triggered by fluid injection into underground formations. Fluid injection can cause large changes in the underground volume that exert stresses on nearby preexisting faults, leading to seismic activity. Assuming an increase in underground development activities in the future, our understanding of the mechanism underlying induced seismicity must be improved, and methods must be developed to properly assess the risk of seismic events. The objective of this study is to develop a seismicity prediction model that calculates the magnitude and timing of triggered earthquakes or seismic events occurring during various subsurface fluid injection activities. We developed an injection-induced seismicity analysis model that predicts the dynamic earthquake nucleation caused by changes in stress and pore pressure that occur during various subsurface activities. The governing equations consisting of the dynamic motion of the poroelastic spring-slider system, rate and state friction laws and pore pressure diffusion equation were solved using the embedded semi-implicit Runge-Kutta (SIRK) method. The dynamic sliding model was also incorporated into the finite element method (FEM) model, considering the variations in the stresses and pore pressures in the formation. A field case study was also conducted to compare the model results with typical microseismicity responses observed from hydraulic fracturing treatments in shale fields. Contrary to the popular understanding derived from Amonton’s law, the dynamic friction model revealed that a large normal stress on the fault leads to rapid sliding. A larger normal stress accumulates a large amount of elastic energy until it slips owing to fluid injection, nucleating large seismic waves. The poroelastic spring-slider model estimated reasonable microseismic magnitudes for hydraulic fracturing treatment but overestimated the time required to trigger a microseismic event under field conditions. To improve the analysis results, the poroelastic spring-slider model was coupled with a linear elastic FEM that considered the complex interplay of stress changes from hydraulic fracturing and the associated pore pressure variation in the formation. Compared with the field data, the coupled simulation model estimated a reasonable timing for the induced microseismic events when the increasing pore pressure during hydraulic fracturing penetrated deep into the formation. These findings suggest the existence of permeable natural fractures in the formation, which intensify early frictional sliding during treatment. The seismicity prediction model presented in this study simulates the magnitude and timing of seismic nucleation, helping to manage and mitigate the environmental impacts of induced seismicity during various subsurface development activities, such as oil and gas extraction, hydraulic fracturing, geothermal, and carbon dioxide sequestration. Moreover, the case study results imply that the time series of seismic events predicted by the model can be used to understand the possible fracture geometry and extent of fluid invasion for field applications. Key Terms and Phrases induced seismicity, subsurface fluid injection, rate- and state-dependent friction law, embedded semi-implicit Runge-Kutta method, finite element analysis

Studying the impact of fluctuations, spikes and rare events in time series through a wavelet entropy predictability measure

Data has become one of the most crucial sources of human life. In particular, the ability to predict the future through data is a widely studied topic. In finance, as an instance, increased volatility, fluctuations, low-frequency events, and rare events negatively affect the predictability of data, thus increasing the level of risk. As a consequence, the inability to make accurate predictions on future events increases the uncertainty and variability of a given scenario, indicating a consequent increase in risk. In this paper, we analyze data predictability introducing a new measure based on entropy and the wavelet transform. In particular, we show that the data are less predictable than one might expect due to the mentioned fluctuations and low-frequency events. Furthermore, we apply our tool to real data, in particular to time series of commodities. As a result, thanks to this new measure, we can observe that the price time series under analysis exhibit a significant level of unpredictability due to increased volatility, fluctuations, and the influence of low-frequency events.

Non fire-adapted dry forest of Northwestern Madagascar: Escalating and devastating trends revealed by Landsat timeseries and GEDI lidar data.

Ankarafantsika National Park (ANP), the last significant remnant of Northwestern Madagascar's tropical dry forests, is facing rapid degradation due to increased incidences of fire. This poses severe threats to biodiversity, local livelihoods, and vital ecosystem services. Our study, conducted on 3,052-ha of ANP's pristine forests, employed advanced remote-sensing techniques to assess fire impacts during the past 37 years. Our aims were to understand historical fire patterns and evaluate forest recovery and susceptibility to repeated fires following initial burns. Using data from multiple Landsat satellite sensors, we constructed a time series of fire events since 1985, which revealed no fire activity before 2014. The Global Ecosystem Dynamics Investigation (GEDI) lidar sensor data were used to observe forest structure in both post-fire areas and undisturbed zones for comparison. We recorded six fire incidents from 2014-2021, during which the fire-affected area exponentially grew. A significant fire incident in October 2021 impacted 1,052 hectares, 59% of which had experienced at least one fire in two-to-four years prior, with 60% experiencing two preceding incidents: one in 2017 and another in 2019. The initial fire drastically reduced plant cover and tree height, with subsequent fires causing minor additional loss. Post-fire recovery was negligible within the initial four years, even in patches without recurrent fires. The likelihood for an initial burn to trigger subsequent fires within a few years was high, leading to larger, more severe fires. We conclude that ANP's dry forests exhibit high vulnerability and low resilience to anthropogenic fires. Prompt preventive measures are essential to halt further fire spread and conserve the park's unique and invaluable biodiversity.

Predicting Typhoon Flood in Macau Using Dynamic Gaussian Bayesian Network and Surface Confluence Analysis

A typhoon passing through or making landfall in a coastal city may result in seawater intrusion and continuous rainfall, which may cause urban flooding. The urban flood disaster caused by a typhoon is a dynamic process that changes over time, and a dynamic Gaussian Bayesian network (DGBN) is used to model the time series events in this paper. The scene data generated by each typhoon are different, which means that each typhoon has different characteristics. This paper establishes multiple DGBNs based on the historical data of Macau flooding caused by multiple typhoons, and similar analysis is made between the scene data related to the current flooding to be predicted and the scene data of historical flooding. The DGBN most similar to the scene characteristics of the current flooding is selected as the predicting network of the current flooding. According to the topography, the influence of the surface confluence is considered, and the Manning formula analysis method is proposed. The Manning formula is combined with the DGBN to obtain the final prediction model, DGBN-m, which takes into account the effects of time series and non-time-series factors. The flooding data provided by the Macau Meteorological Bureau are used to carry out experiments, and it is proved that the proposed model can predict the flooding depth well in a specific area of Macau under the condition of a small amount of data and that the best predicting accuracy can reach 84%. Finally, generalization analysis is performed to further confirm the validity of the proposed model.

A kinematic method for generating earthquake sequences

Computational earthquake sequence models provide generative estimates of the time, location, and size of synthetic seismic events that can be compared with observed earthquake histories and assessed as rupture forecasts. Here we describe a three-dimensional probabilistic earthquake sequence model that produces slip event time series constrained across geometrically complex non-planar fault systems. This model is kinematic in nature, integrating the time evolution of geometric moment accumulation and release with empirical earthquake scaling laws. The temporal probability of event occurrence is determined from the time history of geometric moment integrated with short-term Omori-style rate decay following each earthquake achieving long-term time-averaged moment balance. Similarly, the net geometric moment monotonically controls the probability of event localization, and seismic events release geometric moment with spatially heterogeneous slip on three-dimensional non-planar fault surfaces. We use this model to generate a synthetic earthquake sequence on the Nankai subduction zone over a 1,250-year-long interval, including 700+ MW=5.5−8.5 coseismic events, with decadal-to-centennial scale quiescent intervals quasi-periodic great earthquake clusters followed by aftershock sequences.

Modeling Extreme Events in Time Series and Their Impact on Seasonal Adjustment in the Post-Covid-19 Era

Modeling Extreme Events in Time Series and Their Impact on Seasonal Adjustment in the Post-Covid-19 Era

Time series synchronization in cross-recurrence networks: uncovering a homomorphic law across diverse complex systems

Exploring the synchronicity between time series, especially the similar patterns during extreme events, has been a focal point of research in academia. This is due to the fact that such special dependence occurring between pairs of time series often plays a crucial role in triggering emergent behaviors in the underlying systems and is closely related to systemic risks. In this paper, we investigate the relationship between the synchronicity of time series and the corresponding topological properties of the cross-recurrence network (CRN). We discover a positive linear relationship between the probability of pairwise time series event synchronicity and the corresponding CRN’s clustering coefficient. We first provide theoretical proof, then demonstrate this relationship through simulation experiments by coupled map lattices. Finally, we empirically analyze three instances from financial systems, Earth’s ecological systems, and human interactive behavioral systems to validate that this regularity is a homomorphic law in different complex systems. The discovered regularity holds significant potential for applications in monitoring financial system risks, extreme weather events, and more.

A query-response causal analysis of reaction events in biochemical reaction networks

The stochastic kinetics of biochemical reaction networks is described by a chemical master equation (CME) and the underlying laws of mass action. Assuming network-free simulations of the rule-based models of biochemical reaction networks (BRNs), this paper departs from the usual analysis of network dynamics as the time-dependent distributions of chemical species counts, and instead considers statistically evaluating the sequences of reaction events generated from the stochastic simulations. The reaction event-time series can be used for reaction clustering, identifying rare events, and recognizing the periods of increased or steady-state activity. However, the main aim of this paper is to device an effective method for identifying causally and anti-causally related sub-sequences of reaction events using their empirical probabilities. This allows discovering some of the causal dynamics of BRNs as well as uncovering their short-term deterministic behaviors. In particular, it is proposed that the reaction sub-sequences that are conditionally nearly certain or nearly uncertain can be considered as being causally related. Moreover, since the time-ordering of reaction events is locally irrelevant, the reaction sub-sequences can be transformed into the reaction sets or multi-sets. The distance metrics can be then used to define the equivalences among the reaction events. The proposed method for identifying the causally related reaction sub-sequences has been implemented as a computationally efficient query-response mechanism. The method was evaluated for five models of genetic networks in seven defined numerical experiments. The models were simulated in BioNetGen using the open-source network-free simulator NFsim. This simulator had to be modified first to allow recording the traces of reaction events, and it is available in the Github repository, ploskot/nfsim_1.20. The generated event time-series were analyzed with Python and Matlab scripts. The whole process of data generation, analysis and visualization has been nearly fully automated using shell scripts. This demonstrates the opportunities for substantially increasing the research productivity by creating automated data generation and processing pipelines.

Rockfall monitoring with a Doppler radar on an active rockslide complex in Brienz/Brinzauls (Switzerland)

Abstract. We present and analyze a rockfall catalog from an active landslide complex in Brienz/Brinzauls of the Swiss Alps, collected with a new Doppler radar system. This radar system provides a complete and continuous time series of rockfall events with volumes of 1 m3 and greater since 2018 and serves as automatic traffic control for an important main road. In the period between January 2018 and October 2022, 6743 events were detected, which is 2 orders of magnitude higher activity than in stable continental cliffs. A few percent of all rockfall events reached the shadow area, which hosts an important road and agricultural area. The Doppler radar data set allows us to investigate the triggering factors quantitatively. We found that the background rockfall activity is controlled by seasonal climatic triggers. In winter, more rockfalls are observed during thawing periods, whereas in summer the rockfall activity increases with hourly rainfall intensity. We also found that, due to the geological setting in an active landslide complex, increased rockfall activity occurs clustered in space and time, triggered by local displacement hotspots. Thus, monitoring spatial and temporal variations of slope displacement velocity is crucial for detailed rockfall hazard assessment in similar geological settings.

A machine learning approach for slow slip event detection using GNSS time-series

Extracting tectonic transient displacements on the Earth’s surface from Global Navigation Satellite System (GNSS) time series remains a challenge, because GNSS station displacements depend on multiple processes occurring simultaneously, along with noise that obscures low-magnitude transient signals. We present a novel method for automatic detection of slow slip events (SSEs) in time series of a GNSS network by training a supervised machine learning (ML) model for classification. The proposed methodology detects both temporally and spatially the signatures of SSEs or regional transients within a GNSS network. The time series of a GNSS network were transformed into grayscale images, from which descriptors, including Bag of Visual Words (BoW) and Extended Local Binary Patterns (ELBP), were extracted. These descriptors served as input features for two distinct ML models: Support Vector Machines (SVM) and Artificial Neural Networks (NN). To train and test the ML classification model, two 3-year synthetic datasets were generated, one with GNSS networks featuring slow slip events (SSEs) of varying location, duration, onset time, and magnitude, and the other without SSEs, resulting in positive and negative sets, respectively. For each GNSS network, an image was created by combining the east and north components of the time series, which have been previously detrended and common mode error filtered. Each image is further divided into sub-images corresponding to 60 days time windows, in order to temporarily detect the existence of a transient. For training and testing, the datasets were separated into 75% for training and 25% for testing, each with 50% positive and 50% negative cases. In the final step, we analyze the positively classified images, representing the time windows in which the classifier detected transients. Within each of these windows, we identify the network’s time series with the highest velocity, indicating the stations and geographic area where the detected transients occurred. The test results demonstrate that both ML models achieved high performance using both ELBP and BoW descriptors as features. Finally, our ML models were validated on a real dataset with a transient signal recorded before the 2014 Iquique earthquake in Chile, and they effectively detected this anomalous signal. The proposed method can effectively detect transient signals related to SSEs with high accuracy, sensitivity, and specificity in both the test and instrumentally recorded datasets.

Mitigation of assembly constraints for floating offshore wind turbines using discrete event simulation

There is a large and increasing pipeline of floating offshore wind projects with total global floating offshore wind capacity projected to grow year on year by, on average, between 59 and 104 % in the 2020s. This will lead to competition for infrastructure resources, in particular, port facilities for the construction and marshalling of the floating foundations and turbines. It is likely that multiple ports will need to be combined to provide the necessary fabrication capacity for a floating offshore wind farm of commercial scale. To enable an efficient and coordinated utilisation of multiple fabrication ports, it is crucial to understand the likely duration of different assembly and construction activities at different locations. However, at present this task is difficult to perform using top-down estimation models, as commercial-scale floating offshore wind farms comprising many tens of units have not been built to date. In this work we present a methodology, based on discrete event simulation (DES) and time series analysis, to produce an explicit simulation-based estimate of assembly activity durations, which are sensitive to setting specific factors. Three example case studies are outlined to demonstrate the ability to capture the variation in activity duration due to resource availability, and the season and location of activity. The methodology will be of use to project planners as it can be used at an early stage in the project life-cycle to appraise and adopt different construction strategies.

Complexity Synchronization of Organ Networks.

The transdisciplinary nature of science as a whole became evident as the necessity for the complex nature of phenomena to explain social and life science, along with the physical sciences, blossomed into complexity theory and most recently into complexitysynchronization. This science motif is based on the scaling arising from the 1/f-variability in complex dynamic networks and the need for a network of networks to exchange information internally during intra-network dynamics and externally during inter-network dynamics. The measure of complexity adopted herein is the multifractal dimension of the crucial event time series generated by an organ network, and the difference in the multifractal dimensions of two organ networks quantifies the relative complexity between interacting complex networks. Information flows from dynamic networks at a higher level of complexity to those at lower levels of complexity, as summarized in the 'complexity matching effect', and the flow is maximally efficient when the complexities are equal. Herein, we use the scaling of empirical datasets from the brain, cardiovascular and respiratory networks to support the hypothesis that complexity synchronization occurs between scaling indices or equivalently with the matching of the time dependencies of the networks' multifractal dimensions.

Training LSTMS with circular-shift epochs for accurate event forecasting in imbalanced time series

The use of LSTM models for forecasting has increased substantially because they have proven to yield excellent performance in numerous applications involving time series consisting of normal observations. However, LSTM models tend to perform poorly in forecasting extreme events in imbalanced time series due to the lack of representative extreme event data during training. This study introduces a breakthrough Circular-Shift Epoch Training (CSET) strategy for LSTMs that makes the accurate forecasting of extreme events no more difficult than predicting events in normal time series with similar accuracies. CSET is (a) specifically designed to preserve the natural ordering of observations in epochs during training without any attempt to balance extreme and normal observations, (b) formulated for both univariate and multivariate time series forecasting, (c) applicable to scaler and vector forecasting for the multivariate case and can handle upward, downward, and mixed extremes without any modification, and (d) universal because it can be applied to train LSTMs to forecast events in normal time series or in imbalanced time series in exactly the same manner. Most importantly, the forecasting of normal time series and imbalanced time series do not have to be treated as separate problems. The results from an extensive set of experiments involving forecasting events in simulated and real imbalanced time series confirm that LSTM models trained using CSET dramatically outperform the standard fixed-epoch trained baseline models in (a) forecasting both extreme and normal events and (b) classifying extreme events. Given the ubiquitous use of LSTMs for forecasting, the accurate forecasting of extreme events through the incorporation of the innovative CSET strategy can have a major impact in decreasing the disastrous consequences of dangerous extreme events.