Continuous Series Research Articles

A comparative morphology of trichobothrial bases in araneoid spiders and its significance for the phylogeny and system of the superfamily Araneoidea (Arachnida, Araneae)

Bothrial morphology was studied by SEM in 137 araneoid genera representing all 22 currently recognized extant families and all 42 conventional subfamilies of the Araneoidea. The ancestral type in the superfamily Araneoidea is a ‘hooded’ bothrium with a single well-developed transverse ridge, dividing its proximal and distal plates (‘Erigone-type’); the advanced type is a solid dome-like bothrium without vestiges of the ridge (‘Theridion-type’); there are several intermediate types reflecting various pathways and stages of the ridge reduction (united here as ‘Argiope-type’). The parallel trends in bothrial evolution, recognized as continuous series from the ancestral type up to the advanced one through some intermediate stages, are distinguished in each of the seven main phylogenetic lineages of the superfamily: ‘tetragnathoids’, ‘araneoids’, ‘cyatholipoids’, and ‘theridioids’ possess a complete set of the three types, while ‘malkariods’, ‘symphytognathoids’. and ‘linyphioids’ lack the advanced, dome-like type (‘Theridion-type’). Only three taxa have been proposed earlier as the sister group of the superfamily Araneoidea: Nicodamoidea, Deinopoidea, and Leptonetoidea; morphology of bothria, as well as other cuticular microstructures, clearly supports the araneoid-nicodamoid relationship hypothesis, purely ‘molecular’ to date. Bothrial morphology provides the additional arguments for several taxonomic acts, e.g., for the reranking the Agnarsson’s (2004) ‘clade 35’ (Theonoe, Carniella, Robertus, and Pholcomma) up to the Theonoeinae Simon, 1894, stat. nov., and for the revalidation the micropholcommatid Plectochetos Butler, 1932, gen. revalid. and zygiellid Parazygiella Wunderlich, 2004, gen. revalid.

The annual cycle and sources of relevant aerosol precursor vapors in the central Arctic during the MOSAiC expedition

Abstract. In this study, we present and analyze the first continuous time series of relevant aerosol precursor vapors from the central Arctic (north of 80° N) during the Multidisciplinary drifting Observatory for the Study of Arctic Climate (MOSAiC) expedition. These precursor vapors include sulfuric acid (SA), methanesulfonic acid (MSA), and iodic acid (IA). We use FLEXPART simulations, inverse modeling, sulfur dioxide (SO2) mixing ratios, and chlorophyll a (chl a) observations to interpret the seasonal variability in the vapor concentrations and identify dominant sources. Our results show that both natural and anthropogenic sources are relevant for the concentrations of SA in the Arctic, but anthropogenic sources associated with Arctic haze are the most prevalent. MSA concentrations are an order of magnitude higher during polar day than during polar night due to seasonal changes in biological activity. Peak MSA concentrations were observed in May, which corresponds with the timing of the annual peak in chl a concentrations north of 75° N. IA concentrations exhibit two distinct peaks during the year, namely a dominant peak in spring and a secondary peak in autumn, suggesting that seasonal IA concentrations depend on both solar radiation and sea ice conditions. In general, the seasonal cycles of SA, MSA, and IA in the central Arctic Ocean are related to sea ice conditions, and we expect that changes in the Arctic environment will affect the concentrations of these vapors in the future. The magnitude of these changes and the subsequent influence on aerosol processes remains uncertain, highlighting the need for continued observations of these precursor vapors in the Arctic.

An Improved Transformer Model for Sap Flow Prediction that Efficiently Utilizes Environmental Information

AbstractAs an important reference for assessing plant water consumption and estimating plant transpiration, it is of great significance to achieve accurate prediction of plant sap flow. A number of deep learning models were established and compared using approximately 3 years of continuous eucalyptus flow time series data collected from the SAPFLUXNET open dataset and 6 environmental factors, including shortwave solar incident radiation, air temperature, air relative humidity, net radiation, vapor pressure deficit, and photosynthetic photon flux density. The experimental results show that the improved Transformer model, with the introduction of a two-step self-attention mechanism and simplified design, maintains significant predictive performance advantages compared to the original Transformer model, long short-term memory, gated recurrent unit, and temporal convolutional neural network models. In the shorter 1-h forecast, the mean squared error and coefficient of determination (R2) of the improved Transformer model are 0.0191 and 0.965, respectively. Compared to the suboptimal typical Transformer model, the MSE is reduced by 22.9%, and R2 is increased by 1.0%. Additionally, the improved model maintains stable predictive performance advantages in long-term plant flow prediction. In the longest 8-h advance prediction, the MSE is reduced by 14.9% compared to the suboptimal Transformer model, and R2 increases by 3.0% compared to the Transformer model. The comprehensive experimental results show that the improved Transformer model makes more effective use of environmental information to achieve more accurate and long-term plant flow prediction. This study emphasizes the basic principle and validity of the two-step self-attention network structure and provides a valuable basis for developing more effective methods for predicting plant sap flow.

Real-Time and Long-Term Monitoring of Coastal Water Turbidity Using an Ocean Buoy Equipped with an ADCP.

In this study, an acoustic Doppler current profiler (ADCP) operating at 600 kHz was installed on an ocean data buoy in the Qigu waters, Taiwan, to gather real-time sound echo intensity data. These data were then correlated with turbidity measurements obtained by a turbidimeter mounted on the buoy's mooring line at a water depth of 13 m. The data buoy operated from 6 June to 16 August 2017. During this period, turbidity measurements were recorded from 6 to 21 June 2017. This study established a calibration between the sound echo intensity measured by the ADCP and the turbidity measured using the turbidimeter; a strong linear correlation was discovered between these two variables. This correlation enabled the conversion of echo intensity data into a continuous time series of turbidity measurements, facilitating real-time and long-term monitoring of coastal water turbidity through the deployment of a buoy equipped with an ADCP. The relationships between turbidity and environmental factors such as rainfall, tides, current speeds, and wave activity over an extended period were then investigated. The results revealed that stronger tides and currents in the Qigu waters often lead to higher turbidity, suggesting that these two factors are the primary driving forces for sediment transport in the Qigu waters. Additionally, sampling of water in the Qigu area revealed sediment particles of size ranging from 2 to 120 μm.

Development of laser felling modes of gas-thermal coating

The use of copper and its alloys to create parts for metallurgical equipment is associated with an increase in abrasive wear and high-temperature corrosion. In this regard, there is a need to apply a protective coating. In particular, to prevent wear and premature chipping of the metal of copper tuyeres, the surface is hardened with a coating of zirconium dioxide stabilized with yttria oxide by thermal spraying in an air atmosphere. Due to the difference in the coefficient of thermal expansion of copper (at T = 300 K: 16.7 µm/m оС and at T = 750 K: 19.7 µm/m оС) and its low resistance to gas corrosion, the application of zirconium oxide (produced by a preapplied intermediate layer that plays a role in matching the coefficient of thermal expansion (CTE) between the copper base and the ceramic coating. In addition, the intermediate layer protects copper from gas corrosion. In this case, The use of copper and its alloys to create parts for metallurgical equipment is associated with an increase in abrasive wear and high-temperature corrosion. In this regard, there is a need to apply a protective coating. In particular, to prevent wear and premature chipping of the metal of copper tuyeres, the surface is hardened with a coating of zirconium dioxide stabilized with yttria oxide by thermal spraying in an air atmosphere. Due to the difference in the coefficient of thermal expansion of copper (at T = 300 K: 16.7 pm/m °G and at T = 750 K: 19.7 pm/m оG) and its low resistance to gas corrosion, the application of zirconium oxide (produced by a pre-applied intermediate layer that plays a role in matching the coefficient of thermal expansion (CTE) between the copper base and the ceramic coating. In addition, the intermediate layer protects copper from gas corrosion. In this case, nickel-based alloys were used as intermediate layers. The use of nickel as the basis of intermediate layers is due to the fact that copper and nickel form a continuous series of solid solutions, such as cupronickel or monel metal-like structures. This, in turn, assumes a smooth transition of thermophysical properties from copper to nickel alloy. To ensure increased adhesion of the transition layer to copper by increasing the area of mutual contact between copper and the sublayer (dagger penetration) and significantly increasing the homogeneity of the material of the intermediate layer made of a nickel alloy, laser melting of the intermediate sublayer (Ni–B–Si system) was used on a laser complex based on laser LS-5 with a power of 5 kW with a KUKA KR-60HA robot in an argon atmosphere. To test the modes, experiments were carried out on copper samples of a flat shape and a body of rotation. The optimal parameters for the process of melting flat samples were: processing speed 33 mm/s, power from 400 to 3900 W, focal length from 200 to 230 mm, pitch between tracks: 0.25, 0.5 and 1 mm. The optimal parameters for the process of melting rotating samples were: laser radiation power 400–450 W, processing step 0.125; 0.5, focal length from 200 to 210 mm.

Metagenomic time series reveals a Western English Channel viral community dominated by members with strong seasonal signals.

Marine viruses are key players of ocean biogeochemistry, profoundly influencing microbial community ecology and evolution. Despite their importance, few studies have explored continuous inter-seasonal viral metagenomic time series in marine environments. Viral dynamics are complex, influenced by multiple factors such as host population dynamics and environmental conditions. To disentangle the complexity of viral communities, we developed an unsupervised machine learning framework to classify viral contigs into "chronotypes" based on temporal abundance patterns. Analysing an inter-seasonal monthly time series of surface viral metagenomes from the Western English Channel, we identified chronotypes and compared their functional and evolutionary profiles. Results revealed a consistent annual cycle with steep compositional changes from winter to summer and steadier transitions from summer to winter. Seasonal chronotypes were enriched in potential auxiliary metabolic genes of the ferrochelatases and 2OG-Fe(II) oxygenase orthologous groups compared to non-seasonal types. Chronotypes clustered into four groups based on their correlation profiles with environmental parameters, primarily driven by temperature and nutrients. Viral contigs exhibited a rapid turnover of polymorphisms, akin to Red Queen dynamics. However, within seasonal chronotypes, some sequences exhibited annual polymorphism recurrence, suggesting that a fraction of the seasonal viral populations evolve more slowly. Classification into chronotypes revealed viral genomic signatures linked to temporal patterns, likely reflecting metabolic adaptations to environmental fluctuations and host dynamics. This novel framework enables the identification of long-term trends in viral composition, environmental influences on genomic structure, and potential viral interactions.

A supply-limited torrent that does not feel the heat of climate change

Debris-flow activity in the Alps is anticipated to undergo pronounced changes in response to a warming climate. Yet, a fundamental challenge in comprehensively assessing changes in process activity is the systematic lack of long-term observational debris-flow records. Here, we reconstruct the longest, continuous time series (1626-2020) of debris flows at Multetta, a supply-limited torrential system in the Eastern Swiss Alps. Relying on growth-ring records of trees that were damaged by debris flows, we do not detect significant changes in the frequency or magnitude over time. This seeming absence of a direct climatic influence on debris-flow initiation aligns with the regular distribution of repose time patterns, indicating a dependence of local process activity on sediment discharge and recharge. This stark difference in process behavior between our supply-limited site and transport-limited catchments has implications for assessing torrential hazard and risk mitigation in a context of global warming.

Comparing the hydrological performance of blue green infrastructure design strategies in urban/semi-urban catchments for stormwater management

ABSTRACT Blue green infrastructure (BGI), in recent decades, have been increasingly recognized as robust stormwater control measures to reduce urban flooding, promote infiltration, and restore a catchment's flow to its pre-development stage. However, studies comparing the hydrological benefits of BGI alternatives at catchment scale are often limited to single catchment or single/few BGI options scaled over a catchment. This study designed a set of BGI alternatives as a combination of different BGI facilities in terms of the following: (a) spatial distribution scale (end-of-pipe vs. decentralized) and (b) naturalness scale (less engineered vs. more engineered), in three different urban catchments representing an inner city, a residential suburb, and a new urban housing. In addition, their hydrological performances were compared. A 10-year return period design rain and a continuous rain series of 11 years were modelled for each BGI alternative using the computer model stormwater management model (SWMM). It was observed that in most catchments, decentralized alternatives (both engineered and natural) showed better potential to reduce the magnitude and frequency of flooding than centralized measures. Similarly, the tested decentralized natural, less engineered alternatives showed higher potential to increase infiltration than the decentralized engineered alternatives in all three catchments. Meanwhile, infiltration-based BGI alternatives showed similar potential to mimic pre-development flow as other decentralized BGI alternatives.

A two-step approach for damage identification in bridge structure using convolutional Long Short-Term Memory with augmented time-series data

This paper presents a novel two-step approach to identifying structural damages in bridge structure through the integration of 1D Convolutional Neural Network (1DCNN) and Long Short-Term Memory (LSTM) networks, enhanced by the augmentation and transformation techniques using Symbolic Aggregate approXimation (SAX) for time-series data analysis. In the first step, the time-series data of the bridge is diversified and quantified by augmentation techniques to make the model more robust and increase its generalization capabilities. After that, SAX is implemented to reduce the volume and categorize time series data through the transformation of continuous time series into discrete symbols, thereby decreasing the size of the data for more efficient training performance. In the second step, an advanced DL model combining 1DCNN and LSTM is proposed to tackle the damage identification problems of the processed data. By leveraging the strengths of CNNs in feature extraction and LSTMs in sequence learning, combined with advanced techniques for data augmentation, our methodology offers a robust solution not only for improving the model's training process but also for enabling it to learn from a more diverse and comprehensive dataset that mimics different damage scenarios, allowing more accurate detection of damages within bridge structures. Validation of the proposed method is conducted using time-series data collected from Chuong Duong Bridge structure. The effectiveness of the proposed method is compared with other models, such as 1DCNN, LSTM, and the combined 1DCNN-LSTM. The results show that the proposed 1DCNN-LSTM-SAX outperforms the other methods in terms of accuracy and, thus, can be used extensively to deal with the damage identification problems of bridges using time-series data.

An Evaluation of Sentinel-3 SYN VGT Products in Comparison to the SPOT/VEGETATION and PROBA-V Archives

Sentinel-3 synergy (SYN) VEGETATION (VGT) products were designed to provide continuity to the SPOT/VEGETATION (SPOT VGT) base products archive. Since the PROBA-V mission acted as a gap filler between SPOT VGT and Sentinel-3, and in principle, a continuous series of data products from the combined data archives of SPOT VGT (1998–2014), PROBA-V (2013–2020) and Sentinel-3 SYN VGT (from 2018 onwards) are available to users, the consistency of Sentinel-3 SYN VGT with both the latest SPOT VGT (VGT-C3) and PROBA-V (PV-C2) archives is highly relevant. In past years, important changes have been implemented in the SYN VGT processing baseline. The archive of SYN VGT products is therefore intrinsically inconsistent, leading to different consistency levels with SPOT VGT and PROBA-V throughout the years. A spatio-temporal intercomparison of the combined time series of VGT-C3, PV-C2 and Sentinel-3 SYN VGT 10-day NDVI composite products with an external reference from LSA-SAF, and an intercomparison of Sentinel-3 SYN V10 products with a climatology of VGT-C3 resp. PV-C2 for three distinct periods with different levels of product quality have shown that the subsequent processing baseline updates have indeed resulted in better-quality products. It is therefore essential to reprocess the entire Sentinel-3 SYN VGT archive; a uniform data record of standard SPOT VGT, PROBA-V and Sentinel-3 SYN VGT products, spanning over 25 years, would provide valuable input for a wide range of applications.

3D Surface Velocity Field Inferred from SAR Interferometry: Cerro Prieto Step-Over, Mexico, Case Study

The Cerro Prieto basin, a tectonically active pull-apart basin, hosts significant geothermal resources currently being exploited in the Cerro Prieto Geothermal Field (CPGF). Consequently, natural tectonic processes and anthropogenic activities contribute to three-dimensional surface displacements in this pull-apart basin. Here, we obtained the Cerro Prieto Step-Over 3D surface velocity field (3DSVF) by accomplishing a weighted least square algorithm inversion from geometrically quasi-orthogonal airborne UAVSAR and RADARSAT-2, Sentinel 1A satellite Synthetic Aperture-Radar (SAR) imagery collected from 2012 to 2016. The 3DSVF results show a vertical rate of 150 mm/yr and 40 mm/yr for the horizontal rate, where for the first time, the north component displacement is achieved by using only the Interferometric SAR time series in the CPGF. Data integration and validation between the 3DSVF and ground-based measurements such as continuous GPS time series and precise leveling data were achieved. Correlating the findings with recent geothermal energy production revealed a subsidence rate slowdown that aligns with the CPGF’s annual vapor production.

Investigating a Century of Rainfall: The Impact of Elevation on Precipitation Changes (Northern Tuscany, Italy)

Precipitation is crucial for water resource renewal, but climate change alters their frequency and amounts, challenging societies for correct and effective water management. However, modifications of precipitation dynamics appear to be not uniformly distributed, both in space and time. Even in relatively small areas, precipitation shows the coexistence of positive and negative trends. Local topography seems to be a strong driver of precipitation changes. Understanding precipitation changes and their relationship with local topography is crucial for society’s resilience. Taking advantage of a dense and long-lasting (1920–2019) meteorological monitoring network, we analyzed the precipitation changes over the last century in a sensitive and strategic area in the Mediterranean hotspot. The study area corresponds to northern Tuscany (Italy), where its topography comprises mountain ridges and coastal and river plains. Forty-eight rain gauges were selected with continuous annual precipitation time series. These were analyzed for trends and differences in mean annual precipitation between the stable period of 1921–1970 and the last 30-year 1990–2019. The relationship between precipitation changes and local topography was also examined. The results show the following highlights: (i) A general decrease in precipitation was found through the century, even if variability is marked. (ii) The mountain ridges show the largest decrease in mean annual precipitation. (iii) The precipitation change entity over the last century was not homogenous and was dependent on topography and geographical setting. (iv) A decrease in annual precipitation of up to 400 mm was found for the mountainous sites.

A novel homological approach for the comprehensive study of nonstoichiometric oxide with exceptional oxygen mobility

Actual work devoted to the development of a novel homological approach for describing the kinetic properties of grossly nonstoichiometric oxides ABO3-δ with perovskite structure. The approach consists in considering such oxides having different oxygen stoichiometry 3-δ as chemical homologues involved in the same oxygen exchange reaction and forming a continuous series in δ. The proposed approach is considered using the oxide with exceptional oxygen mobility Ba0·5Sr0·5Co0·8Fe0·2O3-δ (BSCF). The kinetic properties of BSCF were studied by new method of the oxygen partial pressure relaxation. The equilibrium properties of BSCF were reported earlier and obtained by original technique of the quasi-equilibrium oxygen release. The power law dependence of the kinetic parameters on the equilibrium oxygen partial pressure is considered as the consequence of linear free energy relationship similar to the Brønsted-Evans-Polanyi relation.

Vitamin B12 deficiency in newborns: impact on individual's health status and healthcare costs.

The identification of vitamin B12 (B12) deficiency in the newborn may prevent neurological damage and a delay in the normal growth. In this study we characterized the incidence of B12 deficiency in newborns, the costs associated to the clinical diagnosis and management, and the relevance to optimize the use of cobalamin biomarkers during treatment follow-up. Starting from a continuous case series of 146,470 screened newborns (November, 1st 2021- December, 3rd 2023), the Regional Reference Laboratory for Neonatal Screening identified 87 newborns having altered levels of biomarkers of cobalamin metabolism measured by Newborn Screening. These subjects were confirmed with a nutritional B12 deficiency of maternal origin by performing the serum B12 measurements and plasma homocysteine (Hcy) both on the newborns and respective mothers. A cost analysis was performed to characterize the costs/year of identifying and managing B12 deficiency cases. At baseline, median (interquartile range) serum B12 levels of 185.0 (142.3-246.0) ng/L and threefold increased plasma Hcy concentrations above the normal level confirmed a severe condition of deficiency in the newborns. After intramuscular B12 supplementation, serum B12 measured at the first follow up visit showed a fivefold increase, and the levels of Hcy returned to normal. From the healthcare perspective, the costs for diagnosing and managing all newborns with B12 deficiency is 188,480 €/year. Preventing B12 depletion in newborns lowers healthcare costs and likely improves their health outcomes. Further studies are however required to address the clinical pathway to identify, treat and monitor pregnant women with marginal and low B12 status, in order to achieve these goals.

Advanced Wind Power Forecasting

The popularity of LSTM networks in wind prediction is due to their inherent capabilities in handling the long-term dependencies of sequential data. This normally outperforms traditional methods like ARIMA. However, current work has illustrated that EMD-based approaches outperform LSTM models in many critical performance metrics. Although it is very powerful in gathering historical patterns due to the enhanced gating mechanisms, LSTM has a lot of challenges regarding computational efficiency and flexibility. EMD models overwhelmingly outperformed the LSTM model in computational time by probably about 20-25%, thereby solving a weakness that may be one of the intrinsic weaknesses of the LSTM model. In addition, the EMD models improved by about 2-4% in accuracy as compared to LSTM, thus evidencing better ability in the prediction of the wind data. On feature extraction, EMD did better, improving by 7% over LSTM. Moreover, EMD has better robustness and adaptability, with improvements of 6% and 4%, respectively. This introduction of EMD does bring a bit of an increase in model complexity, this disadvantage is rather insignificant when compared with the huge advantages it brings in efficiency and accuracy. While an LSTM network is very good at modeling a long-term relationship, it suffers from some problems in handling nonlinear interactions and continuous series data, thus its effectiveness may be limited in some cases of forecasting. In contrast, EMD has a clear advantage in real-time applications owing to enhanced efficiency and precision. From the results obtained in this study, it becomes easy to envision that the EMD methodology is un-equivocally better than LSTM in all the critical aspects, such as computational efficiency, accuracy in prediction, feature extraction, resilience, and adaptability when used individually. Results indicate that EMD has a more constructive methodology toward wind pattern predictions, beating LSTM on many measures of performance. The advantages presented should be maximized in future studies aimed at further improving and optimizing a wind-forecasting system, with special attention toward EMD's benefits in gaining more accurate and efficient forecasts.

A 40-Year Time Series of Land Surface Emissivity Derived from AVHRR Sensors: A Fennoscandian Perspective

Accurate land surface temperature (LST) retrieval depends on precise knowledge of the land surface emissivity (LSE). Neglecting or inaccurately estimating the emissivity introduces substantial errors and uncertainty in LST measurements. The emissivity, which varies across different surfaces and land uses, reflects material composition and surface roughness. Satellite data offer a robust means to determine LSE at large scales. This study utilises the Normalised Difference Vegetation Index Threshold Method (NDVITHM) to produce a novel emissivity dataset spanning the last 40 years, specifically tailored for the Fennoscandian region, including Norway, Sweden, and Finland. Leveraging the long and continuous data series from the Advanced Very High Resolution Radiometer (AVHRR) sensors aboard the NOAA and MetOp satellites, an emissivity dataset is generated for 1981–2022. This dataset incorporates snow-cover information, enabling the creation of annual emissivity time series that account for winter conditions. LSE time series were extracted for six 15 × 15 km study sites and compared against the Moderate Resolution Imaging Spectroradiometer (MODIS) MOD11A2 LSE product. The intercomparison reveals that, while both datasets generally align, significant seasonal differences exist. These disparities are attributable to differences in spectral response functions and temporal resolutions, as well as the method considering fixed values employed to calculate the emissivity. This study presents, for the first time, a 40-year time series of the emissivity for AVHRR channels 4 and 5 in Fennoscandia, highlighting the seasonal variability, land-cover influences, and wavelength-dependent emissivity differences. This dataset provides a valuable resource for future research on long-term land surface temperature and emissivity (LST&E) trends, as well as land-cover changes in the region, particularly with the use of Sentinel-3 data and upcoming missions such as EUMETSAT’s MetOp Second Generation, scheduled for launch in 2025.

Explainability of CNN-based Alzheimer's disease detection from online handwriting.

With over 55 million people globally affected by dementia and nearly 10 million new cases reported annually, Alzheimer's disease is a prevalent and challenging neurodegenerative disorder. Despite significant advancements in machine learning techniques for Alzheimer's disease detection, the widespread adoption of deep learning models raises concerns about their explainability. The lack of explainability in deep learning models for online handwriting analysis is a critical gap in the literature in the context of Alzheimer's disease detection. This paper addresses this challenge by interpreting predictions from a Convolutional Neural Network applied to multivariate time series data, generated by online handwriting data associated with continuous loop series handwritten on a graphical tablet. Our explainability methods reveal distinct motor behavior characteristics for healthy individuals and those diagnosed with Alzheimer's. Healthy subjects exhibited consistent, smooth movements, while Alzheimer's patients demonstrated erratic patterns marked by abrupt stops and direction changes. This emphasizes the critical role of explainability in translating complex models into clinically relevant insights. Our research contributes to the enhancement of early diagnosis, providing significant and reliable insights to stakeholders involved in patient care and intervention strategies. Our work bridges the gap between machine learning predictions and clinical insights, fostering a more effective and understandable application of advanced models for Alzheimer's disease assessment.

The continuous memory: A neural network with ordinary differential equations for continuous-time series analysis

Continuous-time series analysis has garnered significant research interest due to its extensive applications; however, it remains a challenging endeavor. In recent years, deep learning methods have achieved remarkable success in tasks such as time series classification and forecasting. Nevertheless, the inherent continuity of time series data has not been fully addressed, presenting a persistent obstacle to performance. Recent studies have highlighted a natural similarity between differential equations and continuous-time series, as both inherently encapsulate the concept of continuity. However, several critical issues with differential equations hinder their direct applicability to time series analysis. These issues include the elusive nature of analytical solutions, the indirect observability of the effects of differential equation parameters, and the dependence of numerical solutions on initial conditions. In this context, we propose the integration of differential equations into neural networks to serve as the continuous memory of the model. This integration imparts a continuous nature to the model, resulting in the development of an efficient deep learning architecture known as Long Short Deep Memory (LSDM). Furthermore, we analyze the characteristics of differential equations when employed as the memory component in neural networks, which leads to the proposal of a novel pre-training approach that incorporates an innate memory mechanism into these networks. Additionally, we utilize LSDM to construct a stacked architecture for processing very long time series data. The proposed model can naturally accommodate arbitrary time gaps between observations, thereby enhancing its effectiveness and suitability for continuous time series analysis tasks. Extensive experiments conducted on various real-world datasets demonstrate that the proposed model outperforms existing methods, providing a new solution to the challenges associated with continuous time series analysis.

Simultaneous measurement of temperature and C2H4 concentration in hydrocarbon flames using interference-immune differential absorption spectroscopy at 3.3 μm

In diagnostic applications based on tunable diode laser absorption spectroscopy, the measurement of target substances can be influenced by factors such as background thermal radiation in the combustion environment, extinction caused by solid or liquid particles, and other interfering absorptions. In this work, we developed a differential absorption diagnostic technique based on wavelength pairs, utilizing an interband cascade laser near 3.3 μm to simultaneously measure temperature and C2H4 concentration in hydrocarbon flames. Based on a detailed study of the C2H4 spectrum in this region and considering the optimal standard for spectral lines, two wavelength pairs were selected. The temperature is determined by the ratio of the absorption cross-sections of two wavelength pairs, and the C2H4 concentration is inferred based on the wavelength pair with higher differential absorption. In the initial stage, the system’s accuracy was verified in high-temperature static conditions (T = 300–800 K, P = 1 atm), and continuous time series measurements demonstrated the system’s stability. The limit of detection achieved by Allan-Werle variance analysis is 2.5 ppm at the optimal average time of 100 s. Subsequently, measurements were taken in a hydrocarbon flame. The obtained results indicate an average deviation of 1.021 % between the measured temperature in the flame and the reference value, with a standard deviation of 1.381 % for concentration measurement. All the measurements show that the system can be potentially applied to combustion diagnosis.

Exploring a cost-effective and straightforward mechanism for uninterrupted in situ maximum wave runup measurements.

Wave runup, the excess water level above mean sea level, has been measured using different techniques with varying degrees of precision and associated practical limitations. This critical parameter, typically included in coastal assessment studies, varies temporally and spatially and depends on variables that include beach characteristics and nearshore hydrodynamics. Access to continuous datasets, using efficient mechanisms can assist resource-limited regions, such as Caribbean small-island developing states (SIDS), in overcoming coastal resilience obstacles. Experiments were conducted at University College London (UCL) and the University of the West Indies (UWI), which were designed to explore the temporal behaviour of the water surface within the bed during runup events. The experiments encompassed linear waves impacting a static porous bed (UCL) and a moveable granular beach (UWI), with pressure sensors buried at the base of each beach. The analyses showed that the averaged values of the time-varying water elevations within the bed, when spatially presented, produced a quadratic or cubic polynomial fit, where the curves' stationary points were accurate indicators of the location of the maximum runup position at the surface of the bed. In this way, an arrangement of buried pressure sensors can be used as an efficient means to accurately produce a continuous time series of maximum runup positions.This article is part of the theme issue 'Celebrating the 15th anniversary of the Royal Society Newton International Fellowship'.