Generate Time Series Research Articles

Comparative assessment of synthetic time series generation approaches in healthcare: leveraging patient metadata for accurate data synthesis

BackgroundSynthetic data is an emerging approach for addressing legal and regulatory concerns in biomedical research that deals with personal and clinical data, whether as a single tool or through its combination with other privacy enhancing technologies. Generating uncompromised synthetic data could significantly benefit external researchers performing secondary analyses by providing unlimited access to information while fulfilling pertinent regulations. However, the original data to be synthesized (e.g., data acquired in Living Labs) may consist of subjects’ metadata (static) and a longitudinal component (set of time-dependent measurements), making it challenging to produce coherent synthetic counterparts.MethodsThree synthetic time series generation approaches were defined and compared in this work: only generating the metadata and coupling it with the real time series from the original data (A1), generating both metadata and time series separately to join them afterwards (A2), and jointly generating both metadata and time series (A3). The comparative assessment of the three approaches was carried out using two different synthetic data generation models: the Wasserstein GAN with Gradient Penalty (WGAN-GP) and the DöppelGANger (DGAN). The experiments were performed with three different healthcare-related longitudinal datasets: Treadmill Maximal Effort Test (TMET) measurements from the University of Malaga (1), a hypotension subset derived from the MIMIC-III v1.4 database (2), and a lifelogging dataset named PMData (3).ResultsThree pivotal dimensions were assessed on the generated synthetic data: resemblance to the original data (1), utility (2), and privacy level (3). The optimal approach fluctuates based on the assessed dimension and metric.ConclusionThe initial characteristics of the datasets to be synthesized play a crucial role in determining the best approach. Coupling synthetic metadata with real time series (A1), as well as jointly generating synthetic time series and metadata (A3), are both competitive methods, while separately generating time series and metadata (A2) appears to perform more poorly overall.

Investigating the Role of Byproduct Oxygen in UK-Based Future Scenario Models for Green Hydrogen Electrolysis

Water electrolysis for hydrogen production with renewable electricity is regularly studied as an option for decarbonised future energy scenarios. The inclusion of byproduct electrolytic oxygen capture and sale is of interest for parallel decarbonisation efforts elsewhere in the industry and could contribute to reducing green hydrogen costs. A deterministic hydrogen electrolysis system model is constructed to compare oxygen inclusion/exclusion scenarios. This uses wind and solar-PV electricity generation timeseries, a power-dependent electrolysis model to determine the energy efficiency of gas yield, and power allocation for gas post-processing energy within each hourly timestep. This maintains a fully renewable (and therefore low/zero carbon) electricity source for electrolysis and gas post-processing. The model is validated (excluding oxygen) against an existing low-cost GW-scale solar-hydrogen production scenario and an existing hydrogen production costs study with offshore wind generation at the multi-MW scale. For both comparisons, oxygen inclusion is then evaluated to demonstrate both the benefits and drawbacks of capture and utilisation, for different scenario conditions, and high parameter sensitivity can be seen regarding the price of renewable electricity. This work subsequently proposes that the option for the potential utilisation of byproduct oxygen should be included in future research to exemplify otherwise missed benefits.

Generation of Spatial Time Series Data

In the era of deep learning, global-local deep neural networks are gradually replacing statistical approaches for time-series forecasting, especially for the spatiotemporal modeling field. However, the development of such methods is hindered by the lack of open benchmark datasets in this research domain. Generating synthetic data is an alternative solution to data collection, but prior works focus mainly on generating uncorrelated independent time series. In this work, we present a method for spatially correlated time-series generation. It uses a set of parametric autoregressive models for univariate time series generation in combination with the approach for sampling model parameters which allows one to simulate spatial relationships. We describe its implementation and conduct experiments showing the validity of the data for spatiotemporal modeling.

A NEW TYPE OF DETERMINISTIC CHAOS: DETERMINISTIC CHAOS OF THE SECOND KIND

Abstract. Currently, the phenomena of deterministic (Dynamic) chaos are widely used in science and technology. Therefore, now the urgent task is to study new phenomena and new manifestations of deterministic chaos. This paper examines a fundamentally new manifestation of deterministic chaos (Deterministic chaos of the second kind). In this regard, previously known manifestations of deterministic chaos should be called manifestations of chaos of the first kind (The previously known type of deterministic chaos, in our terminology, is chaos of the first kind). Deterministic chaos of the first kind is chaos that is directly related to the deterministic - chaotic behavior of the dynamic variables of a given dynamic system. At the same time, our work examines the characteristic features of deterministic chaos of the first kind and methods for generating such chaos. In our work, we considered a fundamentally different type of manifestation of dynamic chaos, namely deterministic chaos of the second kind. Deterministic chaos of the second kind is not directly related to the deterministic chaotic behavior of the dynamic variables of a dynamic system and can only exist in characteristics that are not directly related to the dynamic variables. That is, dynamic chaos of the second kind arises only in second-order manifestations: For example, in the Fourier spectrum. In a more general case, this could be any other type of spectrum (Not Fourier) and even another characteristic different from the spectrum. The paper provides two examples of deterministic chaos of the second kind. The work also proposes a general technique for generating time series demonstrating deterministic chaos of the second kind*. It is very important that dynamic chaos in the examples given in this work manifests itself not in the time series themselves (Which are completely random) but in the Fourier spectrum of these time series. Thus, the original series itself (The original signal itself) can be random, but at the same time, the spectrum of this series, if considered as another time series, can exhibit deterministic chaotic behavior. Thus, for the first time we were able to build a bridge between completely random systems and deterministic chaotic systems. That is, in our work we were the first to show that the same system can be considered both random (From the point of view of the behavior of its variables) and at the same time as deterministic-chaotic (From the point of view of its spectrum or other similar characteristics). Goal of the work. Study of a newtype of deterministic chaos discovered by the authors (Deterministic chaos of the second kind) and the study of its properties and methods for generating this type of chaos.

Recruiting neural field theory for data augmentation in a motor imagery brain-computer interface.

We introduce a novel approach to training data augmentation in brain-computer interfaces (BCIs) using neural field theory (NFT) applied to EEG data from motor imagery tasks. BCIs often suffer from limited accuracy due to a limited amount of training data. To address this, we leveraged a corticothalamic NFT model to generate artificial EEG time series as supplemental training data. We employed the BCI competition IV '2a' dataset to evaluate this augmentation technique. For each individual, we fitted the model to common spatial patterns of each motor imagery class, jittered the fitted parameters, and generated time series for data augmentation. Our method led to significant accuracy improvements of over 2% in classifying the "total power" feature, but not in the case of the "Higuchi fractal dimension" feature. This suggests that the fit NFT model may more favorably represent one feature than the other. These findings pave the way for further exploration of NFT-based data augmentation, highlighting the benefits of biophysically accurate artificial data.

Generating Time Series Data With Real-Valued DC-GAN From Complex Time-Frequency Domain: Application to ECG Synthesis

Generating Time Series Data With Real-Valued DC-GAN From Complex Time-Frequency Domain: Application to ECG Synthesis

Changen2: Multi-Temporal Remote Sensing Generative Change Foundation Model.

Our understanding of the temporal dynamics of the Earth's surface has been significantly advanced by deep vision models, which often require a massive amount of labeled multi-temporal images for training. However, collecting, preprocessing, and annotating multi-temporal remote sensing images at scale is non-trivial since it is expensive and knowledge-intensive. In this paper, we present scalable multi-temporal change data generators based on generative models, which are cheap and automatic, alleviating these data problems. Our main idea is to simulate a stochastic change process over time. We describe the stochastic change process as a probabilistic graphical model, namely the generative probabilistic change model (GPCM), which factorizes the complex simulation problem into two more tractable sub-problems, i.e., condition-level change event simulation and image-level semantic change synthesis. To solve these two problems, we present Changen2, a GPCM implemented with a resolution-scalable diffusion transformer which can generate time series of remote sensing images and corresponding semantic and change labels from labeled and even unlabeled single-temporal images. Changen2 is a "generative change foundation model" that can be trained at scale via self-supervision, and is capable of producing change supervisory signals from unlabeled single-temporal images. Unlike existing "foundation models", our generative change foundation model synthesizes change data to train task-specific foundation models for change detection. The resulting model possesses inherent zero-shot change detection capabilities and excellent transferability. Comprehensive experiments suggest Changen2 has superior spatiotemporal scalability in data generation, e.g., Changen2 model trained on 256 2 pixel single-temporal images can yield time series of any length and resolutions of 1,024 2 pixels. Changen2 pre-trained models exhibit superior zero-shot performance (narrowing the performance gap to 3% on LEVIR-CD and approximately 10% on both S2Looking and SECOND, compared to fully supervised counterpart) and transferability across multiple types of change tasks, including ordinary and off-nadir building change, land-use/land-cover change, and disaster assessment. The model and datasets are available at https://github.com/Z-Zheng/pytorch-change-models.

High throughput parameter estimation and uncertainty analysis applied to the production of mycoprotein from synthetic lignocellulosic hydrolysates

The current global food system produces substantial waste and carbon emissions while exacerbating the effects of global hunger and protein deficiency. This study aims to address these challenges by exploring the use of lignocellulosic agricultural residues as feedstocks for microbial protein fermentation, focusing on Fusarium venenatum A3/5, a mycelial strain known for its high protein yield and nutritional quality. We propose a high throughput microlitre batch fermentation system paired with analytical chemistry to generate time series data of microbial growth and substrate utilisation. An unstructured biokinetic model was developed using a bootstrap sampling approach to quantify uncertainty in the parameter estimates. The model was validated against an independent data set of a different glucose-xylose composition to assess the predictive performance. Our results indicate a robust model fit with high coefficients of determination and low root mean squared errors for biomass, glucose, and xylose concentrations. Estimated parameter values provided insights into the resource utilisation strategies of Fusarium venenatum A3/5 in mixed substrate cultures, aligning well with previous research findings. Significant correlations between estimated parameters were observed, highlighting challenges in parameter identifiability. The high throughput workflow presents a novel, rapid methodology for biokinetic model development, enabling efficient exploration of microbial growth dynamics and substrate utilisation. This innovative method directly supports the development of a foundational model for optimising microbial protein production from lignocellulosic hydrolysates, contributing to a more sustainable global food system.

A Synthetic Time-Series Generation Using a Variational Recurrent Autoencoder with an Attention Mechanism in an Industrial Control System.

Data scarcity is a significant obstacle for modern data science and artificial intelligence research communities. The fact that abundant data are a key element of a powerful prediction model is well known through various past studies. However, industrial control systems (ICS) are operated in a closed environment due to security and privacy issues, so collected data are generally not disclosed. In this environment, synthetic data generation can be a good alternative. However, ICS datasets have time-series characteristics and include features with short- and long-term temporal dependencies. In this paper, we propose the attention-based variational recurrent autoencoder (AVRAE) for generating time-series ICS data. We first extend the evidence lower bound of the variational inference to time-series data. Then, a recurrent neural-network-based autoencoder is designed to take this as the objective. AVRAE employs the attention mechanism to effectively learn the long-term and short-term temporal dependencies ICS data implies. Finally, we present an algorithm for generating synthetic ICS time-series data using learned AVRAE. In a comprehensive evaluation using the ICS dataset HAI and various performance indicators, AVRAE successfully generated visually and statistically plausible synthetic ICS data.

ASTMEN: an adaptive spatiotemporal and multi-element fusion network for ocean surface currents forecasting

Accurate forecasting of ocean surface currents is crucial for the planning of marine activities, including fisheries, shipping, and pollution control. Previous studies have often neglected the consideration of spatiotemporal correlations and interdependencies among ocean elements, leading to suboptimal accuracy in medium to long-term forecasts, especially in regions characterized by intricate ocean currents. This paper proposes an adaptive spatiotemporal and multi-element fusion network for ocean surface currents forecasting (ASTMEN). Specifically, we use an improved Swin Transformer (Swin-T) to perform self-attention computation at any given moment, enabling the adaptive generation of multi-element time series with spatial dependencies. Then, we utilize a Long Short-Term Memory network (LSTM) to encode and decode these series in the dimensions of temporal and multi-element features, resulting in accurate forecasts of ocean surface currents. This study takes the Kuroshio region in the northwest Pacific Ocean as the study area with data from the ocean reanalysis dataset. The experimental results show that ASTMEN significantly outperforms the baseline model and the climate state method, and is the only model whose correlation coefficient is still higher than 0.8 at day 12. In the experiments during the summer, when the currents are most variable, ASTMEN provides better forecasts at the sea-land interface and at the junction of different currents, which has the potential to fill the gap of poor forecast performance of previous methods for complex current fields.

Multiomics analysis identifies novel facilitators of human dopaminergic neuron differentiation

Midbrain dopaminergic neurons (mDANs) control voluntary movement, cognition, and reward behavior under physiological conditions and are implicated in human diseases such as Parkinson’s disease (PD). Many transcription factors (TFs) controlling human mDAN differentiation during development have been described, but much of the regulatory landscape remains undefined. Using a tyrosine hydroxylase (TH) human iPSC reporter line, we here generate time series transcriptomic and epigenomic profiles of purified mDANs during differentiation. Integrative analysis predicts novel regulators of mDAN differentiation and super-enhancers are used to identify key TFs. We find LBX1, NHLH1 and NR2F1/2 to promote mDAN differentiation and show that overexpression of either LBX1 or NHLH1 can also improve mDAN specification. A more detailed investigation of TF targets reveals that NHLH1 promotes the induction of neuronal miR-124, LBX1 regulates cholesterol biosynthesis, and NR2F1/2 controls neuronal activity.

INVESTIGATIONS ON INSAR DATA PROCESSING STANDARD FOR VOLCANO ISLAND MONITORING IN INDONESIA

Abstract. In critical situations such as disaster and humanitarian relief, the reliable geospatial data will present an appropriate decision on a timely manner. As fundamental datasets, large-scale topographic maps are are mandatory in order to perform geospatial analysis for many societal challenges. Indeed, Large Scale Topographic Mapping in Indonesia shall be accelerated in an innovative and productive way to consider the production efficiency, especially for the geospatial data acquisition as its primary source. Interferometric Synthetic Aperture Radar (InSAR) data acquisition presents some advantages, especially in conjunction with the security clearance and global weather as some main constraints for the geospatial data acquisition. In terms of disaster preparedness and emergency response, radar interferometry techniques can play an important role by generating a Digital Elevation Model (DEM) as an input to Decision Support Systems (DSS). In addition, the techniques of differential Interferometric Synthetic Aperture Radar (D-InSAR) can provide the earth surface deformation from time series Radar Datasets. A set of standards related to the InSAR data acquisition are essential, especially with regards to the improved sensors, and processing methods implemented by the worldwide space agencies nowadays. Hence, the precise definition of a standard is fundamental in order to prepare the InSAR data as one potential source of geospatial data and information, especially for LSTM data production. Generated time series of Digital Elevation Model (DEM) can identify the potential cause of 2018’s Sunda Strait tsunami (Tampubolon, 2019). It was related with the increasing eruption activity of Anak Krakatau (Child of Krakatoa). From DEM analysis, the Volcanic mudflow material avalanche has been detected before the tsunami hit the coastal surrounding area. In this paper, more advanced techniques was investigated by the combination between Time Series DEM Analysis and Ground Displacement detection. The focus on the integration between reliable TanDEM-X and up-to-date Sentinel 1A/B was proposed on another volcano island of Gamalama in Ternate. As TanDEM-X uses X-band as its medium wave, it can produce more reliable DEM as a reference data. On the other hand, Sentinel 1 uses C-band that can contribute to detect the ground deformation over the aforementioned volcano island. TanDEM-X uses a more robust and consistent approach to scan the earth surface by single data acquisition scheme, which is generally more accurate and reliable for the DEM generation. However, many factors influence the selection of a platform that can be best suited for radar interferometry. Initially, the German TanDEM-X Coregistered Singlelook Slant-range Complex (CoSSC) using bi-static InSAR approach has been implemented in order to generate DEMs comply with the LSTM specification in Indonesia. The basic principle is performing a simultaneous measurement of the same scene and identical doppler spectrum by using 2 sensors, thereby avoiding temporal decorrelation (DLR, 2012). Subsequently, the differential InSAR techniques by using Sentinel 1A/B has been performed in order to detect the ground deformation. Finally, tt can be shown that the reliable TanDEM-X data has increased the ground deformation resolution and accuracy in order to provide potential solution for disaster preparedness.

Fisher-Shannon Investigation of the Effect of Nonlinearity of Discrete Langevin Model on Behavior of Extremes in Generated Time Series.

Diverse forms of nonlinearity within stochastic equations give rise to varying dynamics in processes, which may influence the behavior of extreme values. This study focuses on two nonlinear models of the discrete Langevin equation: one with a fixed diffusion function (M1) and the other with a fixed marginal distribution (M2), both characterized by a nonlinearity parameter. Extremes are defined according to the run theory with thresholds based on percentiles. The behavior of inter-extreme times and run lengths is examined by employing Fisher's Information Measure and the Shannon Entropy. Our findings reveal a clear relationship between the entropic and informational measures and the nonlinearity of model M1-these measures decrease as the nonlinearity parameter increases. Similar relationships are evident for the M2 model, albeit to a lesser extent, even though the background data's marginal distribution remains unaffected by this parameter. As thresholds increase, both the values of Fisher's Information Measure and the Shannon Entropy also increase.

Trend in rainfall associated with tropical cyclones in Mexico attributed to climate change and variability

Abstract The aim of this study was to investigate the existence and the magnitude of trend in different areas and durations of rainfall associated with tropical cyclones (TCR). To achieve this objective, a mixed-method approach was employed using depth–area–duration (DAD) and areal reduction factor (ARFs) curves that can be described as a logarithm equation to generate time series that allows the application of statistical methods such as the Mann–Kendall (MK) and Spearman Rho (SR) to detect trends. Time series are generated by substituting different areas in the logarithmic equations. The evidence presented shows that in Mexico, the TCR lasting 24 h shows an increasing trend for maximum areas between 300 and 1,700 km2 according to the MK and SR tests, respectively; according to these same tests for durations of 48 h, upward trends were observed up to maximum areas between 5,700 and 6,900 km2. The Sen's slope reports annual increases between 0.76 and 1.32 mm and between 1.15 and 2.06 for a duration of 24 and 48 h, respectively. In contrast, no trends were observed in the time series obtained from the ARFs. Finally, the Pettitt test reports an abrupt jump from the year 1997 in all cases.

Investigating laypeople’s short- and long-term forecasts of COVID-19 infection cycles

How do laypeople anticipate the severity of the COVID-19 pandemic in the short and long term? The evolution of COVID-19 infection cases is characterized by wave-shaped cycles, and we examine how individuals make forecasts for this type of time series. Over 42 weeks, we ran forecasting experiments and elicited weekly judgments from the general public to analyze their forecasting behavior (Study 1). We find that laypeople often tend to dampen trends when generating judgmental forecasts, but the degree to which this happens depends on the evolution of the cyclic time series data. The observed forecasting behavior reveals evidence of an optimism bias in that people do not expect the number of infection cases to grow at the observed rate while believing that infection rates would drop at an even faster rate than they are. Also, our results suggest that laypeople’s forecasting judgments are affected by the magnitude of the present wave relative to the previously observed ones. Further, we provide evidence that laypeople rely on a cognitive heuristic for generating long-term forecasts. People tend to rely on a linear discounting rule in that they lower their long-term forecasts proportionally to the interval of the forecast horizon, i.e., from tomorrow to 6 months and from 6 months to 1 year. We also find that this linear discounting rule can change to an exponential one in reaction to externally generated optimistic information signals such as vaccine approval. Furthermore, we replicated the major findings of Study 1 in a more controlled setting with a hypothetical pandemic scenario and artificially generated time series (Study 2). Overall, the current research contributes to the judgmental forecasting literature and provides practical implications for decision-makers in the pandemic.

High-precision reconstruction method based on MTS-GAN for electromagnetic environment data in SAGIoT

Equipment failures and communication interruptions of satellites, aircraft and ground devices lead to data loss in Space-Air-Ground Integrated Internet of Things (SAGIoT). The incomplete data affect the accuracy of data modeling, decision-making and spectrum prediction. Reconstructing the incomplete data of electromagnetic environment is a significant task in the SAGIoT. Most spectral data completion algorithms have the problem of limited accuracy and slow iterative optimization. In light of these challenges, a novel high-precision reconstruction method for electromagnetic environment data based on multi-component time series generation adversarial network (MTS-GAN) is proposed in this paper. MTS-GAN transforms the reconstruction method of electromagnetic environment data into the data generation problem of multiple time series. It extracts the time–frequency joint features and the overall distribution of electromagnetic environment data. To improve the reconstruction precision, MTS-GAN simulates the time irregularity of incomplete time series by applying a gate recursive element to adapt to the attenuation effect of discontinuous time series observations. Experimental results show that the proposed MTS-GAN provides high completion accuracy and achieves better results than competitive data completion algorithms.

Detecting Structural Changes in Time Series by Using the BDS Test Recursively: An Application to COVID-19 Effects on International Stock Markets

Structural change tests aim to identify evidence of a structural break or change in the underlying generating process of a time series. The BDS test has its origins in chaos theory and seeks to test, using the correlation integral, the hypothesis that a time series is generated by an identically and independently distributed (IID) stochastic process over time. The BDS test is already widely used as a powerful tool for testing the hypothesis of white noise in the residuals of time series models. In this paper, we illustrate how the BDS test can be implemented also in a recursive manner to evaluate the hypothesis of structural change in a time series, taking advantage of its ability to test the IID hypothesis. We apply the BDS test repeatedly, starting with a sub-sample of the original time series and incrementally increasing the number of observations until it is applied to the full sample time series. A structural change in the unknown underlying generator model is detected when a change in the trend shown by this recursively computed BDS statistic is detected. The strength of this recursive BDS test lies in the fact that it does not require making any assumptions about the underlying time series generator model. We ilustrate the power and potential of this recursive BDS test through an application to real economic data. In this sense, we apply the test to assess the structural changes caused by the COVID-19 pandemic in international financial markets. Using daily data from the world’s top stock indices, we have detected strong and statistically significant evidence of two major structural changes during the period from June 2018 to June 2022. The first occurred in March 2020, coinciding with the onset of economic restrictions in the main Western countries as a result of the pandemic. The second occurred towards the end of August 2020, with the end of the main economic restrictions and the beginning of a new post-pandemic economic scenario. This methodology to test for structural changes in a time series is easy to implement and can detect changes in any system or process behind the time series even when this generating system is not known, and without the need to specify or estimate any a priori generating model. In this sense, the recursive BDS test could be incorporated as an initial preliminary step to any exercise of time series modeling. If a structural change is detected in a time series, rather than estimating a single predictive model for the full-sample time series, efforts should be made to estimate different predictive models, one for the time before and one for the time after the detected structural change.

Estimating the elastic constants of orthotropic composites using guided waves and an inverse problem of property estimation

The present research focusses on estimating the elastic constants of orthotropic laminates using ultrasonic guided waves (GWs) excited through Lead Zirconate Titanate (PZT) sensors and sensed using one-dimensional laser vibrometer. The elastic constants of a material are crucial for understanding its mechanical behaviour and are typically determined through experimental testing. However, this process can be time-consuming and expensive. We formulate this problem as an inverse problem of property estimation. Thus, in this work, the simulation models with PZT transducers have been employed for generating time series (TS) GWs for the orthotropic material. Then, an inverse machine learning model is trained using a TS dataset pertaining to different elastic constants generated using the simulations. The inverse model consists of deep neural networks and designing a loss function for the specific application. Limited number of unique sets of simulations were conducted. Out of available simulation data, 30% of the sets were used for validation. To further test the model, a blind experimental test was conducted, and the corresponding elastic constants were estimated with a mean absolute percentage error (MAPE) of 12.89% and standard deviation of 5.47%. The results demonstrate that formulation of property estimation as inverse problem is capable of accurately predicting the elastic constants of a material, by using a model solely trained on simulation and a very scare amount of data. This approach has the potential to significantly reduce the computational time for predicting the elastic constants, and thereby could have wide-ranging applications in materials science and engineering.

Toward coherent space–time mapping of seagrass cover from satellite data: an example of a Mediterranean lagoon

Abstract. Seagrass meadows are a highly productive and economically important shallow coastal habitat. Their sensitivity to natural and anthropogenic disturbances, combined with their importance for local biodiversity, carbon stocks, and sediment dynamics, motivate a frequent monitoring of their distribution. However, generating time series of seagrass cover from field observations is costly, and mapping methods based on remote sensing require restrictive conditions on seabed visibility, limiting the frequency of observations. In this contribution, we examine the effect of accounting for environmental factors, such as the bathymetry and median grain size (D50) of the substrate as well as the coordinates of known seagrass patches, on the performance of a random forest (RF) classifier used to determine seagrass cover. Using 148 Landsat images of the Venice Lagoon (Italy) between 1999 and 2020, we trained an RF classifier with only spectral features from Landsat images and seagrass surveys from 2002 and 2017. Then, by adding the features above and applying a time-based correction to predictions, we created multiple RF models with different feature combinations. We tested the quality of the resulting seagrass cover predictions from each model against field surveys, showing that bathymetry, D50, and coordinates of known patches exert an influence that is dependent on the training Landsat image and seagrass survey chosen. In models trained on a survey from 2017, where using only spectral features causes predictions to overestimate seagrass surface area, no significant change in model performance was observed. Conversely, in models trained on a survey from 2002, the addition of the out-of-image features and particularly coordinates of known vegetated patches greatly improves the predictive capacity of the model, while still allowing the detection of seagrass beds absent in the reference field survey. Applying a time-based correction eliminates small temporal variations in predictions, improving predictions that performed well before correction. We conclude that accounting for the coordinates of known seagrass patches, together with applying a time-based correction, has the most potential to produce reliable frequent predictions of seagrass cover. While this case study alone is insufficient to explain how geographic location information influences the classification process, we suggest that it is linked to the inherent spatial auto-correlation of seagrass meadow distribution. In the interest of improving remote-sensing classification and particularly to develop our capacity to map vegetation across time, we identify this phenomenon as warranting further research.

Understanding climate resilience in Scandinavia during the Neolithic and Early Bronze Age

Mid and late-Holocene climate shifts are considered to have profoundly shaped demographic developments and adaptive responses of communities globally. Yet their onset, duration, and impact on Neolithic and Early Nordic Bronze Age communities in the high-latitude ranges of southern and north-western Scandinavia remain a major research gap. Here, we built on an emerging body of archaeological and paleoclimate data, encompassing 20,908 anthropogenic 14C dates and 49 climate records from the Holocene. Additionally, we gathered and correlated a new archaeological dataset of 3649 houses from southern Scandinavia and southern Norway. In this study, we utilised 6268 reliable 14C dates and 2519 dwellings to generate time series and socio-economic trends from ∼4100 to 1100 BCE.Our study revealed three key findings: (1) A distinct lateral zonation, with variations in the duration and timing of the Holocene Thermal Maximum (∼7050–2050 BCE). In Southern Scandinavia, a warmer climate may have facilitated the spread of crop cultivation (3820–3790 BCE), coinciding with significant population growth. Neolithic communities settled in permanent two-aisled houses 90–160 years later (3700–3660 BCE). (2) The 2250 BCE (4.2 ka BP) cooling trend marked the beginning of a climate regime shift with varying duration and timing (∼3450–1450 BCE). This period coincided with demographic growth, migration, crop cultivation diversity, and the development of houses with crop storage facilities (2290–2215 BCE). (3) Severe abrupt cooling periods (∼1850–1450 BCE) corresponded to short-term demographic decline including disruptions in trade networks with continental Europe. However, repopulation and redistribution of wealth (∼1450 BCE), along with the development of stable three-aisled houses (1475–1450 BCE), underscore the resilience of food-producing economies in mitigating environmental disturbances.