Spatiotemporal Autocorrelation Research Articles

High Spatiotemporal Resolution PM2.5 Concentration Estimation with Machine Learning Algorithm: A Case Study for Wildfire in California

As an aggregate of suspended particulate matter in the air, atmospheric aerosols can affect the regional climate. With the help of satellite remote sensing technology to retrieve AOD (aerosol optical depth) on a global or regional scale, accurate estimation of PM2.5 concentration has become an important task to quantify the spatiotemporal distribution of AOD and PM2.5. However, due to the limitations of satellite platforms, sensors, and inversion algorithms, the spatiotemporal resolution of current major AOD products is still relatively low. Meanwhile, for the impact of cloud, the AOD products often have a serious data gap problem, which also objectively limits the spatiotemporal coverage of predicted PM2.5 concentration. Therefore, how to effectively improve the spatiotemporal resolution and coverage of PM2.5 concentration under the requisite accuracy is still a grand challenge. In this study, the fused high spatial-temporal resolution AOD data in our previous study were used to estimate the ground PM2.5 concentration through machine learning algorithms, the deep belief network (DBN). The PM2.5 data had spatiotemporal autocorrelation in geostatistics and followed the Gaussian kernel distribution. Hence, the autocorrelation model modified by Gaussian kernel function integrated with DBN algorithm, named Geoi-DBN, was used to estimate PM2.5 concentration. The cross-validation results showed that the Geoi-DBN (R2 = 0.86, RMSE = 6.84 µg m−3) performed better than the original DBN (R2 = 0.67, RMSE = 10.46 µg m−3). The final high quality PM2.5 concentration data can be applied for urban air quality monitoring and related PM2.5 exposure risk assessment such as wildfire.

Analysis on the characteristics of spatio-temporal evolution and aggregation trend of early COVID-19 in mainland China

To analyze the spatio-temporal aggregation of COVID-19 in mainland China within 20 days after the closure of Wuhan city, and provide a theoretical basis for formulating scientific prevention measures in similar major public health events in the future. Draw a distribution map of the cumulative number of COVID-19 by inverse distance weighted interpolation; analyze the spatio-temporal characteristics of the daily number of COVID-19 in mainland China by spatio-temporal autocorrelation analysis; use the spatio-temporal scanning statistics to detect the spatio-temporal clustering area of the daily number of new diagnosed cases. The cumulative number of diagnosed cases obeyed the characteristics of geographical proximity and network proximity to Hubei. Hubei and its neighboring provinces were most affected, and the impact in the eastern China was more dramatic than the impact in the western; the global spatio-temporal Moran’s I index showed an overall downward trend. Since the 10th day of the closure of Wuhan, the epidemic in China had been under effective control, and more provinces had shifted into low-incidence areas. The number of new diagnosed cases had gradually decreased, showing a random distribution in time and space (P< 0.1), and no clusters were formed. Conclusion: the spread of COVID-19 had obvious spatial-temporal aggregation. China’s experience shows that isolation city strategy can greatly contain the spread of the COVID-19 epidemic.

Geographical inequalities in the decreasing 28-day mortality following incident acute myocardial infarction: a Danish register-based cohort study, 1987\u20132016

BackgroundMortality following acute myocardial infarction (AMI) has decreased in western countries for decades; however, it remains unknown whether the decrease is distributed equally across the population independently of residential location. This study investigated whether the observed decreasing 28-day mortality following an incident AMI in Denmark from 1987 to 2016 varied geographically at municipality level after accounting for sociodemographic characteristics.MethodsA register-based cohort study design was used to investigate 28-day mortality among individuals with an incident AMI. Global spatial autocorrelation (within sub-periods) was analysed at municipality level using Moran's I. Analysis of spatio-temporal autocorrelation before and after adjusting for sociodemographic characteristics was performed using logistic regression and conditional autoregressive models with inference in a Bayesian setting.ResultsIn total, 368,839 individuals with incident AMI were registered between 1987 and 2016 in Denmark; 128,957 incident AMIs were fatal. The 28-day mortality decreased over time at national level with an odds ratio of 0.788 (95% credible interval (0.784, 0.792)) per 5-year period after adjusting for sociodemographic characteristics. The decrease in the 28-day mortality was geographically unequally distributed across the country and in a geographical region in northern Jutland, the 28-day mortality decreased significantly slower (4–12%) than at national level.ConclusionsDuring the period from 1987 to 2016, the 28-day mortality following an incident AMI decreased substantially in Denmark. However, in a local geographical region, the 28-day mortality decreased significantly slower than in the rest of the country both before and after adjusting for sociodemographic differences. Efforts should be made to keep geographical trend inequalities in the 28-day mortality to a minimum.

Electrophysiology and Arrhythmogenesis in the Human Right Ventricular Outflow Tract.

Right ventricular outflow tract (RVOT) is a common source of ventricular tachycardia, which often requires ablation. However, the mechanisms underlying the RVOT's unique arrhythmia susceptibility remain poorly understood due to lack of detailed electrophysiological and molecular studies of the human RVOT. We conducted optical mapping studies in 16 nondiseased donor human RVOT preparations subjected to pharmacologically induced adrenergic and cholinergic stimulation to evaluate susceptibility to arrhythmias and characterize arrhythmia dynamics. We found that under control conditions, RVOT has shorter action potential duration at 80% repolarization relative to the right ventricular apical region. Treatment with isoproterenol (100 nM) shortened action potential duration at 80% repolarization and increased incidence of premature ventricular contractions (P=0.003), whereas acetylcholine (100 μM) stimulation alone had no effect on action potential duration at 80% repolarization or premature ventricular contractions. However, acetylcholine treatment after isoproterenol stimulation reduced the incidence of premature ventricular contractions (P=0.034) and partially reversed action potential duration at 80% repolarization shortening (P=0.029). Immunolabeling of RVOT (n=4) confirmed the presence of cholinergic marker VAChT (vesicular acetylcholine transporter) in the region. Rapid pacing revealed RVOT susceptibility to both concordant and discordant alternans. Investigation into transmural arrhythmia dynamics showed that arrhythmia wave fronts and phase singularities (rotors) were relatively more organized in the endocardium than in the epicardium (P=0.006). Moreover, there was a weak but positive spatiotemporal autocorrelation between epicardial and endocardial arrhythmic wave fronts and rotors. Transcriptome analysis (n=10 hearts) suggests a trend that MAPK (mitogen-activated protein kinase) signaling, calcium signaling, and cGMP-PKG (protein kinase G) signaling are among the pathways that may be enriched in the male RVOT, whereas pathways of neurodegeneration may be enriched in the female RVOT. Human RVOT electrophysiology is characterized by shorter action potential duration relative to the right ventricular apical region. Cholinergic right ventricular stimulation attenuates the arrhythmogenic effects of adrenergic stimulation, including increase in frequency of premature ventricular contractions and shortening of wavelength. Right ventricular arrhythmia is characterized by positive spatial-temporal autocorrelation between epicardial-endocardial arrhythmic wave fronts and rotors that are relatively more organized in the endocardium.

Persistent impacts of the 2018 drought on forest disturbance regimes in Europe

Abstract. Europe was affected by an extreme drought in 2018, compounding with an extensive heat wave in the same and subsequent years. Here we provide a first assessment of the impacts this compounding event had on forest disturbance regimes in Europe. We find that the 2018 drought caused unprecedented levels of forest disturbance across large parts of Europe, persisting up to 2 years post-drought. The 2018 drought pushed forest disturbance regimes in Europe to the edge of their past range of variation, especially in central and eastern Europe. Increased levels of forest disturbance were associated with low soil water availability in 2018 and were further modulated by high vapor pressure deficit from 2018 to 2020. We also document the emergence of novel spatiotemporal disturbance patterns following the 2018 drought (i.e., more and larger disturbances, occurring with higher spatiotemporal autocorrelation) that will have long-lasting impacts on forest structure and raise concerns about a potential loss of forest resilience. We conclude that the 2018 drought had unprecedented impacts on forest disturbance regimes in Europe, highlighting the urgent need to adapt Europe's forests to a hotter and drier future with more disturbance.

Fine-scale movement patterns and habitat selection of little owls (Athene noctua) from two declining populations.

Advances in bio-logging technology for wildlife monitoring have expanded our ability to study space use and behavior of many animal species at increasingly detailed scales. However, such data can be challenging to analyze due to autocorrelation of GPS positions. As a case study, we investigated spatiotemporal movements and habitat selection in the little owl (Athene noctua), a bird species that is declining in central Europe and verges on extinction in Denmark. We equipped 6 Danish food-supplemented little owls and 6 non-supplemented owls in the Czech Republic with high-resolution GPS loggers that recorded one position per minute. Nightly space use, measured as 95% kernel density estimates, of Danish male owls were on average 62 ha (± 64 SD, larger than any found in previous studies) compared to 2 ha (± 1) in females, and to 3 ± 1 ha (males) versus 3 ± 5 ha (females) in the Czech Republic. Foraging Danish male owls moved on average 4-fold further from their nest and at almost double the distance per hour than Czech males. To create availability data for the habitat selection analysis, we accounted for high spatiotemporal autocorrelation of the GPS data by simulating correlated random walks with the same autocorrelation structure as the actual little owl movement trajectories. We found that habitat selection was similar between Danish and Czech owls, with individuals selecting for short vegetation and areas with high structural diversity. Our limited sample size did not allow us to infer patterns on a population level, but nevertheless demonstrates how high-resolution GPS data can help to identify critical habitat requirements to better formulate conservation actions on a local scale.

Physical disorder and crime revisited: New evidence from intensive longitudinal data

ObjectivesThe purpose of this study is to investigate whether the often-debated association between physical disorder and crime exists longitudinally and operates through a specific spatial process. MethodsWe combine four administrative and official databases to analyze patterns of crime occurring each month in New Orleans, LA, between 2012 and 2018 at the block group level. We adopt a generalized additive model (GAM) framework to efficiently account for potential spatial, temporal, and spatiotemporal autocorrelation to investigate this association. ResultsThe resident-driven self-report measure of physical disorder (i.e., 311 calls) consistently predicts neighborhood criminal activities in New Orleans. The spatial effects of 311 calls in adjacent areas on drug violations, property crime, and violence in a focal neighborhood are also observed. In contrast, official reports of physical disorder, captured with code violation, is either weakly related to crime or lacking an empirical relationship. ConclusionOverall, our findings suggest that the association between physical disorder and neighborhood crimes is robust over time. In addition, this association is conditioned by how physical disorder is operationalized, with dimensions such as residential report and official report of physical disorder playing a unique, and occasionally important role in understanding overall patterns of neighborhood criminal activity.

A comparison of multiple neighborhood matrix specifications for spatio-temporal model fitting: a case study on COVID-19 data.

Establishing proper neighbor relations between a set of spatial units under analysis is essential when carrying out a spatial or spatio-temporal analysis. However, it is usual that researchers choose some of the most typical (and simple) neighborhood structures, such as the first-order contiguity matrix, without exploring other options. In this paper, we compare the performance of different neighborhood matrices in the context of modeling the weekly relative risk of COVID-19 over small areas located in or near Valencia, Spain. Specifically, we construct contiguity-based, distance-based, covariate-based (considering mobility flows and sociodemographic characteristics), and hybrid neighborhood matrices. We evaluate the goodness of fit, the overall predictive quality, the ability to detect high-risk spatio-temporal units, the capability to capture the spatio-temporal autocorrelation in the data, and the goodness of smoothing for a set of spatio-temporal models based on each of the neighborhood matrices. The results show that contiguity-based matrices, some of the distance-based matrices, and those based on sociodemographic characteristics perform better than the matrices based on k-nearest neighbors and those involving mobility flows. In addition, we test the linear combination of some of the constructed neighborhood matrices and the reweighting of these matrices after eliminating weak neighbor relations, without any model improvement.

Geographically and temporally weighted neural network for winter wheat yield prediction

Accurate prediction of crop yield is essential for agricultural trading, market risk management and food security. Although various statistical models and machine learning models have been developed to enhance prediction accuracy, spatial and temporal non-stationarity, an intrinsic attribute of many geographical processes, is still rarely considered in crop yield modeling. From a statistical point of view, this study respectively provided evidence for the existence of spatial non-stationarity and temporal non-stationarity in winter wheat yield prediction based on geographically weighted regression (GWR) and temporally weighted regression (TWR). Then, a geographically and temporally weighted neural network (GTWNN) model was proposed by integrating artificial neural network (ANN) into geographically and temporally weighted regression (GTWR) using publicly available data sources, including satellite imagery and climate data. For a more credible evaluation, the leave-one-year-out strategy was adopted to make out-of-sample prediction resulting in a total of 12 test years from 2008 to 2019. The experiment results showed that the proposed GTWNN outperformed ANN, GTWR and support vector regression (SVR) achieving the average coefficient of determination (R2) values of 0.766, 0.759 and 0.720 at the three prediction times of end of July, end of June and end of May. Moreover, an extended Moran's I was adopted to assess the degree of spatiotemporal autocorrelation of the prediction errors. The error aggregation of GTWNN was lower than other models, indicating that GTWNN is applicable to addressing spatial non-stationarity in modeling the relationship between predictors and yield response. The methodology proposed in this paper can be extended to handle spatiotemporal non-stationarity in other crop yield predictions and even other environmental phenomena.

Human or natural? Landscape context improves the attribution of forest disturbances mapped from Landsat in Central Europe

Disturbances have increased in Central Europe's forests, but whether changes in disturbance regimes are driven by natural or human causes remains unclear. Satellite-based remote sensing provides an important data source for quantifying forest disturbance change. Separating causes of forest disturbance is challenging, however, particularly in areas such as Central Europe where disturbance patches are small and disturbance agents interact strongly. Here we present a novel approach for the causal attribution of forest disturbance agents and illustrate its utility for 1.01 million disturbance patches mapped from Landsat data in Austria for the period 1986–2016. We gathered reference data on 2620 disturbance patches by conducting targeted field observations and structured interviews with 21 forest managers. We developed a novel indicator class characterizing the landscape context of a disturbance patch (i.e., the spatio-temporal autocorrelation of disturbance patches on the landscape), and combined it with other predictor variables describing the spectral signal, topography, and patch form of each disturbance patch. We used these predictors to identify the causal agents for disturbances mapped in Austria using Random Forest classification. Landscape context was the most important predictor of disturbance agent, improving model performance by up to 26 percentage points. Wind, bark beetles and timber harvesting were separated with an overall accuracy of 63%. Bark beetle patches were most difficult to identify correctly (producer's accuracy = 15%, user's accuracy = 30%), while regular timber harvesting was classified with highest certainty (producer's accuracy = 68%, user's accuracy = 82%). Harvesting dominates the disturbance regime of Austria's forests, with 70.5% of the disturbed area (76.7% of the disturbed patches) attributed to human causes and 29.5% (23.3%) to natural causes (wind: 23.0% [14.8%], bark beetles: 6.5% [8.5%]). Increases in disturbance since 1986 were driven by natural causes, with wind increasing by 408% and bark beetles increasing by 99% between the first and the second half of the observation period. Wind-disturbed patches were also considerably larger than those caused by bark beetles and harvesting (+102% and + 67%, respectively). Our novel approach to mapping causal agents of forest disturbance, applicable also to highly complex and interactive disturbance regimes, provides an important step towards a comprehensive monitoring and management of forest disturbances in a changing world.

Joint Air Quality and Weather Prediction Based on Multi-Adversarial Spatiotemporal Networks

Accurate and timely air quality and weather predictions are of great importance to urban governance and human livelihood. Though many efforts have been made for air quality or weather prediction, most of them simply employ one another as feature input, which ignores the inner-connection between two predictive tasks. On the one hand, the accurate prediction of one task can help improve another task's performance. On the other hand, geospatially distributed air quality and weather monitoring stations provide additional hints for city-wide spatiotemporal dependency modeling. Inspired by the above two insights, in this paper, we propose the Multi-adversarial spatiotemporal recurrent Graph Neural Networks (MasterGNN) for joint air quality and weather prediction. Specifically, we first propose a heterogeneous recurrent graph neural network to model the spatiotemporal autocorrelation among air quality and weather monitoring stations. Then, we develop a multi-adversarial graph learning framework to against observation noise propagation introduced by spatiotemporal modeling. Moreover, we introduce an adaptive training strategy by formulating multi-adversarial learning as a multi-task learning problem. Finally, extensive experiments on two real-world datasets show that MasterGNN achieves the best performance compared with seven baselines on both air quality and weather prediction tasks.

Spatiotemporal clustering using Gaussian processes embedded in a mixture model

AbstractThe categorization of multidimensional data into clusters is a common task in statistics. Many applications of clustering, including the majority of tasks in ecology, use data that is inherently spatial and is often also temporal. However, spatiotemporal dependence is typically ignored when clustering multivariate data. We present a finite mixture model for spatial and spatiotemporal clustering that incorporates spatial and spatiotemporal autocorrelation by including appropriate Gaussian processes (GP) into a model for the mixing proportions. We also allow for flexible and semiparametric dependence on environmental covariates, once again using GPs. We propose to use Bayesian inference through three tiers of approximate methods: a Laplace approximation that allows efficient analysis of large datasets, and both partial and full Markov chain Monte Carlo (MCMC) approaches that improve accuracy at the cost of increased computational time. Comparison of the methods shows that the Laplace approximation is a useful alternative to the MCMC methods. A decadal analysis of 253 species of teleost fish from 854 samples collected along the biodiverse northwestern continental shelf of Australia between 1986 and 1997 shows the added clarity provided by accounting for spatial autocorrelation. For these data, the temporal dependence is comparatively small, which is an important finding given the changing human pressures over this time.

A Fast Estimator for Binary Choice Models with Spatial, Temporal, and Spatio-Temporal Interdependence

AbstractBinary outcome models are frequently used in the social sciences and economics. However, such models are difficult to estimate with interdependent data structures, including spatial, temporal, and spatio-temporal autocorrelation because jointly determined error terms in the reduced-form specification are generally analytically intractable. To deal with this problem, simulation-based approaches have been proposed. However, these approaches (i) are computationally intensive and impractical for sizable datasets commonly used in contemporary research, and (ii) rarely address temporal interdependence. As a way forward, we demonstrate how to reduce the computational burden significantly by (i) introducing analytically-tractable pseudo maximum likelihood estimators for latent binary choice models that exhibit interdependence across space and time and by (ii) proposing an implementation strategy that increases computational efficiency considerably. Monte Carlo experiments show that our estimators recover the parameter values as good as commonly used estimation alternatives and require only a fraction of the computational cost.

A big-data driven approach to analyzing and modeling human mobility trend under non-pharmaceutical interventions during COVID-19 pandemic

During the unprecedented coronavirus disease 2019 (COVID-19) challenge, non-pharmaceutical interventions became a widely adopted strategy to limit physical movements and interactions to mitigate virus transmissions. For situational awareness and decision-support, quickly available yet accurate big-data analytics about human mobility and social distancing is invaluable to agencies and decision-makers. This paper presents a big-data-driven analytical framework that ingests terabytes of data on a daily basis and quantitatively assesses the human mobility trend during COVID-19. Using mobile device location data of over 150 million monthly active samples in the United States (U.S.), the study successfully measures human mobility with three main metrics at the county level: daily average number of trips per person; daily average person-miles traveled; and daily percentage of residents staying home. A set of generalized additive mixed models is employed to disentangle the policy effect on human mobility from other confounding effects including virus effect, socio-demographic effect, weather effect, industry effect, and spatiotemporal autocorrelation. Results reveal the policy plays a limited, time-decreasing, and region-specific effect on human movement. The stay-at-home orders only contribute to a 3.5%-7.9% decrease in human mobility, while the reopening guidelines lead to a 1.6%-5.2% mobility increase. Results also indicate a reasonable spatial heterogeneity among the U.S. counties, wherein the number of confirmed COVID-19 cases, income levels, industry structure, age and racial distribution play important roles. The data informatics generated by the framework are made available to the public for a timely understanding of mobility trends and policy effects, as well as for time-sensitive decision support to further contain the spread of the virus.

Radiosonde Observations of Environments Supporting Deep Moist Convection Initiation during RELAMPAGO-CACTI

AbstractThe Remote Sensing of Electrification, Lightning, and Mesoscale/Microscale Processes with Adaptive Ground Observations (RELAMPAGO) and Cloud, Aerosol, and Complex Terrain Interactions (CACTI) projects deployed a high-spatiotemporal-resolution radiosonde network to examine environments supporting deep convection in the complex terrain of central Argentina. This study aims to characterize atmospheric profiles most representative of the near-cloud environment (in time and space) to identify the mesoscale ingredients affecting storm initiation and growth. Spatiotemporal autocorrelation analysis of the soundings reveals that there is considerable environmental heterogeneity, with boundary layer thermodynamic and kinematic fields becoming statistically uncorrelated on scales of 1–2 h and 30 km. Using this as guidance, we examine a variety of environmental parameters derived from soundings collected within close proximity (30 km in space and 30 min in time) of 44 events over 9 days where the atmosphere either: 1) supported the initiation of sustained precipitating convection, 2) yielded weak and short-lived precipitating convection, or 3) produced no precipitating convection in disagreement with numerical forecasts from convection-allowing models (i.e., Null events). There are large statistical differences between the Null event environments and those supporting any convective precipitation. Null event profiles contained larger convective available potential energy, but had low free-tropospheric relative humidity, higher freezing levels, and evidence of limited horizontal convergence near the terrain at low levels that likely suppressed deep convective growth. We also present evidence from the radiosonde and satellite measurements that flow–terrain interactions may yield gravity wave activity that affects CI outcome.

Multi-modal transportation recommendation with unified route representation learning

Multi-modal transportation recommendation aims to provide the most appropriate travel route with various transportation modes according to certain criteria. After analyzing large-scale navigation data, we find that route representations exhibit two patterns: spatio-temporal autocorrelations within transportation networks and the semantic coherence of route sequences. However, there are few studies that consider both patterns when developing multi-modal transportation systems. To this end, in this paper, we study multi-modal transportation recommendation with unified route representation learning by exploiting both spatio-temporal dependencies in transportation networks and the semantic coherence of historical routes. Specifically, we propose to unify both dynamic graph representation learning and hierarchical multi-task learning for multi-modal transportation recommendations. Along this line, we first transform the multi-modal transportation network into time-dependent multi-view transportation graphs and propose a spatiotemporal graph neural network module to capture the spatial and temporal autocorrelation. Then, we introduce a coherent-aware attentive route representation learning module to project arbitrary-length routes into fixed-length representation vectors, with explicit modeling of route coherence from historical routes. Moreover, we develop a hierarchical multi-task learning module to differentiate route representations for different transport modes, and this is guided by the final recommendation feedback as well as multiple auxiliary tasks equipped in different network layers. Extensive experimental results on two large-scale real-world datasets demonstrate the performance of the proposed system outperforms eight baselines.

Analysis of HIV/AIDS Epidemic and Socioeconomic Factors in Sub-Saharan Africa.

Sub-Saharan Africa has been the epicenter of the outbreak since the spread of acquired immunodeficiency syndrome (AIDS) began to be prevalent. This article proposes several regression models to investigate the relationships between the HIV/AIDS epidemic and socioeconomic factors (the gross domestic product per capita, and population density) in ten countries of Sub-Saharan Africa, for 2011–2016. The maximum likelihood method was used to estimate the unknown parameters of these models along with the Newton–Raphson procedure and Fisher scoring algorithm. Comparing these regression models, there exist significant spatiotemporal non-stationarity and auto-correlations between the HIV/AIDS epidemic and two socioeconomic factors. Based on the empirical results, we suggest that the geographically and temporally weighted Poisson autoregressive (GTWPAR) model is more suitable than other models, and has the better fitting results.

Spatiotemporal correlation of urban pollutants by long-term measurements on a mobile observation platform

We conducted a three-year campaign of atmospheric pollutant measurements exploiting portable instrumentation deployed on a mobile cabin of a public transport system. Size selected particulate matter (PM) and nitrogen monoxide (NO) were measured at high temporal and spatial resolution. The dataset was complemented with measurements of vehicular traffic counts and a comprehensive set of meteorological covariates. Pollutants showed a distinctive spatiotemporal structure in the urban environment. Spatiotemporal autocorrelations were analyzed by a hierarchical spatiotemporal statistical model. Specifically, particles smaller than 1.1 μm exhibited a robust temporal autocorrelation with those at the previous hour and tended to accumulate steadily during the week with a maximum on Fridays. The smallest particles (mean diameter 340 nm) showed a spatial correlation distance of ≈600 m. The spatial correlation distance reduces to ≈ 60 m for particle diameters larger than 1.1 μm, which also showed peaks at the stations correlated with the transport system itself. NO showed a temporal correlation comparable to that of particles of 5.0 μm of diameter and a correlating distance of 155 m. The spatial structure of NO correlated with that of the smallest sized particles. A generalized additive mixed model was employed to disentangle the effects of traffic and other covariates on PM concentrations. A reduction of 50% of the vehicles produces a reduction of the fine particles of −13% and of the coarse particle number of −7.5%. The atmospheric stability was responsible for the most significant effect on fine particle concentration.

Estimation of hourly PM1 concentration in China and its application in population exposure analysis.

Particulate pollution is closely related to public health. PM1 (particles with an aerodynamic size not larger than 1μm) is much more harmful than particles with larger sizes because it goes deeper into the body and hence arouses social concern. However, the sparse and unevenly distributed ground-based observations limit the understanding of spatio-temporal distributions of PM1 in China. In this study, hourly PM1 concentrations in central and eastern China were retrieved based on a random forest model using hourly aerosol optical depth (AOD) from Himawari-8, meteorological and geographic information as inputs. Here the spatiotemporal autocorrelation of PM1 was also considered in the model. Experimental results indicate that although the performance of the proposed model shows diurnal, seasonal and spatial variations, it is relatively better than others, with a determination coefficient (R2) of 0.83 calculated based on the 10-fold cross validation method. Geographical map implies that PM1 pollution level in Beijing-Tianjin-Hebei is much higher than in other regions, with the mean value of ∼55μg/m3. Based on the exposure analysis, we found about 75% of the population lives in an environment with PM1 higher than 35μg/m3 in the whole study area. The retrieval dataset in this study is of great significance for further exploring the impact of PM1 on public health.

Multi-aspect renewable energy forecasting

The increasing presence of renewable energy plants has created new challenges such as grid integration, load balancing and energy trading, making it fundamental to provide effective prediction models. Recent approaches in the literature have shown that exploiting spatio-temporal autocorrelation in data coming from multiple plants can lead to better predictions. Although tensor models and techniques are suitable to deal with spatio-temporal data, they have received little attention in the energy domain. In this paper, we propose a new method based on the Tucker tensor decomposition, capable of extracting a new feature space for the learning task. For evaluation purposes, we have investigated the performance of predictive clustering trees with the new feature space, compared to the original feature space, in three renewable energy datasets. The results are favorable for the proposed method, also when compared with state-of-the-art algorithms.