Short-term Trends Research Articles

Air pollution in Białowieża forest: Analysis of short-term trends from 2014 to 2021

Air pollution caused by sulphur dioxide (SO2) and nitrogen oxides (NOx) has negative impacts on forest health and can initiate forest dieback. Long-term monitoring and analysis of these pollution are carried out in Białowieża Forest in NE Poland due to the threats from abiotic, biotic and anthropogenic factors. The main objective of our study was to monitor the levels and trends of air pollutant deposition in Białowieża Forest. During a short-term monitoring period over six years (2014–2021), the concentration of SO2 in the air decreased significantly (from 2.03 μg m−3 in December 2015 to 0.20 μg m−3 in July 2016), while the concentration of NO2 in the air showed a non-significant decrease (from 8.24 μg m−3 in December 2015 to 1.61 μg m−3 May 2016). There was no significant linear trend in the wet deposition of S–SO4 anions. Mean monthly S–SO4 deposition varies between 4.54 and 94.14 mg m−2month−1. Wet nitrogen deposition, including oxidized nitrogen (N–NO3) and reduced nitrogen (N–NH4), showed a non-significant increase. Mean monthly precipitation of N–NO3 and N–N H4 ranged from 1.91 to 451.73 mg m−2month−1. Neither did total sulphur deposition nor total nitrogen deposition exceed the mean deposition values for forests in Europe (below 6 ha−1yr−1 and 3–15 ha−1yr−1, respectively). Our results indicate that air pollutants originate from local sources (households), especially from the village of Białowieża, as demonstrated by the level and spatial distribution of air pollutant deposition. This indicates that air pollutants from the village of Białowieża could spread to other parts of Białowieża Forest in the future and may have a negative impact on forest health and can initiate forest dieback. It is therefore important to continue monitoring air pollution to assess the threats to this valuable forest ecosystem.

Impacts of Weather Variability on the International Tourism Receipts—Evidence from Ethiopia (1995–2019)

Every economic sector is susceptible to the direct or indirect effects of weather variability, and the tourism sector is no exception. In fact, the tourism industry is considered to be more vulnerable to the effects of weather variability than the general economy, with changes in weather patterns, extreme events, and environmental degradation offering substantial obstacles. Ethiopia’s tourism industry, like many others, faces challenges from weather variability. This study investigates the short- and long-term effects of weather variability on Ethiopia’s international tourism receipts. Utilizing data from 1995 to 2019, the research employs a vector error correction model to analyze the relationships between weather variables (temperature, rainfall), economic factors (GDP growth, inflation), political stability, and tourist arrivals. The findings reveal that in the long run, higher temperatures, rainfall, and inflation have negative impacts on tourism receipts, while political stability and past tourist arrivals have positive effects. Short-term trends mirror these, with the addition of GDP growth not showing a significant impact. To ensure the sustainability of tourism in Ethiopia, the study emphasizes the importance of understanding weather’s influence, developing adaptation strategies, and promoting sustainable tourism practices.

Determinants of Yearly CO2 Emission Fluctuations: A Machine Learning Perspective to Unveil Dynamics

In order to understand the dynamics in climate change, inform policy decisions and prompt timely action to mitigate its impact, this study provides a comprehensive analysis of the short-term trend of the year-on-year CO2 emission changes across ten countries, considering a broad range of factors including socioeconomic factors, CO2-related industry, and education. This study uniquely goes beyond the common country-based analysis, offering a broader understanding of the interconnected impact of CO2 emissions across countries. Our preliminary regression analysis, using the ten most significant features, could only explain 66% of the variations in the target. To capture the emissions trend variation, we categorized countries by the change in CO2 emission volatility (high, moderate, low with upward or downward trends), assessed using standard deviation. We employed machine learning techniques, including feature importance analysis, Partial Dependence Plots (PDPs), sensitivity analysis, and Pearson and Canonical correlation analyses, to identify influential factors driving these short-term changes. The Decision Tree Classifier was the most accurate model, with an accuracy of 96%. It revealed population size, CO2 emissions from coal, the three-year average change in CO2 emissions, GDP, CO2 emissions from oil, education level (incomplete primary), and contribution to temperature rise as the most significant predictors, in order of importance. Furthermore, this study estimates the likelihood of a country transitioning to a higher emission category. Our findings provide valuable insights into the temporal dynamics of factors influencing CO2 emissions changes, contributing to the global efforts to address climate change.

Analysis of Spatial Differentiation of NDVI and Climate Factors on the Upper Limit of Montane Deciduous Broad-Leaved Forests in the East Monsoon Region of China

The vertical transition zone of mountain vegetation is characterized by high species diversity, and the width of the transition zone may serve as an indirect indicator of climate change. However, research into the differential characteristics of vegetation response to climate changes at the boundary of vertical transition zones has been limited. This study employs MODIS and climate data spanning 2001 to 2018 to investigate spatiotemporal trends in precipitation (PRE), temperature (TMP), radiation (RAD), and Normalized Difference Vegetation Index (NDVI) across nine montane deciduous broad-leaved forests in the eastern monsoon region of China. It explores the time-lag and -accumulation effects of climatic variables on NDVI, quantifying their relative contributions to both its short-term and interannual variations. Results show that, notably, with the Qinling-Daba Mountains as a demarcation, northern regions exhibit significant increases in RAD (0.874–2.047 W m−2/a), whereas southern regions demonstrate notable rises in TMP (0.59–0.73 °C/10a). Areas of lower annual PRE correspond to the most rapid increases in annual average NDVI (5.045 × 10−3/a). NDVI’s lag time and cumulative duration responses to TMP are the shortest (0 and 2~4 periods), while its correlation with RAD is the strongest (0.815–0.975), generally decreasing from higher to lower latitudes. TMP significantly affects NDVI variations, impacting both short-term and interannual trends, with PRE driving short-term fluctuations and RAD dictating long-term shifts. This research provides critical data and a theoretical framework that enhances our understanding of how regional vegetation’s vertical zonation responds to climate change, thereby making a substantial contribution to the study of mountain vegetation’s diverse adaptability to climatic variations.

Unraveling the crystal ball: Machine learning models for crude oil and natural gas volatility forecasting

This study aims to forecast crude oil and natural gas volatility across various forecasting horizons, from daily to quarterly, using diverse machine learning models. It critically analyzes eleven models, including Linear Regression, Elastic Regression, Ridge Regression, Lasso Regression, Huber Regression, Random Forest Regression, SVM, LSTM, GRU, ANN, and XGBoost, using RMSE for accuracy. The study reveals that model performance significantly varies with forecasting horizons; a phenomenon attributed to each model's inherent capabilities in processing short-term versus long-term market trends. This detailed understanding aids in selecting the most appropriate models for specific forecasting needs, essential for policymakers and practitioners in managing volatility effectively. The study concludes with a recommendation for using Random Forest Regression and XGBoost for natural gas volatility forecasting, providing key insights for enhancing economic resilience and stability.

Long-Term Evolution of Significant Wave Height in the Eastern Tropical Atlantic between 1940 and 2022 Using the ERA5 Dataset

Studies on the variability in ocean wave climate provide engineers and policy makers with information to plan, develop, and control coastal and offshore activities. Ocean waves bear climatic imprints through which the global climate system can be better understood. Using the recently updated ERA5 dataset, this study evaluated the spatiotemporal distribution and variability in significant wave height (SWH) in the Eastern Tropical Atlantic (ETA). The short-term trends and rates of change were obtained using the Mann–Kendall trend test and the Theil–Sen slope estimator, respectively, and decadal trends were assessed using wavelet transformation. Significant, positive monthly and yearly trends and a prevailing decadal trend were observed across the domain. Observed trends suggest that stronger waves are getting closer to the coast and are modulated by the Southern and Northern Atlantic mid-latitude storm fields. These observations have implications for the increasing coastal erosion rates on the eastern coast of the Tropical Atlantic.

Adaptive solar power generation forecasting using enhanced neural network with weather modulation

Solar power generation forecasting plays a vital role in optimizing grid management and stability, particularly in renewable energy-integrated power systems. This research paper presents a comprehensive study on solar power generation forecasting, evaluating traditional and advanced machine learning methods, including ARIMA, Exponential Smoothing, Support Vector Regression, Random Forest, Gradient Boosting, and Physics-based Models. Moreover, we propose an innovative Enhanced Artificial Neural Network (ANN) model, which incorporates Weather Modulation and Leveraging Prior Forecasts to enhance prediction accuracy. The proposed model is evaluated using real-world solar power generation data, and the results demonstrate its superior performance compared to traditional methods and other machine learning approaches. The Enhanced ANN model achieves an impressive Root Mean Square Error (RMSE) of 0.116 and a Mean Absolute Percentage Error (MAPE) of 36.26%. The integration of Weather Modulation allows the model to adapt to changing weather conditions, ensuring reliable forecasts even during adverse scenarios. Leveraging Prior Forecasts enables the model to capture short-term trends, reducing forecasting errors arising from abrupt weather changes. The proposed Enhanced ANN model showcases its potential as a promising tool for precise and reliable solar power generation forecasting, contributing to the efficient integration of solar energy into the power grid and advancing sustainable energy practices.

Sustaining Community-Based Research in Computing: Lessons from Two Tech Capacity Building Initiatives for Local Businesses

The field of human-computer interaction (HCI) has traditionally focused on the design of novel technology artifacts. However, ensuring considerations for artifact maintenance and repair is crucial to sustainably supporting the populations they aim to serve over the long term. Drawing on two multi-year programs for tech capacity building in post-industrial U.S. cities, this article presents a comparative analysis to investigate the challenges and strategies for sustained community-based research in computing. In particular, our work detailed three considerations for academic-community partnerships. First, long-term partnerships prioritized transferring trust across academic and community personnel and continually set expectations that responded to evolving community initiatives (i.e., relational sustainability). Second, partnerships used academic support as a way to kickstart community initiatives, and flexibly reframed interventions to stay aligned with evolving community goals (i.e., economic sustainability). Third, partnerships trained personnel to provide technical support alongside interventions and prioritized advice that resisted short-term trends (i.e., technical sustainability). We provide concrete examples of how our two academic-community partnerships carried out such suggestions-such details go unreported in scholarly articles yet are essential for sustainability considerations. We discuss ongoing challenges, such as rethinking when longevity should and should not be the end goal.

Spatiotemporal Variability of Convective Events in Romania Based on METAR Data

Convective weather, through its heavy showers, strong winds and hail, significantly impacts human activities, having the potential to inflict serious damage on social and environmental sectors. Limited research has been conducted on this phenomenon within Romanian territory, and currently there is no referenced climatological study primarily aimed at air traffic management users in this context. This study aims to assess the climatological aspects related to convective events based on sub-hourly observation data recorded at 17 airport weather stations throughout Romania during an 11-year period (2012–2022). The spatiotemporal distribution of convective events was analyzed based on occurrences of Cumulus Congestus (TCU) clouds, Cumulonimbus (CB) clouds, thunderstorms (TSs), heavy showers (+SHs), and hail (GR). With the data being extracted from meteorological aerodrome reports (METARs) and special meteorological aerodrome reports (SPECIs). Short-term trends were determined using Sen’s slope estimator, and statistical significance was assessed through the Mann–Kendall test. The main findings indicated that the highest occurrence of convective events is located over central and western Romania, with June emerging as the extreme month in terms of convective events, while the hourly distribution emphasizes that the highest frequency of convective events occurred in the afternoon. Trend analysis in TCU, CB, and TS show tendencies toward higher frequency of convective events while the results related to +SH and GR indicate a high variability across Romanian territory. Trend analysis disclosed more substantial changes in the TS variable. The results of this study bear potential significance for a broad spectrum of human activities and the management of natural environments.

Analysis of short-term wind speed variation, trends and prediction: A case study of Tamil Nadu, India

Abstract Purpose The purpose of this research article is to analyze the short-term wind speed and develop a framework model to overcome the challenges in the wind power industry. Design/Methodology/Approach Real data with a case study of wind speed is presented to illustrate the advantages of this new wind speed analytical framework. Hourly measurements of wind speed are observed, and the experiments are conducted using tools such as ANOVA, control charts, trend analysis, and predictive models. The August month data for over 13 years from modern era retrospective-analysis for research and applications (MERRA) National aeronautics and space administration (NASA) for Coimbatore and Erode locations in Tamil Nadu, India, have been used. The results were considered for the study to understand the wind speed data and the implementation of new wind power projects in India. Findings The essence of the proposed wind speed analytical framework is its flexible approach, which enables the effective integration of wind firms’ individual requirements by developing tailor-made analytical evaluations. Originality/Value This article derives the wind speed analytical framework with the application of statistical tools and machine learning algorithms.

Research on Machine Learning-based Prediction of Coffee Futures Prices

Coffee is one of the major agricultural commodities in world trade, and its futures are vital tools in the global capital market. The continued vitality of the coffee market and substantial price fluctuations have increased hedging demands among market participants. Consequently, predicting future price trends of coffee futures to yield excess returns has become a focal point in the field of quantitative investment. Machine learning methods are increasingly being applied in the field of quantitative investment due to their performance advantages in complex data classification and regression. This paper analyzes the current state of the coffee futures market and the factors influencing its prices. In this study, five market indicators and one technical indicator, the bias rate, were selected as inputs. The closing price for the subsequent day, along with short-term (50 days) and long-term (200 days) price trends, were forecasted using two machine learning techniques: the linear regression model and the random forest model. The results demonstrated that, of the two predictive models utilized in this study, the random forest model performed better concerning regression prediction evaluation indices. When predicting short-term (50-day) price trends, the linear regression model exhibited superior performance. However, both models revealed significant errors in predicting long-term (200-day) price trends.

Short-term trends of air quality and pollutant concentrations in Nigeria from 2018–2022 using tropospheric sentinel-5P and 3A/B satellite data

This study presents a comprehensive analysis of atmospheric parameters in Nigeria, utilizing Sentinel-5P and 3A/B offline datasets from April 2018 to December 2022. The research evaluates the concentrations of pollutants such as Nitrogen Dioxide (NO2), Sulphur Dioxide (SO2), Ozone (O3), Formaldehyde (HCHO), Methane (CH4), Carbon Monoxide (CO), and Aerosol Indices (AI) across different periods. The data analysis employs multivariate techniques, including Factor Analysis, Pearson Correlation Coefficient, Scatter Plots, and Dendrogram Clustering, providing insights into the interrelationships and potential sources of pollution. Additionally, the study examines monthly trends, temperature variations, and their impacts on human health and climate. The findings reveal stable levels of NO2, SO2, O3, and HCHO with minimal fluctuations, while CH4 and CO exhibit slight variations, possibly influenced by emission reductions or atmospheric changes. Aerosol optical indices consistently indicate aerosol presence, influenced by local emissions and atmospheric processes. Multivariate analyses highlight shared patterns among atmospheric constituents, shedding light on potential common sources and chemical relationships. The Pearson correlation coefficients illustrate the strength and direction of linear relationships, offering valuable information on pollution dynamics. The study's visualizations, including scatter plots and dendrogram clustering, enhance the understanding of complex interactions. Overall, the research underscores the significance of continuous monitoring and research efforts for informed decision-making and environmental management in Nigeria.

Magnitude and determinants of excess total, age-specific and sex-specific all-cause mortality in 24 countries worldwide during 2020 and 2021: results on the impact of the COVID-19 pandemic from the C-MOR project

IntroductionTo examine the impact of the COVID-19 pandemic on mortality, we estimated excess all-cause mortality in 24 countries for 2020 and 2021, overall and stratified by sex and age.MethodsTotal, age-specific...

Assessing emission-driven changes in health risk of source-specific PM2.5-bound heavy metals by adjusting meteorological covariates

Heavy metals (HMs) in PM2.5 gain much attention for their toxicity and carcinogenic risk. This study evaluates the health risks of PM2.5-bound HMs, focusing on how meteorological conditions affect these risks against the backdrop of PM2.5 reduction trends in China. By applying a receptor model with a meteorological normalization technique, followed by health risk assessment, this work reveals emission-driven changes in health risk of source-specific HMs in the outskirt of Tianjin during the implementation of China’ second Clean Air Action (2018–2020). Sources of PM2.5-bound HMs were identified, with significant contributions from vehicular emissions (on average, 33.4 %), coal combustion (26.3 %), biomass burning (14.1 %), dust (11.7 %), industrial boilers (9.7 %), and shipping emission and sea salt (4.7 %). The source-specific emission-driven health risk can be enlarged or dwarfed by the changing meteorological conditions over time, demonstrating that the actual risks from these source emissions for a given time period may be higher or smaller than those estimated by traditional assessments. Meteorology contributed on average 56.1 % to the interannual changes in source-specific carcinogenic risk of HMs from 2018 to 2019, and 5.6 % from 2019 to 2020. For the source-specific noncarcinogenic risk changes, the contributions were 38.3 % and 46.4 % for the respective periods. Meteorology exerts a more profound impact on daily risk (short-term trends) than on annual risk (long-term trends). Such meteorological impacts differ among emission sources in both sign and magnitude. Reduced health risks of HMs were largely from targeted regulatory measures on sources. Therefore, the meteorological covariates should be considered to better evaluate the health benefits attributable to pollution control measures in health risk assessment frameworks.

Short-Term Exhaust Gas Temperature Trend Prediction of a Marine Diesel Engine Based on an Improved Slime Mold Algorithm-Optimized Bidirectional Long Short-Term Memory—Temporal Pattern Attention Ensemble Model

As the core component of a ship’s engine room, the operation of a marine diesel engine (MDE) directly affects the economy and safety of the entire vessel. Predicting the future changes in the status parameters of a MDE helps to understand the operational status, enabling timely warnings to the engine crew, and to ensure the safe navigation of the vessel. Therefore, this paper combines the temporal pattern attention mechanism with the bidirectional long short-term memory (BiLSTM) network to propose a novel trend prediction method for short-term exhaust gas temperature (EGT) forecasting. First, the Pearson correlation analysis (PCA) is conducted to identify input feature variables that are strongly correlated with the EGT. Next, the BiLSTM network models input feature variables such as load, fuel oil pressure, and scavenging air pressure and capture the interrelationships between different vectors from the hidden layer matrix within the BiLSTM network. This allows the selection of valuable information across different time steps. Meanwhile, the temporal pattern attention (TPA) mechanism has the ability to explore complex nonlinear dependencies between different time steps and series. This assigns appropriate weights to the feature variables within different time steps of the BiLSTM hidden layer, thereby influencing the input effect. Finally, the improved slime mold algorithm (ISMA) is utilized to optimize the hyperparameters of the prediction model to achieve the best level of short-term EGT trend prediction performance based on the ISMA-BiLSTM-TPA model. The prediction results show that the mean square error, the mean absolute percentage error, the root mean square error and the coefficient of determination of the model are 0.4284, 0.1076, 0.6545 and 98.2%, respectively. These values are significantly better than those of other prediction methods, thus fully validating the stability and accuracy of the model proposed in this paper.

Cliff recession geodynamics variability and constraints within poorly consolidated landslide-prone coasts in the southern Baltic Sea, Poland

Abstract This study identifies the reasons for geodynamics variability of the coastal system within two cliff-shore sections of the southern Baltic Sea (SBS). The comparative analysis included distinct moraines and their foregrounds near the open sea (S1) and within the Gulf of Gdańsk (S2). Short-term trends indicate a direct link between landslide occurrence and increased cliff retreat. Long-term (total) values were obtained by developing the 4F MODEL for large-scale applications, based on the analysis of remote sensing and hydroacoustic data (to determine the extent of shore platforms), the modelling of higher-order polynomial functions describing their extent, followed by the integral calculus of the indicated functions within the open-source Desmos environment. The retreat dynamics for individual landslides (S1) was an order of magnitude higher (m/yr) than the average for the whole cliff section (0.17 ± 0.008 m/yr), which correlates well with medium- and long-term development tendencies and recession dynamics, revealed by the numerical modelling method, since approximately 8 ka b2k, years before 2000 CE (at S1 = 0.17 ± 0.020 m/yr, at S2 = 0.11 ± 0.005 m/yr). While the approach described in this paper can reveal, project, and simulate the dynamics of past and future trends within other cliffed coasts shaped in tideless conditions, it also proves stable moraine erosional responses to sea-level rise since the Mid-Holocene.

Trajectories of relationship and sexual satisfaction over 2 years in the Covid-19 pandemic: A latent class analysis.

Previous research on the impact of the Covid-19 pandemic on romantic relationships has mainly concentrated on short-term effects and average trends of change. This study aimed to explore different trajectories of relationship and sexual satisfaction from April 2020 to March 2022. Including a cross-national sample of 2859 individuals, a latent class approach was applied to identify subgroups of sexual and relationship satisfaction trajectories. Participants' satisfaction levels, attachment style, psychological symptoms, life satisfaction, and living arrangements were taken into account. Three latent classes were identified for both relationship and sexual satisfaction: (1) a group with high satisfaction levels but a slightly declining trajectory, (2) a group with low satisfaction levels but an ascending trajectory, and (3) a fluctuating group. Living apart, having children, psychological symptoms, low life satisfaction, and avoidant attachment characterized the two groups with low or fluctuating trajectories. The findings suggest that there were different trajectories of relationship and sexual satisfaction during the pandemic, which can be predicted by individual dispositions. Even 2 years after the pandemic, its effects had not disappeared.

LSTTN: A Long-Short Term Transformer-based spatiotemporal neural network for traffic flow forecasting

Accurate traffic forecasting is a fundamental problem in intelligent transportation systems and learning long-range traffic representations with key information through spatiotemporal graph neural networks (STGNNs) is a basic assumption of current traffic flow prediction models. However, due to structural limitations, existing STGNNs can only utilize short-range traffic flow data; therefore, the models cannot adequately learn the complex trends and periodic features in traffic flow. Besides, it is challenging to extract the key temporal information from the long historical traffic series and obtain a compact representation. To solve the above problems, we propose a novel LSTTN (Long-Short Term Transformer-based Network) framework comprehensively considering the long- and short-term features in historical traffic flow. First, we employ a masked subseries Transformer to infer the content of masked subseries from a small portion of unmasked subseries and their temporal context in a pretraining manner, forcing the model to efficiently learn compressed and contextual subseries temporal representations from long historical series. Then, based on the learned representations, long-term trend is extracted by using stacked 1D dilated convolution layers, and periodic features are extracted by dynamic graph convolution layers. For the difficulties in making time-step level prediction, LSTTN adopts a short-term trend extractor to learn fine-grained short-term temporal features. Finally, LSTTN fuses the long-term trend, periodic features and short-term features to obtain the prediction results. Experiments on four real-world datasets show that in 60-minute-ahead long-term forecasting, the LSTTN model achieves a minimum improvement of 5.63% and a maximum improvement of 16.78% over baseline models. The source code is availble at https://github.com/GeoX-Lab/LSTTN.

A mixed-effects model to predict COVID-19 hospitalizations using wastewater surveillance

During the COVID-19 pandemic, many countries and regions investigated the potential use of wastewater-based disease surveillance as an early warning system. Initially, methods were created to detect the presence of SARS-CoV-2 RNA in wastewater. Investigators have since conducted extensive studies to examine the link between viral concentration in wastewater and COVID-19 cases in areas served by sewage treatment plants over time. However, only a few reports have attempted to create predictive models for hospitalizations at a county-level based on SARS-CoV-2 RNA concentrations in wastewater. This study implemented a linear mixed-effects model that evaluates the association between levels of virus in wastewater and county-level hospitalizations. The model was then utilized to predict short-term county-level hospitalization trends in 21 counties in California based on data from March 21, 2022, to May 21, 2023. The modeling framework proposed here permits repeated measurements, as well as fixed and random effects. The model that incorporated wastewater data as an input variable rather than cases or test positivity rate exhibited robust performance and effectively captured discernible trends in hospitalizations. Additionally, the model allows for the prediction of SARS CoV-2 hospitalizations two weeks ahead. Forecasts of COVID-19 hospitalizations could provide crucial information for hospitals to better allocate resources and prepare for potential surges in patient numbers.

Attention Mechanism with Spatial-Temporal Joint Deep Learning Model for the Forecasting of Short-Term Passenger Flow Distribution at the Railway Station

Accurate understanding of passenger flow distribution is crucial for effective station crowd management. However, due to the complexity and randomness of passenger flow and the unclear spatial-temporal correlation between functional areas within the station, predicting the spatiotemporal distribution dynamics of inflow and future short-term distribution trends is challenging. Emerging deep learning models offer valuable insights for accurately predicting passenger flow distribution. Thus, we propose a deep learning architecture, named “ST-Bi-LSTM,” which combines a bidirectional long short-term memory network with a spatial-temporal attention mechanism. Initially, we outline the methodologies of Bi-LSTM, the DeepWalk-based spatial attention mechanism, and the temporal attention mechanism. The spatial attention mechanism is employed to extract station spatial network topology information and enhance the representation of passenger flow characteristics in highly correlated areas during the forecasting process. Simultaneously, the temporal attention Bi-LSTM is utilized for capturing temporal correlations. The architecture comprises four branches dedicated to station real-time video monitoring data, spatial network topology, function area attributes, and train timetables. Subsequently, leveraging in-station CCTV data, passenger travel behavior data, and train timetables, we apply the architecture to the Tianjin West High-Speed Railway Station. We conduct a comparative analysis of the prediction performance and time complexity of the proposed architecture against existing baseline models, demonstrating superior performance and robustness exhibited by the ST-Bi-LSTM model (achieving a reduction in RMSE of over 10%). This study facilitates the transition of station management from passive response to active prediction of station passenger flow dynamics.