Air Temperature Prediction Research Articles

Employing Sequence-to-Sequence Stacked LSTM Autoencoder Architecture to Forecast Indian Weather

One of the buzz words in current research depends for a weather forecast is the atmospheric temperature, and this value may change at any time. Predicting air temperature exactly is necessary since it is particularly important for various communities and events. In today’s world of massive amounts of data and unforeseen data fluctuations, long-short term memory (LSTM), a sort of neural network method, is becoming more and more popular. The present work employed a sequence to sequence LSTM autoencoder to forecast Indian weather patterns. The prediction model was trained using historical data obtained from the Khordha district administration. The proposed method primarily emphasizes two phases of the procedure. The present study involves an examination of a proposed model, wherein a comparison is made between four distinct variants of LSTM techniques: vanilla LSTM, stacked LSTM, bidirectional LSTM, and convLSTM. The objective of this analysis is to predict temperature patterns. Additionally, this paper utilizes an LSTM autoencoder model to forecast the temperature for the subsequent two-year period in the Khordha district of Odisha, India. The experimental investigation demonstrates that the utilization of Stacking encoder-decoders with bidirectional LSTM cells leads to a notable enhancement in the accuracy of the model, resulting in maximum efficacy. A notably lower Mean Absolute Error (MAE) result suggests that the 9-layer stacked autoencoder model has superior performance in predicting both the minimum and maximum temperature.

Method of Cumulative Error Estimation for Surrogate Model in Indoor Air Temperature Prediction Task

Method of Cumulative Error Estimation for Surrogate Model in Indoor Air Temperature Prediction Task

“Hospital’s Thermo-neutral Zone for Patient Safety and Climate Change Sustainability

The efficacy of Iraq's direct heat recovery system is influenced by the nation's architectural design, particularly in view of the escalating climate conditions. There is an increasing worry on the necessity to openly address the local building patterns in Iraqi cities due to the challenging task of mechanically cooling buildings in the absence of electricity. The model was built using data obtained from field research carried out from January 2021 to March 2022. The field measurements of microclimatological parameters were carried out between March and August 2022 in order to evaluate the accuracy of the models. This methodology showcases the potential to make accurate predictions of future air temperatures by utilising variables as low as 5 percent. The air temperature decreases by 0.5 °C for every G increase across an area of 0.03-0.20 m2/m2, which is equivalent to a Leaf Area Index (LAI) of 4.5 m2/ha. This work introduces a standardized approach for assessing thermal comfort in real-world settings. The study examines the continuous use of hybrid ventilation systems at Kadhimiya Teaching Hospital to ensure a consistent air quality in the wards. Installing temperature control systems in these sites is vital to provide thermal comfort, as it serves as a potentially effective but expensive measure to address the impacts of climate change. Installing fans in military fortifications in Iraq seems to be a cost-efficient and uncomplicated approach to enhance their resilience against the escalating heat caused by the projected effects of global warming. However, when assessing the thermal comfort of buildings in Iraq's present and future climates, the established method proves to be considerably more advantageous.

Pentad-mean air temperature prediction using spatial autocorrelation and attention-based deep learning model

Pentad-mean air temperature prediction using spatial autocorrelation and attention-based deep learning model

Spatiotemporal prediction of daily air temperature using remote sensing and machine learning in Morocco

Spatiotemporal prediction of daily air temperature using remote sensing and machine learning in Morocco

Short-Term Prediction for Indoor Temperature Control Using Artificial Neural Network

Recently, data-based artificial intelligence technology has been developing dramatically, and we are considering how to model, predict, and control complex systems. Energy system modeling and control have been developed in conjunction with building technology. This study investigates the use of an artificial neural network (ANN) for predicting indoor air temperature in a test room with windows on an entire side. Multilayer perceptron (MLP) models were constructed and trained using time series data obtained at one-second intervals. Several subsampling time steps of 1 s, 60 s, 300 s, 600 s, 900 s, 1800 s, and 3600 s were performed by considering the actual operation control mode in which the time interval is important. The performance indices of the neural networks were evaluated using various error metrics. Successful results were obtained and analyzed based on them. It was found that as the multi-step time interval increases, performance degrades. For system control designs, a shorter prediction horizon is suggested due to the increase in computational time, for instance, the limited computing capacity in a microcontroller. The MLP structure proved useful for short-term prediction of indoor air temperature, particularly when control horizons are set below 100. Furthermore, highly reliable results were obtained at multi-step time intervals of 300 s or less. For the multivariate model, both calculation time and data dispersion increased, resulting in worsened performance compared to the univariate model.

Super-resolution-assisted rapid high-fidelity CFD modeling of data centers

Data center thermal management requires a good understanding of critical cooling airflow path. While CFD modeling excels at portraying airflow and temperature fields, it is often computationally intensive. Recent advancement in deep learning offers a rapid prediction of key parameters at several points of interest based on operation conditions of data centers. However, these models fail to deliver the comprehensive flow physics of an entire domain that CFD provides. This paper presents an innovative super-resolution approach, originating from computer vision, to model the air flow and temperature field in the cold aisle of a realistic data center. The proposed model reconstructs a high-fidelity CFD-generated flow field that typically takes several hours by using a low-fidelity CFD-generated flow field which only takes several minutes. The drastic reduction in computational time makes real-time prediction feasible, while providing detailed information for data center engineers to understand the flow field. Two variations of the framework are proposed and compared with the ground truth generated by the high-fidelity CFD model. A Mean Absolute Error less than 0.5 °C is achieved for air temperature prediction across all test scenarios, and less than 0.1 m/s for air velocity prediction. A sensitivity study is also conducted to determine the importance of input features.

Data driven modeling advancements for air temperature predictions in data centers

While data center cooling energy usage optimization studies have been performed through computational fluid dynamics/heat transfer (CFD/HT), and heuristic methods, data driven modeling techniques are now also being used for these applications. This paper investigates the air temperature prediction capabilities of static artificial neural network (ANN), Gaussian progress regression (GPR), support vector regression (SVR), relevance vector machine (RVM), linear regression, and regression trees; and transient long-short term memory (LSTM), and nonlinear autoregressive neural network with external input (NARX)) data driven modeling frameworks. The static study compared various models and found that GPR provided the best results (average error of 0.56 °C), closely followed by the ANN and SVR (average error of 0.60 °C and 0.68 °C respectively) methods. The transient study compared models based on an experimental data set and found that NARX outperforms LSTM for normal operations (0.83 °C and 1.07 °C average error respectively), and that data driven models are able to provide relatively good predictions, even if the input variables are slightly outside the training domain.

Application of artificial neural network to screen out the dominant meteorological parameters for prediction of air temperature

Application of artificial neural network to screen out the dominant meteorological parameters for prediction of air temperature

Absolute heat sources as a method to check the accuracy of temperature prediction in underground structures within cryolithozone

Forecasting the thermal regime of mine workings and the surrounding rock mass is a necessary element of the design of underground structures in cryolithic zone. This is particularly necessary when substantiating and selecting reliable methods and means of rock supporting, in order to ensure safe operation of underground structures during the entire standard service life. Changes in the temperature of discontinuous permafrost rocks in the range of negative values (below the ice point in the rock) can lead to a decrease in their strength characteristics, and consequently to a decrease in the stability of workings. The aim of the research was to compare two ways of considering absolute heat sources (point sources and sources uniformly distributed along the length of a mine working) when forecasting the thermal regime in mine workings of underground structures. The dependencies used to determine temperature differences in various methods of considering absolute heat sources were established. For the sake of generality, the dependencies were produced in dimensionless (criterial) form. The variants were calculated, and the results are presented in the form of graphs. The aim is to visually present the influence of the method of heat sources when considering the accuracy of air temperature prediction in an underground facility. Key qualitative and quantitative features of the formation of thermal regime in workings at different methods of considering absolute heat sources were established. It was shown in particular that during the transition from a negative temperature in a working to a positive one, incorrect consideration of the action of absolute heat sources can lead to an almost 30 % (1.26 times) difference (i.e., error) in the calculated depth of thawing of discontinuous rocks. It was also established that at a positive temperature, when the initial air temperature in a structure is more than 7.5 oC, there is no fundamental difference in engineering calculations results depending on the method of considering of absolute heat sources.

COMPARISON OF WEIGHTED MARKOV CHAIN AND FUZZY TIME SERIES-MARKOV CHAIN METHODS IN AIR TEMPERATURE PREDICTION IN BANDA ACEH CITY

Air temperature prediction is needed for various needs such as helping plan daily activities, agricultural planning, and disaster prevention. In this research, Weighted Markov Chain (WMC) method and Fuzzy Time Series-Markov Chain (FTS-MC) method are applied to predict the weekly air temperature in Banda Aceh city. The purpose of this study is to find out how the results of the application and comparison of the accuracy of the WMC method and the FTS-MC method on weekly air temperature prediction in Banda Aceh City. The prediction result of air temperature in Banda Aceh city using the WMC method for the next three weeks obtained an air temperature of 26,5℃. The prediction results of air temperature in Banda Aceh city using the FTS-MC method for the next three weeks obtained predicted values of 26,66℃ for the 105th week, 26,79℃ for the 106th week, and 26,83℃ for the 107th week. The MAPE accuracy level of the WMC method is 1,5% and the FTS-MC method is 1,7%. This shows that the MAPE of the WMC method is smaller than the FTS-MC method so it can be concluded that air temperature prediction using the WMC method is better than the FTS-MC method.

A perceptible stacking ensemble model for air temperature prediction in a tropical climate zone

Bangladesh is one of the world’s most susceptible countries to climate change. Global warming has significantly increased surface temperatures worldwide, including in Bangladesh. According to meteorological observations, the average temperature of the world has risen approximately 1.2 °C to 1.3 °C over the last century. Researchers and decision-makers have recently paid attention into the climate change studies. Climate models are used extensively throughout the nation in studies on global climate change to determine future estimates and uncertainties. This paper outlines a perceptible stacking ensemble learning model to estimate the temperature of a tropical region—Cox’s Bazar, Bangladesh. The next day’s temperature, maximum temperature, and minimum temperature are estimated based on the daily weather database collected from the weather station of Cox’s Bazar for a period of 20 years between 2001 and 2021. Five machine learning (ML) models, namely linear regression (LR), ridge, support vector regression (SVR), random forest (RF), and light gradient boosting machine (LGBM) are selected out of twelve ML models and combined to integrate the outputs of each model to attain the desired predictive performance. Different statistical schemes based on time-lag values play a significant role in the feature engineering stage. Evaluation metrics like mean absolute error (MAE), mean squared error (MSE), mean absolute percentage error (MAPE), and coefficient of determination (R2) are determined to compare the predictive performance of the models. The findings imply that the stacking approach presented in this paper prevails over the standalone models. Specifically, the study reached the highest attainable R2 values (0.925, 0.736, and 0.965) for forecasting temperature, maximum temperature, and minimum temperature. The statistical test and trend analysis provide additional evidence of the excellent performance of the suggested model.

Mitigating undercooling and overheating risk in existing desert schools under current and future climate using validated building simulation model

In Palestine and neighboring regions, buildings are constructed without insulation and mechanical heating and cooling systems, leading to significant thermal discomfort for occupants. To address this issue, the paper introduces a robust methodology that utilizes a validated building simulation model (BSM) created based on hourly indoor air temperature to assess indoor thermal comfort during winter and summer seasons under both current and future climates. This methodology is applied to an existing school built in 1990 in Jericho-Palestine, which has a hot desert climate. Classrooms rely on the thermal mass to increase the indoor temperature in the winter, and natural ventilation to reduce it in the summer. The effect of climate change on indoor thermal conditions is evaluated using typical warmer and colder future years based on the latest SSP5-8.5 scenario. The results showed that the calibrated and validated BSM achieved a highly accurate prediction of indoor air temperature compared to indoor air measured temperature. The validated BSM showed extreme cold for 880 h (91% of wintertime) and excessive heat for 90 h (19% of summertime) during school days. Passive winter measures, including high insulation and airtightness level, and using double glass windows, reduce undercooling to less than 40 h but increase overheating to 180 h. Passive summer measures, including night cooling and exterior shading, are necessary to reduce overheating to around 40 h. These measures are still effective in resisting the cold future years, but more creative passive summer measures and/or a mechanical cooling system are needed.

Influence of climate change on asphalt binder selection in China

ABSTRACT The impact of climate change on asphalt pavement infrastructure has become increasingly apparent. This paper investigated how climate change impacts the selection of asphalt binder for freeway pavement in China. The climate data predicted by the Global Climate Model were employed to select the appropriate performance grade for asphalt binder. The predicted air temperature converted pavement temperature, and performance grade distribution of the country was analysed. It is suggested that the performance grade of asphalt binder for freeways’ pavement should be determined based on ten years’ predicted climate data. This study concluded the following four major findings. Firstly, the selection of the appropriate asphalt binder performance grade needs to take into account the effects of future climate change. Secondly, the change in the standard deviation of temperature has a greater impact on the change in asphalt performance grade, as compared to the change in average air temperature. Thirdly, climate change affects the low-temperature performance grade more than the high-temperature performance grade in China. Finally, by 2050s, 15.2% of freeway asphalt pavement needs to upgrade high-temperature performance grade limit, and 17.3% of freeway asphalt pavement needs to upgrade to low-temperature performance grade limit.

Prediction of maximum air temperature for defining heat wave in Rajasthan and Karnataka states of India using machine learning approach

Temperature rise, associated with global warming, has increased the severity and frequency of heat waves around the world. Rajasthan and Karnataka are two major states of India and these have been facing frequent heat waves in the past few years. The present study used machine learning approach to predict the maximum air temperature (AT) for defining the heat wave occurrences in these two states. The analysis was based on the monthly data of 13 parameters, collected from NASA's Giovanni and ERA5 reanalysis data during the summer season for the past 10-years (2013–2022). The data obtained were at different resolutions, which was resampled to 5 km × 5 km spatial resolution. Pearson's correlation and sensitivity analysis were used to check the dominant input parameters. Three machine learning approaches were used in the study: multiple linear regression (MLR), support vector regression (SVR), and random forest (RF). 3-fold cross validation was used to evaluate the model performance. The data was divided into three periods for the purpose of testing, training and validation. It was observed that the maximum AT in both states was above the limits of IMD criteria for defining heat wave occurrences. The performance of the models was evaluated using statistical metrics, comprising of root mean square error (RMSE) and coefficient of determination (R2). Good correlation of AT with land surface temperature (LST), black carbon (BC), AOD, and CO was found for both the states. AT was found to be more sensitive to the change in LST and BC compared to other parameters. The machine learning results indicated that RF outperformed MLR and SVR in predicting AT in both states. The adjusted R2 was 0.90 and 0.92, for Rajasthan and Karnataka, respectively while RMSE was 2.36% and 1.44%, for Rajasthan and Karnataka, respectively. Overall, the study shows that machine learning-based approaches can predict maximum AT for defining heat wave conditions with high accuracy. The study can have significantly applications in different fields like climate modelling studies, urban planning and infrastructure, agriculture etc. and it can help to implement appropriate measures to mitigate the adverse impacts of temperature rise.

Sowing time of winter crops in Belarus and recommendation for optimization in conditions of modern climate change

Assessment of agroclimate conditions of the autumn period on the territory of Belarus under the conditions of modern climate change with the distribution of cumulative temperatures during autumn period for 2000–2020 is presented. The results of analysis of estimates of moisture indicators for the autumn period (precipitation amounts, Selyaninov’s hydrothermal moisture coefficient, and productive moisture reserves in the arable soil layer of 0–20 cm) are provided. Based on literary sources, recent research by Belarusian scientists, data from observation points of the state hydrometeorological network for the phases of development of winter crops (wheat, triticale, rye) in autumn, results of studies on tillering of winter crops depending on the sowing time and weather conditions in the autumn period, an analysis was carried out and criteria for cumulative temperatures were clarified to calculate the optimal sowing time for winter crops on the territory of Belarus. The optimal time for sowing winter grain crops has been calculated for administrative regions and points of the state network of hydrometeorological observations, taking into account the predicted air temperature for September. Recommendations are given for calculating the optimal sowing time for winter crops for points with no hydrometeorological observations, as well as with possible consideration of soil types and field location.

Sensor-based indoor air temperature prediction using deep ensemble machine learning: An Australian urban environment case study

Accurate prediction of indoor temperature is critical for climate change adaptation and occupant health. The aim of this study is to investigate an improved deep ensemble machine learning framework (DEML), by adjusting the model architecture with several machine learning (ML) and deep learning (DL) approaches to forecast the sensor-based indoor temperature in the Australian urban environment. We collected ambient station-based temperatures, satellite-based outdoor climate characteristics, and low-cost sensor-based indoor environmental metrics from 96 devices from August 2019 to November 2022, and established DEML with a rolling windows approach to assess the prediction stability over time. The DEML model was compared with several benchmark models, including Random Forest (RF), Support Vector Machine (SVM), eXtreme Gradient Boosting (XGboost), Long-short term memory (LSTM), and Super Learner model (SL). A total of 13,715 days [median: 341 days; IQR (the interquartile range): 221–977 days] of low-cost sensor-based indoor temperature were included in 25 commercial and residential buildings across eight cities. The prediction performance of DEML was superior to the other five benchmark models in most of the sensors [coefficients of determination (R2) of 0.861–0.990 and root mean square error (RMSE) of 0.125–0.886 °C], followed by RF and SL algorithms. DEML consistently achieved high accuracy across different climate zones, seasons, and building types, which could be used as a crucial tool for optimizing energy use, maintaining occupant comfort and health, and adapting to the impacts of climate change.

Boosting the prediction accuracy of a process-based greenhouse climate-tomato production model by particle filtering and deep learning

By generating high quality data without the big time investment and economic cost of real experiments, dynamic greenhouse climate and crop simulation models can support decisions on greenhouse climate control, crop management and greenhouse design. The reliability of simulation-based decisions depends on both the prediction accuracy and interpretability of simulation models. The prediction accuracy of these simulation models can be increased by: 1) improving mechanisms in process-based models; 2) calibrating process-based model parameters; 3) deriving black-box relationships from data. Considering the descending interpretability from (1) to (3), this study presents a knowledge-based data-driven modelling approach where firstly a process-based model is selected and modified based on domain knowledge, then data-driven improvement is applied including two steps: parameter value estimation by particle filter (PF) and further black-box improvement by deep neural networks (DNN). The approach was tested with an example of greenhouse climate-tomato production system modelling. Modules from GreenLight (Katzin et al., 2020) and TOMSIM (Heuvelink, 1995, 1996) were selected, modified and integrated into a process-based greenhouse climate-tomato model. Validation showed that PF-calibration of five greenhouse parameters decreased the seasonal relative root mean squared error (RRMSE) of indoor air vapor pressure predictions from 40.7% of that before PF-calibration to 16.4%, while it did not decrease the RRMSE of indoor air temperature predictions. Combining the PF-calibrated model with a DNN trained on a season of data decreased the RRMSE of indoor air temperature from 15.0% without DNN to 6.7%, and decreased the RRMSE of indoor air vapor pressure to 12.6%. The knowledge-based data-driven greenhouse climate-tomato model had a relative error of 0.9% for seasonal total fresh yield, and an RRMSE of 6.6% for the cumulative yield throughout the season. If process-based model parameters were not calibrated before combining the model with DNNs, the required amount and diversity of DNN training data increased because more information needed to be learnt from data by the DNNs. Without PF-calibration, combining a DNN trained on 50 days of data with the process-based model resulted in RRMSEs of 44.8% and 31.8% for indoor air temperature and vapor pressure prediction, respectively; with PF-calibration, the RRMSEs were decreased to 13.1% and 17.9%. The proposed three-step knowledge-based data-driven approach can not only improve the model prediction accuracy, but can also help to track and interpret the improvements.

Investigating patterns of air pollution in metropolises using remote sensing and neural networks during the COVID-19 pandemic

The purpose of this study is to determine the amount of air pollution in Tehran, Isfahan, Semnan, Mashhad, Golestan, and Shiraz during the Corona era and before. For this purpose, Sentinel satellite images were used to investigate the concentration of Methane (CH4), Carbon Monoxide (CO), Carbon Dioxide (CO2), Nitrogen Dioxide (NO2), Ozone (O3), Sulfur Dioxide (SO2), aerosol pollutants in In the era before and during Corona. Furthermore, greenhouse effect-prone areas were determined in this study. In the following, the state of air inversion in the studied area was determined by taking the temperature on the surface of the earth and in the upper atmosphere, as well as the wind speed into account. In this research, the prediction of air temperature for the year 2040 was conducted using the Markov and Cellular Automaton (CA)-Markov methods, considering the impact of air pollution on the air temperature of metropolises. Additionally, the Radial Basis Function (RBF) and Multilayer Perceptron (MLP) methods have been developed to determine the relationship between pollutants, areas prone to air inversions, and temperature values. According to the results, pollution caused by pollutants has decreased in the Corona era. According to the results, there is more pollution in Tehran and Isfahan metropolises. In addition, the results showed that air inversions in Tehran is the highest. Additionally, the results showed a high correlation between temperature and pollution levels (R2 = 0.87). Thermal indices in the studied area indicate that Isfahan and Tehran, with high values of Surface Urban Heat Island (SUHI) and being in the 6th class of thermal comfort (Urban Thermal Field Variance Index (UTFVI)), are affected by thermal pollution. The results showed that parts of southern Tehran province, southern Semnan and northeastern Isfahan will have higher temperatures in 2040 (class 5 and 6). Finally, the results of the neural network method showed that the MLP method with R2 = 0.90 is more accurate than the RBF method in predicting pollution amounts. This study significantly contributes by introducing innovative advancements through the application of RBF and MLP methods to assess air pollution levels during the COVID-19 and pre-pandemic periods, while also investigating the intricate relationships among greenhouse gases, air inversion, air temperature, and pollutant indices within the atmosphere. The utilization of these methods notably enhances the accuracy and reliability of pollution predictions, amplifying the originality and significance of this research.

Development of a hybrid model to predict air temperature over an urban area: A case study over Ahmedabad, India

The present paper discusses development of a hybrid model for predicting near surface air temperature (2 m) over the city scale (∼100 m). The model is based on the premise that temperature at each place within the city is correlated with its mean temperature. The proposed model estimates the mean temperature of a city, using the forecast generated (5 km × 5 km resolution) by regional weather prediction model viz., Weather and Research Forecast (WRF) and downscale it to city scale, using the correlations estimated from the satellite based high-resolution (1 km × 1 km) surface characteristics i.e. Land Surface Temperature (LST). The downscaling coefficients at each grid (1 km x1 km) are generated by formulating the statistical relationship between the LST over the grid to the mean temperature of a city. LST from Moderate Resolution Imaging Spectro-radiometer (MODIS) sensor on-board Aqua satellite during the Indian summer season (April–June) of the year 2017–2018 has been employed for this purpose. The relationship between LST and air temperature is established using the collocated satellite estimated LST and monitored air temperature observations. As a case study, the model is demonstrated for the Indian city Ahmedabad, but the same procedure can be extended for any other cities. The existence of strong urban heat island during night hours over Ahmedabad city with a temperature gradient of 4-5 °C has been found in the study. During day hours the different parts of the city depicts different temperature based on their surface characteristics. The results have been validated using the in-situ data based on the ground based observational network temporally installed at different locations of the city for the year 2022. The developed model shows that it has a potential to be used in operational mode for generating location specific extreme temperature or heat wave warning within a city.