Ground-based Stations Research Articles

Assessment of current methane emission quantification techniques for natural gas midstream applications

Abstract. Methane emissions from natural gas systems are increasingly scrutinized, and accurate reporting requires quantification of site- and source-level measurement. We evaluate the performance of 10 available state-of-the-art CH4 emission quantification approaches against a blind controlled-release experiment at an inerted natural gas compressor station in 2021. The experiment consisted of 17 blind 2 h releases at a single exhaust point or multiple simultaneous ones. The controlled releases covered a range of methane flow rates from 0.01 to 50 kg h−1. Measurement platforms included aircraft, drones, trucks, vans, ground-based stations, and handheld systems. Herewith, we compare their respective strengths, weaknesses, and potential complementarity depending on the emission rates and atmospheric conditions. Most systems were able to quantify the releases within an order of magnitude. The level of errors from the different systems was not significantly influenced by release rates larger than 0.1 kg h−1, with much poorer results for the 0.01 kg h−1 release. It was found that handheld optical gas imaging (OGI) cameras underestimated the emissions. In contrast, the “site-level” systems, relying on atmospheric dispersion, tended to overestimate the emission rates. We assess the dependence of emission quantification performance on key parameters such as wind speed, deployment constraints, and measurement duration. At the low wind speeds encountered (below 2 m s−1), the experiments did not reveal a significant dependence on wind speed. The ability to quantify individual sources degraded during multiple-source releases. Compliance with the Oil and Gas Methane Partnership's (OGMP 2.0) highest level of reporting may require a combination of the specific advantages of each measurement technique and will depend on reconciliation approaches. Self-reported uncertainties were either not available or were based on the standard deviation in a series of independent realizations or fixed values from expert judgment or theoretical considerations. For most systems, the overall relative errors estimated in this study are higher than self-reported uncertainties.

A GIS-big data model for improving the coverage and analysis processes of territory observation, and integrating ground-based observations with retrospective meteorological data

This paper presents Kawak, a GIS-bigdata model for merging retrospective meteorological data with ground-based observations to achieve comprehensive territorial coverage. This model identifies areas without active ground-based stations and creates virtual stations for these areas. Kawak incorporates AI algorithms for conducting spatio-temporal studies, enabling seamless merging and exploration of climate patterns. We implemented Kawak in a prototype and conducted an exploratory case study by fusing temperature records from MERRA-2 with EMAS ground-based observations in Mexico over 33 years. The findings include: (i) Missing data and outliers in ground-based observations increased over time; (ii) Ground-station territorial coverage gradually reduced; (iii) Differences between ground-based and reanalysis observations were 1.95 and 1.91 Celsius degrees for maximum and minimum yearly temperatures; (iv) Kawak replaces inactive ground stations with virtual stations based on MERRA data, ensuring complete territorial coverage. The implementation of this GIS-bigdata fusion model is a viable solution for achieving whole coverage and supporting decision-making processes based on temperature territory observations, such as defining plans for allocating or replacing ground stations, climate data quality assessment, refining climate models and enhancing the accuracy and reliability of temperature predictions, ultimately supporting effective climate monitoring, adaptation, and mitigation strategies.

Comparative Study of Aerosol Optical Properties along a Megacity

Interest in the study and characterization of aerosols has been growing given its multiple connections with air quality and climate. However, the limited ground-based stations for measurements in many countries are not enough to completely study a big city or a region. Low-cost technology arises as a viable alternative to improve the spatial resolution of measurements. This study aims to demonstrate that the properties of aerosols may change within a megacity and that is advisable to segment measurements according to their characteristics. We measured the optical properties of aerosols in the outskirts of Buenos Aires and compared the data obtained with AERONET data from the city center. This is the first study of its kind in the region. The comparison of concentration, sizes, and types of aerosols shows differences between the sites regardless of the background levels that they share regionally. While the city center has a strong influence from traffic and sea particles and levels of pollution typical of a dense city, the outskirts present several aerosol sources and characteristics of a semi-rural site with a moderate influence of anthropogenic sources.

UAV-assisted fair communications for multi-pair users: A multi-agent deep reinforcement learning method

Unmanned aerial vehicle (UAV) plays an important role in scenarios like search and rescue, remote communication relay, battlefield mobile networks, etc. In this paper, we investigate multiple UAV relays providing real-time transmission for multi-pair ground users (GUs) in the absence of ground-based stations (GBSs). Due to the limited load capacities and energy resources on-board, fairness between multi-pair users, throughput maximization, as well as the connectivity maintenance between UAVs are jointly considered. We formulate the energy-efficient fair throughput objective function, which turns to be non-convex with hybrid variables. To solve this intractable problem, we utilize the power of neural networks in function approximation and propose a multi-agent deep reinforcement learning (MADRL)-based algorithm. Different from traditional MADRL algorithms, we utilize the method, named independent proximal policy optimization (IPPO), which allows agents to update according to their own observations memory and encourages more explorations to some extend, as the base of our solution. Simulation results demonstrate that our algorithm outperforms some baselines in terms of fairness, throughput as well as energy consumption.

Exposure to particulate matter and ozone, locations of regulatory monitors, and sociodemographic disparities in the city of Rio de Janeiro: Based on local air pollution estimates generated from machine learning models

South America is underrepresented in research on air pollution exposure disparities by sociodemographic factors, although such disparities have been observed in other parts of the world. We investigated whether exposure to and information about air pollution differs by sociodemographic factors in the city of Rio de Janeiro, the second most populous city in Brazil with dense urban areas, for 2012–2017. We developed machine learning-based models to estimate daily levels of O3, PM10, and PM2.5 using high-dimensional datasets from satellite remote sensing, atmospheric and land variables, and land use information. Cross-validations demonstrated good agreement between the estimated levels and measurements from ground-based monitoring stations: overall R2 of 76.8 %, 63.9 %, and 69.1 % for O3, PM2.5, and PM10, respectively. We conducted univariate regression analyses to investigate whether long-term exposure to O3, PM2.5, PM10 and distance to regulatory monitors differs by socioeconomic indicators, the percentages of residents who were children (0-17 years) or age 65+ years in 154 neighborhoods. We also examined the number of days exceeding the Brazilian National Air Quality Standard (BNAQS). Long-term exposures to O3 and PM2.5 were higher in more socially deprived neighborhoods. An interquartile range (IQR) increment of the social development index (SDI) was associated with a 3.6 μg/m3 (95 % confidence interval [CI]: 2.9, 4.4; p-value≤0.001) decrease in O3, and 0.3 μg/m3 (95 % CI: 0.2, 0.5; p-value = 0.010) decrease in PM2.5. An IQR increase in the percentage of residents who are children was associated with a 4.1 μg/m3 (95 % CI: 3.1, 5.0; p-value≤0.001) increase in O3, and 0.4 μg/m3 (95 % CI: 0.3, 0.6; p-value = 0.009) increase in PM2.5. An IQR increase in the percentage of residents age ≥65was associated with a 3.3 μg/m3 (95 % CI: 2.4, 4.3; p-value=<0.001) decrease in O3, and 0.3 μg/m3 (95 % CI: 0.1, 0.5; p-value = 0.058) decrease in PM2.5. There were no apparent associations for PM10. The association for daily O3 levels exceeding the BNAQS daily standard was 0.4 %p–0.8 %p different by the IQR of variables, indicating a 7–15 days difference in the six-year period. The association for daily PM2.5 levels exceeding the BNAQS daily standard showed a 0.7–1.5 %p difference by the IQR, meaning a 13–27 days difference in the period. We did not find statistically significant associations between the distance to monitors and neighborhood characteristics but some indication regarding SDI. We found that O3 levels were higher in neighborhoods situated farther from monitoring stations, suggesting that elevated levels of air pollution may not be routinely measured. Exposure disparity patterns may vary by pollutants, suggesting a complex interplay between environmental and socioeconomic factors in environmental justice.

A Deep-Learning and Transfer-Learning Hybrid Aerosol Retrieval Algorithm for FY4-AGRI: Development and Verification over Asia

The Advanced Geosynchronous Radiation Imager (AGRI) is a mission-critical instrument for the Fengyun series of satellites. AGRI acquires full-disk images every 15 min and views East Asia every 5 min through 14 spectral bands, enabling the detection of highly variable aerosol optical depth (AOD). Quantitative retrieval of AOD has hitherto been challenging, especially over land. In this study, an AOD retrieval algorithm is proposed that combines deep learning and transfer learning. The algorithm uses core concepts from both the Dark Target (DT) and Deep Blue (DB) algorithms to select features for the machine-learning (ML) algorithm, allowing for AOD retrieval at 550 nm over both dark and bright surfaces. The algorithm consists of two steps: ① A baseline deep neural network (DNN) with skip connections is developed using 10 min Advanced Himawari Imager (AHI) AODs as the target variable, and ② sunphotometer AODs from 89 ground-based stations are used to fine-tune the DNN parameters. Out-of-station validation shows that the retrieved AOD attains high accuracy, characterized by a coefficient of determination (R2) of 0.70, a mean bias error (MBE) of 0.03, and a percentage of data within the expected error (EE) of 70.7%. A sensitivity study reveals that the top-of-atmosphere reflectance at 650 and 470 nm, as well as the surface reflectance at 650 nm, are the two largest sources of uncertainty impacting the retrieval. In a case study of monitoring an extreme aerosol event, the AGRI AOD is found to be able to capture the detailed temporal evolution of the event. This work demonstrates the superiority of the transfer-learning technique in satellite AOD retrievals and the applicability of the retrieved AGRI AOD in monitoring extreme pollution events.

An early warning approach for the rapid identification of extreme weather disasters based on phased array dual polarization radar cooperative network data.

In recent years, X-band phase array dual polarization weather radar technology has matured. The cooperative networking data from X-band phase array dual polarization weather radar have many advantages compared with traditional methods, namely, high spatial and temporal resolution (approximately 70 seconds in one scan, 30 m in radial distance resolution), wide coverages that can compensate for the observation blind spots, and data fusion technology that is used in the observation overlap area to ensure that the observed precipitation data have spatial continuity. Based on the above radar systems, this study proposes an improved hail and lightning weather disaster rapid identification and early warning algorithm. The improved thunderstorm identification, tracking, analysis, and nowcasting (TITAN) algorithm is used to quickly identify three-dimensional strong convective storm cells. Large sample observation experiment data are used to invert the localized hail index (HDR) to identify the hail position. The fuzzy logic method is used to comprehensively determine the probability of lightning occurrence. The comparative analysis experiment shows that, compared with the live observation data from the ground-based automatic station, the hail and lightning disaster weather warning algorithm developed by this study can increase warning times by approximately 7 minutes over the traditional algorithm, and its critical success index (CSI), false alarm ratio (FAR) and omission alarm ratio (OAR) scores are better than those of the traditional method. The average root mean square error (ARMSE) for identifying hail and lightning locations by this improved method is also significantly better than that of traditional methods. We show that our method can provide probabilistic predictions that improve hail and lightning identification, improve the precision of early warning and support operational utility at higher resolutions and with greater lead times that traditional methods struggle to achieve.

Retrieving the 3D Tropospheric Wet Refractivity Field From a Standalone Ground-based GNSS Station with a priori Information: Theory and Simulation

Retrieving the 3D Tropospheric Wet Refractivity Field From a Standalone Ground-based GNSS Station with <i>a priori</i> Information: Theory and Simulation

Enhancing Precipitation Prediction in the Ziz Basin: A Comprehensive Review of Traditional and Machine Learning Approaches

Accurate precipitation forecasting is paramount for various sectors. Traditional methods for rainfall prediction involve understanding physical processes, historical weather data, and statistical models. These methods utilize observations from ground-based weather stations, satellites, and weather radars to assess current conditions and predict future precipitation. However, accurate precipitation prediction remains challenging due to the intricate and non-linear characteristics of rainfall. Over the past few years, machine learning (ML) algorithms have shown promise in improving precipitation prediction accuracy. This research provides an overview of both traditional methods and advanced ML models applicable to rainfall prediction, including regression, classification, and time series models. We conducted a comprehensive review of related works that explore the impact of using ML algorithms for rainfall estimation. Through this analysis, we identified the strengths and limitations of ML models in this context and highlighted advancements in rainfall prediction using these algorithms. We possess a comprehensive dataset, spanning data from 1996 to 2015, comprising historical weather data from the Ziz basin, our designated study area. This dataset contains five key meteorological features: precipitation, humidity, wind, temperature, and evaporation. In terms of perspective, we plan to utilize this dataset and conduct a comprehensive comparative study to evaluate the performance of different ML models. Our objective is to demonstrate the effectiveness and potential of these algorithms in improving weather forecasting capabilities and enhancing the accuracy of rainfall estimation methods in the specific study area.

Evaluating the freeze-thaw vulnerability of soapstone monuments and geoheritage sites in the Parco del Paradiso (Chiavenna, Italy).

This study investigates the vulnerability of soapstone monuments and ancient quarries in the Parco del Paradiso, Chiavenna, to freeze-thaw cycles. It employs climate-based damage functions to assess the risk of weathering-induced damage. A crucial aspect of the assessment involves the study of the homogeneity and trends in the selected climate data sources in the area of the case study. These data sources are ground-based weather stations, and in-situ installed temperature-humidity data loggers. The findings of this contribution underscore the significance of appropriately chosen reference climatic series and validation tests in ensuring the accuracy of the climate signal to analyze freeze-thaw statistics for specific cultural heritage sites. These insights are essential not only for supporting reliable assessment of the vulnerability of these invaluable heritage sites, which have played a vital role in local architecture and trade since Roman times, and for the successive development of long-term preventive conservation and maintenance strategies.

Development of a project to create a road transport infrastructure using small unmanned aircraft

The infrastructure for small unmanned aircraft has not yet been formed: there are no organized ground-based stations, there are no clear operating regulations for performing specialized types of work. It is proposed that when creating projects for an updated road transport infrastructure using ground-based autonomous vehicles, take into account the ability to maintain and maintain the operability of unmanned aircraft systems. Unmanned aerial vehicles can be used not only when conducting an inspection of the accident site, but also when checking on-site readings, investigative experiments, determining visibility on the road, etc. In the course of these investigative and procedural actions. The use of traffic control methods using small unmanned aircraft will make a significant contribution to ensuring road safety and reducing the level of accidents on highways. It is proposed that when creating projects for an updated road transport infrastructure using ground-based autonomous vehicles, take into account the ability to maintain and maintain the operability of unmanned aircraft systems.

ELSEM-Net, a network of ground-based telemetric stations for the monitoring of fracture-induced electromagnetic emissions in Greece: Instrumentation, management and analysis of recent observations associated with strong earthquakes

The hELlenic Seismo-ElectroMagnetics Network (ELSEM-Net, http://elsem-net.uniwa.gr) is a network of telemetric stations spanning all over Greece for the monitoring of fracture-induced electromagnetic emissions (FEME), often also referred to as electromagnetic radiation (FEMR). At the laboratory scale, it has been proved that FEME/FEMR are produced by opening cracks in a wide frequency spectrum ranging from the MHz to the kHz bands as the general (macroscopic) fracture is approaching, rendering them precursors of general fracture that permit the monitoring of the gradual damage of stressed materials. FEME/FEMR is also observed at the geophysical scale, preceding strong earthquakes (EQs) with epicenters on land or near the coast-line. Based on the idea that MHz-kHz FEME/FEMR should also permit the monitoring of the gradual damage of stressed materials in the Earth’s crust, as it happens in the laboratory experiments, ELSEM-Net was gradually developed from 1992 to 1998 and currently comprises 11 stations installed in various locations in Greece. Here we present the instrumentation that has been specifically designed for ELSEM-Net, both hardware and firmware/software, as well as a web-based system for the management of the field stations, the recorded data, and the automated preliminary analysis results. Finally, we present the analysis of selected observations associated with recent strong earthquakes that hit Greece, using a most recently introduced time series analysis method. Our presentation aims to communicate in detail the experimental infrastructure behind our almost 30 years of research in geophysical scale FEME/FEMR so that other interested research groups around the world can take advantage of it and have the opportunity to install similar stations/networks in other locations of the world for the study of pre-EQ processes associated with natural or man-induced seismicity.

Validation of precipitation obtained from GPM and H-SAF satellite data with relation to Ukrainian ground weather stations' data

The main objective of this study is to validate H-SAF and GPM satellite precipitation data for the territory of Ukraine. Validation was carried out by comparing satellite data with data from ground weather stations of Ukrainian Hydrometeorological Center and included the following stages: data collection and processing, bringing the data to the same spatial and temporal resolution, calculation of correlation coefficients and construction of confusion matrices. Correlation coefficients were calculated for each of the stations for two studied periods (the first period - April-September 2020, the second period - April-September 2021). In addition, to ensure a more detailed analysis, confusion matrices were constructed based on division of precipitation into classes of different intensity. The comparative analysis showed that a larger number of stations (117) has a correlation coefficient of 0.5 - 0.8 with relation to GPM data, and, as for HSAF precipitation data, a larger number of stations (59) has a low correlation coefficient (0 - 0.3). Confusion matrices were calculated for the following four classes of rainfalls (mm): “0-2”, “2-5”, “5-10”, “&gt;10”. The confusion matrices coefficients showed that low-intensity precipitation or “no precipitation”, as well as high-intensity precipitation, can be determined by a satellite with high accuracy as opposed to the ground-based weather station measurements. Thus, according to the "specificity" indicator, a precipitation class &gt;10 mm has the highest level of correspondence of satellite data (H-SAF, GPM Late, GPM Early) to ground data. The low value of the "specificity" for the range of 0-2 mm is explained by the fact that satellite methods are able to detect very low values of precipitation intensity, while the station shows their absence – 0 mm. Thus, according to the values of the confusion matrices, we see that a large part of the values of ground measurements of precipitation in the range of 0-2 mm are shown by the satellite method as “2-5 mm” class. We can also see that the “2-5 mm” class by satellite has the most confusion with the 0-2 mm class by station (for H-SAF, GPM Late, GPM Early). But based on the high values of the "recall" for the range of 0-2 mm (for H-SAF, GPM Late, GPM Early), we can conclude that when the station does not show this amount of precipitation (0-2 mm), then the satellite also shows another class of values. The values of "precision" coefficient show that the satellite measurements are actually reliable. This means that presence or absence of precipitation may be determined precisely, since the highest value of the "precision" coefficient is typical for low and high amounts of precipitation (on average 0.76 and 0.66, respectively). Detected by satellites, heavy precipitation or no precipitation are also confirmed in most cases by ground stations. Based on the results of the validation, GPM and H-SAF satellite precipitation data products can be used for rainfall monitoring and considered as an additional source of meteorological information.

Refractivity Observations from Radar Phase Measurements: The 22 May 2002 Dryline Case during IHOP Project

The dryline, often associated with the development of severe storms in the Southern Great Plains of the United States of America, is a boundary layer phenomenon that occurs when a warm and moist air mass from the Gulf of Mexico meets a hot and dry air mass from the southwest desert area. An accurate knowledge of the water vapor spatio-temporal variability in the lower part of the atmosphere is crucial for a better understanding of the evolution of the dryline. The tropospheric refractivity, directly related to water vapor content, is a proxy for the water vapor content of the troposphere. It has already been demonstrated that the refractivity and the refractivity vertical gradient can be jointly estimated from radar phase measurements. In fact, it has been shown that using kriging interpolation techniques, accurate refractivity maps within the coverage area of the radar can be obtained with high temporal resolution. In this paper, a detailed analysis of the time series of radar-based refractivity maps obtained during a dryline that occurred on the afternoon of 22 May 2002 during the International H2O Project (IHOP_2002) is presented. Comparisons between the time series of radar refractivity maps, obtained with the NCAR S-Pol radar, and the refractivity measurements derived from automatic ground-based weather stations and the AERI instrument, placed at different locations within the coverage area of the NCAR S-Pol radar, demonstrate the accuracy of radar refractivity estimates even for highly variable conditions, both in time and space, in the troposphere. Correlation coefficients higher than 0.95 are obtained in all weather station locations. Regarding the RMSE, errors less than 6 N-units are obtained for all cases, being even as low as 2.92 N-units at some locations.

High-Spatiotemporal-Resolution Estimation of Ground-Level Ozone in China Based on Machine Learning

High concentrations of ground-level ozone (O3) pose a significant threat to human health. Obtaining high-spatiotemporal-resolution information about ground-level O3 is of paramount importance for O3 pollution control. However, the current monitoring methods have a lot of limitations. Ground-based monitoring falls short in providing extensive coverage, and remote sensing based on satellites is constrained by specific spectral bands, lacking sensitivity to ground-level O3. To address this issue, we combined brightness temperature data from the Himawari-8 satellite with meteorological data and ground-based station data to train four machine learning models to obtain high-spatiotemporal-resolution information about ground-level O3, including Categorical Boosting (CatBoost), eXtreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine (LGBM), and Random Forest (RF). Among these, the CatBoost model exhibited superior performance, achieving a ten-fold cross-validation R2 of 0.8534, an RMSE of 17.735 μg/m3, and an MAE of 12.6594 μg/m3. Furthermore, all the selected feature variables in our study positively influenced the model. Subsequently, we employed the CatBoost model to estimate averaged hourly ground-level O3 concentrations at a 2 km resolution. The estimation results indicate a close relationship between ground-level O3 concentrations and human activities and solar radiation.

Метод определения параметров движения космических аппаратов на основе радиоинтерферометрических измерений со сверхдлинной базой

The proposed method for measuring current navigational parameters of spacecrafts makes it possible to control them to a certain extent using a single ground-based automated station. With this the following specifics of a single-station control are taken into account: location of spacecraft spatial-time monitoring stations, geometrical interpretation of these measurements, difficulties in processing of the received information and proper selection of radio equipment required for performing such measurements. The proposed method for trajectory tracking is based on the ultra-long-baseline interferometry. Measuring base here is the distance between a ground-based radio complex and onboard radio complex of the reference spacecraft that is constantly visible by the ground-based monitoring station. Onboard radio complex can include an array of spacecrafts traveling either along elongated elliptical orbit with an apogee height of more than twenty thousand kilometers or along a 36000 km high geostationary orbit. Application of the proposed method will help extend capabilities of the space monitoring system in terms of prompt clarification of public and private catalogues of spacecrafts orbiting the Earth.

Characteristics of cloud-to-ground lightning (CG) and differences between +CG and −CG strokes in China regarding the China National Lightning Detection Network

Abstract. A lightning location system consisting of multiple ground-based stations is an effective means of lightning observation. The dataset from CNLDN (China National Lightning Detection Network) in 2016–2022 is employed to analyze the temporal and spatial lightning distributions and the differences between +CG (positive cloud-to-ground lightning) and −CG (negative cloud-to-ground lightning) strokes in China. On the annual scale, lightning activity is most prevalent during the summer months (June, July, and August), accounting for 72.6 % of the year. Spring sees more lightning than autumn, and winter has only a small amount in southeastern coastal areas. During the day, the frequency of lightning peaks at 15:00–17:00 CST (China standard time) and is lowest at 8:00–10:00 CST. For the period with high CG stroke frequency (summer of a year or afternoon of a day), the proportion of +CG strokes and the discharge peak current are relatively small. Winter in a year and morning or midnight in a day correspond to a greater +CG stroke proportion and discharge current. Spatially, low latitudes, undulating terrain, the seaside, and humid surfaces are favorable factors for lightning occurrence. Thus, the southeast coastland has the largest lightning stroke density, while the northwest deserts and basins and the western and northern Tibetan Plateau, with altitudes over 6000 m, have almost no lightning. The proportion of +CG strokes and the peak current are low in the southern region with high density but diverse in other regions. The Tibetan Plateau causes the diversity of lightning activity in China and lays the foundation for studying the impact of surface elevation on lightning. Results indicate that the +CG stroke proportion on the eastern and southern Tibetan Plateau is up to 15 %, larger than the plain regions. The peak current of −CG strokes is positively correlated with altitude, but +CG strokes show a negative correlation, resulting in a large difference in peak current between +CG and −CG on the plain and a small difference on the plateau.

Remote sensing retrieval of aerosol types in China using geostationary satellite

Aerosol types have crucial and distinguished impacts on climate effects and cause the difference of air pollution. However, direct retrieval of aerosol type by satellite remote sensing is difficult, and the satellite products of aerosol type are seriously lacking. At present, the identification of aerosol type is mainly based on ground-based observation data, but the coverage area of ground-based stations is limited, and large-scale spatial continuous monitoring cannot be carried out. The existing aerosol type identification methods based on satellite data need to combine multiple satellites data (e.g. Moderate-resolution Imaging Spectroradiometer (MODIS), Ozone Monitoring Instrument (OMI), etc.) which leads to a low temporal resolution. The research about aerosol type identification only based on geostationary satellite data is lacking. Therefore, aerosol type identification based on the multi-parameter threshold method and the neural network method are proposed and compared in this paper. The neural network jointly using spectral and temporal information was trained and tested by the measurements from 26 sites of Aerosol Robotic Network (AERONET) and Sun-Sky Radiometer Observation Network (SONET). For the overall validation, the aerosol type identification based on the neural network is better than that based on the multi-parameter threshold method, and the accuracy reaches 71.44%. Atmospheric environment monitoring and quantitative remote sensing can be supported by this study.

Machine learning-based global air quality index development using remote sensing and ground-based stations

Air pollution refers to the presence of hazardous substances in the air that has adverse effects on health, causing millions premature deaths annually. Ground-based stations can provide accurate measurements for monitoring air pollution. However, the spatial coverage of air pollution measurements is limited by the number of measurement instruments available in specific hotspot areas. Satellite remote sensing can reduce spatial uncertainty; however, the measurement results are mostly limited to the upper atmosphere with high measurement sensitivity. To better represent surface conditions, this study aims to model the Air Quality Index for pollutants CO, NO2, SO2, PM2.5, and PM10 in the global region using remotely sensed data. To support this study, 425 data points from air pollution stations distributed globally are combined using Machine Learning Linear Regression methods. Furthermore, socioeconomic and environmental data are combined with air pollution satellite to form Multiple models. According to the results of this study, the Multiple Models are more accurate than the single models, showing that the addition of socioeconomic and environmental data can enhance accuracy. The results of this study are expected to help regions without air pollution monitoring stations to estimate the air quality index using satellite data, in turn, preventing air pollution disasters.

Interpolation, Satellite-Based Machine Learning, or Meteorological Simulation? A Comparison Analysis for Spatio-temporal Mapping of Mesoscale Urban Air Temperature

AbstractFine-resolution spatio-temporal maps of near-surface urban air temperature (Ta) provide crucial data inputs for sustainable urban decision-making, personal heat exposure, and climate-relevant epidemiological studies. The recent availability of IoT weather station data allows for high-resolution urban Ta mapping using approaches such as interpolation techniques or machine learning (ML). This study is aimed at executing these approaches and traditional numerical modeling within a practical and operational framework and evaluate their practicality and efficiency in cases where data availability, computational constraints, or specialized expertise pose challenges. We employ Netatmo crowd-sourced weather station data and three geospatial mapping approaches: (1) Ordinary Kriging, (2) statistical ML model (using predictors primarily derived from Earth Observation Data), and (3) weather research and forecasting model (WRF) to predict/map daily Ta at nearly 1-km spatial resolution in Warsaw (Poland) for June–September and compare the predictions against observations from 5 meteorological reference stations. The results reveal that ML can serve as a viable alternative approach to traditional kriging and numerical simulation, characterized by reduced complexity and higher computational speeds within the domain of urban meteorological studies (overall RMSE = 1.06 °C and R2 = 0.94, compared to ground-based meteorological stations). The results have implications for identifying the urban regions vulnerable to overheating and evidence-based urban management in response to climate change. Due to the open-sourced nature of the applied predictors and input parsimony, the ML method can be easily replicated for other EU cities.