Towards interpretable machine learning for observational quantification of soil heavy metal concentrations under environmental constraints

Monitoring heavy metal concentrations in soils is central to assessing agricultural production safety. Satellite observations permit inferring concentrations from spectrum, thereby contributing to the prevention and control of soil heavy metal pollution. However, heavy metals exhibit weak spectral responses, particularly at low and medium concentrations, and are predominantly influenced by other soil components. Machine learning (ML)-driven modelling can produce predictions but lacks interpretability. Here, we present an interpretable ML framework for concentration quantification modelling and investigated the contributions of spectral and environmental factors-pH and organic carbon-to the estimation of metals with multiple concentration gradients, as analysed through SHAP (SHapley Additive exPlanations) data derived from four learning-based scenarios. The results indicated that scenarios SHC (spectral, pH, and organic carbon) and SH (spectral and pH) were the most optimal for chromium (Cr) [RPD = 1.42, Adj R2 = 0.62], and cadmium (Cd) [RPD = 1.80, Adj R2 = 0.80]. Under environmental constraints, the spectral predictability for Cr and Cd was improved by 67 % and 87 %, respectively. We concluded that interpretable modelling, utilising both spectral and soil environmental factors, holds significant potential for estimating heavy metals across concentration gradients. It is recommended that samples with higher organic carbon content and lower pH be selected to enhance Cr and Cd predictions. An advanced grasp of interpretable predictions facilitates earlier warning of heavy metal contamination and guides the formulation of robust sampling strategies.

Geometric working volume of a satellite positive displacement machine

This article describes a method for determining the geometric working volume of satellite positive displacement machines (pump and motor). The working mechanism of these machines is satellite mechanism consisting of two non-circular gears (rotor and curvature) and circular gears (satellites). Two variants of the satellite mechanism are presented. In the first mechanism, the rolling line of the rotor is a sinusoid "wrapped" around a circle. In the second mechanism, the rolling line of the rotor is a double sinusoid "wrapped" around a circle. A method for calculating the area of the working chamber as a function of the rotor rotation angle is presented, based on mathematical formulae of the rotor, the curvature and the satellite rolling lines. It has been shown that the second variant of the satellite mechanism is advantageously characterised by a larger difference between the maximum area of the working chamber and the minimum area of this chamber. New mathematical formulas have been proposed to calculate the area of the working chamber for any angle of rotation of the shaft (rotor) based on the maximum and minimum values of the area of this chamber. It was thus confirmed that the geometric working volume depends on the maximum and minimum area of a working chamber and on the height of the satellite mechanism. The analyses of the area of the working chamber were carried out both for the mechanism without gears (the area delimited by the rolling lines of the elements of the mechanism) and for the real mechanism with gears. Differences in the values of these fields were also detected.

Upscaling vascular aboveground biomass and topsoil moisture of subarctic fens from Unoccupied Aerial Vehicles (UAVs) to satellite level

Arctic and subarctic ecosystems are experiencing rapid changes in vegetation composition and productivity due to global warming. Tundra wetlands are especially susceptible to these changes, which may trigger shifts in soil moisture dynamics. It is therefore essential to accurately map plant biomass and topsoil moisture. In this study, we mapped total, wood, and leaf above ground biomass and topsoil moisture in subarctic tundra wetlands located between Norway and Finland by linking models derived from Unoccupied Aerial Vehicles with multiple satellite data sources using the Extreme Gradient Boosting algorithm. The most accurate predictions for topsoil moisture (R2 = 0.73) used a set of red edge-based vegetation indices with a spatial resolution of 20 m per pixel. On the contrary, wood biomass showed the lowest accuracies across all tested models (R2 = 0.38). We found a trade-off between the spatial resolution and the performance of upscaling models, where smaller pixel sizes generally led to lower accuracies. However, we were able to compensate for reduced accuracy at smaller pixel sizes using Copernicus phenology metrics. A modelling uncertainty assessment revealed that the uncertainty of predictions increased with decreasing pixel sizes and increasing number of co-predictors. Our approach could improve efforts to map and monitor changes in vegetation at regional to pan-Arctic scales.

Influence of operating pressure on the durability of a satellite hydraulic motor supplied by rapeseed oil

This article describes the results of a durability test of a hydraulic satellite motor supplied by rapeseed oil. The tests were carried out on a test stand in a power recuperation system. The tests of the motor were carried out at a constant shaft speed for three fixed pressure drops in the motor. This made it possible to demonstrate the influence of the motor operating pressure on the durability of the satellite mechanism. The influence of the pressure drop in the motor and the influence of the operating time on the motor absorbency, on the torque on the motor shaft and the influence on the volumetric and hydraulic-mechanical efficiency are also shown. The basic relationship between the efficiency of the motor and the temperature rise in the motor is also described. The results of the calculations of the temperature rise in the motor are compared with the experimental results. The article also shows which components of the motor's working mechanism wear out the fastest. The cause of the wear and failure is also explained.

Unveiling global land fine- and coarse-mode aerosol dynamics from 2005 to 2020 using enhanced satellite-based monthly inversion data

Accurate fine-mode and coarse-mode aerosol knowledge is crucial for understanding their impacts on the climate and Earth's ecosystems. However, current satellite-based Fine-Mode Aerosol Optical Depth (FAOD) and Coarse-Mode Aerosol Optical Depth (CAOD) methods have drawbacks including inaccuracies, low spatial coverage, and limited temporal duration. To overcome these issues, we developed new global-scale FAOD and CAOD from 2005 to 2020 using a novel deep learning model capable of the synergistic retrieval of two aerosol sizes. After validation with the aerosol robotic network (AERONET) and sky radiometer network (SKYNET), the new monthly FAOD and CAOD showed significant improvements in accuracy and spatial coverage. From 2005 to 2020, the new data showed that China had the greatest decrease in FAOD and CAOD. In contrast, India and South Latin America had a significant increase in FAOD versus North Africa in CAOD. FAOD in the regions of Argentina, Paraguay, and Uruguay in South America have shown an upward trend. The results reveal that FAOD and CAOD display distinct patterns of change in the same regions, particularly on the west coast of the United States where FAOD is increasing, while CAOD is decreasing. Aside from the year 2020 due to the global COVID-19 pandemic, the analysis showed that although China has seen at least an +85% increase in energy consumption and urban expansion in 2019 compared to 2005 due to the needs of development and construction, the implementation of China's air pollution control policies has led to a significant decrease in FAOD (-46%) and CAOD (-65%) after 2013. This research enriches our comprehension of global fine and coarse aerosol patterns, additional investigations are needed to determine the potential global implications of these changes.

Assessment of machine learning algorithms and new hybrid multi-criteria analysis for flood hazard and mapping

Floods in Iran cause a lot of damage in different places every year. The 2019 floods of the Gorgan and Atrak rivers basins in the north of Iran were one of the most destructive events in this country. Therefore, investigating the flood hazard of these areas is very necessary to manage probable future floods. For this purpose, in the present study, the capability of Random Forest (RF) and Support Vector Machine (SVM) algorithms was investigated in combination with Sentinel series and Landsat-8 images to prepare the 2019 flood map. Then, the flood hazard map of these areas was prepared using the new hybrid Fuzzy Best Worse Model-Weighted Multi-Criteria Analysis (FBWM-WMCA) model. According to the results of the FBWM-WMCA model, 38.58%, 50.18%, 11.10%, and 0.14% of the Gorgan river basin and 45.11%, 49.96%, 4.17%, and 0.076% of the Atrak river basin are in high, medium, low, and no hazards, respectively. The highest flood hazard areas in Gorgan and Atrak rivers basins in the north, northwest, west, and east, and south and southwest are mostly at medium flood hazard. Also, the results of RF and SVM algorithms with an overall accuracy of more than 85% for Sentinel-1, Sentinel-2, and Landsat-8 images and 80% for Sentinel-3 images indicate that the flooding is related to the western, southwestern, and northern regions including agricultural, bare lands and built up. According to the obtained results and the efficiency of the FBWM-WMCA model, the Gorgan and Atrak rivers basins need proper planning for flood hazard management.

Geospatial mapping of carbon estimates for forested areas using the InVEST model and Sentinel-2: A case study in Galicia (NW Spain)

CO2 emissions have increased exponentially in recent years, so measuring and quantifying carbon sequestration is a step towards sustainable forest management and combating climate change. The overall goal of this study is to develop an accurate model for estimating carbon storage and sequestration for forest areas of the Atlantic Biogeographic Region. Specifically, the modelling and field sampling are carried out in the municipality of Baiona (Galicia, NW Spain), which was selected as a representative biome of this region. The methodology consists of carrying out two object-based image analysis (OBIA) classifications in spring and autumn to observe possible stocks of seasonal differences. Two carbon storage and sequestration models are built up (model 1 and model 2): model 1 for forest areas only and model 2 including all other land cover in the study area. Sentinel-2 geospatial data for 2021, Integrated Valuation of Ecosystem Services and Tradeoffs (InVEST) tools and geographic information systems (GIS) are used. A Kappa index of 0.92 is obtained for both classifications, thus ruling out any notable seasonal differences in the images used. The results from both models indicate that it is land covers associated with forest uses which store the most carbon in the study area, accounting for >50 % more than the other land covers. It is concluded that the methodology and data used are very useful for quantifying ecosystem services, which will help the governance of the region by implementing measures to mitigate some of the effects of climate change and help to create silvicultural models for the sustainable management of the Atlantic Biogeographic Region.

Estimation of pan-European, daily total, fine-mode and coarse-mode Aerosol Optical Depth at 0.1° resolution to facilitate air quality assessments

Aerosol Optical Depth (AOD) data derived from satellites is crucial for estimating spatially-resolved PM concentrations, but existing AOD data over land remain affected by several limitations (e.g., data gaps, coarser resolution, higher uncertainty or lack of size fraction data), which weakens the AOD-PM relationship. We developed a 0.1° resolution daily AOD data set over Europe over the period 2003-2020, based on two-stage Quantile Machine Learning (QML) frameworks. Our approach first fills gaps in satellite AOD data and then constructs three components' models to obtain reliable full-coverage AOD along with Fine-mode AOD (fAOD) and Coarse-mode AOD (cAOD). These models are based on AERONET (AErosol RObotic NETwork) observations, Gap-filled satellite AOD, climate and atmospheric composition reanalyses. Our QML AOD products exhibit better quality with an out-of-sample R2 equal to 0.68 for AOD, 0.66 for fAOD and 0.65 for cAOD, which is 23-92 %, 11-13 % and 115-132 % higher than the corresponding satellite or reanalysis products, respectively. Over 91.6 %, 81.6 %, and 88.9 % of QML AOD, fAOD and cAOD predictions fall within ±20 % Expected Error (EE) envelopes, respectively. Previous studies reported that a weak satellite AOD-PM correlation across Europe (Pearson correlation coefficient (PCC) around 0.1). Our QML products exhibit higher correlations with ground-level PMs, particularly when broadly matched by size: AOD with PM10, fAOD with PM2.5, cAOD with PM coarse (R = 0.41, 0.45 and 0.26, respectively). Different AOD fractions more effectively distinct PM size fractions, than total AOD. Our QML aerosol dataset and models pioneer full-coverage, daily high-resolution monitoring of fine-mode and coarse-mode aerosols, effectively addressing existing AOD challenges for further PMs exposures' estimations. This dataset opens avenues for more in-depth exploration of the impacts of aerosols on human health, climate, visibility, and biogeochemical processes, offering valuable insights for air quality management and environmental health risk assessment.

Identifying the spatial pattern and the drivers of the decline in the eastern English Channel chlorophyll-a surface concentration over the last two decades

It has been established from previous studies that chlorophyll-a surface concentration has been declining in the eastern English Channel. This decline has been attributed to a decrease in nutrient concentrations in the rivers. However, the decrease in river discharge could also be a cause. In our study, rivers outflows and in-situ data have been compared to time series of satellite-derived chlorophyll-a concentrations. Dynamic Linear Model has been used to extract the dynamic and seasonally adjusted trends of several environmental variables. The results showed that, for the 1998-2019 period, chlorophyll-a levels stayed significantly lower than average and satellite images revealed a coast to offshore gradient. Chlorophyll-a concentration of coastal stations appeared to be related to the declining fluxes of phosphate while offshore stations were more related to nitrate-nitrite. Therefore, we can exclude that the climate variability, through river flows alone, has a dominant effect on the decline of chlorophyll-a concentration.

Environmental impacts of earth observation data in the constellation and cloud computing era

Numbers of Earth Observation (EO) satellites have increased exponentially over the past decade reaching the current population of 1193 (January 2023). Consequently, EO data volumes have mushroomed and data storage and processing have migrated to the cloud. Whilst attention has been given to the launch and in-orbit environmental impacts of satellites, EO data environmental footprints have been overlooked. These issues require urgent attention given data centre water and energy consumption, high carbon emissions for computer component manufacture, and difficulty of recycling computer components. Doing so is essential if the environmental good of EO is to withstand scrutiny. We provide the first assessment of the EO data life-cycle and estimate that the current size of the global EO data collection is ~807 PB, increasing by ~100 PB/year. Storage of this data volume generates annual CO2 equivalent emissions of 4101 t. Major state-funded EO providers use 57 of their own data centres globally, and a further 178 private cloud services, with considerable duplication of datasets across repositories. We explore scenarios for the environmental cost of performing EO functions on the cloud compared to desktop machines. A simple band arithmetic function applied to a Landsat 9 scene using Google Earth Engine (GEE) generated CO2 equivalent (e) emissions of 0.042-0.69 g CO2e (locally) and 0.13-0.45 g CO2e (European data centre; values multiply by nine for Australian data centre). Computation-based emissions scale rapidly for more intense processes and when testing code. When using cloud services such as GEE, users have no choice about the data centre used and we push for EO providers to be more transparent about the location-specific impacts of EO work, and to provide tools for measuring the environmental cost of cloud computation. The EO community as a whole needs to critically consider the broad suite of EO data life-cycle impacts.

A novel chlorophyll-a retrieval model based on suspended particulate matter classification and different machine learning

Chlorophyll-a (Chla) in inland waters is one of the most significant optical parameters of aquatic ecosystem assessment, and long-term and daily Chla concentration monitoring has the potential to facilitate in early warning of algal blooms. MOD09 products have multiple observation advantages (higher temporal, spatial resolution and signal-to-noise ratio), and play an extremely important role in the remote sensing of water color. For developing a high accuracy machine learning model of remotely estimating Chla concentration in inland waters based on MOD09 products, this study proposed an assumption that the accuracy of Chla concentration retrieval will be improved after classifying water bodies into three groups by suspended particulate matter (SPM) concentration. A total of 10 commonly used machine learning models were compared and evaluated in this study, including random forest regressor (RFR), deep neural networks (DNN), extreme gradient boosting (XGBoost), and convolutional neural network (CNN). Altogether, 41 basic bands and 820 band ratios between the 41 bands were filtered by measuring their correlation with Ln(Chla) and several bands brought into different machine learning models. Results demonstrated that the construction of Chla concentration remote estimation model based on SPM classification could significantly improve the correlation between Ln(Chla) and 41 basic spectral band combinations, the correlation between Ln(Chla) and 820 band ratios, and the model verification R2 from 0.41 to 0.83. Furthermore, B3, B20, and B32 were finally selected based on correlation with SPM to classify SPM and the classification accuracy could reach 0.9. Finally, we concluded that RFR model performed best via comparing the R2, RMSE, and MAPE. By comparing the relative contribution of input bands in different groups, B3 contributed most to three groups. The model constructed in this study has promising prospects for promotion and application in other inland waters, and could provide systematic research reference for subsequent research.

A new approach to quantify chlorophyll-a over inland water targets based on multi-source remote sensing data

Chlorophyll-a (Chl-a) concentration is a reliable indicator of phytoplankton biomass and eutrophication, especially in inland waters. Remote sensing provides a means for large-scale Chl-a estimation by linking the spectral water-leaving signal from the water surface with in situ measured Chl-a concentrations. Single-sensor images cannot meet the practical needs for long-term monitoring of Chl-a concentrations due to cloud cover and satellite operational lifetimes. However, quantifying long-term inland water Chl-a concentrations using multi-source remote sensing data remains a problem, as improper input of satellite reflectance products will affect the accuracy of Chl-a over inland waters, as well as existing models cannot meet the need for multi-source remote sensing data to retrieve high precision Chl-a. To explore these problems towards a solution, four reflectance data derived from Ocean and Land Colour Instrument (OLCI), MultiSpectral Instrument (MSI), and Operational Land Imager (OLI) were evaluated against in situ measurements of Erhai Lake. Reflectance data from these sensors were assessed to determine their consistency. Results indicate that R_rhos products (i.e., surface reflectance, a semi-atmospheric correction reflectance) that controlled for the atmospheric diffuse transmittance were highly correlated with the measured reflectance values. The in situ reflectance also confirmed the higher fidelity of satellite reflectance in the green-red band. Subsequently, a new extreme gradient boosting (XGB) model applied to multi-source remote sensing data is proposed to estimate long-term inland water Chl-a concentrations. Comparative experiments showed the XGB model with R_rhos products outperformed other solutions, providing accurate estimates for daily, monthly, and long-term trends in Erhai Lake. The XGB model was finally processed 3954 R_rhos reflectance data derived from OLCI, ENVISAT Medium Resolution Imaging Spectrometer (MERIS), MSI, and OLI sensors, mapping Chl-a concentrations in Erhai Lake over a 20-year period. This study could serve as a reference for the long-term Chl-a monitoring using multi-source remote sensing data to support inland lake management and future water quality evaluation.

Recent changes in cyanobacteria algal bloom magnitude in large lakes across the contiguous United States

Cyanobacterial blooms in inland lakes produce large quantities of biomass that impact drinking water systems, recreation, and tourism and may produce toxins that can adversely affect public health. This study analyzed nine years of satellite-derived bloom records and compared how the bloom magnitude has changed from 2008-2011 to 2016-2020 in 1881 of the largest lakes across the contiguous United States (CONUS). We determined bloom magnitude each year as the spatio-temporal mean cyanobacteria biomass from May to October and in concentrations of chlorophyll-a. We found that bloom magnitude decreased in 465 (25 %) lakes in the 2016-2020 period. Conversely, there was an increase in bloom magnitude in only 81 lakes (4 %). Bloom magnitude either didn't change, or the observed change was in the uncertainty range in the majority of the lakes (n = 1335, 71 %). Above-normal wetness and normal or below-normal maximum temperature over the warm season may have caused the decrease in bloom magnitude in the eastern part of the CONUS in recent years. On the other hand, a hotter and dryer warm season in the western CONUS may have created an environment for increased algal biomass. While more lakes saw a decrease in bloom magnitude, the pattern was not monotonic over the CONUS. The variations in temporal changes in bloom magnitude within and across climatic regions depend on the interactions between land use land cover (LULC) and physical factors such as temperature and precipitation. Despite expectations suggested by recent global studies, bloom magnitude has not increased in larger US lakes over this time period.

Unbalanced impacts of nature and nurture factors on the phenology, area and intensity of algal blooms in global large lakes: MODIS observations

Under the influence of climate warming and human activities, many large lakes have experienced an increase in eutrophication and algal blooms. Although these trends have been identified using low temporal resolution (~16 days) satellites such as those of the Landsat missions, the opportunity to compare high-frequency spatiotemporal variations of algal bloom characteristics between lakes has not been explored. In the present study, we explore daily satellite observations by developing a universal, practical, and robust algorithm to identify the spatiotemporal distribution of algal bloom dynamics in large lakes (>500 km²) across the globe. Data from 161 lakes, taken from 2000 to 2020 showed an average accuracy of 79.9 %. Algal blooms were detected in 44 % of all lakes, with a higher incidence in temperate lakes (67 % of all temperate lakes), followed by tropical lakes (59 %) compared to lakes in arid climates (23 %). We found positive trends in bloom area and frequency (p < 0.05), as well as an earlier bloom time (p < 0.05). Climate factors were found to be linked to changes in annual initial bloom time (44 %); while an increase in human activities was associated to bloom duration (49 %), area (max percent: 53 %, mean percent: 45 %), and frequency (46 %). The study shows the evolution of daily algal blooms and their phenology in global large lakes for the first time. Such information enhances our understanding of algal bloom dynamics and their drivers, with important considerations to improve the management of large lake ecosystems.

Using climate reanalysis and remote sensing-derived data to create the basis for predicting the occurrence of algal blooms, harmful algal blooms and toxic events in Santa Catarina, Brazil

This study aimed to form a basis for future predictive modeling efforts in support of the harmful algal blooms (HAB) surveillance program currently in force in the Brazilian State of Santa Catarina (SC). Data from monitoring toxin-producing algae were merged with both meteorological and oceanographic data and analyzed. Data from four sources were used in this study: climate reanalysis (air temperature, pressure, cloud cover, precipitation, radiation, U and V winds); remote sensing (chlorophyll concentration and sea surface temperature); Oceanic Niño Index; and HAB monitoring data (phytoplankton counts and toxin levels in shellfish samples obtained from 39 points located in shellfish farms distributed along the SC coastline). This study analyzed the period from 2007-01-01 to 2019-12-31 (7035 records in the HAB database) and used descriptive, bivariate and multivariate analyses to draw correlations among environmental parameters and the occurrence of algal blooms (AB), HAB and toxic events. Dinophysis spp. AB were the most registered type of event and tended to occur during the late autumn and winter months. These events were associated with high atmospheric pressure, predominance of westerly and southerly winds, low solar radiation and low sea and air temperature. An inverted pattern was observed for Pseudo-nitzschia spp. AB, which were mostly registered during the summer and early autumn months. These results give evidence that the patterns of occurrence of highly prevalent toxin-producing microalgae reported worldwide, such as the Dinophysis AB during the summer, differ along the coast of SC. Our findings also show that meteorological data, such as wind direction and speed, atmospheric pressure, solar radiation and air temperature, might all be key predictive modeling input parameters, whereas remote sensing estimates of chlorophyll, which are currently used as a proxy for the occurrence of AB, seem to be a poor predictor of HAB in this geographic area.

"Satellite pedicle screws" - A novel technique of pedicle screw insertion in obese patients undergoing lumbar fusion

The presence of thick sub-cutaneous fat and bulky paraspinal musculature mandates extensive surgical dissection in obese patients undergoing open Transforaminal lumbar interbody fusion surgery. Securing a 'converging' pedicle screw trajectory becomes difficult by the counterforces of the erector spinae muscles and thick sub-cutaneous fat in obese patients, especially at the L5-S1 level. We describe the use of a limited standard posterior midline exposure and a separate, far lateral 'satellite' incision to insert pedicle screws in an optimal trajectory in obese patients. Through proper pre-operative planning of the axial and sagittal MRI, the appropriate entry site is determined which is executed intra-operatively to insert pedicle screws freehand. Through a single 1.5 cm incision, both L5-S1 screws were inserted. Fourteen obese patients (mean BMI was 30.5 ± 1.1) received 56 satellite pedicle screws for TLIF at L5-S1 level. The mean age was 48.3 ± 9.7 years. The mean blood loss was 244.8 ± 114 ml and the mean operative time was 126.7 ± 82.8 min. In all patients, the screws were inserted as per pre-operative planning without any difficulties. All wounds healed well without wound complications. There were no screw related complications, and in the antero-posterior and lateral radiographs, there were no screw breaches. Satellite free-hand pedicle screws are safe and easily reproducible. They enable limited dissection of the main surgical wound and well-medialised converging pedicle screws in obese patients.

From UAV to PlanetScope: Upscaling fractional cover of an invasive species Rosa rugosa

Invasive plant species pose a direct threat to biodiversity and ecosystem services. Among these, Rosa rugosa has had a severe impact on Baltic coastal ecosystems in recent decades. Accurate mapping and monitoring tools are essential to quantify the location and spatial extent of invasive plant species to support eradication programs. In this paper we combined RGB images obtained using an Unoccupied Aerial Vehicle, with multispectral PlanetScope images to map the extent of R. rugosa at seven locations along the Estonian coastline. We used RGB-based vegetation indices and 3D canopy metrics in combination with a random forest algorithm to map R. rugosa thickets, obtaining high mapping accuracies (Sensitivity = 0.92, specificity = 0.96). We then used the R. rugosa presence/absence maps as a training dataset to predict the fractional cover based on multispectral vegetation indices derived from the PlanetScope constellation and an Extreme Gradient Boosting algorithm (XGBoost). The XGBoost algorithm yielded high fractional cover prediction accuracies (RMSE = 0.11, R² = 0.70). An in-depth accuracy assessment based on site-specific validations revealed notable differences in accuracy between study sites (highest R² = 0.74, lowest R² = 0.03). We attribute these differences to the various stages of R. rugosa invasion and the density of thickets. In conclusion, the combination of RGB UAV images and multispectral PlanetScope images is a cost-effective method to map R. rugosa in highly heterogeneous coastal ecosystems. We propose this approach as a valuable tool to extend the highly local geographical scope of UAV assessments into wider areas and regional evaluations.

Cooperative simultaneous inversion of satellite-based real-time PM2.5 and ozone levels using an improved deep learning model with attention mechanism

Ground-level fine particulate matter (PM_2.5) and ozone (O₃) are air pollutants that can pose severe health risks. Surface PM_2.5 and O₃ concentrations can be monitored from satellites, but most retrieval methods retrieve PM_2.5 or O₃ separately and disregard the shared information between the two air pollutants, for example due to common emission sources. Using surface observations across China spanning 2014-2021, we found a strong relationship between PM_2.5 and O₃ with distinct spatiotemporal characteristics. Thus, in this study, we propose a new deep learning model called the Simultaneous Ozone and PM_2.5 inversion deep neural Network (SOPiNet), which allows for daily real-time monitoring and full coverage of PM_2.5 and O₃ simultaneously at a spatial resolution of 5 km. SOPiNet employs the multi-head attention mechanism to better capture the temporal variations in PM_2.5 and O₃ based on previous days' conditions. Applying SOPiNet to MODIS data over China in 2022, using 2019-2021 to construct the network, we found that simultaneous retrievals of PM_2.5 and O₃ improved the performance compared with retrieving them independently: the temporal R² increased from 0.66 to 0.72 for PM_2.5, and from 0.79 to 0.82 for O₃. The results suggest that near-real time satellite-based air quality monitoring can be improved by simultaneous retrieval of different but related pollutants. The codes of SOPiNet and its user guide are freely available online at https://github.com/RegiusQuant/ESIDLM.

Short-term effects of individual exposure to PM2.5 on hospital admissions for myocardial infarction and stroke: a population-based case-crossover study in Guangzhou, China

Some studies have investigated the effects of PM_2.5 on cardiovascular diseases based on the population-average exposure data from several monitoring stations. No one has explored the short-term effect of PM_2.5 on cardiovascular hospitalizations using individual-level exposure data. We assessed the short-term effects of individual exposure to PM_2.5 on hospitalizations for myocardial infarction (MI) and stroke in Guangzhou, China, during 2014-2019. The population-based data on cardio-cerebrovascular events were provided by Guangzhou Center for Disease Control and Prevention. Average annual percent changes (AAPCs) were used to describe trends in the hospitalization rates of MI and stroke. The conditional logistic regression model with a time-stratified case-crossover design was applied to estimate the effects of satellite-retrieved PM_2.5 with 1-km resolution as individual-level exposure. Furthermore, we performed stratified analyses by demographic characteristics and season. There were 28,346 cases of MI, 188,611, and 36,850 cases of ischemic stroke (IS) and hemorrhagic stroke (HS), respectively, with an annual average hospitalization rate of 37.2, 247, and 48.4 per 100,000 people. Over the six-year study period, significant increasing trends in the hospitalization rates were observed with AAPCs of 12.3% (95% confidence interval [CI]: 7.24%, 17.6%), 13.1% (95% CI: 9.54%, 16.7%), and 9.57% (95% CI: 6.27%, 13.0%) for MI, IS, and HS, respectively. A 10 μg/m³ increase in PM_2.5 was associated with an increase of 1.15% (95% CI: 0.308%, 1.99%) in MI hospitalization and 1.29% (95% CI: 0.882%, 1.70%) in IS hospitalization. A PM_2.5-associated reduction of 1.17% (95% CI: 0.298%, 2.03%) was found for HS hospitalization. The impact of PM_2.5 was greater in males than in females for MI hospitalization, and greater effects were observed in the elderly (≥ 65 years) and in cold seasons for IS hospitalization. Our study added important evidence on the adverse effect of PM_2.5 based on satellite-retrieved individual-level exposure data.

Diverse begomovirus-betasatellite complexes cause tomato leaf curl disease in the western India

In the Indian sub-continent, tomato leaf curl disease (ToLCD) of tomato caused by begomoviruses has emerged as a major limiting factor for tomato cultivation. Despite the spread of this disease in the western India, a systematic study on the characterization of virus complexes with ToLCD is lacking. Here, we report the identification of a complex of begomoviruses including 19 DNA-A and 4 DNA-B as well as 15 betasatellites with ToLCD in the western part of the country. Additionally, a novel betasatellite and an alphasatellite were also identified. The recombination breakpoints were detected in the cloned begomoviruses and betasatellites. The cloned infectious DNA constructs cause disease on the tomato (a moderately virus-resistant cultivar) plants, thus fulfilling Koch's postulates for these virus complexes. Further, the role of non-cognate DNA B/betasatellite with ToLCD-associated begomoviruses on disease development was demonstrated. It also emphasizes the evolutionary potential of these virus complexes in breaking disease resistance and plausible expansion of its host range. This necessitates to investigate the mechanism of the interaction between resistance breaking virus complexes and the infected host.

Multilevel thresholding satellite image segmentation using chaotic coronavirus optimization algorithm with hybrid fitness function

Image segmentation is a critical step in digital image processing applications. One of the most preferred methods for image segmentation is multilevel thresholding, in which a set of threshold values is determined to divide an image into different classes. However, the computational complexity increases when the required thresholds are high. Therefore, this paper introduces a modified Coronavirus Optimization algorithm for image segmentation. In the proposed algorithm, the chaotic map concept is added to the initialization step of the naive algorithm to increase the diversity of solutions. A hybrid of the two commonly used methods, Otsu's and Kapur's entropy, is applied to form a new fitness function to determine the optimum threshold values. The proposed algorithm is evaluated using two different datasets, including six benchmarks and six satellite images. Various evaluation metrics are used to measure the quality of the segmented images using the proposed algorithm, such as mean square error, peak signal-to-noise ratio, Structural Similarity Index, Feature Similarity Index, and Normalized Correlation Coefficient. Additionally, the best fitness values are calculated to demonstrate the proposed method's ability to find the optimum solution. The obtained results are compared to eleven powerful and recent metaheuristics and prove the superiority of the proposed algorithm in the image segmentation problem.

Evaluation of the simulated aerosol optical properties over India: COALESCE model inter-comparison of three GCMs with ground and satellite observations

Within the framework of COALESCE project (Carbonaceous aerosol emissions, source apportionment, and climate impacts) initiative, spatio-temporal distribution of aerosol optical properties from three general circulation models are evaluated against aerosol data from satellite observations (MODIS and CALIPSO) and ground-based measurements (AERONET) for the period 2005-2014. The GCMs, NICAM-SPRINTARS (N-S), ECHAM6.3-HAM2.3 (E-H), CAM5.3 (CAM), input with identical emissions from the SMoG-India-v1 emission inventory over India nested in the CEDS global inventory, including all emission sectors except sea salt and soil dust. The annual mean total aerosol optical depth (AOD) averaged over the Indian land region is 0.38, 0.27, and 0.17 from the N-S, CAM, and E-H models respectively, while the annual mean value from the MODIS observational dataset is 0.43. Single scattering albedo predicted by E-H is lower compared to CAM and N-S while model predictions of Angstrom exponent are closer to MERRA2 dataset. However, the average total aerosol column burden over Indian landmass simulated by the models is very close and comparable to the reanalysis results. Statistical analysis of AOD between model and AERONET measurements at nine sites shows that the root mean square error varies from 0.1 to 0.4 and the index of agreement (average value) is ∼0.4. The aerosol emission and transport models, methodology for calculation of aerosol optical properties and their mixing states contributes to the diversity in the results from various models. The present study provides an analysis of limitations and uncertainties contributing to the differences between the simulations and observations, and the inter-model diversity.

Comparison of iCOR and Rayleigh atmospheric correction methods on Sentinel-3 OLCI images for a shallow eutrophic reservoir

Remote sensing of inland waters is challenging, but also important, due to the need to monitor the ever-increasing harmful algal blooms (HABs), which have serious effects on water quality. The Ocean and Land Color Instrument (OLCI) of the Sentinel-3 satellites program is capable of providing images for the monitoring of such waters. Atmospheric correction is a necessary process in order to retrieve the desired surface-leaving radiance signal and several atmospheric correction methods have been developed through the years. However, many of these correction methods require programming language skills, or function as commercial software plugins, limiting their possibility of use by end users. Accordingly, in this study, the free SNAP software provided by the European Space Agency (ESA) was used to evaluate the possible differences between a partial atmospheric correction method accounting for Rayleigh scattering and a full atmospheric correction method (iCOR), applied on Sentinel-3 OLCI images of a shallow, highly eutrophic water reservoir. For the complete evaluation of the two methods, in addition to the comparison of the band reflectance values, chlorophyll (CHL) and cyanobacteria (CI) indices were also calculated and their values were intercompared. The results showed, that although the absolute values between the two correction methods did not coincide, there was a very good correlation between the two methods for both bands' reflectance (r > 0.73) and the CHL and CI indices values (r > 0.95). Therefore, since iCOR correction image processing time is 25 times longer than Rayleigh correction, it is proposed that the Rayleigh partial correction method may be alternatively used for seasonal water monitoring, especially in cases of long time-series, enhancing time and resources use efficiency. Further comparisons of the two methods in other inland water bodies and evaluation with in situ chlorophyll and cyanobacteria measurements will enhance the applicability of the methodology.

Large-scale demonstration of machine learning for the detection of volcanic deformation in Sentinel-1 satellite imagery

Radar (SAR) satellites systematically acquire imagery that can be used for volcano monitoring, characterising magmatic systems and potentially forecasting eruptions on a global scale. However, exploiting the large dataset is limited by the need for manual inspection, meaning timely dissemination of information is challenging. Here we automatically process ~ 600,000 images of > 1000 volcanoes acquired by the Sentinel-1 satellite in a 5-year period (2015-2020) and use the dataset to demonstrate the applicability and limitations of machine learning for detecting deformation signals. Of the 16 volcanoes flagged most often, 5 experienced eruptions, 6 showed slow deformation, 2 had non-volcanic deformation and 3 had atmospheric artefacts. The detection threshold for the whole dataset is 5.9 cm, equivalent to a rate of 1.2 cm/year over the 5-year study period. We then use the large testing dataset to explore the effects of atmospheric conditions, land cover and signal characteristics on detectability and find that the performance of the machine learning algorithm is primarily limited by the quality of the available data, with poor coherence and slow signals being particularly challenging. The expanding dataset of systematically acquired, processed and flagged images will enable the quantitative analysis of volcanic monitoring signals on an unprecedented scale, but tailored processing will be needed for routine monitoring applications.

SUPPLEMENTARY INFORMATION

The online version contains supplementary material available at 10.1007/s00445-022-01608-x.

Improving the accuracy of nonpoint-source pollution estimates in inland waters with coupled satellite-UAV data

Quantitatively and accurately analyzing nonpoint-source (NPS) pollution is essential for efficiently preventing the input of NPS loads into inland waters. However, the accuracy of previous NPS pollution models is limited by the accuracy of ground parameter data. In addition, there are few effective methods that thoroughly verify modeling results at large scales. This paper presents a framework for accurate NPS pollution estimation by coupling satellite and unmanned aerial vehicle (UAV) monitoring data, and the results are verified by both field sampling and a newly developed inlet NPS pollution "observation" simulation method. Fractional vegetation coverage (FVC) data obtained by satellite were used to improve the accuracy of the runoff module of the framework. Satellite and UAV data were coupled to acquire livestock data, determine inlets, and identify reservoir buffer zones and vegetation types. These new data were then used to improve the accuracy of the livestock and runoff modules in the framework. The results show that the estimation accuracy of total nitrogen, total phosphorus, ammonia nitrogen, and chemical oxygen demand with FVC were improved by 39.96%, 69.29%, 54.05% and 47.22% (in relative error), respectively. The high-resolution livestock data acquisition improved the estimation accuracy of the NPS pollution load by 7-53%. The high-resolution inlet extraction improved the accuracy by 3-24%. The high-resolution buffer zone identification improved the accuracy with the estimated NPS pollutant concentration into reservoir decreasing by 60-99%. Finally, the high-resolution vegetation type identification improved the accuracy by 10-72%. The framework performs satisfactorily, which was verified based on the simulated NPS observations with an average relative error of 11.54-24.31%. We found that the FVC, livestock number, and inlet number are key parameters for NPS pollution modeling; the introduction of monthly variation in the FVC makes the modeled NPS pollution load much higher in areas with mature complex forested ecosystems or densely distributed vegetation but much lower in areas with sparsely distributed vegetation. The above methods provide a scientific reference for high-efficiency NPS pollution prevention in inland waters, laying a solid basis for decision-making regarding water quality management in data-scarce regions around the world.

Satellite remote sensing of environmental variables can predict acoustic activity of an orthopteran assemblage

Passive acoustic monitoring (PAM) is a promising method for biodiversity assessment, which allows for longer and less intrusive sampling when compared to traditional methods (e.g., collecting specimens), by using sound recordings as the primary data source. Insects have great potential as models for the study and monitoring of acoustic assemblages due to their sensitivity to environmental changes. Nevertheless, ecoacoustic studies focused on insects are still scarce when compared to more charismatic groups. Insects' acoustic activity patterns respond to environmental factors, like temperature, moonlight, and precipitation, but community acoustic perspectives have been barely explored. Here, we provide an example of the usefulness of PAM to track temporal patterns of acoustic activity for a nocturnal assemblage of insects (Orthoptera). We integrate satellite remote sensing and astronomically measured environmental factors at a local scale in an Andean Forest of Colombia and evaluate the acoustic response of orthopterans through automated model detections of their songs for nine weeks (March and April of 2020). We describe the acoustic frequency range and diel period for the calling song of each representative species. Three species overlapped in frequency and diel acoustics but inhabit different strata: canopy, understory, and ground surface level. Based on the acoustic frequency and activity, we identified three trends: (i) both sampled cricket species call at lower frequency for shorter periods of time (dusk); (ii) all sampled katydid species call at higher frequency for longer time periods, including later hours at night; and (iii) the diel acoustic activity span window seems to increase proportionally with dominant acoustic frequency, but further research is required. We also identified a dusk chorus in which all the species sing at the same time. To quantify the acoustic response to environmental factors, we calculated a beta regression with the singing activity as a response variable and moon phase, surface temperature and daily precipitation as explanatory variables. The response to the moon phase was significant for the katydids but not for the crickets, possibly due to differences in diel activity periods. Crickets are active during dusk, thus the effects of moonlight on acoustic activity are negligible. The response to precipitation was significant for the two crickets and not for the katydids, possibly because of higher likelihood of rain interrupting crickets' shorter diel activity period. Our study shows how the local survey of orthopteran acoustic assemblages, with a species taxonomic resolution coupled with remote-sensing environmental measurements can reveal responses to environmental factors. In addition, we demonstrate how satellite data might prove to be a useful alternative source of environmental data for community studies with geographical, financial, or other constraints.

Estimating ground-level PM2.5 over Bangkok Metropolitan Region in Thailand using aerosol optical depth retrieved by MODIS

A number of previous studies have shown that statistical model with a combination of satellite-derived aerosol optical depth (AOD) and PM_2.5 measured by the monitoring stations could be applied to predict spatial ground-level PM_2.5 concentration, but few studies have been conducted in Thailand. This study aimed to estimate ground-level PM_2.5 over the Bangkok Metropolitan Region in 2020 using linear regression model that incorporates the Moderate Resolution Imaging Spectroradiometer (MODIS) AOD measurements and other air pollutants, as well as various meteorological factors and greenness indicators into the model. The 12-fold cross-validation technique was used to examine the accuracy of model performance. The annual mean (standard deviation) concentration of observed PM_2.5 was 22.37 (± 12.55) µg/m³ and the mean (standard deviation) of PM_2.5 during summer, winter, and rainy season was 18.36 (± 7.14) µg/m³, 33.60 (± 14.48) µg/m³, and 15.30 (± 4.78) µg/m³, respectively. The cross-validation yielded R ² of 0.48, 0.55, 0.21, and 0.52 with the average of predicted PM_2.5 concentration of 22.25 (± 9.97) µg/m³, 21.68 (± 9.14) µg/m³, 29.43 (± 9.45) µg/m³, and 15.74 (± 5.68) µg/m³ for the year round, summer, winter, and rainy season, respectively. We also observed that integrating NO₂ and O₃ into the regression model improved the prediction accuracy significantly for a year round, summer, winter, and rainy season over the Bangkok Metropolitan Region. In conclusion, estimating ground-level PM_2.5 concentration from the MODIS AOD measurement using linear regression model provided the satisfactory model performance when incorporating many possible predictor variables that would affect the association between MODIS AOD and PM_2.5 concentration.

SUPPLEMENTARY INFORMATION

The online version contains supplementary material available at 10.1007/s11869-022-01238-4.

Perivascular infiltration reflects subclinical lymph node metastasis in invasive lobular carcinoma

Invasive lobular carcinoma (ILC) is characterized by discohesive cells due to irreversible loss of E-cadherin expression and multiple satellites, where individual cell migration is evident without disturbance of the stroma. Neoplastic cells sometimes infiltrate the surrounding vessel in satellites. Here, we aimed to clarify the specific role of perivascular infiltration (PVI) and ameboid migration, characterized by nondisturbance of the background stromal structure, in ILCs. A total of 139 cases with ILC and 122 cases with invasive breast carcinoma of no special type (IBC-NST) were evaluated retrospectively. PVI was significantly more common in ILC than in IBC-NST (50% [70 of 139 cases] vs. 9% [11 of 122 cases], p < 0.001). ILC cases with PVI showed a larger pathological tumour size than clinical tumour size (p < 0.01), a higher frequency of pathological node status pN2-pN3 when limited to clinically node-negative cases (p < 0.01) and lower circularity of tumour morphology on imaging (p < 0.01) than ILC cases without PVI. In the pathological evaluation, the intensity and occupancy of tumour cells expressing phospho-myosin light chain 2, which is a hallmark of ameboid migration, were significantly higher in ILC cases with PVI than in those without PVI at the tumour margins (p < 0.05). ILC with PVI is associated with irregular, poorly defined tumour margins and lymph node metastasis without adenopathy, which is difficult to assess using imaging. PVI may be caused by ameboid migration, as shown by the positive expression of phospho-myosin light chain 2. The presence of PVI may be a predictor for clinically node-negative pN2-pN3 in ILC patients.

Geostationary satellite-derived ground-level particulate matter concentrations using real-time machine learning in Northeast Asia

Rapid economic growth, industrialization, and urbanization have caused frequent air pollution events in East Asia over the last few decades. Recently, aerosol data from geostationary satellite sensors have been used to monitor ground-level particulate matter (PM) concentrations hourly. However, many studies have focused on using historical datasets to develop PM estimation models, often decreasing their predictability for unseen data in new days. To mitigate this problem, this study proposes a novel real-time learning (RTL) approach to estimate PM with aerodynamic diameters of <10 μm (PM₁₀) and <2.5 μm (PM_2.5) using hourly aerosol data from the Geostationary Ocean Color Imager (GOCI) and numerical model outputs for daytime conditions over Northeast Asia. Three schemes with different weighting strategies were evaluated using 10-fold cross-validation (CV). The RTL models, which considered both concentration and time as weighting factors (i.e., Scheme 3) yielded consistent improvement for 10-fold CV performance on both hourly and monthly scales. The real-time calibration results for PM₁₀ and PM_2.5 were R² = 0.97 and 0.96, and relative root mean square error (rRMSE) = 12.1% and 12.0%, respectively, and the 10-fold CV results for PM₁₀ and PM_2.5 were R² = 0.73 and 0.69 and rRMSE = 41.8% and 39.6%, respectively. These results were superior to results from the offline models in previous studies, which were based on historical data on an hourly scale. Moreover, we estimated PM concentrations in the ocean without using land-based variables, and clearly demonstrated the PM transport over time. Because the proposed models are based on the RTL approach, the density of in-situ monitoring sites could be a major uncertainty factor. This study identified that a high error occurred in low-density areas, whereas a low error occurred in high-density areas. The proposed approach can be operated to monitor ground-level PM concentrations in real-time with uncertainty analysis to ensure optimal results.

lakes

Cyanobacteria harmful algal blooms (cyanoHABs) negatively affect ecological, human, and animal health. Traditional methods of validating satellite algorithms with data from water samples are often inhibited by the expense of quantifying cyanobacteria indicators in the field and the lack of public data. However, state recreation advisories and other recorded events of cyanoHAB occurrence reported by local authorities can serve as an independent and publicly available dataset for validation. State recreation advisories were defined as a period delimited by a start and end date where a warning was issued due to detections of cyanoHABs over a state's risk threshold. State reported events were defined as any event that was documented with a single date related to cyanoHABs. This study examined the presence-absence agreement between 160 state reported cyanoHAB advisories and 1,343 events and cyanobacteria biomass estimated by a satellite algorithm called the Cyanobacteria Index (CI_cyano). The true positive rate of agreement with state recreation advisories was 69% and 60% with state reported events. CI_cyano detected a reduction or absence in cyanobacteria after 76% of the recreation advisories ended. CI_cyano was used to quantify the magnitude, spatial extent, and temporal frequency of cyanoHABs; each of these three metrics were greater (r > 0.2) during state recreation advisories compared to non-advisory times with effect sizes ranging from small to large. This is the first study to quantitatively evaluate satellite algorithm performance for detecting cyanoHABs with state reported events and advisories and supports informed management decisions with satellite technologies that complement traditional field observations.

A higher-order robust correlation Kalman filter for satellite attitude estimation

A higher-order robust correlation Kalman filtering approach is presented to achieve attitude estimation for satellites with unknown modeling errors. A robust correlation Kalman filter (RCKF) is preliminarily derived by using the sequence orthogonal principle. To improve its performance further, a higher-order sigma version of the RCKF is designed by incorporating a novel sigma point generation algorithm. This modified filter can capture the third and the fourth central moment's information of the system posteriori probability density function. It is proved that the proposed filter achieves better estimation accuracy and robustness. The effectiveness of the developed filter is demonstrated by simulation results with it applied to the satellite attitude estimation problem.

Wetland monitoring technification for the Ecuadorian Andean region based on a multi-agent framework

Remote sensing using satellites and unmanned aerial vehicles (UAVs) has become an important tool for wetland delimitation and saturation assessment since they enable patterns identification and wetland saturation data collection in an agile and optimum way. However, their deployment and operative costs limit their implementation in harsh environments, such as the ones presented in the high Andean wetlands. In this context, this work presents a framework to monitor cost-effectively high Andean wetlands using a multi-agent approach based on: field testing, UAV orthomosaics, and satellite imagery. The method developed comprises two stages: i) definition of the monitoring agent (field testing, satellite, UAV) and ii) image processing. For these stages, semi-empirical and statistical models, which were developed in previous works are incorporated in an open-source framework to tailor each monitoring approach accordingly to the seasonality of a representative Andean wetland. The application of the method and its results highlight the suitability of using visual spectrum low-cost remote sensing approach to compute wetlands saturation percentage. In addition, the methodology proposed allowed the development of a temporal monitoring scheme, where the viability of each monitoring agent is examined. In order to validate the method, field data and multispectral imagery were employed using as case of study the Pugllohuma wetland located in the Antisana Reserve. Thus, the main contribution of this work lies in establishing a technified monitoring framework for the Ecuadorian high Andean wetlands, which can be scaled up and extrapolated to other wetlands with similar harsh environmental conditions, helping to their management and protection policies decision-making.

Merging of the Case 2 Regional Coast Colour and Maximum-Peak Height chlorophyll-a algorithms: validation and demonstration of satellite-derived retrievals across US lakes

Water quality monitoring is relevant for protecting the designated, or beneficial uses, of water such as drinking, aquatic life, recreation, irrigation, and food supply that support the economy, human well-being, and aquatic ecosystem health. Managing finite water resources to support these designated uses requires information on water quality so that managers can make sustainable decisions. Chlorophyll-a (chl-a, µg L^-1) concentration can serve as a proxy for phytoplankton biomass and may be used as an indicator of increased anthropogenic nutrient stress. Satellite remote sensing may present a complement to in situ measures for assessments of water quality through the retrieval of chl-a with in-water algorithms. Validation of chl-a algorithms across US lakes improves algorithm maturity relevant for monitoring applications. This study compares performance of the Case 2 Regional Coast Colour (C2RCC) chl-a retrieval algorithm, a revised version of the Maximum-Peak Height (MPH_(P)) algorithm, and three scenarios merging these two approaches. Satellite data were retrieved from the MEdium Resolution Imaging Spectrometer (MERIS) and the Ocean and Land Colour Instrument (OLCI), while field observations were obtained from 181 lakes matched with U.S. Water Quality Portal chl-a data. The best performance based on mean absolute multiplicative error (MAE_mult) was demonstrated by the merged algorithm referred to as C₁₅-M₁₀ (MAE_mult = 1.8, bias_mult = 0.97, n = 836). In the C₁₅-M₁₀ algorithm, the MPH_(P) chl-a value was retained if it was > 10 µg L^-1; if the MPH_(P) value was ≤ 10 µg L^-1, the C2RCC value was selected, as long as that value was < 15 µg L^-1. Time-series and lake-wide gradients compared against independent assessments from Lake Champlain and long-term ecological research stations in Wisconsin were used as complementary examples supporting water quality reporting requirements. Trophic state assessments for Wisconsin lakes provided examples in support of inland water quality monitoring applications. This study presents and assesses merged adaptations of chl-a algorithms previously reported independently. Additionally, it contributes to the transition of chl-a algorithm maturity by quantifying error statistics for a number of locations and times.

Individual exposure to ambient PM2.5 and hospital admissions for COPD in 110 hospitals: a case-crossover study in Guangzhou, China

Few studies have evaluated the short-term association between hospital admissions and individual exposure to ambient particulate matter (PM_2.5). Particularly, no studies focused on hospital admissions for chronic obstructive pulmonary disease (COPD) at the individual level. We assessed the short-term effects of PM_2.5 on hospitalization admissions for COPD in Guangzhou, China, during 2014-2015, based on satellite-derived estimates of ambient PM_2.5 concentrations at a 1-km resolution near the residential address as individual-level exposure for each patient. Around 40,002 patients with COPD admitted to 110 hospitals were included in this study. A time-stratified case-crossover design with conditional logistic regression models was applied to assess the effects of PM_2.5 based on a 1-km grid data of aerosol optical depth provided by the National Aeronautics and Space Administration on hospital admissions for COPD. Further, we performed stratified analyses by individual demographic characteristics and season of hospital admission. Around 10 μg/m³ increase in individual-level PM_2.5 was associated with an increase of 1.6% (95% confidence interval [CI]: 0.6%, 2.7%) in hospitalization for COPD at a lag of 0-5 days. The impact of PM_2.5 on hospitalization for COPD was greater significantly in males and patients admitted in summer. Our study strengthened the evidence for the adverse effect of PM_2.5 based on satellite-based individual-level exposure data.

Year-to-year variability of oil pollution along the Eastern Arabian Sea: The impact of COVID-19 imposed lock-downs

This study investigated the year-to-year variability in the occurrence, abundance and sources of oil spills in the Eastern Arabian Sea (EAS) using sentinel-1 imagery and identified the potential oil spills vulnerable zones. The four consecutive year's data acquired from 2017 to 2020 (March-May) reveal three oil spill hot spot zones. The ship-based oil spills were dominant over zone's-1 (off Gujarat) and 3 (off Karnataka and Kerala), and the oil field based over zone-2 (off Maharashtra). The abundance of oil spills was significantly low in zone-1, only 14.30km² (1.2%) during lock-down due to the covid-19 pandemic. Whereas, the year-to-year oil spills over zone's 2 and 3 are not significantly varied (170.29 km² and 195.01 km²), further suggesting the influence of oil exploration and international tanker traffic are in operation during the lock-down. This study further recommends that manual clustering is the best method to study the distribution of unknown oil spills.

Latitude impact on Pandemic Sars-Cov-2 2020 outbreaks and possible utility of UV indexes in predictions of regional daily infections and deaths

The importance of two related factors ^_ latitude and solar ultraviolet radiation ^_ has been insufficiently recognized as determining the spread of pandemic Sars-CoV-2 outbreaks across the globe. In this study we provide evidence of the impact of latitude and investigate how daily RT-PCR diagnosed infections and deaths are quantitively correlated with the UV component of Solar light. Here, we present regression analyses using daily national numbers from Austria and from Portugal with daily ultraviolet indexes of two selected locations in these territories, obtained from a Satellite source. These countries, have similar surfaces areas and population size but Austria's mean latitude is 9° up-north. The equations derived from regression analyses of those two variables are comparable for both countries, fit best the fall (2^nd) pandemic wave and can be a useful non-R(t) (rate of transmission) dependent predictive tool. Similar equations were derived for deaths that follow infections within a few weeks delay. Strong correlations depend on the size of the region/country from which infections are collected, the robustness of screening practices, ideally kept through weekends and holidays. Besides the forecasting usefulness of such correlations, these findings also suggest that covid-19 transmission co-exists with a Sars-Cov-2 specific UV-induced immunosuppression response. While in 2020, intensity of pandemic spring and fall waves reflect a Solar UV-light modulation, we relate exceptional low temperature and humidity with additional waves, as the winter 2020/2021 3^rd wave, felt in the western European countries. This work may help understanding this Pandemic phenomenon and dealing with similar catastrophes in the future.

Marine heatwaves of sea surface temperature off south Java

The frequency of marine heatwaves (MHWs) events has been rising globally in recent years, and this trend is expected to continue in the region off the coast of south Java Island. These oceanic extreme events may have the potential to devastate marine habitats, ecosystems and fisheries. This paper characterized MHWs off south Java from 1982 to 2019 using satellite-observed sea surface temperature. The aim of this study was to examine the dynamics of MHWs in one of Indonesia's most important fisheries hotspots, located in the southeast of the tropical Indian Ocean. We have identified two strong MHWs events in 1998 and 2016, both of which started in the austral winter months. Both events were lasted through the spring before dissipating in the early austral winter. These intense MHWs were likely related to a strong El Niño and decreased monsoon activity.

•

First assessment of MHWs south off Java using high-resolution satellite SST products.

•

The most intense and longest MHWs events are identified in 1998 and 2016.

•

Record-breaking MHWs occurred during strong El Niño and weakened monsoon wind.

FSS superstrate antenna for satellite cynosure on IoT to combat COVID-19 pandemic

The global pandemic, COVID-19 needs joint techniques and technology to combat it. The internet of things (IoT) has been at the forefront in solving problems, not only in the health care sector but in other sectors. It delivers accuracy with robustness in the developing service and application. However, it remains clear that the use of IoT is limited to coverage, longevity, security, connectivity issue, immediacy, and multicasting, we proposed in this paper frequency selective surface (FSS) as superstrate for rectangular microstrip antenna. An FSS design combine with the rectangular microstrip antenna for better performance is placed over FSS parallel configuration. The rectangular microstrip antenna was titled 45 degrees to change the band-stop. Analysis of the proposed performance in terms of gain, return loss, and directivity shows that the FSS structure's integration brings better results. With the help of a 3D electromagnetic computer simulation technology CST studio suite, we model the proposed antenna, perform the simulation with a frequency-domain solver, and validate it with a time-domain solver. The proposed impressive result is suitable for satellite networks, which hybrid with IoT can provide a sustainable long-time solution in fighting the COVID-19 pandemic.

Mate-guarding duration is mainly influenced by the risk of sperm competition and not by female quality in a golden orb-weaver spider

Males are expected to mate with as many females as possible, but can maximize their reproductive success through strategic mating decisions. For instance, males can increase their own fitness by mating with high quality females that produce more offspring. Additionally, males can adjust mating effort based on the relative distribution of females and male competitors. To test factors that influence male mate choice, we assessed male mating decisions in the golden silk orb-weaver spider, Trichonephila clavipes (Nephilidae), a species in which females are polyandrous, males guard females before and after copulation occurs and large males are the most successful at guarding mates. We tested the hypothesis that males spend more time guarding high quality females that are spatially isolated, and when the risk of sperm competition is higher. We also hypothesized that this effect increases with male body size. We assessed solitary and aggregated female webs in the field and quantified female quality (i.e., female body condition), male size (i.e., male body size), the risk of sperm competition (i.e., number of males in each female web), and mate-guarding duration (i.e., number of days each male spent in each web). We found that mate-guarding behaviour is largely influenced by the presence of male competitors. In addition, male body size seems to moderately influence male guarding decisions, with larger males guarding for a longer time. Finally, female body condition and type of web (i.e., solitary or aggregated) seem to play small roles in mate-guarding behaviour. As mate-guarding duration increased by 0.718 day per each additional male competitor in the web, and guarding behaviour prevents males from seeking additional mates, it seems that guarding females can be considerably costly. We conclude that failing to guard a sexual partner promotes high costs derived from sperm competition, and a male cannot recover his relative loss in fertilization success by seeking and fertilizing more females. In addition, the search for more sexual partners can be constrained by possible high costs imposed by weight loss and fights against other males, which may explain why the type of web only moderately influenced male mate choice. Following the same rationale, if high-quality females are not easy to find and/or mating with a high-quality female demands much effort, males may search females and guard them regardless of female quality. In conclusion, the factor that most influences male mate-guarding behaviour among T. clavipes in the field is the risk of sperm competition.

Estimation of surface-level NO2 and O3 concentrations using TROPOMI data and machine learning over East Asia

In East Asia, air quality has been recognized as an important public health problem. In particular, the surface concentrations of air pollutants are closely related to human life. This study aims to develop models for estimating high spatial resolution surface concentrations of NO₂ and O₃ from TROPOspheric Monitoring Instrument (TROPOMI) data in East Asia. The machine learning was adopted by fusion of various satellite-based variables, numerical model-based meteorological variables, and land-use variables. Four machine learning approaches-Support Vector Regression (SVR), Random Forest (RF), Extreme Gradient Boost (XGB), and Light Gradient Boosting Machine (LGBM)-were evaluated and compared with Multiple Linear Regression (MLR) as a base statistical method. This study also modeled the NO₂ and O₃ concentrations over the ocean surface (i.e., land model for scheme 1 and ocean model for scheme 2). The estimated surface concentrations were validated through three cross-validation approaches (i.e., random, temporal, and spatial). The results showed that the NO₂ model produced R² of 0.63-0.70 and normalized root-mean-square-error (nRMSE) of 38.3-42.2% and the O₃ model resulted in R² of 0.65-0.78 and nRMSE of 19.6-24.7% for scheme 1. The indirect validation based on the stations near the coastline for scheme 2 showed slight decrease (~0.3-2.4%) in nRMSE when compared to scheme 1. The contributions of input variables to the models were analyzed based on SHapely Additive exPlanations (SHAP) values. The NO₂ vertical column density among the TROPOMI-derived variables showed the largest contribution in both the NO₂ and O₃ models.

DeepGRP: engineering a software tool for predicting genomic repetitive elements using Recurrent Neural Networks with attention

BACKGROUND

Repetitive elements contribute a large part of eukaryotic genomes. For example, about 40 to 50% of human, mouse and rat genomes are repetitive. So identifying and classifying repeats is an important step in genome annotation. This annotation step is traditionally performed using alignment based methods, either in a de novo approach or by aligning the genome sequence to a species specific set of repetitive sequences. Recently, Li (Bioinformatics 35:4408-4410, 2019) developed a novel software tool dna-brnn to annotate repetitive sequences using a recurrent neural network trained on sample annotations of repetitive elements.

RESULTS

We have developed the methods of dna-brnn further and engineered a new software tool DeepGRP. This combines the basic concepts of Li (Bioinformatics 35:4408-4410, 2019) with current techniques developed for neural machine translation, the attention mechanism, for the task of nucleotide-level annotation of repetitive elements. An evaluation on the human genome shows a 20% improvement of the Matthews correlation coefficient for the predictions delivered by DeepGRP, when compared to dna-brnn. DeepGRP predicts two additional classes of repeats (compared to dna-brnn) and is able to transfer repeat annotations, using RepeatMasker-based training data to a different species (mouse). Additionally, we could show that DeepGRP predicts repeats annotated in the Dfam database, but not annotated by RepeatMasker. DeepGRP is highly scalable due to its implementation in the TensorFlow framework. For example, the GPU-accelerated version of DeepGRP is approx. 1.8 times faster than dna-brnn, approx. 8.6 times faster than RepeatMasker and over 100 times faster than HMMER searching for models of the Dfam database.

CONCLUSIONS

By incorporating methods from neural machine translation, DeepGRP achieves a consistent improvement of the quality of the predictions compared to dna-brnn. Improved running times are obtained by employing TensorFlow as implementation framework and the use of GPUs. By incorporating two additional classes of repeats, DeepGRP provides more complete annotations, which were evaluated against three state-of-the-art tools for repeat annotation.

Clinical characteristics and therapy response in unresectable melanoma patients stage IIIB-IIID with in-transit and satellite metastases

INTRODUCTION

Cutaneous melanoma is notorious for the development of in-transit metastases (ITM). For unknown biological reasons, ITM remain the leading tumour manifestation without progression to distant sites in some patients.

METHODS

In total, 191 patients with initially unresectable stage III ITM and satellite metastases from 16 skin cancer centres were retrospectively evaluated for their tumour characteristics, survival and therapy response. Three groups according to disease kinetics (no distant progress, slow (>6 months) and fast (<6 months) distant progression) were analysed separately.

RESULTS

Median follow-up time was 30.5 (range 0.8-154.0) months from unresectable ITM. Progression to stage IV was observed in 56.5% of cases. Patients without distant metastasis were more often female, older (>70 years) and presented as stage III with lymph node or ITM at initial diagnosis in 45.7% of cases. Melanoma located on the leg had a significantly better overall survival (OS) from time of initial diagnosis compared to non-leg localised primaries (hazard ratio [HR] = 0.61, 95% confidence interval [CI] 0.40-0.91; p = 0.017), but not from diagnosis of unresectable stage III (HR = 0.67, 95% CI 0.45-1.02; p = 0.06). Forty percent of patients received local therapy for satellite and ITM. Overall response rate (ORR) to all local first-line treatments was 38%; disease control rate (DCR) was 49%. In total, 72.3% of patients received systemic therapy for unresectable stage IIIB-D. ORR for targeted therapy (n = 19) was highest with 63.2% and DCR was 84.2% compared to an ORR of 31.4% and a DCR of 54.3% in PD-1 treated patients (n = 70). Patients receiving PD-1 and intralesional talimogene laherparepvec (n = 12) had an ORR of 41.7% and a DCR of 75%.

CONCLUSION

Patients with unresectable ITM and without distant progression are more often female, older, and have a primary on the leg. Response to PD-1 inhibitors in this cohort was lower than expected, but further investigation is required to elucidate the biology of ITM development and the interplay with the immune system.

Assessing the reliability of peatland GPP measurements by remote sensing: From plot to landscape scale

Estimates of peatland carbon fluxes based on remote sensing data are a useful addition to monitoring methods in these remote and precious ecosystems, but there are questions as to whether large-scale estimates are reliable given the small-scale heterogeneity of many peatlands. Our objective was to consider the reliability of models based on Earth Observations for estimating ecosystem photosynthesis at different scales using the Forsinard Flows RSPB reserve in Northern Scotland as our study site. Three sites across the reserve were monitored during the growing season of 2017. One site is near-natural blanket bog, and the other two are at different stages of the restoration process after removal of commercial conifer forestry. At each site we measured small (flux chamber) and landscape scale (eddy covariance) CO₂ fluxes, small scale spectral data using a handheld spectrometer, and obtained corresponding satellite data from MODIS. The variables influencing GPP at small scale, including microforms and dominant vegetation species, were assessed using exploratory factor analysis. A GPP model using land surface temperature and a measure of greenness from remote sensing data was tested and compared to chamber and eddy covariance CO₂ fluxes; this model returned good results at all scales (Pearson's correlations of 0.57 to 0.71 at small scale, 0.76 to 0.86 at large scale). We found that the effect of microtopography on GPP fluxes at the study sites was spatially and temporally inconsistent, although connected to water content and vegetation species. The GPP fluxes measured using EC were larger than those using chambers at all sites, and the reliability of the TG model at different scales was dependent on the measurement methods used for calibration and validation. This suggests that GPP measurements from remote sensing are robust at all scales, but that the methods used for calibration and validation will impact accuracy.

A comprehensive review of catchment water quality monitoring using a tiered framework of integrated sensing technologies

Due to the growing threat of climate change, new advances in water quality monitoring strategies are needed now more than ever. Reliable and robust monitoring practices can be used to improve and better understand catchment processes affecting the water quality. In recent years the deployment of long term in-situ sensors has increased the temporal and spatial data being obtained. Furthermore, the development and research into remote sensing using satellite and aerial imagery has been incrementally integrated into catchments for monitoring areas that previously might have been impossible to monitor, producing high-resolution data that has become imperative to catchment monitoring. The use of modelling in catchments has become relevant as it enables the prediction of events before they occur so that strategic plans can be put in place to deal with or prevent certain threats. This review highlights the monitoring approaches employed in catchments currently and examines the potential for integration of these methods. A framework might incorporate all monitoring strategies to obtain more information about a catchment and its water quality. The future of monitoring will involve satellite, in-situ and air borne devices with data analytics playing a key role in providing decision support tools. The review provides examples of successful use of individual technologies, some combined approaches and identifies gaps that should be filled to achieve an ideal catchment observation system.

Understanding global changes in fine-mode aerosols during 2008-2017 using statistical methods and deep learning approach

Despite their extremely small size, fine-mode aerosols have significant impacts on the environment, climate, and human health. However, current understandings of global changes in fine-mode aerosols are limited. In this study, we employed newly developed satellite retrieval data and an attentive interpretable deep learning model to explore the status, changes, and association factors of the global fine-mode aerosol optical depth (fAOD) and aerosol fine-mode fraction (FMF) from 2008 to 2017. At the global scale, the results show a significant increasing trend in land FMF (2.34 × 10^-3/year); however, the FMF over the ocean and the fAOD over land and ocean did not reveal significant trends. Between 2008 and 2017, high levels of both fAOD (>0.30) and FMF (>0.75) were identified over China, southeastern Asia, India, and Africa. Seasonally, global land FMF showed high values in summer (>0.70) and low values in spring (<0.65), while land fAOD was high in summer (>0.15) but low in winter (<0.13). Importantly, Australia and Mexico experienced significant increasing trends in FMF during all four seasons. At the regional scale, a significant decline in fAOD was identified in China, which indicates that government emission controls and reductions have been effective in recent decades. The deep learning model was used to interpret the result and showed that O₃ was significantly associated with changes in both the FMF and fAOD. This finding suggests the importance of synergizing the regulations for both O₃ and fine particles. Our work comprehensively examined global spatial and seasonal fAOD and FMF changes and provides a holistic understanding of global anthropogenic impacts.

Remote Delivery of the Satellite Virtual Fracture Clinic - a Pilot Report of the First 500 Cases

Presenting to the fracture clinic carries economic, social and societal consequences. The virtual fracture clinic (VFC) has proven to be a safe, patient-focused, cost-effective means of delivering trauma care, whilst reducing unnecessary clinic attendances. Within our institution, a Satellite VFC was established, so as to accommodate an offsite referring emergency department. The VFC database was accessed to identify the first 500 patients who were referred to the Satellite VFC. The decision made for each patient, the rate of returns to the clinic, and the rate of referrals requiring surgical intervention, following discussion at the VFC, ,were identified. A cost analysis and cost comparison was carried out between the Satellite VFC and the traditional "face to face" fracture clinic. There were 500 patients referred to the Satellite VFC within the study period. Of such patients, 288 (58%) were discharged directly following review at the Satellite VFC, 141 patients (28%) were referred to physiotherapy, 50 (10%) were redirected to the trauma clinic, 11 (2%) were sent directly to hand therapy, and 10 (2%) were sent to the ED review clinic. Patients who returned to the fracture clinic accounted for 3.8% of all referrals, and 0.2% of all referrals necessitated surgical intervention. This pilot initiative saved the Dublin Midlands Hospital Group over €50,000. The Satellite VFC is the first of its kind in the literature. Rural communities worldwide would benefit from remote orthopaedic management of suitable fracture patterns. The true value of the Satellite VFC process comes from its use of robust patient care pathways, rationalising resource use and minimising patient travel, whilst demonstrating reliable outcomes and promoting safety.

Surface and satellite observations of air pollution in India during COVID-19 lockdown: Implication to air quality

The strict nationwide lockdown imposed in India starting from 25^th March 2020 to prevent the spread of COVID-19 disease reduced the mobility and interrupted several important anthropogenic emission sources thereby creating a temporary air quality improvement. This study conducts a multi-scale (national-regional-city), multi-species, and multi-platform analysis of air pollutants and meteorological data by synergizing surface and satellite observations. Our analysis suggests a significant reduction in surface measurements of nitrogen dioxide (NO₂) (46-61 %) and fine particulate matter (PM_2.5) (42-60 %) during the lockdown period that are also corroborated by the reduction in satellite observed aerosol optical depth (AOD) (3-56 %) and tropospheric NO₂ column density (25-50 %) data over multiple cities. Other species, namely coarse particulate matter (PM₁₀) (24-62 %), ozone (22-56 %) also showed a substantial reduction whereas carbon monoxide (16-46 %), exhibited a moderate decline. In contrast, sulfur dioxide (SO₂) levels did not show any defined reduction trend but rather increased in Mumbai, Bengaluru, and Kolkata. The temporary air quality improvement achieved by the painful natural experiment of this pandemic has helped demonstrate the importance of reducing emissions from other sectors along with transportation and industry to achieve the national air quality targets in the future.

A dataset composed of multiangular spectral libraries and auxiliary data at tree, leaf, needle, and bark level for three common European tree species

This article describes a dataset of multiangular scattering properties of small trees (height = 0.38–0.7 m) at visible, near-infrared, and shortwave-infrared wavelengths (350–2500 nm), and provides supporting auxiliary data that comprise leaf, needle, and bark spectra, and structural characteristics of the trees. Multiangular spectra were measured for 18 trees belonging to three common European tree species: Scots pine (Pinus sylvestris L.), Norway spruce (Picea abies (L.) H. Karst), and sessile oak (Quercus petraea (Matt.) Liebl.). The measurements were performed in 47 different view angles across a hemisphere, using a laboratory goniometer and a non-imaging spectrometer. Leaf and needle spectra were measured for each tree, using a non-imaging spectrometer coupled to an integrating sphere. Bark spectra were measured for one sample tree per species. In addition, leaf and needle fresh mass, surface area of leaves, needles, and woody parts, silhouette area, and spherically averaged silhouette to total area ratio (STAR) for each tree were measured or derived from the measurements. The data are useful for modeling the shortwave reflectance characteristics of small trees and potentially forests, and thus benefit climate modeling or interpretation of remote sensing data.

Water vapor satellite products in the European Arctic: An inter-comparison against GNSS data

The European Arctic is a region of high interest for climate change. Water vapor plays a fundamental role in global warming; therefore, high-quality water vapor monitoring is essential for assimilation in forecast simulations. The seven analyzed instruments on-board satellite platforms are: Atmospheric Infrared Sounder (AIRS), Global Ozone Monitoring Instrument 2 (GOME-2), Moderate-Resolution Imaging Spectroradiometer (MODIS), Ozone Monitoring Instrument (OMI), SCanning Imaging Absorption Spectrometer for Atmospheric Carthography (SCIAMACHY) and Polarization and Directionality of the Earth's Reflectances (POLDER). The GNSS data from Ny-Ålesund are matched to satellite observations of IWV in a 30-min temporal window, and 100-km radius. Then, statistics and the distribution of satellite-ground differences under different conditions are studied. The correlation coefficient (R²) with ground-based measurements is about 0.7 for all products except OMI (R²=0.5), and MODIS NIR and POLDER (R²=0.3). OMI shows high bias and variability compared to the rest of products. RMSE values are of the order of 3 mm for all satellites, except OMI (7 mm) and POLDER (5 mm). Bias (MBE) is negligible for AIRS, close to +1.6 mm for GOME-2 and MODIS IR, +0.8 mm for MODIS NIR, +5.9 mm for OMI, -2.7 mm for POLDER and -1.2 mm for SCIAMCHY. All satellite products tend to overestimate small IWV values and underestimate large IWV values. Variability also increases with IWV. An underestimation of the satellite products and an increase on the variability is generally observed for large Solar Zenith Angle (SZA) values. Under cloudy conditions, underestimation and variability are increased. Seasonal behavior is driven by the typical cloud cover (CC), SZA, and IWV values. In summer, it is typical to find conditions with large IWV, small SZA and large CC values. Therefore, in summer months satellite products are more biased (either positively or negatively) and with more variability, but in relative terms they are less biased and exhibit less variability.

New interpretable deep learning model to monitor real-time PM2.5 concentrations from satellite data

Particulate matter with a mass concentration of particles with a diameter less than 2.5 μm (PM_2.5) is a key air quality parameter. A real-time knowledge of PM_2.5 is highly valuable for lowering the risk of detrimental impacts on human health. To achieve this goal, we developed a new deep learning model-EntityDenseNet to retrieve ground-level PM_2.5 concentrations from Himawari-8, a geostationary satellite providing high temporal resolution data. In contrast to the traditional machine learning model, the new model has the capability to automatically extract PM_2.5 spatio-temporal characteristics. Validation across mainland China demonstrates that hourly, daily and monthly PM_2.5 retrievals contain the root-mean-square errors of 26.85, 25.3, and 15.34 μg/m³, respectively. In addition to a higher accuracy achievement when compared with various machine learning inversion methods (backpropagation neural network, extreme gradient boosting, light gradient boosting machine, and random forest), EntityDenseNet can "peek inside the black box" to extract the spatio-temporal features of PM_2.5. This model can show, for example, that PM_2.5 levels in the coastal city of Tianjin were more influenced by air from Hebei than Beijing. Further, EntityDenseNet can still extract the seasonal characteristics that demonstrate that PM_2.5 is more closely related within three month groups over mainland China: (1) December, January and February, (2) March, April and May, (3) July, August and September, even without meteorological information. EntityDenseNet has the ability to obtain high temporal resolution satellite-based PM_2.5 data over China in real-time. This could act as an important tool to improve our understanding of PM_2.5 spatio-temporal features.