A document by Jing Wei. Click on the document to view its contents.

Surface ozone (O3) is a crucial ambient pollutant gas that poses substantial risks to both human health and ecosystems. Nonetheless, there is a scarcity of high-spatial-resolution hourly surface O3 data, particularly during the day when this information is needed due to the strong diurnal variations of O3. We thus determined a best-performing artificial intelligence model to derive 24-hourly 1-kmresolution surface O3 concentrations in China from a large array of satellite and surface data, which can portray well the diurnal variations of O3 concentration. The overall sample-based crossvalidated coefficients of determination (root-mean-square error) are 0.89 (15.74 μg/m 3), 0.91 (14.91 μg/m 3), and 0.85 (16.31 μg/m 3) during the full day (00:00-23:00 local time, or LT), daytime (08:00-17:00 LT), and nighttime (18:00-07:00 LT), respectively. The surface O3 level generally rises from sunrise, around 07:00 LT, reaching a peak at ~15:00 LT, then continuously declining overnight. The magnitude of the diurnal variation amounts to 180% relative to its diurnal mean level. During daytime, solar radiation in the ultraviolet and shortwave spectral bands, along with temperature, explain more than half (32% and 24%) of the diurnal variations using the interpretable SHapley Additive exPlanations (SHAP) method, while nighttime O3 levels are dominated by temperature (31%) and relative humidity (16%). In 2018, approximately 59%, 93%, and 100% of populated areas were susceptible to O3 exposure risk for at least one day, with the maximum daily average 8-h O3 levels surpassing the World Health Organization's recommended daily air quality standards of 160 µg/m³, 120 µg/m³, and 100 µg/m³, respectively. Approximately 65%, 70%, and 99% of vegetated areas in China exceed the minimum critical levels for O3 mixing ratios, as determined by the sum of all hourly values ≥ 0.06 μmol mol-1 (SUM06), the sigmoidally weighted sum of all hourly values (W126), and accumulates over the threshold of 40 nmol mol-1 (AOT40), respectively. Notably, gross primary productivity stands out as the most responsive indicator to surface O3 pollution across various vegetated types in China, especially concerning the Hourly O3 Accumulates without Threshold (AOT0, R =-0.37-0.53, p < 0.001).

J. Wei and Z. Wang made equal contributions to this work.*Corresponding authors:zhanqing@umd.edu; sunlin6@126.com; weijing@umd.eduAbstractLandsat imagery offers remarkable potential for various applications, including land monitoring and environmental assessment, thanks to its high spatial resolution and over 50 years of data records. However, the presence of atmospheric aerosols greatly hinders the precision of land classification and the quantitative retrieval of surface parameters. Notably, there has been no global retrieval of aerosol optical depth (AOD) from Landsat imagery that is needed for atmospheric correction, among other applications. To address this issue, this paper presents an innovative global AOD retrieval framework for Landsat imagery, propelled by atmospheric radiative transfer (ART) and enhanced GeoChronoTransformers (GCT) models incorporating multidimensional spatiotemporal sequence information and executed on the Google Earth Engine (GEE) cloud platform. We gathered all Landsat 8 and 9 images from their respective launch dates (February 2013 and September 2021) up to 2022, which were used to construct a robust ART-GCT-GEE model, and then rigorously validated the model performance across ~470 monitoring stations over land using diverse spatiotemporally independent methods. Leveraging information from multiple spectral channels, contributing to 58% according to the SHapley Additive exPlanation (SHAP) method, our results are highly consistent with observations (e.g., correlation coefficient = 0.863 and root-mean-square error = 0.096), suggesting that accurate historical and future AOD levels can be obtained. Around 81% and 50% of our AOD predictions meet the criteria of Moderate Resolution Imaging Spectroradiometer (MODIS) expected errors [±(0.05+20%)] and the Global Climate Observation System {[max(0.03, 10%)]}, respectively. Additionally, our model is less influenced by changes in surface conditions like topography and land cover. This allows us to generate spatially continuous AOD distributions with highly detailed and fine-scale information from dark to bright surfaces, especially for densely populated urban areas and expansive deserts with high aerosol loadings from both anthropogenic and natural sources.

A document by Jing Wei. Click on the document to view its contents.

Background: Exposure to ambient particulate matter (PM) has been associated with an increased risk of allergic rhinitis in children. However, it is unclear whether food allergy modifies the association between PM exposure and childhood allergic rhinitis. Objectives: We aimed to evaluate the modification of food allergy on the association between PM exposure and allergic rhinitis in preschool children. Methods: We adopted a cross-sectional study and conducted a questionnaire survey among preschool children aged 3 to 6 years in 7 cities in China from June 2019 to June 2020 to collect information on allergic rhinitis. A mature machine learning-based space-time extremely randomized trees model was applied to estimate early-life, prenatal, and first-year exposure of PM 1, PM 2.5 and PM 10 at 1 × 1-km resolution. We used a combination of multilevel logistic regression and restricted cubic spline functions to quantitatively assess whether food allergy modifies the associations between size-specific PM exposure and the risk of childhood allergic rhinitis. Results: The adjusted ORs for childhood allergic rhinitis among the children with food allergy as per interquartile range (IQR) increase in early-life PM 1, PM 2.5 and PM 10 were significantly higher than the corresponding ORs among the children without food allergy [e.g. OR: 1.57, 95% CI (1.32, 1.87) vs. 1.29, 95% CI (1.18, 1.41), for per IQR increase in PM 1 (9.8 μg/m 3)]. The similar patterns were observed for both prenatal and first-year size-specific PM exposure. The interactions between food allergy and size-specific PM exposure on childhood allergic rhinitis were statistically significant (all p- int < 0.001). Conclusions: Food allergy, as an important part of the allergic disease progression, may modify the association between ambient PM exposure and the risk of childhood allergic rhinitis. Children with food allergy should pay more attention to minimize outdoor air pollutants exposure to prevent the further progression of allergic diseases.

Ozone (O3) is an important trace and greenhouse gas in the atmosphere yet, and it threatens the ecological environment and human health at the ground level. Large-scale and long-term studies of O3 pollution in China are few due to highly limited direct measurements whose accuracy and density vary considerably. To overcome these limitations, we employed the ensemble learning method of the extremely randomized trees model by utilizing the spatiotemporal information of a large number of input variables from ground-based observations, remote sensing, atmospheric reanalysis, and model simulation products to estimate ground-level O3. This method yields uniform, long-term and continuous spatiotemporal information of daily maximum eight-hour average (MDA8) O3 over China (called ChinaHighO3) from 2013 to 2020 at a 10 km resolution without any missing values (spatial coverage = 100%). Evaluation against observations indicates that our O3 estimations and predictions are reliable with an average out-of-sample (out-of-station) coefficient of determination (CV-R2) of 0.87 (0.80) and root-mean-square error of 17.10 (21.10) μg/m3 [units here are at standard conditions (273K, 1013hPa)], and are also robust at varying spatial and temporal scales in China. This high-quality and full-coverage O3 dataset allows us to investigate the exposure and trends in O3 pollution at both long- and short-term scales. Trends in O3 concentrations varied substantially but showed an average growth rate of 2.49 μg/m3/yr (p < 0.001) from 2013 to 2020 in China. Most areas show an increasing trend since 2015, especially in summer ozone over the North China Plain. Our dataset accurately captured a recent national and regional O3 pollution event from 23 April to 8 May in 2020. Rapid increase and recovery of O3 concentrations associated with variations in anthropogenic emissions were seen during and after the COVID-19 lockdown, respectively. This carefully vetted and smoothed dataset is valuable for studies on air pollution and environmental health in China.