Search

It is unclear how to develop a model based on the combined satellite data and ground monitoring data to accurately estimate daily NO2 levels. Furthermore, the conventional cross-validation (CV) results represent average levels but the model performance may vary greatly from grid to grid. It is an essential issue to evaluate model's prediction ability in different grids and determine the factors affecting model extrapolating ability, which have never been well examined to date. The aim of this study was to compare the ability of three different methods to estimate the daily NO2 across mainland China during 2014–2016; and to develop a novel two-stage meta-analysis method for exploring the influence of the number and the distribution of nearby sites on grid-level prediction accuracy. For better estimating the daily NO2 level, we developed and compared three methods, including universal kriging model, satellite-based method (Non-linear exposure-lag-response model & Extreme gradient boosting combined technique) and the kriging-calibrated satellite method. For exploring influencing factors, the two-stage meta-analysis method was purposed. The kriging-calibrated satellite method had an overall CV R-square and root mean square error (RMSE) of 0.85 and 7.87μg/m3, better than the Universal Kriging model and the satellite-based method (CV R2 = 0.57 and 0.81). The two-stage meta-analysis method revealed that the model performance did decrease with the sparser distribution of nearby sites. And adding 5 sites within 50 km in the random mode can bring 17.51% improvement in model extrapolating ability. The kriging-calibration can help satellite-based machine learning to provide more accurate NO2 prediction. Our novel evaluation method can provide the suggestion of adding new sites effectively within a limit budget.

Background: In the context of global warming, most researches have been conducted on health influences of heat waves, with limited understanding of health impacts of cold spells, especially for developing countries. Methods: We collected daily mortality and meteorological data for 31 capital cities across China during the maximum period of 2007–2013. A quasi-Poisson regression model combined with a distributed lag non-linear model was used to estimate the short-term effects of cold spells on mortality in cold seasons (November to March). 19 definitions of cold spell were clearly compared, including three definitions from the China Meteorological Administration (CMA) and 16 definitions by combining two temperature indicators (daily minimum and mean temperature), two temperature thresholds (3rd and 5th percentile) and four durations of at least 2–5 days. Then, a random effect meta-analysis was applied to pool the effect estimates at national level. Furthermore, a stratified analysis was constructed to identify the vulnerable subpopulations to cold spells. Results: The definition, in which daily mean temperature falls below 5th percentile for at least two consecutive days, produced the optimum model fit performance. Generally, the mortality risk increased to the maximum after 3–6 days' exposure to cold spell and then leveled off in the next 3 weeks. The pooled relative risks (RR) of non-accidental mortality for cold spells were 1.03 (95% CI: 1.01–1.05), 1.27 (1.19–1.35) and 1.55 (1.40–1.70) at lag 0, lag 0–14 and lag 0–27 days, respectively. The greatest effect estimates of cold spells were found among total respiratory diseases and COPD, with RR of 1.88 (1.65–2.11) and 1.88 (1.58–2.19), respectively. The elderly, less-educated individuals and residents in southern China were more vulnerable to cold spells. Conclusion: There are remarkable mortality effects of cold spells, with effect estimates varying with the definition of cold spell and subpopulations. Using the official definition of cold spells may fail to capture the mortality risk associated with cold spells. These findings may facilitate the development of cold alert warning system and preventive actions to the vulnerable populations. Keywords: Cold spell, Definition, Cause-specific mortality, Vulnerable population, China

Background Drought is a major natural disaster that causes severe social and economic losses. The prediction of regional droughts may provide important information for drought preparedness and farm irrigation. The existing drought prediction models are mainly based on a single weather station. Efforts need to be taken to develop a new multistation-based prediction model. Objectives This study optimizes the predictor selection process and develops a new model to predict droughts using past drought index, meteorological measures and climate signals from 32 stations during 1961 to 2016 in Shaanxi province, China. Methods We applied and compared two methods, including a cross-correlation function and a distributed lag nonlinear model (DLNM), in selecting the optimal predictors and specifying their lag time. Then, we built a DLNM, an artificial neural network model and an XGBoost model and compared their validations for predicting the Standardized Precipitation Evapotranspiration Index (SPEI) 1–6 months in advance. Results The DLNM was better than the cross-correlation function in predictor selection and lag effect determination. The XGBoost model more accurately predicted SPEI with a lead time of 1–6 months than the DLNM and the artificial neural network, with cross-validation R2 values of 0.68–0.82, 0.72–0.89, 0.81–0.92, and 0.84–0.95 at 3-, 6-, 9- and 12-month time scales, respectively. Moreover, the XGBoost model had the highest prediction accuracy for overall droughts (89%–97%) and for three specific drought categories (i.e., moderate, severe, and extreme) (76%–94%). Conclusion This study offers a new modeling strategy for drought predictions based on multistation data. The incorporation of nonlinear and lag effects of predictors into the XGBoost method can significantly improve prediction accuracy of SPEI and drought.

Several studies have attempted to predict ground PM2.5 concentrations using satellite aerosol optical depth (AOD) retrieval. However, over 70%–90% of aerosol retrievals are non-random missing, which limits and biases the estimation. To the best of our knowledge, this issue has not been well resolved to date. The aim of this study was to develop an interpolation technique to handle the missing data retrieval problem and to estimate the daily PM2.5 for a high coverage dataset with 3-km resolution in China by fitting the complex temporal and spatial variations. We developed a two-step interpolation method (i.e., the mixed-effect model and inverse distance weighting technology) to replace the missing values in AOD. Next, the extreme gradient boosting (XGBoost) technique that includes a non-linear exposure-lag-response model (NELRM) was proposed and validated to estimate the daily levels of PM2.5 across China during 2014–2015. After two steps of interpolation, the missing value rate of daily AOD data was reduced from 87.91% to 13.83%. The cross-validation (CV) R-square, root mean square error (RMSE) and mean absolute percentage prediction error (MAPE) of the interpolation were 0.76, 0.10 and 21.41%, respectively. The cross-validation for the prediction of daily PM2.5 resulted in R2 = 0.86, RMSE = 14.98, and MAPE = 23.72%. The results of this study indicate that the two-step interpolation method can largely resolve the non-random missing data problem and that the combined XGBoost methods have a good ability to estimate fine particulate matter concentrations.

The immense problem of missing satellite aerosol retrievals (Aerosol Optical Depth, (AOD)) detrimentally affects the prediction ability of ground-level PM2.5 concentrations and may lead to unavoidable biases. An appropriate missing-imputation method has not been well developed to date. This study developed a two-stage approach (AOD-imputation stage and PM2.5-prediction stage) to predict short-term PM2.5 exposure in mainland China from 2013–2018. At the AOD-imputation stage, geostatistical methods and machine learning (ML) algorithms were examined to interpolate 1 km satellite aerosol retrievals. At the PM2.5-prediction stage, the daily levels of PM2.5 were predicted at a resolution of 1 km, based on interpolated AOD and meteorological data. The statistical performances of the different interpolation methods were comprehensively compared at each stage. The original coverage of retrieved AOD was 15.46% on average. For the AOD-imputation stage, ML methods produced a higher coverage (98.64%) of AOD than geostatistical methods (21.43–87.31%). Among ML algorithms, random forest (RF) or extreme gradient boosted (XG-interpolated) AOD produced better interpolated quality (CV R2 = 0.89 and 0.85) than other algorithms (0.49–0.78), but XGBoost required only 15% of the computing time of RF. For the PM2.5 predicted stage, neither RF-AOD nor XG-AOD could guarantee higher accuracy in PM2.5 estimations (CV R2 = 0.88 (RF or XG-AOD) compared to 0.85 (original)), or more stable spatial and temporal extrapolation (spatial, (temporal) CV R2 = 0.83 (0.83), 0.82 (0.82), and 0.65 (0.61) for RF, XG, and original). For the AOD-imputation stage, the missing-filled efficiency depended more on external information, while the missing-filled accuracy relied more on model structure. For the PM2.5 predicted stage, efficient AOD interpolation (or the ability to eliminate the missing data) was a precondition for the stable spatial and temporal extrapolation, while the quality of interpolated AOD showed less significant improvements. It was found that XG-AOD is a better choice to estimate daily PM2.5 exposure in health assessments.

Background The accurate measurement of particulate matter (PM) provides a crucial basis for health impact assessment and pollution management and control. However, monitoring stations of air pollution are limited worldwide. Recently, some researchers have attempted to estimate the levels of PM based on remote sensing data, but the methods still need to be validated and further improved. Objectives This study aimed to develop a new model, to estimate daily ground-level PM2.5 concentrations using the fused aerosol optical depth (AOD) retrieved by the Moderate Resolution Imaging Spectro radiometer and meteorological information. Methods We combined generalized additive mixed-effects model with the log-linked Gaussian error distribution and non-linear exposure-lag-response model for AOD and meteorological measures, to estimate daily ground-level PM2.5 concentrations in 2014–2015 in Guangzhou, China. Results The PM2.5 concentration was significantly associated with AOD and meteorological measures. Compared to the log-linear model, the non-linear exposure-lag-response model had better model performance with a higher temporal (spatial) cross-validation R–square (0.81 (0.81) vs 0.67 (0.67)), and a smaller mean absolute percentage error (17.65% (16.90%) vs 21.22% (21.01%)). AOD explained about 15% variations of PM2.5 in the mixed-effect model. The planetary-boundary -layer-height-revised AOD and relative-humidity-revised PM2.5 did not significantly improve the model performance. Conclusion Considering the non-linear exposure-lag-response association between PM2.5 and AOD and meteorological factors can significantly increase the modelling ability to estimate PM2.5 concentrations.

Due to the COVID-19 outbreak, the Chinese government implemented nationwide traffic restrictions and self-quarantine measures from January 23 to April 8 (in Wuhan), 2020. We estimated how these measures impacted ambient air pollution and the subsequent consequences on health and the health-related economy in 367 Chinese cities. A random forests modeling was used to predict the business-as-usual air pollution concentrations in 2020, after adjusting for the impact of long-term trend and weather conditions. We calculated changes in mortality attributable to reductions in air pollution in early 2020 and health-related economic benefits based on the value of statistical life (VSL). Compared with the business-as-usual scenario, we estimated 1239 (95% CI: 844-1578) PM2.5-related deaths were avoided, as were 2777 (95% CI: 1565-3995) PM10-related deaths, 1587 (95% CI: 98-3104) CO-related deaths, 4711 (95% CI: 3649-5781) NO2-related deaths, 215 (95% CI: 116-314) O3-related deaths, and 1088 (95% CI: 774-1421) SO2-related deaths. Based on the reduction in deaths, economic benefits for in PM2.5, PM10, CO, NO2, O3, and SO2 were 1.22, 2.60, 1.36, 4.05, 0.20, and 0.95 billion USD, respectively. Our findings demonstrate the substantial benefits in human health and health-related costs due to improved urban air quality during the COVID lockdown period in China in early 2020.

Background: Limited evidence is available on the health effects of particulate matter (PM including PM2.5 with an aerodynamic diameter ≤ 2.5 μm; PM10, ≤ 10 μm; PM2.5–10, 2.5–10 μm) during the pandemic of COVID-19 in Italy. The aims of the study were to examine the associations between all-cause mortality and PM in the pandemic period and compare them to the normal periods (2015–2019). Methods: We collected daily data regarding all-cause mortality (stratified by age and gender), and PM concentrations for 107 Italian provinces from 1 January 2015 to 31 May 2020. A time-stratified case-cross design with the distributed lag non-linear model was used to examine the association between PM and all-cause mortality. We also compared the counts and fractions of death attributable to PM in two periods. Results: Italy saw an increase in daily death counts while slight decreases in PM concentrations in pandemic period. Each 10 µg/m3 increase in PM was associated with much higher increase in daily all-cause mortality during the pandemic period compared to the same months during 2015–2019 (increased mortality rate: 7.24% (95%CI: 4.84%, 9.70%) versus 1.69% (95%CI: 1.12%, 2.25%) for PM2.5; 3.45% (95%CI: 2.58%, 4.34%) versus 1.11% (95%CI: 0.79%, 1.42%) for PM10; 4.25% (95%CI: 2.99%, 5.52%) versus 1.76% (95%CI: 1.14%, 2.38%) for PM2.5–10). The counts and fractions of deaths attributable to PM were higher in 2020 for PM2.5 (attributable death counts: 20,062 versus 3927 per year in 2015–2019; attributable fractions: 10.2% versus 2.4%), PM10 (15,112 versus 3999; 7.7% versus 2.5%), and PM2.5–10 (7193 versus 2303; 3.7% versus 1.4%). Conclusion: COVID-19 pandemic increased the vulnerability and excess cases of all-cause mortality associated with short-term exposure to PM2.5, PM2.5–10, and PM10 in Italy, despite a decline in air pollution level.

The global climate has changed significantly, characterized by the warming of the surface air temperature, which seriously affects public health. We examined the trends of extreme temperatures, heat waves and cold spells in a subtropical city of Guangzhou, China, during 1951–2015. Specifically, the relationship between ENSO (El Niño–Southern Oscillation) events and heat waves/cold spells was discussed. The results of linear regression showed the annual mean temperature and extreme warm days increased (0.14 °C/decade and 6.26 days/decade) while extreme cold days decreased significantly (1.77 days/decade). Heat waves were more frequent, longer lasting and had stronger intensity over the past 65 years. In addition, the frequency, duration and intensity of heat waves were correlated with annual Atlantic Multi-decadal Oscillation (AMO) and Indian Ocean Basin-wide Warming (IOBW), while there were no significant differences in the characteristics of heat waves among an El Niño year, a La Niña year and a Neutral year. In contrast, neither significant trend nor association with ENSO events was observed for cold spells. In conclusion, our study indicated an obvious increasing trend for all aspects of heat waves in Guangzhou, China.

Objective Ambient particulate pollution, especially PM2.5, has adverse impacts on health and welfare. To manage and control PM2.5 pollution, it is of great importance to determine the factors that affect PM2.5 levels. Previous studies commonly focused on a single or several cities. This study aims to analyze the impacts of meteorological and socio-economic factors on daily concentrations of PM2.5 in 109 Chinese cities from January 1, 2015 to December 31, 2015. Methods To evaluate potential risk factors associated with the spatial and temporal variations in PM2.5 levels, we developed a Bayesian spatio-temporal model in which the potential temporal autocorrelation and spatial autocorrelation of PM2.5 levels were taken into account to ensure the independence of the error term of the model and hence the robustness of the estimated parameters. Results Daily concentrations of PM2.5 peaked in winter and troughed in summer. The annual average concentration reached its highest value (79 μg/m3) in the Beijing-Tianjin-Hebei area. The city-level PM2.5 was positively associated with the proportion of the secondary industry, the total consumption of liquefied petroleum gas and the total emissions of industrial sulfur dioxide (SO2), but negatively associated with the proportion of the primary industry. A reverse U-shaped relationship between population density and PM2.5 was found. The city-level and daily-level of weather conditions within a city were both associated with PM2.5. Conclusion PM2.5 levels had significant spatio-temporal variations which were associated with socioeconomic and meteorological factors. Particularly, economic structure was a determinant factor of PM2.5 pollution rather than per capita GDP. This finding will be helpful for the intervention planning of particulate pollution control when considering the environmental and social-economic factors as part of the strategies.