Geostatistical uncertainty of assessing air quality using high-spatial-resolution lichen data: A health study in the urban area of Sines, Portugal

Geographic uncertainties in external exposome studies: A multi-scale approach to reduce exposure misclassification

BACKGROUND

Many studies on environment-health associations have emphasized that the selected buffer size (i.e., the scale of the geographic context when exposures are assigned at people's address location) may affect estimated effect sizes. However, there is limited methodological progress in addressing these buffer size-related uncertainties.

AIM

We aimed to 1) develop a statistical multi-scale approach to address buffer-related scale effects in cohort studies, and 2) investigate how environment-health associations differ between our multi-scale approach and ad hoc selected buffer sizes.

METHODS

We used lacunarity analyses to determine the largest meaningful buffer size for multiple high-resolution exposure surfaces (i.e., fine particulate matter [PM2.5], noise, and the normalized difference vegetation index [NDVI]). Exposures were linked to 7.7 million Dutch adults at their home addresses. We assigned exposure estimates based on buffers with fine-grained distance increments until the lacunarity-based upper limit was reached. Bayesian Cox model averaging addressed geographic uncertainties in the estimated exposure effect sizes within the exposure-specific upper buffer limits on mortality. Z-tests assessed statistical differences between averaged effect sizes and those obtained through pre-selected 100, 300, 1200, and 1500 m buffers.

RESULTS

The estimated lacunarity curves suggested exposure-specific upper buffer size limits; the largest was for NDVI (960 m), followed by noise (910 m) and PM2.5 (450 m). We recorded 845,229 deaths over eight years of follow-up. Our multi-scale approach indicated that higher values of NDVI were health-protectively associated with mortality risk (hazard ratio [HR]: 0.917, 95 % confidence interval [CI]: 0.886-0.948). Increased noise exposure was associated with an increased risk of mortality (HR: 1.003, 95 % CI: 1.002-1.003), while PM2.5 showed null associations (HR:0.998, 95 % CI: 0.997-1.000). Effect sizes of NDVI and noise differed significantly across the averaged and prespecified buffers (p < 0.05).

CONCLUSIONS

Geographic uncertainties in residential-based exposure assessments may obscure environment-health associations or risk spurious ones. Our multi-scale approach produced more consistent effect estimates and mitigated contextual uncertainties.

An iterative optimization scheme to accommodate inequality constraints in air quality geostatistical estimation of multivariate PM

The kriging-based estimation of the different types of atmospheric particulate matter (PM) pollutions defined in the air quality regulation raises some operational problems because the (co)kriging equations are obtained by minimizing a linear combination of the estimation variances subject to unbiasedness constraints. As a consequence, the estimation process can result in total PM₁₀ concentrations that are less than the PM_2.5 concentrations which would be physically impossible. In a previous publication, it was shown that a convenient external drift modeling can reduce the number of spatial locations where the inequality constraint is not satisfied, without completely solving the problem. In this work, the formulation of the cokriging system is modified, inspired by previous works focusing on positive kriging. The introduction of additional constraints on the cokriging weights are presented, leading to a unique and optimal solution to the problem of cokriging under inequality constraints between two variables. Some computational and algorithmic details are introduced. An evaluation of the penalized cokriging is provided by using the European PM monitoring sites dataset: some maps and performance scores are given to assess the relevance of our iterative optimization scheme.

Comparison of machine learning and deep learning techniques for the prediction of air pollution: a case study from China

The adverse effect of air pollution has always been a problem for human health. The presence of a high level of air pollutants can cause severe illnesses such as emphysema, chronic obstructive pulmonary disease (COPD), or asthma. Air quality prediction helps us to undertake practical action plans for controlling air pollution. The Air Quality Index (AQI) reflects the degree of concentration of pollutants in a locality. The average AQI was calculated for the various cities in China to understand the annual trends. Furthermore, the air quality index has been predicted for ten major cities across China using five different deep learning techniques, namely, Recurrent Neural Network (RNN), Bidirectional Gated Recurrent unit (Bi-GRU), Bidirectional Long Short-Term Memory (BiLSTM), Convolutional Neural Network BiLSTM (CNN-BiLSTM), and Convolutional BiLSTM (Conv1D-BiLSTM). The performance of these models has been compared with a machine learning model, eXtreme Gradient Boosting (XGBoost) to discover the most efficient deep learning model. The results suggest that the machine learning model, XGBoost, outperforms the deep learning models. While Conv1D-BiLSTM and CNN-BiLSTM perform well among the deep learning models in the estimation of the air quality index (AQI), RNN and Bi-GRU are the least performing ones. Thus, both XGBoost and neural network models are capable of capturing the non-linearity present in the dataset with reliable accuracy.

Spatially variable pesticide application in olive groves: Evaluation of potential pesticide-savings through stochastic spatial simulation algorithms

Site-specific management using spatial crown volume characterization can greatly reduce the amount of pesticides applied in agricultural treatments performed with air-assisted sprayers, while helping farmers achieve the European legislation on safe use of pesticides. Nevertheless, variable rate treatments in olive groves have received little attention. Thus, field research was conducted in a 20.6-ha traditional olive grove. Two attributes of the trees - tree crown volume (V) and tree projected area - were determined, using 67 samples for V and all trees of the field (1433) for tree projected area. Spatial continuity of both attributes was modelled with exponential variograms. To gain a measure of local uncertainty, stochastic simulation algorithms were applied. One hundred simulated images were obtained for tree projected area using direct sequential simulation. Tree projected area simulations were used to improve spatial prediction of V, more difficult and more expensive to obtain, taking advantage of the high linear correlation between both variables (r_xy = 0.72,p < 0.001). Thus, direct sequential cosimulation was employed to predict the spatial distribution of V, obtaining 100 geostatistical realizations of V. In order to estimate the potential reduction of pesticide use in the farm with variable rate treatments, two cut-off values of V were considered (50 and 100 m³_{crown volume}). Local uncertainty, understood as the probability of each tree belonging to a given crown volume interval was determined. Probability maps were further transformed to morphological maps and finally to variable prescription maps. Two scenarios with 2 and 3 management zones (MZs) were obtained. In comparison with a conventional phytosanitary application, the variable rate treatments could reduce the pesticide amounts by 21.3% with 2 MZs, and by 38% with 3 MZs. The joint use of V and tree projected area in stochastic sequential simulation algorithms has shown to be useful to determine MZs in olive groves.

A novel optimal-hybrid model for daily air quality index prediction considering air pollutant factors

Accurate and reliable air quality index (AQI) forecasting is extremely crucial for ecological environment and public health. A novel optimal-hybrid model, which fuses the advantage of secondary decomposition (SD), AI method and optimization algorithm, is developed for AQI forecasting in this paper. In the proposed SD method, wavelet decomposition (WD) is chosen as the primary decomposition technique to generate a high frequency detail sequence WD(D) and a low frequency approximation sequence WD(A). Variational mode decomposition (VMD) improved by sample entropy (SE) is adopted to smooth the WD(D), then long short-term memory (LSTM) neural network with good ability of learning and time series memory is applied to make it easy to be predicted. Least squares support vector machine (LSSVM) with the parameters optimized by the Bat algorithm (BA) considers air pollutant factors including PM_2.5, PM₁₀, SO₂, CO, NO₂ and O₃, which is suitable for forecasting WD(A) that retains original information of AQI series. The ultimate forecast result of AQI can be obtained by accumulating the prediction values of each subseries. Notably, the proposed idea not only gives full play to the advantages of conventional SD, but solve the problem that the traditional time series prediction model based on decomposition technology can not consider the influential factors. Additionally, two daily AQI series from December 1, 2016 to December 31, 2018 respectively collected from Beijing and Guilin located in China are utilized as the case studies to verify the proposed model. Comprehensive comparisons with a set of evaluation indices indicate that the proposed optimal-hybrid model comprehensively captures the characteristics of the original AQI series and has high correct rate of forecasting AQI classes.

Do Different Teams Produce Different Results in Long-Term Lichen Biomonitoring?

Lichen biomonitoring programs focus on temporal variations in epiphytic lichen communities in relation to the effects of atmospheric pollution. As repeated surveys are planned at medium to long term intervals, the alternation of different operators is often possible. This involves the need to consider the effect of non-sampling errors (e.g., observer errors). Here we relate the trends of lichen communities in repeated surveys with the contribution of different teams of specialists involved in sampling. For this reason, lichen diversity data collected in Italy within several ongoing biomonitoring programs have been considered. The variations of components of gamma diversity between the surveys have been related to the composition of the teams of operators. As a major result, the composition of the teams significantly affected data comparability: Similarity (S), Species Replacement (R), and Richness Difference (D) showed significant differences between “same” and “partially” versus “different” teams, with characteristics trends over time. The results suggest a more careful interpretation of temporal variations in biomonitoring studies.

Selecting lichen functional traits as ecological indicators of the effects of urban environment

Air pollution and the urban heat island effect are known to directly affect ecosystems in urban areas. Lichens, which are widely known as good ecological indicators of air quality and of climatic conditions, can be a valuable tool to monitor environmental changes in urban environments. The objective of this work was to select lichen functional traits and functional groups that can be used as ecological indicators of the effects of urbanization, with emphasis in the Southern subtropics, where this had never been done. For that, we assessed lichen functional composition in urban sites with different population density, which was considered as proxy for grouping sites in two levels of urbanization (low and medium/high). This a priori grouping was based on their significantly differences on air pollutants and land cover. Urbanization and air pollution showed to affect all lichen functional traits, with different responses depending on the functional group. Medium/high density urbanization was associated to an increase on the mean relative abundance of lichens with chlorococcoid green algae, foliose narrow lobes, soredia as the main reproduction strategy, pruinose thallus and containing secondary metabolites for chemical protection. Lower density urbanization showed a higher relative frequency of cyanolichens and lichens with Trentepohlia as the main algae, loosely attached crustose thallus and isidia as the main reproductive structure. The differences found on photobiont and growth form traits in response to the environmental variables used as proxies of microclimatic conditions (forest cover and number of trees around the sampling units), enabled us to detect the urban heat island effect (drier conditions in more urbanized sites).

Modelling local uncertainty in relations between birth weight and air quality within an urban area: combining geographically weighted regression with geostatistical simulation

In this study, we combine known methods to present a new approach to assess local distributions of estimated parameters measuring associations between air quality and birth weight in the urban area of Sines (Portugal). To model exposure and capture short-distance variations in air quality, we use a Regression Kriging estimator combining air quality point data with land use auxiliary data. To assess uncertainty of exposure, the Kriging estimator is incorporated in a sequential Gaussian simulation algorithm (sGs) providing a set of simulated exposure maps with similar spatial structural dependence and statistical properties of observed data. Following the completion of the simulation runs, we fit a geographically weighted generalized linear model (GWGLM) for each mother's place of residence, using observed health data and simulated exposure data, and repeat this procedure for each simulated map. Once the fit of GWGLM with all exposure maps is finished, we take the distribution of local estimated parameters measuring associations between exposure and birth weight, thus providing a measure of uncertainty in the local estimates. Results reveal that the distribution of local parameters did not vary substantially. Combining both methods (GWGLM and sGs), however, we are able to incorporate local uncertainty on the estimated associations providing an additional tool for analysis of the impacts of place in health.

Predicting polycyclic aromatic hydrocarbons using a mass fraction approach in a geostatistical framework across North Carolina

Currently in the United States there are no regulatory standards for ambient concentrations of polycyclic aromatic hydrocarbons (PAHs), a class of organic compounds with known carcinogenic species. As such, monitoring data are not routinely collected resulting in limited exposure mapping and epidemiologic studies. This work develops the log-mass fraction (LMF) Bayesian maximum entropy (BME) geostatistical prediction method used to predict the concentration of nine particle-bound PAHs across the US state of North Carolina. The LMF method develops a relationship between a relatively small number of collocated PAH and fine Particulate Matter (PM2.5) samples collected in 2005 and applies that relationship to a larger number of locations where PM2.5 is routinely monitored to more broadly estimate PAH concentrations across the state. Cross validation and mapping results indicate that by incorporating both PAH and PM2.5 data, the LMF BME method reduces mean squared error by 28.4% and produces more realistic spatial gradients compared to the traditional kriging approach based solely on observed PAH data. The LMF BME method efficiently creates PAH predictions in a PAH data sparse and PM2.5 data rich setting, opening the door for more expansive epidemiologic exposure assessments of ambient PAH.

Mapping Exposure to Multi-Pollutants Using Environmental Biomonitors-A Multi-Exposure Index

Atmosphere is a major pathway for transport and deposition of pollutants in the environment. In industrial areas, organic compounds are released or formed as by-products, such as polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/F's). Inorganic chemical elements, including lead and arsenic, are also part of the pollutants mixture, and even in low concentrations may potentially be toxic and carcinogenic. However, assessing the spatial pattern of their deposition is difficult due to high spatial and temporal heterogeneity. Lichens have been used as biomonitors of atmospheric deposition, because these organisms encompass greater spatial detail than air monitoring stations and provide an integration of overall pollution. Based upon the ability of lichens to concentrate pollutants such as PCDD/F and chemical elements, the main objectives of this study were to develop a new semi-quantitative multi-pollutant toxicity exposure index (TEQ-like), derived from risk estimates, in an attempt to correlate several atmospheric pollutants to human exposure levels. The actual pollutant concentrations were measured in the environment, from biomonitors (organisms that integrate multi-pollutants), enabling interpolation and mapping of contaminant deposition within the region. Thus, the TEQ-like index provides a spatial representation not from absolute accumulation of the different pollutants, but from the accumulation weighted by their relative risk. The assessment of environmental human exposure to multi-pollutants through atmospheric deposition may be applied to industries to improve mitigation processes or to health stakeholders to target populations for a comprehensive risk assessment, epidemiological studies, and health recommendations.

Traffic represents the main source of pollution in small Mediterranean urban areas as seen by lichen functional groups

The land-use type (residential, green areas, and traffic) within relatively small Mediterranean urban areas determines significant changes on lichen diversity, considering species richness and functional groups related to different ecological factors. Those areas with larger volume of traffic hold lower species diversity, in terms of species richness and lichen diversity value (LDV). Traffic areas also affect the composition of the lichen community, which is evidenced by sensitive species. The abundance of species of lichens tolerant to low levels of eutrophication diminishes in traffic areas; oppositely, those areas show a higher abundance of species of lichens tolerating high levels of eutrophication. On the other hand, residential and green areas have an opposite pattern, mainly with species highly tolerant to eutrophication being less abundant than low or moderate ones. The characteristics of tree bark do not seem to affect excessively on lichen composition; however, tree species shows some effect that should be considered in further studies.

The role of forest in mitigating the impact of atmospheric dust pollution in a mixed landscape

Atmospheric dust pollution, especially particulate matter below 2.5 μm, causes 3.3 million premature deaths per year worldwide. Although pollution sources are increasingly well known, the role of ecosystems in mitigating their impact is still poorly known. Our objective was to investigate the role of forests located in the surrounding of industrial and urban areas in reducing atmospheric dust pollution. This was tested using lichen transplants as biomonitors in a Mediterranean regional area with high levels of dry deposition. After a multivariate analysis, we have modeled the maximum pollution load expected for each site taking into consideration nearby pollutant sources. The difference between maximum expected pollution load and the observed values was explained by the deposition in nearby forests. Both the dust pollution and the ameliorating effect of forested areas were then mapped. The results showed that forest located nearby pollution sources plays an important role in reducing atmospheric dust pollution, highlighting their importance in the provision of the ecosystem service of air purification.

A novel hybrid model for air quality index forecasting based on two-phase decomposition technique and modified extreme learning machine

The randomness, non-stationarity and irregularity of air quality index (AQI) series bring the difficulty of AQI forecasting. To enhance forecast accuracy, a novel hybrid forecasting model combining two-phase decomposition technique and extreme learning machine (ELM) optimized by differential evolution (DE) algorithm is developed for AQI forecasting in this paper. In phase I, the complementary ensemble empirical mode decomposition (CEEMD) is utilized to decompose the AQI series into a set of intrinsic mode functions (IMFs) with different frequencies; in phase II, in order to further handle the high frequency IMFs which will increase the forecast difficulty, variational mode decomposition (VMD) is employed to decompose the high frequency IMFs into a number of variational modes (VMs). Then, the ELM model optimized by DE algorithm is applied to forecast all the IMFs and VMs. Finally, the forecast value of each high frequency IMF is obtained through adding up the forecast results of all corresponding VMs, and the forecast series of AQI is obtained by aggregating the forecast results of all IMFs. To verify and validate the proposed model, two daily AQI series from July 1, 2014 to June 30, 2016 collected from Beijing and Shanghai located in China are taken as the test cases to conduct the empirical study. The experimental results show that the proposed hybrid model based on two-phase decomposition technique is remarkably superior to all other considered models for its higher forecast accuracy.