Urban Air Quality Forecasting Research Articles

A deep spatio-temporal learning network for continuous citywide air quality forecast based on dense monitoring data

As urban air pollution becomes a severe environmental and societal issue globally, there is an increasing need on making air quality forecasts to prevent health and capital loss. Currently, most of the studies only provide air pollution forecasts on single or multiple sites, rather than citywide information. Due to complex urban structure and spatial transportation of air pollutants, the distribution of citywide air quality can be highly varying, which makes it extremely challenging to make accurate forecasts. Here, we address this challenge by exploring the potential of combining the dense air quality monitoring data with the deep learning technique. By using the fine-granular PM2.5 concentrations data collected from 417 micro monitoring stations in Lanzhou City, China, we proposed an end-to-end method to predict future citywide PM2.5 concentrations at 500 m × 500 m × 1 h resolution. This method consists of a spatio-temporal transformation component and a spatio-temporal prediction component, of which the former is responsible for converting PM2.5 concentration data into spatial-maps, while the spatio-temporal prediction network (i.e., Air-PredNet in our study) based on ConvLSTM in the latter is responsible for capturing the spatio-temporal characterizations in the spatial-map sequences and finally generating citywide air quality forecasts. The results show that our method can accurately predict the spatial distribution and variations of PM2.5 concentration with an RMSE of 2.12 μg/m3 for 1-h prediction and an RMSE of 4.09 μg/m3 for 6-h prediction. In addition, we also proved the robustness of Air-PredNet based on the prediction performance for O3 and NO2. Our results demonstrate the potential and feasibility of dense monitoring data combined with advanced data science methods for citywide urban air quality forecast, which supports the urban atmospheric environment fine management and decision-making.

Roadside Air Quality Forecasting in Shanghai with a Novel Sequence-to-Sequence Model.

The establishment of an effective roadside air quality forecasting model provides important information for proper traffic management to mitigate severe pollution, and for alerting resident’s outdoor plans to minimize exposure. Current deterministic models rely on numerical simulation and the tuning of parameters, and empirical models present powerful learning ability but have not fully considered the temporal periodicity of air pollutants. In order to take the periodicity of pollutants into empirical air quality forecasting models, this study evaluates the temporal variations of air pollutants and develops a novel sequence to sequence model with weekly periodicity to forecast air quality. Two-year observation data from Shanghai roadside air quality monitoring stations are employed to support analyzing and modeling. The results conclude that the fine particulate matter (PM2.5) and carbon monoxide (CO) concentrations show obvious daily and weekly variations, and the temporal patterns are nearly consistent with the periodicity of traffic flow in Shanghai. Compared with PM2.5, the CO concentrations are more affected by traffic variation. The proposed model outperforms the baseline model in terms of accuracy, and presents a higher linear consistency in PM2.5 prediction and lower errors in CO prediction. This study could assist environmental researchers to further improve the technologies for urban air quality forecasting, and serve as tools for supporting policymakers to implement related traffic management and emission control policies.

Urban air pollution estimation using unscented Kalman filtered inverse modeling with scaled monitoring data

The increasing rate of urbanization requires effective and reliable techniques for air quality monitoring and control. For this, the Air Pollution Model and Chemical Transport Model (TAPM-CTM) has been developed and used in Australia with emissions inventory data, synoptic data and terrain data used as its input parameters. Since large uncertainties exist in the emissions inventory (EI), further refinements and improvements are required for accurate air quality prediction. This study evaluates the performance of urban air quality forecasting, using TAPM-CTM, and improves accuracy of air pollution estimation by using a two-stage optimization technique to upgrade EI with validation from monitoring data. The first stage is based on statistical analysis for EI correction and the second stage is based on the unscented Kalman filter (UKF) to take into account the spatio-temporal distributions of air pollutant levels utilizing a Matérn covariance function. The predicted nitrogen monoxide (NO) and nitrogen dioxide (NO2) concentrations with apriori emissions are first compared with observations at monitoring stations in the New South Wales (NSW). Ozone (O3) is also considered since at the ground level it represents a major air pollutant affecting human health and the environment. In the second stage, with the improved EI, TAPM-CTM model errors are reduced further by using the UKF to calibrate EI. Results obtained show effectiveness of the proposed technique, which is promising for air quality inverse modeling, an important aspect of air pollution control in smart cities to achieve environmental sustainability.

Revisiting urban air quality forecasting: a regression approach

We address air quality (AQ) forecasting as a regression problem employing computational intelligence (CI) methods for the Gdańsk Metropolitan Area (GMA) in Poland and the Thessaloniki Metropolitan Area (TMA) in Greece. Linear Regression as well as Artificial Neural Network models are developed, accompanied by Random Forest models, for five locations per study area and for a dataset of limited feature dimensionality. An ensemble approach is also used for generating and testing AQ forecasting models. Results indicate good model performance with a correlation coefficient between forecasts and measurements for the daily mean hbox {PM}_{10} concentration one day in advance reaching 0.765 for one of the TMA locations and 0.64 for one of the GMA locations. Overall results suggest that the specific modelling approach can support the provision of air quality forecasts on the basis of limited feature space dimensionality and by employing simple linear regression models.

SURF: Understanding and Predicting Urban Convection and Haze

AbstractUrbanization modifies atmospheric energy and moisture balances, forming distinct features [e.g., urban heat islands (UHIs) and enhanced or decreased precipitation]. These produce significant challenges to science and society, including rapid and intense flooding, heat waves strengthened by UHIs, and air pollutant haze. The Study of Urban Impacts on Rainfall and Fog/Haze (SURF) has brought together international expertise on observations and modeling, meteorology and atmospheric chemistry, and research and operational forecasting. The SURF overall science objective is a better understanding of urban, terrain, convection, and aerosol interactions for improved forecast accuracy. Specific objectives include a) promoting cooperative international research to improve understanding of urban summer convective precipitation and winter particulate episodes via extensive field studies, b) improving high-resolution urban weather and air quality forecast models, and c) enhancing urban weather forecasts for societal applications (e.g., health, energy, hydrologic, climate change, air quality, planning, and emergency response management). Preliminary SURF observational and modeling results are shown (i.e., turbulent PBL structure, bifurcating thunderstorms, haze events, urban canopy model development, and model forecast evaluation).

Real-Time Forecasting Method of Urban Air Quality Based on Observation Sites and Thiessen Polygons

Abstract This paper proposes a real-time urban air quality forecasting method based on monitoring sites and Thiessen polygon. Firstly, the concentration of pollutants affecting air quality is obtained through the real-time observations of the monitoring sites deployed in wireless sensor network, according to which the air quality index (AQI) can be calculated and air quality levels and categories can be graded. Then, Thiessen polygons are constructed based on the monitoring sites and the air quality conditions from the only site within any polygon will be the representative to that of any other points within this polygon. Finally, the monitoring sites and Thiessen polygons rendered with standard air quality colors will be visualized on a geographical map for the realization of real-time forecasting method taking Thiessen polygons as forecasting units. Taking Shanghai city as an example, a real-time and visualized air quality prediction platform has been constructed in downtown Shanghai city, the real-time visualization of urban air quality forecasting and early warnings under the constraints of Thiessen polygon have been realized

A semiparametric statistical approach for forecasting SO2 and NO x concentrations

This study develops a new semiparametric statistical approach for urban air quality forecasting. Compared to conventional approaches, the semiparametric approach allows the model users to benefit from the positive aspects and alleviate the negative ones of parametric and nonparametric approaches. Two advantages of the approach lie in (1) the interpretation of the data set being easily decoded and used by the model and (2) its capability in dependence on prior assumption. To illustrate the performance of the proposed approach, three semiparametric regression models (i.e., linear-, quadratic-, and interactive-based semiparametric regression) are applied to an air quality forecasting problem in the city of Xiamen, China, and satisfactory training and prediction performance are obtained. The three models are also compared to three parametric and two nonparametric regression models. The results indicate that the predictive accuracy of semiparametric regression models is higher than those obtained from the parametric and stepwise cluster analysis models. However, the proposed three semiparametric regression models could be much favored, since they can be achieved more easily and rapidly than the artificial neural network model.

An Air Quality Predictive Model of Licang of Qingdao City Based on BP Neural Network

In order to obtain high precision results of urban air quality forecast, we propose a short-term predictive model of air quality in this paper, which is on the basis of the ambient air quality monitoring data and relevant meteorological data of a monitoring site in Licang district of Qingdao city in recent three years. The predictive model is based on BP neural network and used to predict the ambient air quality in the next some day or within a certain period of hours. In the design of the predictive model, we apply LM algorithm, Simulated Annealing algorithm and Early Stopping algorithm into BP network, and use a reasonable method to extract the historical data of two years as the training samples, which are the main reasons why the prediction results are better both in speed and in accuracy. And when predicting within a certain period of hours, we also adopt an average and equivalent idea to reduce the error accuracy, which brings us good results.

Regression Analysis and urban air quality forecasting: an application for the city of Athens

Air pollution forecasts in major urban areas are becoming a problem concerning the day to day environmental management for city authorities. This paper describes the development of an application to forecast the peak ozone levels with the aid of meteorological and air quality variables, in the Greater Athens Area. For this purpose, a number of regression models were considered, while the selection of the final model was based on extensive analysis and on literature. The model adapted includes variables that are available on a daily basis, so as daily operational maximum ozone concentration level forecast can be achieved.

Urban air quality and regional haze weather forecast for Yangtze River Delta region

Air pollution and haze weather have become more and more important environmental and meteorological issues in Yangtze River Delta (YRD) region of China. In order to foster urban and regional air quality management and realize operational prediction and early warning of air pollution and haze weather episode, an urban air quality forecasting system based on the new generation of weather research forecast and chemistry model WRF-Chem and a regional haze weather forecasting system based on Regional Atmospheric Environment Modeling System (RegAEMS) were applied in Shanghai, Nanjing and YRD area. More than one year runs and typical case studies show that WRF-Chem performed well in urban air quality forecast on surface concentrations of air pollutants such as SO2, NO2 and PM10. The accuracy rate of prediction on urban Air Pollution Index (API) is 50–83% and 80% for Shanghai and Nanjing, respectively. RegAEMS presents relatively good ability in forecast on regional haze weather. A new classification standard on haze level was proposed, which take the key parameters such as relative humidity, PM2.5 and visibility into account. It is estimated that RegAEMS predicts haze level with accuracy rate of 58 and 77% for Nanjing and Shanghai in YRD region. Many factors, including meteorology, emission inventory and chemical processes can be attributable to the forecast bias. However, from this study, it is highly suggested that improvements of emission inventory from construction dust, fugitive dust, soil dust, transportation and biomass burning are very crucial to get better performance on air quality and haze weather prediction.

Towards improving the simulation of meteorological fields in urban areas through updated/advanced surface fluxes description

Abstract. The increased resolution of numerical weather prediction models allows nowadays addressing more realistically urban meteorology and air pollution processes. This has triggered new interest in modelling and describing experimentally the specific features and processes of urban areas. Recent developments and results performed within the EU-funded project FUMAPEX on integrated systems for forecasting urban meteorology and air pollution are reported here. Sensitivity studies with respect to optimum resolution, parametrisation of urban roughness and surface exchange fluxes and the role of urban soil layers are carried out with advanced meso- or sub-meso meteorological models. They show that sensible improvements can be achieved by higher model resolution that is accompanied with better description of urban surface features. Recommendations, especially with respect to advanced urban air quality forecasting and information systems, are given together with an assessment of the needed further research and data.