Forecasting peak pollutant levels from meteorological variables

Ziomas, I.; Melas, Dimitrios; Zerefos, C. S.; Bais, A. F.; Παλιατσός, Α. Γ.

doi:10.1016/1352-2310(95)00131-h

Cited by 133 publications

(67 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The selected input variables and their aggregation are presented in Table 1. In principle, it might be surprising to use a linear approach, i.e., cross-correlation analysis, in order to select the inputs of non-linear models; nevertheless, such a procedure leads generally to satisfactory outcomes and constitutes a de facto standard for this kind of studies (see, for instance, Viotti et al, 2002;Nunnari et al, 1998;Ziomas et al, 1995;Balaguer Ballester et al, 2002). The average goodness of the prediction is firstly evaluated by a series of indicators, such as the true/predicted correlation ρ, the mean absolute error MAE = (1/D)( |y(t) −ŷ(t)|), the mean bias error MBE = (1/D)( y(t) −ŷ(t)), and the index of agreement d = 1 − ( (y(t) −ŷ(t)) 2 / (|ŷ(t) − y(t)| − |y(t) −ȳ (t)|)).…”

Section: Results: Ozonementioning

confidence: 99%

Air quality prediction in Milan: feed-forward neural networks, pruned neural networks and lazy learning

Corani

2005

Ecological Modelling

216

View full text Add to dashboard Cite

Ozone and PM 10 constitute the major concern for air quality of Milan. This paper addresses the problem of the prediction of such two pollutants, using to this end several statistical approaches. In particular, feed-forward neural networks (FFNNs), currently recognized as state-of-the-art approach for statistical prediction of air quality, are compared with two alternative approaches derived from machine learning: pruned neural networks (PNNs) and lazy learning (LL). PNNs constitute a parameterparsimonious approach, based on the removal of redundant parameters from fully connected neural networks; LL, on the other hand, is a local linear prediction algorithm, which performs a local learning procedure each time a prediction is required. All the three approaches are tested in the prediction of ozone and PM 10 ; predictors are trained to return at 9 a.m. the concentration estimated for the current day.No strong differences are found between the forecast accuracies of the different models; nevertheless, LL provides the best performances on indicators related to average goodness of the prediction (correlation, mean absolute error, etc.), while PNNs are superior to the other approaches in detecting of the exceedances of alarm and attention thresholds. In some cases, datadeseasonalization is found to improve the prediction accuracy of the models.Finally, some striking features of lazy learning deserve consideration: the LL predictor can be quickly designed, and, thanks to the simplicity of the local linear regressors, it both gets rid of overfitting problems and can be readily interpreted; moreover, it can be also easily kept up-to-date.

show abstract

Section: Results: Ozonementioning

confidence: 99%

Air quality prediction in Milan: feed-forward neural networks, pruned neural networks and lazy learning

Corani

2005

Ecological Modelling

216

View full text Add to dashboard Cite

show abstract

“…In general, although the characteristics of regional air quality depend on many complex elements, the major contributors are the emission of the pollutants and favourable largescale meteorological conditions (Z. H. ; X. L. . Ziomas et al (1995) pointed out that in an urban environment, the serious air pollution episodes are not attributed to sudden increases in the emission of pollutants but caused by meteorological conditions that are unfavourable for dispersion. Normally, the anthropogenic emissions of widespread pollutant sources are quasi-stable in east-…”

Section: Introductionmentioning

confidence: 99%

Impact of atmospheric circulations on aerosol distributions in autumn over eastern China: observational evidence

Zheng

Yang

et al. 2015

Atmos. Chem. Phys.

View full text Add to dashboard Cite

Abstract. Regional heavy pollution events in eastern China (110-122 • E, 28-40 • N) are causing serious environmental problems. In this study, the relationship between the degree of regional pollution and the patterns of large-scale atmospheric circulation over eastern China in October is investigated using 10-year (2001-2010) Terra/MODIS aerosol optical depth and NCEP reanalysis data by both case study and composite analysis. Eighteen polluted and 10 clean episodes are selected and categorised into six polluted types and three clean types respectively. Generally speaking, weather patterns such as a uniform surface pressure field in eastern China or a steady straight westerly in the middle troposphere, particularly when being at the rear of the anticyclone at 850 hPa, are typically responsible for heavy pollution events. Meanwhile, clean episodes occur when strong southeastward cold air advection prevails below the middle troposphere or air masses are transported from sea to land. Uniform descending motion prevails over the study region, trapping pollutants in the lower atmosphere. Therefore, the value of vertical velocity averaged from 1000 to 100 hPa and divergence of wind field in the lower troposphere are used in this study to quantify the diffusion conditions in each circulation type. The results reveal that it is often a clean episode when both the mean downward motion (larger than 2.56 × 10 −2 Pa s −1 ) and the divergence of lowlevel winds (larger than 1.79 × 10 −2 s −1 ) are strong. Otherwise, it is more likely to be a polluted episode.

show abstract

“…1a, b. The most frequent meteorological parameters determining the horizontal transport and dispersion of air particles are the mean wind speed and the wind direction (Ziomas et al 1995). The average monthly wind speed was 4.4 m s -1 , with values ranging from 2.9 to 6 m s -1 .…”

Section: Metrological Conditionsmentioning

confidence: 99%

Characteristics and composition of the falling dust in urban environment

Al‐Harbi

2013

Int. J. Environ. Sci. Technol.

View full text Add to dashboard Cite

The monthly total amount of dust fall, as well as its chemical and physical characteristics, was systematically investigated in Shuwaikh city, Kuwait. Dust samples were collected on a monthly basis for the entire year of 2009 and analyzed in the laboratory for water-soluble and water-insoluble matter. Water-insoluble matter represented the major portion of the total annual dust. ANOVAs showed significant temporal variation in the concentration of dust fall over the months (p \ 0.05). Higher dust deposits were encountered between June and August and ranged from 76.4 to 97.6 ± 2.5 (SD) ton km -2 month -1 , where dusty winds and low humidity are a common attribute in such arid areas. The main three soluble matter species measured are nitrate, sulfate, and chloride, and sulfate was found to be the most abundant inorganic species, ranging from 0.72 ± 0.13 to 4.1 ± 0.3 ton km -2 -month -1 . Major insoluble matter species measured are ash, silica, combustible, and tarry. Ash, silica, and combustible account for 63, 19, and 17.8 % of total insoluble dust, and 58.4, 17.7, and 16.6 % of total dust, respectively. Particle size distribution was also investigated, and results showed that dust particles[7 lm were the highest concentration of falling dust. Metrological conditions were found to play a vital role in temporal variations in falling dust.

show abstract

Forecasting peak pollutant levels from meteorological variables

Cited by 133 publications

References 8 publications

Air quality prediction in Milan: feed-forward neural networks, pruned neural networks and lazy learning

Air quality prediction in Milan: feed-forward neural networks, pruned neural networks and lazy learning

Impact of atmospheric circulations on aerosol distributions in autumn over eastern China: observational evidence

Characteristics and composition of the falling dust in urban environment

Contact Info

Product

Resources

About