Wyniki wyszukiwania - BazTech

Ograniczanie wyników

Znaleziono wyników: 2

Liczba wyników na stronie

Wyniki wyszukiwania

Wyszukiwano:
w słowach kluczowych: forecasting of air pollution

Sortuj według:

Ogranicz wyniki do:

Ocena wpływu stężeń zanieczyszczeń powietrza w GOP-ie na jakość powietrza w rejonie Opola i Kędzierzyna-Koźla

Tumidajski T., Siewior J., Foszcz D., Niedoba T.

Rocznik Ochrona Środowiska

2014

Tom 16, cz. 1

519--533

The stochastic models of air pollution spreading may be used as good description of its changeability considering mean daily changes of mean temperature, winds velocity and their directions. Applying the earlier elaborated forms of models (equations 1 and 2) and the collected data the equations for Opole and Kędzierzyn-Koźle were determined. The obtained results were very satisfying (high statistical significance of models and their coefficients) for models a posteriori (equation 1) and models with applied dummy variables (equation 2) describing directions of air flow. It should be stated that the elaborated type of model is correct for every type of conditions for which it was designed what means evaluation of average daily concentrations of pollution in selected areas. The correctness of the model form should be verified by the more precise analyzes of this type of modeling conducted for other areas, not necessary located in the neighborhood of such polluted region as Upper Silesia.

Prognozowanie stężeń zanieczyszczeń powietrza w GOP-ie modelami statystycznymi

Siewior J., Tumidajski T., Foszcz D., Niedoba T.

Rocznik Ochrona Środowiska

2011

Tom 13

1261-1274

Propagację zanieczyszczeń powietrza można opisywać za pomocą modeli deterministycznych, wykorzystujących równania różniczkowe fizyki atmosfery lub modeli fenomenologicznych analizujących statystycznie zebrane dane pomiarowe. Zgodność obu typów modeli z danymi rzeczywistymi jest różna, zakres i sens ich stosowalności także są inne i wielokrotnie wybór modelu jest określony celem jego stosowania oraz posiadanym zbiorem danych. W przypadku analizy propagacji zanieczyszczeń w dużych aglomeracjach przemysłowych stosowanie modeli deterministycznych jest praktycznie niemożliwe ze względu na potrzeby w zakresie danych (charakterystyka źródeł topografii terenu, charakterystyk meteorologicznych itp.) i dlatego próbuje się wykorzystać fenomenologiczne podejście do problemu opisu propagacji z zastosowaniem różnych technik obliczeniowych - od prostych metod statystycznych do sieci neuronowych.

Predicting of pollutants concentrations is very crucial for big cities and industrial agglomerations. It allows introduction of possible preventive activities with purpose of prevention to accumulation of concentrations as well warning people of possible states being dangerous to health for people from risk group, i.e. elders, people with cardiovascular diseases or people with respiratory system diseases. Upper Silesian Industrial Region (GOP) is one of the most polluted regions in Poland. Because of the location of several important heavy industrial plants it is necessary to constantly monitor concentrations of various dust and gas pollutants in this area. The paper presents the possibilities of stochastic modeling of air pollutants on the basis of data collected by monitoring stations and their application to pollutants concentrations forecasting. The data from following heating seasons from monitoring station in Zabrze concerning SO2 concentrations was applied to the analysis. The obtained models were statistically evaluated. Efficient modeling of daily SO2 concentrations should be based on regressive models with weights as well on adaptive approach to concentrations modeling dependably on forecasted synoptic situations. Documented heuristically statistical models allow to state that at unambiguous synoptic situation distinguished by low temperatures and windless weather, average limit values of concentrations of pollutants are expected to be exceeded. It is a manifestation of the compatibility of models with reality. An example of this is smog in Cracow recorded before the end of 2010, at meteorological conditions described above. The use of statistical methods of description of measurement data from the station allows for more accurate interpretation and documentation of the existing quantitative and even qualitative dependencies, which is not provided by purely numerical methods of modeling and predicting (e.g. neural networks).