Summary
The aim of this thesis was verify the possibilities of fog modelling in GIS (geographic informational system). One of the research direction on the Department of geoinformatics on the Palacky University is the spatial modelling of geographic phenomenon in GIS. The fog is one of these phenomena and fog modelling is necessary particularly because of its negative effects on surroundings. Understanding the complexity of fog formation is a basis for building a suitable model for fog modelling.
Theoretical part of the thesis describe fog definition, inclusion, examples of fog occurrence in Czech Republic, conditions for creation, types according to intensity and formation, forecast. Other chapter summarize present models for fog forecast and studies with fog modelling. There are also basic information about the airport Brno-Tuřany and the airport Ostrava, its climatic classification, devices for visibility measurement and measuring principle and data description.
The first step of the practical part was pre-processed provided data. It includes description of each parameter of basic meteorological report METAR with concrete example on record of the airport data. These parameters are header, wind, visibility, runway visibility, type of the actual weather, air pressure, additional information and forecast, air temperature and temperature of dew point. There were generated only records with the state of the fog according to column "type of weather". The airport Brno_Tuřany has 178 record with fog from total number of records 8576 for 2010. And the Airport Ostrava has 268 records with fog from 8753 records. The frequency of fog in individual values of all parameters was detected in ArcGIS 10 and resulting table were adjusted in MS Excel, where I created also graphs.
There was selected variables for modelling, which has available data from airports. Data is related to one point, so it was chosen a statistical processing. Variables were the air, temperature, the air pressure, the temperature of dew point and the wind speed. They were examined in statistical program RStudio, where it were detected its influence on the values of the visibility. There was used the correlation analysis with Spearman coefficient, where was the low dependence between variables on both airports. Other used methods were the correlation matrix and GLM models, which find influence of combination of parameters on the visibility. The resulting success of the Brno data model was about 8% and Ostrava data about 16%. It means that models are not suitable. This results is affected by non-inclusion of all parameters which influence the fog creation. For example detailed relief, which is important because literature say that greater fog occurrence is in valley. This thesis doesn't work with relief because data are related to the point and the interpolation method can't be used in this case.
The main parameter, which influences the fog is a humidity, but this parameter isn't measured within METAR report. Further There aren't included movement of the front, vertical movement, movement of the dirty air from surroundings and type of surface, which is important because the literature state that greater fog occurrence is near water surface.
The main problem of this thesis is the lack of data for suitable algorithm, which could include all parameters which influence the fog. The thesis evaluate only parameters, which has a data in METAR report from the airport Ostrava and Brno. It would have more detail network of points with meteorological records for interpolation of values for spatial modelling. So the detailed research of the atmosphere could help to create better model. For example one study in the USA is based on two-year observation the weather in New York with methods of remote sensing. It allowed to determined patterns in weather for fog creation for this area and GIS were used for analyse local factors as land use, altitude and slope of relief.