Determination of back-trajectories is carried out by using the HYSPLIT model, which is widely used by air quality researchers to locate the source of a pollutant. There are several meteorology archives and vertical velocity methods that the model offers to its users.

Previous researches have shown that two different meteorology archives, GDAS (Global Data Assimilation System) and Reanalysis, and two different vertical velocity methods, kinematic and isentropic are mostly used in the studies performed in Turkey. It is not certain that which meteorology archive or vertical velocity method gives better results. Back trajectories calculated for the same region can be quite different due to the model inputs. Even the fact that some studies do not even need to give any information about the model inputs reveals the necessity of doing a study on this subject. Therefore, the differences in the model inputs should be analyzed and it is necessary to highlight the difference.

To date, no study has been conducted in Turkey. The most important contribution of the study is numerically presenting the results of two different model inputs for the first time in the Mediterranean Region which is the most studied region in Turkey.

In this study, to identify the effects of different model inputs on results, HYSPLIT back trajectories have been visualized by using MATLAB, cluster analyses were performed by using SPSS, and then clusters were visualized by using MATLAB.

Consequently, the study focused on a non-consensus practice that which meteorology archive and vertical velocity methods should be used among practitioners. In this way, researchers can reach a consensus if it is determined that the model inputs have a significant effect on the results.