Applications of Statistical Models and Artificial Neural Networks to Investigate Thermals in Turkey
This paper presents some applications of statistical models and artificial neural networks (ANN) to investigate thermals for two gliding regions in Turkey. ANN was used for modeling and prediction of favorable dry and wet thermal conditions. Long-term monthly, seasonal and annual averages of meteorological data (air temperature, solar radiation and sunshine duration) were used, in part, to ‘teach’ the ANN model. The ANN model exhibited skill with one-month-in-advance predictions; longer-range predictions were not attempted. By considering climate-change scenarios, predictions were made of these variables for the year of 2025.
CLEARANCE AND LICENSE TO PUBLISH:
This paper is UNCLASSIFIED (for public reasons) and has been cleared by the appropriate agencies, company and government. This paper represents original work by the author(s). No portion of the material is covered by a prior copyright; or for any portion copyrighted, the author has obtained permission for its use.
I hereby license OSTIV to publish this paper and to use it for all of OSTIV's current and future publications uses.