Özet:

Dew point temperature plays a significant role in determining many activities such as ecological, hydrological, and climatological models, especially agriculture. Dew point is an important factor for the growth and development of plants. Plants are negatively affected when the dew point temperature is high and varies. For these reasons, the dew point temperature prediction has become important in arid regions with sparse rainfall and in areas where greenhouse cultivation is intense. In this study, the artificial neural network (ANN) method was used to estimate the dew point temperature in Mugla Province based on relative humidity, pressure and temperature data from January 1, 2019 to December 31, 2019. The MATLAB programming language is used for this. The Levenberg-Marquardt (LM) algorithm and the Feed-Forward Back Propagation neural network model give the best results in ANN training, where there are 7 neurons in the hidden layer. The RMSE error value is 2.112987863, the R2 value is 0.887436814, and the cov coefficient of variation is 0.170439905. These values have been found to be at acceptable levels, and this method can be successfully used to estimate the dew point temperature. The artificial neural network model can be used to estimate the dew point temperature for different regions and different meteorological parameters. These estimates will be very helpful, especially for those engaged in agriculture.

Anahtar Kelimeler: