Abstract:
The industrialization process which began in Western countries in the 19th century has brought about the problem of urbanization in the following years. Urban population is increasing rapidly in comparison to the rural population. Today, in almost any country industrialization and rapid urbanization adversely affect many of our environmental values, such as our core ecosystem, regional climate variations and global diversity. In this study, the effects of regional urbanization on vegetation were examined by using satellite data and atmospheric variables. In the vegetation analysis, 2005-2018 multi-time index values obtained from TERRA-MODIS satellite, EVI (Enhanced Vegetation Index) and LST (Land Surface Temperature) were taken into account. In the analysis, temperature and precipitation were chosen as atmospheric variables. The expected variations of EVI values of Catalca, one of the districts of Istanbul with the highest rates of population increase, until the year 2030 were estimated. NARX (nonlinear autoregressive exogenous) neural network, which gives successful results in the estimation of nonlinear data sets, was used for analysis. In addition, a hybrid W-ANN (Wavelet-Artificial Neural Network) model was developed using NARX and DWT (Discrete Wavelet Transform) models to increase estimated performance. In the light of the information obtained, W-ANN estimates improved by %4.3 compared to the prediction conditions made only with the NARX model.
