versión impresa ISSN 0187-6236
Global climate change causes a decrease of precipitation in Turkey, as in many other parts of the world. As a result, droughts have now occurred over a larger area and in a more drastic way than in the past. Determining the factors in the formation and early prediction of drought will allow required measures of prevention to be taken in time. The present study evaluates drought conditions on monthly and yearly bases, with the measurements of precipitation, wind, humidity and temperature taken in the Ankara region between 1926 and 2006 using the techniques of decision trees. The obtained results demonstrated that the province of Ankara has a generally normal and near-normal arid climate and that the precipitation amounts in all months and precipitation and wind in January, should be taken into consideration to determine such aridity.
Palabras llave : Standardized precipitation index; clustering; C5.0 algorithm; water shortage; time series; drought modeling; risk assessment.