Forecasting of Drought Using Larg Scale Climate Signal by Neural Network.
2011
Jafari, Behnosh | Fatehi, Ahmad | Golam Ali, Shaban Ali
Now there are many research in climate change and drought in Iran and also in the world. The number of research in this field is low in Iran than other countries and mostly focused on simulation of rainfall and streamflow using local data. Using climatical signal in orther to monitoring and forecasting of rainfall and streamflow is new espesialy in Iran. In this research a new technique in Artificial Neural Network (ANN) and teleconnection data that is measured and saved regulatory far from us. The object of this research is seasonal forecasting of rainfall and streamflow (climatical drought and hydrological drought) using large scale climatical signals in North of Iran. The main part of the models use for simulation and forecasting is ANN. In the main part of this model for selection of input data, filtering network MIMO, MISO, SISO was used. The filtering network selects the best combination of data for input layer of ANN. The result show this ANN model with high ability in simulation can forecast seasonal rainfall with high reliability. Drought conditions assessment based on standardized precipitation index showed that Characteristics of drought of weather stations in the north of the country is almost identical. Two extreme drought events in the study area, at two stations can be seen in the period 2008-1968. Rainfall prediction results for the region 2 for the winter of 2008 showed that the whole study area, have near normal precipitation amount with a slight tendency towards a lower than normal. Key Words: Large Scale Climate Signal, Neural Network, River Flow, Forecasting, North of IRAN
显示更多 [+] 显示较少 [-]