Effect of poultry farms contamination to underground water in Çorum and modelling of artificial neural network

Water has an indispensable importance for human beings and other animates. Today, in our country, especially in big settlement places, there has been a fresh water problem. The existing water sources are not able to be saved from the impact of the contaminators. 20 out of 107 poultry farms as sampling stations having representative ability have been chosen which work in an area near Derinçay Creek(Çorum Creek) and which are within the boundaries of Çorum. The poultry farms provide their own water from the wells that they had drilled. The depth of these wells vary between 15?90 m. The poultry farms collect poultry manure (containing 85% water) in the waste collecting store and then transport it to the storage field. In the farm field, the pre-stored water mixes into the surface and underground water due to the reasons like carrying, flood and leakage. During the study, coordinate determination has been made in 20 poultry farms having representative ability. The data like distance to the river, farm type, number of the poultry, well depth, the way of waste storage and daily waste amount have been taken. Later, in order to follow the seasonal changes, 4 well water samples have been taken in different dates and the analysis of amonia, nitrite, nitrate, phosphate, pH, conductivity, salinity, total dissolved solid, turbidity, total hardness and total coliform have been made. The contamination factor coliform has been found in all samples. The other contamination factor which is amonia varies between 0?4,02 mg/L. By spreading amonia on to the field waste depot has been observed. This situation shows that waste depot form affects the leaking amount. The acquired results took shape with the Artificial Neural Network (ANN) and this models has been tested. The accuracy test made by 60 samples have been tested by 20 samples and 100% accuracy has been found as a result. When well depth, soil structure, poultry number, farm type, waste amount are given as inputs for a farm whose analysis has not been made, ANN and total coliform amount can be determined with high accuracy. In addition to this, from Derinçay Creek, on 19 samples out of 12 determined sampling stations receiver milieu, water analysis has been made. In the study, due to high ratio of contamination, there has not been able to have an evaluation about the impact of poultry farms.