Comparison of Annual Prediction Methods for Spring Flow in the Aggtelek Region

Abstract

Karst spring flow plays an increasing important role in groundwater resources in Hungary. This paper evaluates three different estimation methods to predict mean annual spring flows in the Aggtelek region, using GIS based catchment area. Annual spring flow was predicted by two regional regression equations, by applying the Budyko equations, and by the original and modified Maucha method. Using measured spring flows, precipitation and temperature data between 1975-1992 each method was evaluated and compared for 12 spring location in the Aggtelek region. Neither method was found significantly better than the others. The Budyko curves gave a good estimation for annual spring flows, with average variance. The non-linear regression method gave the best result, with the smallest median error, and error variance.