Identifikacija hidroloških režima otjecanja u kršu konceptualnim i parametarskim modelima

Abstract

Water balance characteristic of karst aquifer consist of different water regime flows influenced not only with amount of precipitation but especially with heterogenic and hydrogeological processes which can’t be directly measured. Aside from measurement problems, identification of hydrological regime becomes more difficult if the available data is scarce and discontinuous. Although constant uncertainty is always present in predicting natural phenomena, it is possible to model estimators which can be used to analyse current state of the system and give indication trends of their main features. Relations between rainfall and runoff, the RR model, are becoming more prominent because of their capabilities to implement noisy data. RR model represents an assembly of different mathematical principles trying to describe functioning of river catchment. RR models use certain assumptions about shapes of recession curve, catchment area, base flow-quick flow ratio, soilmoisture content and so on, to determine key parameters of the water circulation process. If the parameters are successfully calibrated and tested, they can be used as new insight to the properties of regional karst. Then, calibrated model represents hydrological behaviour under conditions determined in available database. Simulation of complex karst aquifer requires conceptual and computational learning approach. Current state of computer technology is becoming ever more useful in analysing processes of limited knowledge especially for the fact that computational time is becoming shorter and available software is more flexible and understandable to end user. Aim of proposed models is defining key factors for successful representation of observed runoff. Conceptual model combines transfer function and imbeds these key factors. Parametric model uses transformation which is not transparent but gives better approximation of cost function. Systematic approach to architectural design of parametric model overcomes these problems. Implementation of new findings from both models gives better hydrological understanding of investigated case study in karst.