MODELLING APPROACHES IN KARST AQUIFERS - SYSTEM UNDERSTANDING, CHARACTERISATION AND PREDICTION -

The paper presents a critical review and systematic classification of global (lumped parameter) and distributive methods for the modelling of karst hydrodynamics. Global methods are based on the analysis of spring discharge – precipitation time series. These data are believed to reflect the overall hydraulic characteristics of karst hydrogeological systems. Although these methods are considered to reflect the physical functioning of karst hydrogeological systems, they only allow qualitative interpretation because they neglect the spatial heterogeneity and the specific internal structure of karst. Distributed parameter models incorporate two concepts. Discrete approaches describe flow within networks of fractures or conduits. Continuum approaches treat heterogeneities in terms of effective model parameters and their spatial distribution. The Combined Discrete-Continuum (CDC) approach constitutes a hybrid version between continuum and discrete modeling approaches. The Equivalent Porous Medium Approach (EPM) neglects strong heterogeneity of the hydraulic parameter fields, and has only limited capabilities to model groundwater flow in karst aquifers. In the Double Continuum concept, the conduit network and the fissured medium are represented as two separate and hydraulically coupled continua. Sparse data demand of DC models makes this method a useful tool for modeling water resources or water budget problems. The CDC approach can handle discontinuities at all scales in karst systems. Although data concerning the hydraulic and geometric parameters of the conduit network are difficult to obtain, spring hydrograph analysis may provide indirect information necessary to build reliable discrete-continuum models. Furthermore, a systematic analysis of typical hydrogeological problem settings in regional karst groundwater catchments are presented. These problems are analysed in terms of the relative scale hierarchy of the various length scales of the investigated flow or transport domain, the flow dominating heterogeneities (e.g. conduits), the length scales of the investigation method and the source area. Appropriate distributed parameter model approaches, designed for the respective problem class are suggested.