Modelling the hydrological functioning of karst aquifer systems in Slovenia using geomorphological features and the random forest algorithm

19. December, 2025 in Znanstvene objave

Modelling the hydrological functioning of karst aquifer systems in Slovenia using geomorphological features and the random forest algorithm

Karst aquifers are important sources of drinking water, which are highly vulnerable in terms of both quality and quantity. Understanding their behaviour under different hydrological conditions is crucial for the effective management of these water resources. In the study described here, we investigated the link between the hydrological behaviour of karst aquifer systems and the geomorphological characteristics of their recharge backwaters. We analysed flow time series from 15 karst springs and their catchments in Slovenia. By analysing the recession curves of the hydrographs, we determined 11 key hydrological parameters describing the functioning of the aquifer system. By spatial processing of morphological, geological and hydrological data, we have identified 7 aggregated geomorphological characteristics of the catchments. These characteristics (independent variables) and key hydrological parameters (dependent variables) were used in a random forest model. By analysing the influence of the variables, we found that the area, density of karst caves and slope gradient are the most important geomorphological parameters of the backwatershed for predicting the hydrological characteristics of karst spring flow. The developed approach provides a methodological framework for predicting the hydrological characteristics of karst hinterland without measuring spring flows.

Summarised from the article JANŽA, Mitja, HUDOVERNIK, Valter, SERIANZ, Luka, STROJ, Andrej. Modeling hydrological functioning of karst aquifer systems in Slovenia using geomorphological features and random forest algorithm. Journal of hydrology. Regional studies. 2025, vol. 62, [article no.] 102774, 16 p. ISSN 2214-5818. DOI: 10.1016/j.ejrh.2025.102774. [COBISS.SI-ID 249995267]