The effect of seasonal variation of precipitation/recharge on karst genesis behaviors in different climatic contexts

Recharge is an important factor controlling dissolution processes during the speleogenesis of karst aquifers. Previous numerical studies commonly assume a fixed, non-fluctuating potential (maximum) recharge rate. Thus, the karst genesis is controlled by two successive processes characterized respectively by a hydraulic head limitation (hydraulic control) followed by a flow rate limitation (catchment control). In this study, we simulate karst genesis under a fluctuating potential recharge from seasonal precipitation variations. This may lead to frequent switches between hydraulic control and catchment control during incipient karst genesis, depending on the size of hydrological catchment. Our simulation results show that, excluding recharge fluctuations may underestimate the size of dissolution area but overestimate fracture aperture enlargement over a long-term evolution. We define a characteristic time t c as the time required to establish the final dissolution pattern (i.e., the final geometry of the dissolving fracture network). With varying sizes of the hydrological catchment area, the predicted t c obtained under the assumption of fluctuating potential recharge rate is different from the one obtained for a fixed mean potential recharge rate. We further find that the intensity and total duration of precipitation, rather than the precipitation frequency, are the key parameters that control karst genesis behaviors. Merging the oscillating precipitation/recharge periods into an equivalent periodic function that has a larger period gives similar results. This simplification reduces computational cost for long-term (e.g., millions of years) simulations of karst genesis. In addition, the flow focusing during the karst genesis may be interrupted during the dry seasons, which can stall flow focusing and channelization. These findings have important implications for karst aquifer evolution and dame site leakage risk prediction.