Seasonal dynamics of curve numbers in mountain forest catchments of Czech Republic: Insights from 42 years of rainfall–runoff–snowmelt data

Study Region The study was conducted in the Červík experimental catchment, a 1.83 km² fully forested mountain catchment located in the Moravian-Silesian Beskids, Czech Republic. The catchment has been monitored since 1953 as part of a long-term forestry experiment and is characterized by seasonal snow cover that strongly influences winter hydrological processes. Study Focus Reliable estimation of runoff in mountain forest catchments with seasonal snow cover remains challenging, particularly under the increasing frequency of ROS episodes. This study evaluates the seasonal behaviour of Curve Number (CN) values using a unique 42-year daily record (1980–2022) of precipitation, snow water equivalent (SWE), and streamflow. A total of 862 runoff events were identified. Effective water input was defined as rainfall during the growing season (April–October) and as rainfall plus meltwater derived from changes in SWE during the dormant season (November–March). Event-based CN values were back-calculated and analysed according to season and event type. To place the catchment-scale results into a broader context, a nationwide analysis of snow-season conditions (November–April) was performed using daily SWE and precipitation data from 689 meteorological stations across the Czech Republic for the period 2000–2022. New Hydrological Insights for the Region Mean CN increased from 63.98 during the growing season to 83.88 during the dormant season, with ROS episodes contributing substantially to high winter event water inputs. Asymptotic fitting of the CN–(P + M) relationship yielded an asymptotic CN value of 52.6, indicating a standard hydrological response with decreasing CN values at higher event inputs. No significant long-term trend in annual mean CN was detected during 1980–2022. The nationwide analysis of SWE and precipitation further showed that seasonal snow accumulation, snowmelt, and ROS events occur in a large number of forest-dominated catchments across the Czech Republic. Given that more than 30% of Czech fourth-order catchments are forest-dominated, these findings highlight the importance of considering snow-related processes when applying the CN method in Czech forested catchments. The results demonstrate that standard rainfall-based CN applications underestimate event water inputs and direct runoff in forested regions affected by seasonal snow. Incorporating snowmelt and ROS contributions into CN estimation can improve runoff predictions and provide more realistic design values for winter-affected catchments.