A simple model structure enhances parameter identification and improves runoff prediction in ungauged high-latitude catchments

Abstract

Using data from 22 catchments in Northern Finland, this study demonstrates how runoff prediction in ungauged high-latitude catchments can be improved using simplified conceptual models tailored to readily available data in the region. The newly developed current precipitation index (CPIsnow) model provides a parsimonious tool to predict streamflow in data-limited high-latitude regions. Use of current precipitation index (CPI) and runoff coefficient formulation in the soil moisture and response routine conceptualization makes soil moisture and response routine parameters in the model identifiable with catchment properties influencing storage, evapotranspiration, and infiltration losses. Most of the model parameters showed a significant relationship with observable physical catchment properties, climate properties, or both. This made it easier to estimate the values of the model parameters in an ungauged catchment with minimal uncertainty. The parameter functional relationships derived with catchment attributes produced equally good model results when applied to independent test catchments. Inclusion of snow-water equivalent records and use of multiple objective functions for both snow-water equivalent and runoff simulations in the model optimization process helped reduce the effect of parameter equifinality, thereby making it easier to determine the optimal parameter values. Ranges of parameter values are proposed for the CPIsnow model in relation to key observable catchment characteristics in the boreal zone, enabling application of the model in ungauged catchments in the region.