Distributed runoff modelling: Wetland runoff and its importance for spring-flood predictions
2007 (English)Report (Other (popular science, discussion, etc.))
The focus of this study was on the development and testing of a new fully-distributed modelling approach to improve simulations of stream discharge and spatial-temporal patterns of snow and soil moisture at the catchment scale. In this report we describe the implemented hydrological routines in detail and the novel conceptualisation of soil water dynamics. The latter routine was also compared to the commonly-used soil routine in the HBV model. Additionally we describe a new approach for simulating the effect of forests on snow accumulation using a GIS forest database available for all of Sweden. The basic utilized model concepts are introduced: Interception, snow accumulation and –melt, two alternative soil- and groundwater storage conceptualisations, potential and actual evaporation and routing of water flows. Furthermore the processing of meteorological data is described and regionalisation techniques for model parameters are discussed especially in regard to distributed modelling of snow accumulation. Techniques and general problems associated with the solution of ordinary differential equations are brought to mind and the solution strategy in the current modelling context is elucidated. The first part of this document is concluded by a brief note on the way the model was implemented and on its interactions with the user and third party software. Calibration and evaluation of the model was performed based on data from the Krycklan research catchment in Northern Sweden. The model was first calibrated on discharge data recorded at two internal points within the Kallkälsbäcken catchment (which is a nested subcatchment of the Krycklan catchment) and evaluated against discharge registered at the subcatchment outlet. Spatial predictions of soil moisture were compared to mapped wetlands by applying the model to the entire 68 km2 large Krycklan catchment. It was found that the new soil routine performed better than the standard HBV routine when wetland and forest discharges were to be simulated simultaneously using the same set of parameters. Spatial wetness patterns produced by the dynamic model compared well to mapped wetland and outperform the results achieved using the common static topographic wetness, TWI (Beven and Kirkby, 1979). The distributions of simulated water storage values were compared for wetlands and forested areas. This analysis also indicated that the new modelling approach might be more appropriate than the traditional HBV formulation.
Place, publisher, year, edition, pages
Elforsk, Stockholm , 2007. , 60 p.
, Elforsk rapport, 07:16
IdentifiersURN: urn:nbn:se:su:diva-21234OAI: oai:DiVA.org:su-21234DiVA: diva2:187760