Some Norwegian high-altitude catchments demonstrate a different hydrological response and runoff patterns compared to more lowland catchments within a region with a relatively uniform precipitation forcing. The runoff patterns are characterised by a modest but noticeable, snowmelt spring flood and then a steadily increasing runoff during spring and summer which declines in autumn and has a minimum in winter. This characteristic runoff pattern was brought to our attention when an otherwise well-performing rainfall-runoff model failed to give good simulation results for several high-elevation catchments in this region.
The rainfall-runoff model (the Distance Distribution Dynamics model, Skaugen and Onof, 2014) is of a quite sufficiently parsimonious nature so it was not possible to manipulate model parameters to obtain a better fit with observed runoff. The current hypothesis as to why simulations by the DDD model fail is that the model has a structure that cannot account for the hydrological processes in play.
Since the catchments that exhibit the described runoff pattern are located at high elevations where permafrost may be present, it is natural to investigate the possible influence of permafrost processes on these runoff patterns. In this master thesis, we want to examine the following hypothesis: The observed runoff pattern at high-altitude catchments, is caused by the presence of permafrost. In early spring, snowmelt is impeded by freezing onto the cold surface and as spring and early summer proceed an increasing active layer releases previously frozen water and increases relatively slow subsurface flow. In autumn the active layer is reduced and the subsurface flow decreases.
In the current version of the DDD model, runoff response from both impermeable (IP) and permeable (P) areas can be modelled. To mimic the process of a temporal varying active layer we can, in the model, vary the fractions of IP and P areas according to, for example, as a function of daily, weekly or monthly mean temperatures.
NVE will provide meteorological and hydrological data along with a starting point for the Julia model code of the DDD model.