Testing approaches to simulate drainage from the canopy in interception models

Drainage from the canopy is one of the sub-processes described in conceptual interception models. In theory, this refers to rainfall water that is temporally stored on the canopy and starts to drip down, when the canopy is saturated. In most Rutter-type models, storage in the canopy is described as a single linear storage (in multilayer models as a storage cascade) from which the drainage occurs after a saturation value is reached. For a precise simulation of the timing and amount of canopy drainage, this approach appears to be insufficient. E.g., with large time-steps and highly filled storage, drainage rates can become larger than the precipitation intensity, leading to a negative storage.

In our study, we present a review of different approaches to simulate drainage according to literature. We test those approaches using the interception model CanWat, which allows temporally and spatially resolved simulations of the relevant processes. CanWat relies on a vegetation model derived from terrestrial laser scans for a detailed description of the stand. CanWat is written in R language allowing to process multi-annual time series for a study area up to 1 km² with a spatial resolution of 1 m³ on a PC. For validation purposes, we use rain events that are available from long-term continuous measurements of interception in a spruce stand (mainly Picea abies; a continuous flux site since started within EUROFLUX in 1996) within the Tharandter Wald Southwest of Dresden.

We outline benefits and limitations of each approach and evaluate which one best fits the needs for a precise simulation of the timing and amount of canopy drainage. The implementation and refinement of the most suitable drainage simulation approach into CanWat is part of the overall attempt to improve the model description of the interception process being the focus of a German Science Foundation (DFG BE-1721/23-1) interception project.