Predicting Tile Drainage Discharge.

More than 50 % of Danish agricultural areas are expected to be artificial tile drained. Transport of water and nutrients through the tile drain system to the aquatic environment is expected to be significant. For different mitigation strategies such as constructed wetlands an exact knowledge of the water load coming from the tile drainage system is therefore essential. This work aims at predicting tile drainage discharge using dynamic as well as a statistical predictive models. A large dataset of historical tile drain discharge data, daily discharge values as well as yearly average values were used in the analysis. For the dynamic modelling, a simple linear reservoir model was used where different outlets in the model represented tile drain as well as groundwater discharge outputs. This modelling was based on daily measured tile drain discharge values. The statistical predictive model was based on a polynomial regression predicting yearly tile drain discharge values using site specific parameters such as soil type, catchment topography, etc. as predictors. Values of calibrated model parameters from the dynamic modelling were compared to the same site specific parameter as used in the statistical predictive model. Both the dynamic as well as the statistical model was validated from a second dataset based on a number newly established tile drain discharge stations. The work opens of for the possibilities to understand the dynamic of tile drain discharge on a temporal as well as spatial scale.