Mapping drainage ditches in agricultural landscapes using LiDAR data

The aim of this study is to develop a method for identification of the drainage ditch network, which can be used for surface runoff modeling and to increase accuracy of estimation of greenhouse gas (GHG) emissions in croplands and grasslands, using remote sensing data. The study area consists of 11 objects throughout Latvia with a total area of 145 km². Digital elevation models (DEMs) in two resolutions, which were created using three different interpolation methods, were used for the analysis. Several multi-level data filtering methods were applied to identify ditch network, including flow patterns, which can be used in surface runoff process. The method we developed correctly identified 85–89% of ditches, depending on the DEM used, in comparison to the reference data. Mapped ditches are located within 3 m range of the reference data in 89–93% of cases. The elaborated model is robust and uses openly available source data and can be used for large scale ditch mapping with sufficient accuracy necessary for hydrological modelling and GHG accounting in the national inventories.