Drainage network detection andquantification of water storage capacity within drainage channels in alluvial plains through LiDAR derived DTMs

Drainage networks in agrarian landscape within floodplains constitute man-made surfaces discontinuities, and they are expected to affect hydrological response during flood events. Drainage network recognition and quantification of water storage capacity within drainage channels, becomes therefore crucial for watershed planning and management. These evaluations require accurate spatial information for the area of interest and in most cases, when studying large catchments, broad datasets of ditches locations and descriptions are not available. In order to characterize drainage networks for large areas, the availability of high resolution topography (DTMs) derived by airborne laser scanner (LiDAR) represents a new tool for drainage systems recognition. LiDAR derived DTMs have been proven to be reliable and accurate for hydrological applications in steep mountain lanscape in several recent researches (Sofia et al. 2011, Tarolli and Dalla Fontana, 2009), where the topographic information has been considered at different resolution scales. Only a few studies have been conducted to take into account the specific characteristics of agrarian landscapes, and drainage network identification for agrarian/urbanized areas actually represent the next challenge when using high-resolution topography. The accuracy with which a DTM is able to detect and correctly represent the hydrological asset of a catchment is determined by the strength of the landscape gradient (i.e. flatness and/or slope changes). Alluvial plains are therefore more challenging even when high-resolution DTMs are available: for these environments, network extraction through GRID-derived flow-direction matrices is not applicable due to the low relief of the landscapes. The goals of this research were to propose a method able to (1) identify drainage networks in agrarian/floodplain context, and (2) to estimate some of the network summary statistics (i.e. network length, width, drainage density and storage capacity). A morphometric and semi-automatic methodology is proposed, that relies on consideration of having to extract local small-scale, low-relief features (ditches and channels) and to eliminate as far as possible the large-scale landscape forms from the data. The procedure is applied in two typical alluvial plains areas in the North East of Italy, and tested comparing automatically derived network information with field surveyed ones. The results underline the capability of high resolution topography for drainage network extraction and characterization in the context of agrarian landscapes within floodplains, opening at the same time a new challenge to evaluate hydrological processes in these areas.