Identification of Tile Drain Networks using Thermal and RGB Data from Unmanned Aircraft

The use of tile drainage is documented as far back as 200 B. C. and continues to be used in poorly drained agricultural regions throughout the world. Recent increases in annual precipitation throughout the mid-western United States, the potential for future regulation of tile, and more efficient installation methods for plastic tile have accelerated tile installation across the region. While good for crop production, the eco-hydrologic impacts of this modification have been shown to adversely affect natural drainage networks. Knowing the location of tile drain networks is essential to developing groundwater and surface water models. The geometry of tile networks installed decades ago has often been lost with time or was never well documented in the first place. Previous work has recognized that tiles can be observed for certain soil types in visible remote sensing data due to changes in soil albedo. The soil surface directly above the tile appears to have a lower soil moisture content due to strong water table gradients adjacent to tiles, causing a detectable color contrast at the surface. In this work, small Unmanned Aerial Systems (sUAS) were used to collect high resolution visible and thermal data to map tile drain patterns. Within less than 96 hours of a 12 mm rain event, a total of approximately 60 hectares of sUAS thermal and RGB data were acquired at two different locations at the Intensively Managed Lands Critical Zone Observatory in Illinois. Selected thermal images were co-registered with RGB images at known tile locations. The thermal imagery showed limited evidence of thermal contrast related to the tile, however, it is possible that a contrast could have been detected sooner after the rain event when greater thermal contrasts due to lower soil moisture proximal to tile would be expected. The RGB data, however, elucidated the tile entirely at one site and provided traces of the tile at the other site. These results illustrate the importance of the timing of sUAS data collection with respect to the precipitation event. Ongoing related work focusing on laboratory and numerical experiments to better quantify feedbacks between albedo, soil moisture, and heat transfer will help predict the optimal timing of data collection for applications such as tile mapping.

Raw project data is available by contacting [e-mail]