Forestry drainage has been used to improve timber growth and reforestation in wooded wetlands, primarily in northern European boreal forests, as well as North America. Called ditching, the process helps lower groundwater levels to improve soil aeration and promote tree growth. While this has merit, long-term use of ditching has an impact on the larger ecosystem. Environmental risks can include degradation of wetlands and soil, greenhouse gas emissions, an increase of nutrient and sediment loads in water bodies, as well as losses in biodiversity. There is, however, a lack of accurate data on ditch networks, with Swedish maps only identifying 9% of ditches.
Researchers sought to identify these ditch networks to prioritize sustainable land and hydrological management. Traditional satellite and aerial photography are not an option in a forest region where drainage ditches are obscured by the tree canopy, but airborne laser scanning integrated with powerful machine learning and statistical tools could work.
By combining deep learning, a technique of teaching computers to learn by example, with ALS, the authors could more accurately detect drainage ditches in forests. In “Mapping Drainage Ditches in Forested Landscapes Using Deep Learning and Aerial Laser Scanning,” by William Lidberg, Siddhartho Shekhar Paul, Florian Westphal, Kai Florian Richter, Niklas Lavesson, Raitis Melniks, Janis Ivanovs, Mariusz Ciesielski, Antti Leinonen, and Anneli M. Ågren for the Journal of Irrigation and Drainage Engineering, the authors used their methodology to manually digitize ditches from across 10 regions in Sweden. Learn more about this novel technique to improve hydrology management at https://doi.org/10.1061/JIDEDH.IRENG-9796. The abstract is below.
Extensive use of drainage ditches in European boreal forests and in some parts of North America has resulted in a major change in wetland and soil hydrology and impacted the overall ecosystem functions of these regions. An increasing understanding of the environmental risks associated with forest ditches makes mapping these ditches a priority for sustainable forest and land use management. Here, we present the first rigorous deep learning–based methodology to map forest ditches at regional scale. A deep neural network was trained on airborne laser scanning data (ALS) and 1,607 km of manually digitized ditch channels from 10 regions spread across Sweden. The model correctly mapped 86% of all ditch channels in the test data, with a Matthews correlation coefficient of 0.78. Further, the model proved to be accurate when evaluated on ALS data from other heavily ditched countries in the Baltic Sea Region. This study leads the way in using deep learning and airborne laser scanning for mapping fine-resolution drainage ditches over large areas. This technique requires only one topographical index, which makes it possible to implement on national scales with limited computational resources. It thus provides a significant contribution to the assessment of regional hydrology and ecosystem dynamics in forested landscapes.
Learn about the potential in this new technique in the ASCE Library: https://doi.org/10.1061/JIDEDH.IRENG-9796.