Search

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. [ABSTRACT FROM AUTHOR]

Novel mapping methods using AI have led to improved mapping of the extent of drainage systems, but the full scope of the effects of drainage on ecosystems has yet to be understood. By combining ditches mapped with remote sensing and AI methods with soil data from the Swedish Forest Soil Inventory, and vegetation data from the National Forest Inventory we identified 4 126 survey plots within 100 m of a ditch. The inventory data span across three biomes; the northern boreal zone, the hemiboreal zone, and the temperate zone. We explored if soils and vegetation close to ditches were indeed different from the surrounding landscape. The large number of plots spread widely across the Swedish forest landscape spanning different physiographic regions, climates, topography, soils, and vegetation made it possible to identify the general effect of drainage on soil properties, tree productivity, and plant species composition. We found a surprisingly large amount of ditches on mineral soils (50–70%, depending on the definition of peatlands). Forest growth was affected, with higher growth rates of trees closer to ditches, particularly Norway spruce. Sphagnum mosses – a key indicator of wet soils - were less common near ditches, where they were replaced by feather mosses. The soil bulk density was higher closer to ditches, as was the concentration of metals that are typically associated with organic matter (Al), while concentrations of metals with a lower affinity for organic material decreased toward ditches (Na, K, Mg). The results from mineral soils and peat soils often differed. For example, N and tree volume increased toward ditches, but on different levels for peat and mineral soils, while the thickness of the humus layer and Pleurozium schreberi cover showed opposite patterns for the different soils. Clearly, ditches have affected the entire Swedish forest landscape, driving it towards a drier, more spruce-dominated productive forested ecosystem and away from wetland ecosystems like mires and littoral areas along streams. Furthermore, the biogeochemistry of the soils and understory species cover near ditches have changed, potentially irreversibly, at least within human time frames, and have implications for restoration goals and the future of forestry. • 50-70% of the drainage systems were found on mineral soils. • Ditches have changed the soil biogeochemistry. • Ditches have increased the tree volume and growth. • Ditches have changed understory species coverage. [ABSTRACT FROM AUTHOR]

Comparisons between field data and available maps show that 64% of wet areas in the boreal landscape are missing on current maps. Primarily forested wetlands and wet soils near streams and lakes are missing, making them difficult to manage. One solution is to model missing wet areas from high-resolution digital elevation models, using indices such as topographical wetness index and depth to water. However, when working across large areas with gradients in topography, soils and climate, it is not possible to find one method or one threshold that works everywhere. By using soil moisture data from the National Forest Inventory of Sweden as a training dataset, we show that it is possible to combine information from several indices and thresholds, using machine learners, thereby improving the mapping of wet soils (kappa = 0.65). The new maps can be used to better plan roads and generate riparian buffer zones near surface waters. [ABSTRACT FROM AUTHOR]