Acid sulfate (a.s.) soil mapping constitutes a fundamental step in order to plan and carry out effective mitigation at catchment scale. The main goal of this study was to assess the use of an artificial neural network (ANN) based on a Radial Basis Function (RBF) for a.s. soil mapping and characterization of soil properties relevant for environmental planning. This method was applied on the Sirppujoki River catchment (c. 440 km(2)), located in southwestern Finland. It required using various evidential datalayers (quaternary geology, slope and aerogeophysics) and point datasets (i.e. soil profiles) and enabled the creation of probability maps for a.s. soil occurrence, as well as predictive maps for different soil properties (sulfur content, organic matter content and critical sulfur depth). For the most accurate a.s. soil probability map, the high and very high probability areas cover about 10% of the whole study area (c. 42 km(2)) and contain all the known as, soil occurrences used as validation points. When considering the areas overlapping with the high and very high a.s. soil probability zones on the most accurate soil property predictive maps: (a) about 82% of these most probable areas display a predicted sulfur content between 03 and 1%, which is consistent with the values typically measured in the sulfidic horizons (i.e. between 02 and 1% in Finland); (b) the predicted organic matter content ranges between 5 and 15% in 98% of the areas of interest, indicating that sulfur contents greater than 03% are often associated with organic matter contents larger than 5%; (c) the very high as, soil probability areas mostly concur with the shallowest critical sulfide depth classes (0 to 0.4 m). Notably, the use of laser scanning (i.e. Light Detection And Ranging, LiDAR) data enabled a more precise definition of the different as, soil probability areas, as well as the sulfur content and critical sulfide depth predictive modeling classes. Therefore, the RBF-based ANN method represents a promising approach for a.s. soil mapping and characterization, enabling the creation of reliable as. soil probability maps and soil property predictive maps at catchment scale. (C) 2015 Elsevier B.V. All rights reserved.
- Acid sulfate soils
- Artificial neural network
- Radial Basis Function
- Soil probability mapping
- Soil property predictive modeling