Self-organizing map analysis of planted soil Infiltration systems for treatment of log yard runoff

Infiltration through a soil system planted with four different species was assessed as a sustainable treatment method for logyard runoff at a field-scale experimental site in Sweden. Runoff was infiltrated through soil planted with Alnus glutinosa (L.) Gärtner (common alder), Salix schwerinii X viminalis (willow variety "Gudrun"), Lolium perenne (L.) (rye grass), and Phalaris arundinacea (L.) (reed canary grass). Analyses of variance and t tests showed no significant differences in treatment when the different plants were compared with each other, and there were also no significant differences between the tree and the grass species. The infiltration treatment was effective in reducing total organic carbon and total phosphorus in the runoff even when the loads on the infiltration system increased over time. The self-organizing map (SOM) model was applied to assess the relationships between different water quality variables and to predict total nitrogen (TN) and total phosphorus (TP) concentrations. A supplementary correlation analysis confirmed the key relationships among variables revealed by the SOM. The SOM model performed very well in predicting TN and TP concentrations with the help of other water quality variables, which can be measured more cost effectively.