Fig. 9 | Scientific Reports

Fig. 9

From: Development and optimization of a morphodiversity model for mountainous areas using supervised classification and artificial neural networks

Fig. 9

Patterns in the data used for training the neural networks in the RCMSC-ANN variants of the analysis (pattern locations after Mastej & Bartuś36. SC-ANN pattern classes: 1—non-morphodiversity class (NMDRCMSC-ANN); 2—morphodiversity class (MDRCMSC-ANN). Symbols within the objects indicate: T—training sample; U—test sample; V—validation sample. Terrain Relief (TR) model classes: 3—non-morphodiversity class (NMDTR); 4—morphodiversity class (MDTR). Peak types: 5—dome-shaped; 6—cone-shaped. Ridge types: 7—broad, rounded; 8—narrow, rounded. Valley types: 9—flat-bottomed; 10—V-shaped; 11—antecedent gorge; 12—alluvial terrace. Other landforms and features: 13—flattened surface; 14—landslide; 15—stone rubble; 16—tors and klippen; 17—rivers and streams; 18—water bodies; 19—Pieniny National Park; 20—national border.

Back to article page