Pixel Purity Index Applied to the Mapping of Degraded Soils by the Presence of Cangahuas in the Ilaló Volcano, Ecuador

Soil degradation is a severe problem in the northern region of Ecuador. Due to deforestation, expansion of the agricultural frontier, and poor tillage practices, the outcrop of cangahuas further aggravates. Remote Sensing allows mapping this type of subsoil, which is often related to eroded areas; the spatial resolution of free multispectral images contains more than one coverage. This means that techniques to discern the pure spectral signature of the object of interest are required. Pixel Purity Index (PPI) is an endmember extraction algorithm capable of selecting the pure pixel and classifying it better than object-oriented techniques. The study's objective was to map soils with outcrops of cangahua, by PPI applied to Landsat 8 images in Ilalo volcano and later performed a physicochemical characterization to know the magnitude of the soil degradation in the mapped areas. We used two models with PPI: SAM and LSU; both were compared with classifications based on three vegetation indexes. LSU obtained the best result (91.2% accuracy and 0.81 Kappa coefficient). The mapped cangahua was approximately 806.85 ha. The soil had an average porosity of 45%, a relative density of 2.271 g/cm3, low concentrations of nitrates, phosphates, and sulfates, electrical conductivity <500 µS/cm, and alkaline pH, this means there is soil degradation. The PPI method had good accuracy and was achieved in identifying cangahua outcrops, which demonstrated its potentiality in mapping land cover.