Infiltration characterization using principal component analysis and K-means cluster analysis on quaternary volcanic landscape at the southern flank of Rinjani Volcano, Lombok Island, Indonesia

Abstract

ABSTRACTComprehending the characteristics of infiltration is essential for conserving spring recharge areas, particularly on hillslopes of a Quaternary volcanic landscape. Quantitative investigations were carried out to identify soil physical properties and morphological features related to infiltration behavior. We used 15 transects with 70 sampling points in total to collect soil samples, observed morphological characteristics, and measured the infiltration. Data analysis employed statistical methods, including principal component analysis (PCA) and K-means cluster analysis. PCA revealed that infiltration was strongly correlated with slope, slope length, surface curvature, and slope position and that it was a function of soil physical properties, i.e., soil moisture, particle size distribution, and bulk density. The K-means cluster analysis classified the sampling points into four clusters that corresponded to three slope positions, namely upper slope, middle slope, and footslope. The high infiltration rate in the middle slope was caused by the relatively steep slope, sandy soil texture, low soil moisture, and low bulk density. On the contrary, the relatively flat slope, high clay content, high soil moisture, and high bulk density of the footslope created a low infiltration rate. According to these infiltration parameters, the preferred area for spring water catchment conservation was the middle slope.