Investigating geometrical characteristics of collapsed pipes and the changing role of driving factors

Determination of the amount (i.e., area and volume) of soil losses due to erosional landforms, especially collapsed pipes, plays a considerable role in different decision-making approaches. Further, mapping the spatial distribution and predicting the volumetric and areal losses of collapsed pipes (CPs) are essential for supporting ecosystem health. The study was conducted in relation to the area and volume of CPs and their related covariables. It focused on the estimation of soil losses due to collapsed pipes using unmanned aerial vehicle (UAV) images as well as field covariates at the Chatal Watershed, Golestan Province, Iran. A total of 481 soil samples were collected from homogeneous units with an area of approximately 1,410 ha. The potential relationship between the area/volume of collapsed pipes and land use, several topographic attributes (i.e., altitude, slope, and aspect), and soil properties, including soil stability, soil organic matter, clay, silt, and sand contents were analyzed using five distance-based methods (i.e., kernel density (KD), average nearest neighbor (ANN), spatial autocorrelation, hotspot analysis (HSA), and ordinary least square (OLS) analysis. The average nearest neighbor (Ratio = 0.12, Z score = -20.30, p-value < 0.05) and Moran space solidarity (Moran index = 0.258, Z score = 5.50, p-value < 0.05) showed the cluster distribution of area and volume of CPs. Hot spots and cold spots in the southwestern part of the study area were identified using KD and HSA. The relationship between existing independent and dependent variables (area of CPs) using regression analysis of OLS showed that slope and aggregate stability (>2.5 standard deviation) had the highest positive relationship with the dependent variable. Regarding the volume of CPs, land use (especially agricultural lands) had the strongest relationship with the dependent variable. Thus, geometrical characteristics of collapsed pipes can be applied as a quantitative indicator for the identification of hotspot zones (hazardous areas), land use planning, and erosion hazard mitigation. However, more studies are required to measure geometrical characteristics of soil landforms.