Water repellency prediction in high-organic Greenlandic soils: Comparing vis-NIRS to pedotransfer functions

Soil water repellency (SWR) is a common phenomenon across agricultural soils of South Greenland that can negatively affect soil functions. Existing methods to measure SWR as a function of water content (w) are laborious. This study was conducted to compare the potential of visible-near-infrared spectroscopy (vis-NIRS) as an alternative method to pedotransfer functions (PTFs) for predicting four SWR indices in 143 agricultural soils from South Greenland (clay, 0.016-0.172 kg kg(-1); organic carbon (OC), 0.009-0.241 kg kg(-1)). Pedotransfer functions were established by multiple linear regression based on OC, clay, and pH. Partial least squares regression (PLS-R) and interval PLS-R were applied to build vis-NIRS prediction models (vis-NIR range 400-2,500 nm). The area under the SWR-w curve (SWRarea) and the critical soil water content (w(non)) were accurately predicted by PTFs (R-2 = .90; R-adj(2) = .91) while the SWR after 60 degrees C pretreatment (SWR60), and the integrative repellency dynamic index (IRDI) were predicted less accurately (R-adj(2) = .36; R-adj(2) = .27). Vis-NIRS models with variable selection performed at a better or close to the same level of accuracy as PTFs (SWRarea, R-2 = .88; w(non), R-2 = .90; SWR60, R-2 = .63; IRDI, R-2 = .54). This study demonstrated vis-NIRS as a valuable alternative to PTFs for rapid assessment of SWR and as a tool for SWR mitigation for farmers in South Greenland. The results may well apply to other regions with similar texture and OC ranges, but further testing is required.