Pedotransfer functions for estimating soil water retention curve of Sicilian soils. Issue 10 (24th August 2019)

• Record Type: Journal Article
• Authors: Castellini, Mirko; Iovino, Massimo
• Is Part Of: Archives of agronomy and soil science. Volume 65:Issue 10(2019)
• Page Start: 1401
• Page End: 1416
• Publication Date: 2019-08-24
• Subjects: Parametric pedotransfer functions -- water retention models -- evaporation method -- extended nonlinear regression technique -- pedotransfer functions recalibration; Horticulture -- Periodicals; Soils -- Periodicals; 630.5
• DOI: 10.1080/03650340.2019.1566710
• Languages: English
• ISSNs: 0365-0340

Abstract

ABSTRACT: Pedotransfer functions (PTFs) make use of routinely surveyed soil data to estimate soil properties but their application to soils different from those used for their development can yield inaccurate estimates. This investigation aimed at evaluating the water retention prediction accuracy of eight existing PTFs using a database of 217 Sicilian soils exploring 11 USDA textural classes. PTFs performance was assessed by root mean square differences (RMSD) and average differences (AD) between estimated and measured data. Extended Nonlinear Regression (ENR) technique was adopted to recalibrate or develop four new PTFs and Wind's evaporation method was applied to validate the effectiveness of the relationships proposed. PTFs evaluation resulted in RMSD and AD values in the range 0.0630–0.0972 and 0.0021–0.0618 cm 3 cm –3, respectively. Best and worst performances were obtained respectively by PTF-MI and PTF-ZW. ENR allowed to recalibrate PTF-MI and PTF-ZW with improvements of RMSD (0.0594 and 0.0508 cm 3 cm –3 ) and to develop two relationships that improved RMSD by 75–78% as compared to PTF-MI. The results confirmed the potential of ENR technique in calibrating existing PTFs or developing new ones. Validation conducted with an independent dataset suggested that recalibrated/developed PTFs represent a viable alternative for water retention estimation of Sicilian soils.