The aim of the research was to compare the reliability of the methods used to estimate the parameters of the soil water retention curve (SWRC) from Hungarian soil map information and to investigate how the methods could be improved, using 11,470 soil horizon data series from the Hungarian Soil Hydrophysical Dataset (the MARTHA dataset). Among the methods found in the literature, the SWRC estimation method has only yet been tested in Hungary for the Kreybig Digital Soil Information System (BAKACSI et al., 2012). These authors determined the FAO texture class (FAO, 1995) of the given soil on the basis of soil hygroscopic data (hy). Then class pedotransfer functions (class PTFs) derived on the HYPRES dataset by WÖSTEN et al. (1999) and on the HUNSODA dataset by NEMES (2003) were used to estimate van Genuchten parameters of the SWRC for the mapped texture classes (HYPRES_hy and HUNSODA_hy). The relationship between hy and the five FAO texture classes was then tested on the MARTHA dataset following the procedure of BAKACSI et al. (2012). Texture was also estimated on the basis of the upper limit of plasticity according to Arany (KA). The van Genuchten parameters of the characteristic SWRC for each FAO texture class were calculated on the training set of MARTHA using the method of WÖSTEN et al. (1999). The calculation was first carried out for soil samples having at least three measured water retention values (MARTHA_min3pF) and then only for those where at least five θ(h) data pairs were available (MARTHA_min5pF). It was found that the FAO texture class of soil samples could be assigned more efficiently on the basis of KA than using hy. In cases where data on the particle size distribution were not available and FAO texture class was given on the basis of soil hygroscopicity, the reliability of SWRC estimation was significantly worse. For Hungarian soil samples, SWRC estimation methods derived on the MARTHA dataset were found to be significantly more reliable than the HYPRES and HUNSODA methods. The SWRC estimations calculated from hy were significantly more reliable for this dataset than those of HYPRES method of WÖSTEN et al. (1999), despite the fact that the latter was not influenced by errors in texture classification.

TÓTH Brigitta;  MAKÓ András;  TOTH Gergely;  CSILLA Farkas;  KÁLMÁN Rajkai; 

2014-08-18

AKADEMIAI KIADO

JRC83462

0002-1873,   

http://www.akademiai.com/content/6m6g837223310471/,    https://publications.jrc.ec.europa.eu/repository/handle/JRC83462,   

10.1556/Agrokem.62.2013.1.1,