In:
Journal of Plant Nutrition and Soil Science, Wiley, Vol. 185, No. 3 ( 2022-06), p. 417-426
Abstract:
Though soil texture is one of the most basic soil characteristics its quantification needs still laborious procedures. A commercially available, efficient approach has been introduced as Pario classic method where silt and clay fractions are calculated by inverse fitting of transducer‐measured suspension pressure curves to modeled Stokes’ law of sedimentation. However, comparison of Pario ‐measured textures of 64 samples with Köhn method revealed unsatisfactory bias and random error of fractions except for medium silt. Aims The goal of our study was to improve precision and accuracy of Pario classic measurements by multiple linear regression models using regressors that were anyway available in the measurement procedure. Methods For the model we included two groups of regressors: (1) Pario ‐estimated clay, fine, medium, and coarse silt fractions to cover measurement data and dependencies between fractions; (2) parameters assessed during sample preparation including residual moisture of the air‐dried sample (ϴ res ), soil‐organic carbon (SOC), and pH‐value. The choice of regressors has been optimized according to the Aikaike Information Criterion (AIC). Results The final models yielded unbiased estimations and strongly reduced root‐mean square errors below 5 mass‐%. In case of clay and coarse silt, the intrinsic Pario estimated clay and coarse silt contributed strongest to the prediction, but all fractions of the Pario method have been included in the models. The most important external regressor was the residual moisture. It contributed positively to the clay model and negatively to the silt models as expected because it is a known proxy of clay content. Conclusions The suggested semi‐empirical correction allows to benefit from the efficient Pario classic method without loss of data quality. We suggest considering ϴ res as a standard parameter in soil analysis to improve the quality of texture estimations.
Type of Medium:
Online Resource
ISSN:
1436-8730
,
1522-2624
DOI:
10.1002/jpln.202100213
Language:
English
Publisher:
Wiley
Publication Date:
2022
detail.hit.zdb_id:
1481142-X
detail.hit.zdb_id:
1470765-2
detail.hit.zdb_id:
200063-5
SSG:
12
SSG:
13
Bookmarklink