Variability and compatibility in determining soil particle size distribution by sieving, sedimentation and laser diffraction methods

A range of methods and applications are in use to determine soil particle size distribution. Due to the differences in measurement technology, the analytical results may deviate more or less from each other, which has implications for the matching with historical soil databases. There is a need for studies to critically evaluate their results, both concerning subsample variabilities and compatibilities. In the present study the more recent integral suspension pressure (ISP) and laser diffraction (LDM) methods were compared with the reference sieve and pipette (SPM) method. Samples from topsoil and subsoil of four agricultural soils with sandy clay loam to clay textures were analyzed. A protocol, comparing alternative pre-sievings at the meshes 0.063 (ps0.063), 0.2 (ps0.2) and 2.0 mm (ps2) for the sedimentation (SPM, ISP) and laser diffraction (LDM) measurements, was used. Here we report, based on particle size fraction contents for clay (<0.002 mm), silt (0.002-0.063 mm) and sand (0.063-2.0 mm), i) apparent deviations between pre-sieving options for each method, ii) variabilities between sample replicates (three subsamples), and iii) relationships (linear regression) and iv) texture class differences between SPM, ISP and LDM analyses. Overall, SPM showed smallest deviations between pre-sieving options, LDM largest, and ISP intermediate. Higher silt content, for ISP, and higher sand content, for LDM, seemed to be critical in the choice of optimum pre-sieving. Regarding variabilities between replicates, SPM showed smallest variabilities, ISP (especially ISP-ps0.2 and ISP-ps2) and LDM-ps2 largest, and LDM-ps0.063 and LDM-ps0.2 intermediate. SPM-ps0.063, SPM-ps2, ISP-ps2 and ISP-ps0.2 showed strongest relationships (i.e. largest R2) with the reference SPM-ps0.2, LDM-ps0.063 intermediate and LDM-ps2 weakest. Regarding texture classification, compared to the reference SPM-ps0.2, SPM-ps2 and ISP-ps2 showed largest (good, i.e. 80-100% of the cases) agreement, whereas LDM pre-sievings showed smallest (LDM-ps0.063, poor agreement, i.e. <55%). Lineartransfer transformed LDMt-ps0.063 improved the texture compatibility with SPM-ps0.2 to intermediate (63%) agreement, and SPMt-ps0.063 and ISPt-ps0.2 from intermediate (75%) to good (88%) agreement. Also clay-silt cutoff modified LDMc-ps0.063 and LDMc-ps0.2 improved the texture compatibility with SPM-ps0.2, to intermediate (63%) agreement. There is a need to continue fine-tuning methodologies to align particle size distribution composition from one method to the other, especially regarding the influence of equivalent and efficient and on the results.