Comparison of soil quality indexes calculated by network and principal component analysis for carbonated soils under different uses

• Published in: Ecological indicators Vol. 143; p. 109374
• Main Authors: Martín-Sanz, Juan Pedro, de Santiago-Martín, Ana, Valverde-Asenjo, Inmaculada, Quintana-Nieto, José Ramón, González-Huecas, Concepción, López-Lafuente, Antonio L.
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.10.2022; Elsevier
• Subjects: bulk density; data collection; electrical conductivity; enzymes; forests; Land use; Mediterranean calcareous soil; Network analysis; olives; Principal component analysis; sand; soil; Soil quality; total organic carbon; GMEAN; BD; aglu; FeA; Ncw; Pav; dh; FeF; Soil quality; FS; WHC; Chw; pac; CaCO3; nag; Network analysis; FeX; pak; Nhw; EC; bglu; MnA; aryls; MnF; aryln; phos; TOC; ure; Ninor; Land use; CS; MnX; Ccw; TN; Mediterranean calcareous soil; Principal component analysis
• DOI: 10.1016/j.ecolind.2022.109374
• ISSN: 1470-160X

•Indicators of soil quality index (SQI) were selected by Network analysis (NTA).•SQI from NTA were compare with those obtained by principal component analysis (PCA).•NTA generated SQIs that reflected more accurately the different soil uses.•Agricultural soils quality was deteriorated by the loss of organic matter. There is an urgent need to conserve and improve the quality of agricultural soils in the coming decades. Decision tools capable of providing reliable information about soil quality are needed, and soil quality index (SQI) is one of the most used. Principal component analysis (PCA) is the common methodology to calculate it, however in some cases fails to differentiate soil quality properly. Therefore, the aim of this work is to assess a SQI through a different methodology as network analysis (NTA) and compare it with PCA, assuming that soil uses affect soil qualities differently. From soils with different uses (rainfed, olive grove and forest) network analysis and principal component analysis have been used to select a minimum dataset (MDS) to generate SQI from 36 physical, chemical and biological soil variables. Using NTA, geometric mean of the enzyme activities (GMEAN), bulk density (BD) and phosphatase activity (phos) where selected as indicators, while PCA selected total organic carbon (TOC), free Fe oxides (FeF), crystalline Mn oxides (MnX), pH, electrical conductivity (EC) and percentage of coarse sand (CS). Four SQI were calculated from each MDS through linear and non-linear scoring equations and by additive integration and weights. The SQI generated by NTA were more useful than those generated by PCA, as in addition to having fewer indicators they were able to better differentiate the uses in the study. This greater resolution capacity of the NTA would be the consequence of a better selection of indicators using this method than using PCA.