Open AccessCan Apparent Electrical Conductivity Improve the Spatial Characterization of Soil Organic Carbon?
Author(s) -
Martinez Gonzalo,
Vanderlinden Karl,
Ordóñez Rafaela,
Muriel José L.
Publication year - 2009
Publication title -
vadose zone journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.036
H-Index - 81
ISSN - 1539-1663
DOI - 10.2136/vzj2008.0123
Subject(s) - soil carbon , kriging , soil science , tillage , total organic carbon , environmental science , electrical resistivity and conductivity , variogram , mathematics , chemistry , soil water , statistics , environmental chemistry , physics , agronomy , biology , quantum mechanics
Ancillary information, such as apparent electrical conductivity (EC a ), can improve the spatial and temporal estimation of soil properties. The purpose of this study was to determine if EC a could be used for the spatial characterization of soil organic C (SOC) within a long‐term tillage experiment. Apparent electrical conductivity was measured using an electromagnetic induction sensor, the EM38DD, and its predictive potential for mapping SOC was evaluated. The EC a maps showed clear differences between the conventional tillage and direct drilling plots, with higher EC a and SOC in the direct drilling plots. A normalized EC a difference (ΔEC a ), calculated as the difference between the normalized vertical and horizontal dipole EC a values (EC aV and EC aH , respectively) successfully classified the SOC observations according to their corresponding management systems. Maps of ΔEC a (FKM1) and EC aV and EC aH (FKM2) classified by fuzzy k ‐means accounted for 30% of the total SOC variability, whereas the individual plots and management strategy explained 44 and 41%, respectively. Simple kriging with local varying means using either FKM2 or plot‐average SOC as secondary information reduced the RMSE by 8% and increased the efficiency index by about 70% compared with ordinary kriging. Despite the low point‐to‐point correlation between EC a and SOC, EC a was shown to be useful for the spatial estimation of SOC.