Proximal soil sensing for Precision Agriculture: Simultaneous use of electromagnetic induction and gamma radiometrics in contrasting soils F.A. Rodrigues Jr. a,b, ⁎, R.G.V. Bramley a , D.L. Gobbett a a CSIRO, Waite Campus, PMB 2, Glen Osmond, SA 5064, Australia b School of Agricultural Engineering, University of Campinas (FEAGRI-UNICAMP), Campinas, SP 13083875, Brazil abstract article info Article history: Received 4 July 2014 Received in revised form 21 December 2014 Accepted 7 January 2015 Available online xxxx Keywords: Data fusion Weighted principal component analysis Within-field variation Digital soil mapping The use of high spatial resolution, on-the-go proximal soil sensing of apparent electrical conductivity (EC a ) through electromagnetic induction (EMI) is increasingly common, in concert with yield mapping, to assist in the delineation of management zones for Precision Agriculture (PA). Less common, but gaining in popularity, is the use of gamma-radiometric (γ) soil sensing. Using contrasting sites in South Australia and Queensland, the specific objectives of the study were to assess for each site, region or all sites together, how well soil cation exchange capacity (CEC) and clay content may be predicted by EMI and γ sensing; to see whether the predictions were improved when both sensors were used, compared to a single sensor; and to evaluate the potential utility of the multi-sensor data in terms of understanding the variation in observed crop yield within sites. Of particular interest was evaluating a generic, as opposed to site-specific, approach to the simultaneous use and calibration of EMI and γ sensing at contrasting sites chosen across a dispersed geography and pedology. EMI and γ soil surveys were carried out at five sites across three cereal growing regions in South Australia, and at three sites in Queensland used for sugarcane production. Soil samples were also collected from each site for lab- oratory analysis. Data analysis comprised simple correlation analysis between soil sensor data and soil proper- ties; fusion of sensor data by region and across all sites using weighted principal component analysis (PCA), with the data weighted on the basis of the two source sensors (weight of 0.5 assigned to EC a and the remaining weight divided equally amongst 238 U, 232 Th, 40 K and ‘total count’ (CPS); weights of 0.125 to each). The output from the PCA was used to predict maps of CEC and clay using multiple regression. Simple correlation analysis showed the expected potential utility of both sensors for predicting soil properties by site and by region. The first three principal components (PCs) explained 98% of the data variation across regions and all sites. Models for the prediction of CEC and clay content, derived from the all sites PCs, were significant (p b 0.05) at five of the eight study sites. Overall, the results show that PCA may be used as a generic approach to the fusion of EMI and γ sensor across dispersed geography and contrasting pedology and farming systems and that maps of predicted CEC and clay content were potentially helpful in understanding within-paddock yield variation. © 2015 Published by Elsevier B.V. 1. Introduction Proximal soil sensing at high spatial resolution has become an in- creasingly common and essential element in the delineation of ‘man- agement zones’ for Precision Agriculture (PA; Bramley, 2009; Bramley and Trengove, 2013). Most commonly, this has involved on-the-go measurement of apparent electrical conductivity (EC a ; Corwin and Plant, 2005 and references therein) using either electromagnetic induc- tion (EMI; e.g. Hedley et al., 2004), or measurement of resistivity (e.g. Kitchen et al., 2003). Recent interest has also been shown in on-the-go gamma radiometry (γ; e.g. Loonstra and van Egmond, 2009; Wong et al., 2008, 2009). In the absence of salinity, EC a responds primarily to the amount of clay and moisture (McBratney et al. 2005) whilst gamma radiometry responds more subtly to variation in clay mineralo- gy (e.g. Cattle et al., 2003)—which is a key driver of CEC. Of particular in- terest in this study is the idea that EMI and γ sensing might be used simultaneously and in complimentary fashion as a means of enhancing the value of such high resolution soil survey (Castrignanò et al., 2012), especially in landscapes where important features may not be captured by one or other sensor alone (Wong et al., 2008). Adoption of PA infers that some form of site-specific management will be implemented (Whelan and McBratney, 2000). For such a strate- gy to be optimal, the derivation of management recommendations may also need to be site-specific(Bramley, 2009); conversion of soil test data into a fertilizer recommendation is one such example (Bramley and Geoderma 243–244 (2015) 183–195 ⁎ Corresponding author at: CIMMYT, Global Conservation Agriculture Program, 56237, Col. El Batán, Texcoco, Mexico. E-mail address: f.a.rodrigues@cgiar.org (F.A. Rodrigues). http://dx.doi.org/10.1016/j.geoderma.2015.01.004 0016-7061/© 2015 Published by Elsevier B.V. Contents lists available at ScienceDirect Geoderma journal homepage: www.elsevier.com/locate/geoderma