Characterizing Soil Spatial Variability For Precision Agriculture Using Geophysical Measurements

Key components of precision agriculture are (i) identifying the site-specific factors that influence<br>within-field crop yield variation and (ii) spatially characterizing those factors. Geo-referenced<br>measurements of apparent soil electrical conductivity (ECa) provide a potential means of characterizing<br>the spatial variability of edaphic properties that influence crop yield. It is the objective (i) to utilize an<br>intensive ECa survey to direct soil sampling, (ii) to identify soil properties that influence cotton yield,<br>and (iii) to use this spatial information to make site-specific management recommendations to increase<br>cotton yield. A 32.4-ha field in California’s San Joaquin Valley was used as a study site. Cotton yield<br>monitoring data were collected in August 1999 followed by an intensive ECa survey of 4000+<br>measurements using electrical resistivity. Sixty soil sample sites were selected based upon a responsesurface<br>sampling design utilizing the spatial ECa measurements. Scatter plots were obtained and<br>correlation and regression analyses were performed to assess the relationship between cotton yield and<br>the properties of pH, boron (B), nitrate-nitrogen (NO3-N), chloride (Cl-), salinity (i.e., electrical<br>conductivity of the saturation extract; ECe), leaching fraction (LF), water content (θg), bulk density (ρb),<br>% clay, and saturation percentage (SP). Correlation coefficients of -0.01, 0.50, -0.03, 0.25, 0.53, -0.49,<br>0.42, -0.29, 0.36, and 0.38, respectively, were found. The correlation coefficient between yield and ECa<br>was 0.51. A site-specific response model of cotton yield based on ordinary least squares (OLS) and<br>adjusted for spatial autocorrelation using restricted maximum likelihood was developed. The response<br>model indicated that leaching fraction, salinity, water content, and pH were the most significant soil<br>properties influencing cotton yield: cotton yield (Mg ha-1) = 19.28 + 0.22 ECe – 0.02 ECe<br>2 – 4.42 LF2 –<br>1.99 pH + 6.93 θg. The spatial information and response model provide sufficient information to make<br>site-specific management recommendations to increase cotton yield.