Identifying soil management zones in a sugarcane field using proximal sensed electromagnetic induction and gamma‐ray spectrometry data

Abstract

AbstractSpatial variation in soils is required to supply ameliorants and fertilisers in the Australian sugarcane industry. However, traditional approaches are cost‐prohibitive. We investigated how a digital soil mapping (DSM) approach could be used to identify management zones. First, ancillary data including electromagnetic induction and gamma‐ray spectrometry data were collected. Using fuzzy k‐means (FKM) clustering, two to six management zones were identified. A similar approach was used to cluster percentage yield variations (2014, 2015 and 2016). Using restricted maximum likelihood analysis of topsoil (0–0.3 m) and subsoil (0.6–0.9 m) physical (e.g. clay) and chemical (e.g. exchangeable sodium percentage [ESP], and exchangeable calcium and magnesium) properties, three zones were found to minimise the mean squared prediction error (). By comparison, the three zones obtained using the percentage yield variation only minimised for subsoil ESP, which suggested it had some influence on sugarcane yield and productivity. Different rates of gypsum were required to manage the moderately sodic topsoil ESP for each zone. This was similarly the case with lime to overcome deficiencies in exchangeable calcium and magnesium. The results were consistent with yield variance, suggesting the smaller yield in some zones was due to topsoil sodicity and strongly sodic subsoil with a greater clay content. We concluded that the DSM approach was successful in identifying soil management zones and can be used to improve soil structural stability and fertility. The zones also had ramifications for strip trials to determine yield increases by comparing variable rate applications of gypsum and lime.