Remote detection of Fusarium crown rot in broadacre bread wheat and durum wheat through use of aerial imagery

Abstract

<p><b>Context.</b> The cereal disease Fusarium crown rot (FCR), caused by the fungal pathogen <i>Fusarium pseudograminearum</i>, is a worldwide major constraint to winter cereal production, especially in Australia’s northern grain region of New South Wales and Queensland. <b>Aims.</b> Detection of the disease is labour-intensive and often not spatially quantifiable; hence, the aim of this study was to provide methods for in-crop FCR detection on a broadacre scale. <b>Methods.</b> A replicated field experiment across three locations in northern New South Wales explored the use of thermal and multispectral imagery and hyperspectral reflectance data for the spatial detection of FCR in three bread wheat (<i>Triticum aestivum</i> L.) and three durum wheat (<i>T. durum</i> Desf.) varieties in the presence and absence of inoculation with <i>F. pseudograminearum</i>. <b>Key results.</b> Canopy temperature was 0.30–0.90°C higher in two-thirds of field sites inoculated with the pathogen during early wheat growth in a slightly wetter than normal season. Some multispectral indices including normalised difference red edge, normalised difference vegetation index, near infrared and red edge also demonstrated the ability to identify inoculated versus uninoculated treatments as early as the first node stage (GS31). <b>Conclusions.</b> Although positive identification was achieved with remote detection, environmental conditions (i.e. soil-water availability and ambient temperature) and physiological maturity influenced the accuracy of the technology for detecting FCR infection, particularly in wetter early-season conditions. <b>Implications.</b> Early spatial detection of FCR infection on a broadacre scale could allow producers to manage this disease spatially through better agronomic decisions.</p>