Fast-spreading Xylella fastidiosa is one of the world’s most dangerous plant pathogens, devastating tree species from citrus to oak trees across the globe. Now, hyperspectral cameras mounted on drones can provide an early warning system by detecting subtle changes in leaf colour long before symptoms are visible to the human eye.
X. fastidiosa has been identified as Australia’s number one national priority plant pest by the Australian Government Department of Agriculture and Water Resources Plant Health Committee. This means it is not present or established in Australia, but poses a major threat to our agricultural and food production industries.
X. fastidiosa is a bacterium that can destroy plant crops of high economic value (such as olive trees, grapes, stone fruits and almonds) and has wreaked destruction for a long time in the Americas and more recently in Europe, Iran and Taiwan. It has wiped out more than a million olive trees in southern Italy and costs the Californian grape industry an estimated US$104 million a year.
The bacterium lives in the plant xylem tissue and it is normally spread by insect vectors feeding from the plant xylem.
There is no cure. Culling trees is the only way to halt an outbreak, but trees can show no visible symptoms for up to a year after being infected. During this time, a wide range of sap-sucking insects can spread the disease. A major limitation of current large-scale mapping methods based on red and near-infrared is that they are useful only for detecting the advanced stages of disease damage; that is, when canopy defoliation, leaf wilting and chlorosis are apparent.
The new research demonstrates the first way to spot infected trees months before signs are visible to farmers.
Pablo Zarco-Tejada from the European Commission’s Joint Research Centre in Ispra, Italy, led the new study. The researchers concede that pre-visual detection is only part of the solution, but that robust monitoring of global trade in plants will also be required, however this is a major step forward.
Over a two-year period the researchers surveyed 15 olive groves on the ground and from the air. They used a drone carrying both a thermal camera and a hyperspectral sensor, which takes high-resolution measurements across a wide range of wavelengths.
The sensor system detects biochemical and physiological changes that can indicate onset of the disease in trees.
The hyperspectral camera, flown 500 metres over the groves, analysed 250 bands of light from visible to infrared, far more than the red-blue-green seen by the human eye. With a resolution of 40-60 centimetres, each tree could be assessed for the damage X. fastidiosa causes to photosynthesis and transpiration (how plants draw water up to their leaves).
Analysis of this data showed subtle shifts in colour caused by changes in pigments the plants use for photosynthesis, such as chlorophyll. Likewise, changes in transpiration could be detected by thermal infrared data because plants that transpire less get hotter.
The researchers said they had more than 80 per cent success rate in detecting the disease before it was visible to the human eye.
Source: Nature Plants 4:432–439
Photo Credit: Juan A. Navas-Cortes
