Near-infrared spectroscopy for discrimination of Huanglongbing-infected citrus leaves from unifected leaves

Authors: Windham, William; Poole, Gavin; Park, Bosoon; Heitschmidt, Gerald - Jerry; Gottwald, Timothy; Lawrence, Kurt

Submitted to: American Society of Agricultural and Biological Engineers
Publication Type: Abstract Only
Publication Acceptance Date: 6/21/2010
Publication Date: 6/21/2010
Citation: Windham, W.R., Poole, G.H., Park, B., Heitschmidt, G.W., Gottwald, T.R., Lawrence, K.C. 2010. Near-infrared spectroscopy for discrimination of Huanglongbing-infected citrus leaves from unifected leaves. American Society of Agricultural and Biological Engineers.

Interpretive Summary: Citrus greening, also called Huanglongbing (HLB) or yellow dragon disease, is one of the more serious diseases of citrus. The disease is primarily spread by two species of psyllid insects, Diaphorina citri and Candidatus Liberibacter asiaticus. Citrus greening disease is a threat to the U.S. citrus industry. An infected tree produces fruit that is unsuitable for sale as fresh fruit or for juice. The only definitive method of diagnosis of trees suspected of infection by citrus greening pathogens is by analysis of DNA. The purpose of the research was to investigate the use of near infrared spectroscopy to differentiate HLB infected leaves from uninfected leaves. Leaf samples (N=153) were collected from orange and grapefruit trees on the U. S. Horticultural Research Laboratory's research farm that showed the various symptoms of HLB. Leaves (N=30) were also collected from citrus trees in a controlled quarantined greenhouse and were used as a “negative” control. Leaves were analyzed for citrus greening pathogens DNA by real-time polymerase chain reaction (PCR). The value assigned from the PCR test was a Critical Threshold (CT) value, which indicates whether a sample is either positive or negative. A value above 30 is considered negative for citrus greening disease, whereas, below 30 is considered positive. The remaining leaves were dried, ground and analyzed by near-infrared reflectance spectroscopy. Partial least squares regression was used to regress the CT values on spectra of ground leaves. Standard error of cross-validation (SECV) for the CT value was lower for a first derivative data pre-treatment of visible plus NIR spectra compared to log 1/reflectance spectra. The SECV and R2 was 2.4 and 0.90, respectively. The model correctly differentiates HLB infected leaves from uninfected leaves with a standard error of prediction and an r2 of 2.5 and 0.88, respectively. Spectra of uninfected leaves compared to HLB infected leaves indicated obvious differences in C-H absorbance due to cuticle waxes and O-H due to carbohydrate. The visible region alone cannot be used to differentiate HLB infected leaves from uninfected leaves because uninfected leaves can have a yellow appearance due to other agronomic conditions, such as a zinc deficiency. The results indicate that NIR reflectance spectroscopy has the potential to be used as a screen method for HLB infected citrus leaves.

Technical Abstract: n/a