Phototactic response of angoumois grain moth Sitotroga cerealella (olivier) to visible spectrum: Attraction analysis and modelling using machine learning

Abstract

The Angoumois grain moth Sitotroga cerealella (Olivier), a primary pest of food grain storage, poses serious problems in India and many other countries. The insect's stronger phototaxis behaviour has been studied using monochromatic wavelengths in the visible spectrum. Among the different wavelengths (250, 330, 410, 480, 530, 580, and 680 nm) assessed for attraction under laboratory conditions, the highest attraction was observed for the blue wavelength (480 nm). Subsequently, the impact of luminance intensities (60, 80, and 100 lux) at each wavelength was also determined. At 60 lux, the 480 and 250 nm wavelengths attracted significantly more adults, i.e., 91.67% and 85%, respectively. The 680 nm wavelength exhibited the lowest attraction rate of only 15% at 100 lux. A gender-based analysis of S. cerealella adults demonstrated that female moths showed a significantly lower attraction (1:3.5) than male moths. This attraction behaviour was also predicted using machine learning. Among the different machine learning algorithms used for modelling insect behaviour, artificial neural network produced the highest prediction with an R2 value of 0.96. In conclusion, the results suggest that using the 480 nm wavelength may be an effective option for mass trapping and monitoring of S. cerealella.