Monitoring agricultural pest insects using dual-Polarization radars

Abstract

The Fall Armyworm (FAW), Spodoptera frugiperda, is a migratory pest originally native to tropical and subtropical regions of the Americas, but it has emerged as a global threat to agriculture, particularly in Africa since its arrival in 2016. FAW larvae feed on more than 80 different crops, including staple foods such as maize, rice, sorghum, and sugarcane, making it a significant concern for food security across the continent. In Rwanda, where FAW was first detected in 2017, conventional pest monitoring and control methods have often proven insufficient in preventing large-scale crop damage due to delayed detection of infestations. This research introduces a novel approach to FAW monitoring by leveraging dualpolarization weather radar technology traditionally used for meteorological applications to detect and classify airborne biological entities, including insect swarms. Specifically, this study employs the Meteo Rwanda C-band polarimetric Doppler radar to track FAW infestations and outbreaks in southern Rwanda. Dual-polarization radars are uniquely suited for such applications because they can detect both the size and shape of targets, allowing them to distinguish between meteorological (e.g., rain) and biological (e.g., insects) targets using key radar parameters such as horizontal reflectivity (DBZH), differential reflectivity (ZDR), correlation coefficient (RHOHV), and specific differential phase (KDP). A significant contribution of this research is the development and application of a neurofuzzy logic-based classification algorithm, which combines fuzzy logic with neural networks to improve the accuracy of FAW detection from radar data. This neuro-fuzzy approach enables more precise classification of FAW by processing radar signals and accounting for the inherent uncertainties in radar measurements. The introduction of a new radar parameter, the Depolarization Ratio (DR), further enhances the ability to distinguish biological targets like FAW from meteorological phenomena. By integrating DR into the fuzzy logic classification algorithm, the study achieved significant improvements in insect identification accuracy, particularly in separating FAW from other biological entities such as birds and smaller insects. Through extensive analysis of radar data collected during the FAW infestation period from September 2020 to January 2021 across the Nyanza, Huye, and Gisagara districts of 2 Monitoring Agricultural Pest Insects Using Dual-Polarization Weather Radars RadarsRadars Rwanda, the research demonstrates that dual-polarization weather radar can detect FAW adult moths up to four weeks earlier than ground-based monitoring techniques. Early detection is crucial for enabling timely interventions, such as targeted pesticide application or biological control measures, which can substantially reduce crop damage. This proactive approach to pest management not only helps mitigate the economic losses associated with FAW infestations but also promotes more efficient use of resources, reducing the reliance on blanket pesticide applications. The study also provides a comprehensive analysis of FAW migration and infestation patterns using advanced radar data processing tools like the bioRad R package. Metrics such as Vertically Integrated Reflectivity (VIR), Vertically Integrated Density (VID), and Migration Traffic Rate (MTR) were used to quantify the spatial and vertical distribution of FAW swarms. The results highlight distinct nocturnal activity patterns of FAW, with peak migration occurring at specific altitudes, typically between 300 and 700 meters above ground level. These findings offer valuable insights into the behavior of FAW, enabling the development of more targeted pest control strategies that are aligned with the pest's movement and activity patterns. This research advances the use of meteorological radar technology in agricultural pest management, showing that dual-polarization weather radar can effectively monitor pests like the Fall Armyworm (FAW). This cost-effective method has proven successful in Rwanda and has the potential to be scaled to other regions facing similar pest challenges. Integrating radar-based detection into agricultural monitoring frameworks can provide real-time data, improving pest management and contributing to more sustainable agricultural practices. This study highlights how technology can reduce crop losses and enhance food security in vulnerable areas.