Journal of Environmental Biology

Development and validation of predictive regression model for Brown planthopper population in rice

S. Rajna¹*, A. Vishwakarma¹, V.K. Padala², K.V. Raghavendra³ and S. Chander³     

¹Division of Entomology, ICAR- Indian Agricultural Research Institute, New Delhi-110012, India

²ICAR-National Research Institute for Makhana, Darbhanga-800 014, India

³ICAR-Natioal Research Centre for Integrated Pest Management, New Delhi-110 068, India

Received: 15 May 2024                   Revised: 29 January 2025                   Accepted: 28 May 2025

*Corresponding Author Email : rajnasalim@gmail.com                  *ORCiD: https://orcid.org/0009-0001-9729-4759

Abstract

Aim: To develop and validate a multi-linear regression model correlating brown planthopper (Nilaparvata lugens) population with weather parameters specific to New Delhi conditions.

Methodology: A weather-based forecasting model was developed for predicting brown planthopper (BPH) infestation for early, normal and late transplanted rice. The field data during 2017 to 2021 were utilized for the model development and was validated during 2022 and 2023.

Results: In early transplanted rice, a significant negative correlation with BPH infestation and minimum temperature (r=-0.462) and evening relative humidity (-0.387) as well as a significant positive correlation with total sunshine hours (r=0.447) were observed. For normal transplanted rice, BPH population was significantly and negatively correlated with minimum temperature (r=-0.526), evening relative humidity (r=-0.559) and rainfall (r=-0.411) while it was significantly positively correlated with sunshine hours (r=0.390). In case of late transplanted rice, the abiotic factor, sunshine hours (r = 0.355) alone showed a positive correlation with BPH population. Multiple linear regression (MLR) models were developed using data from 2017 to 2021 and validated with 2022 and 2023 data. The models were evaluated based on mean bias error (MBE), mean absolute error (MAE), and root mean square error (RMSE). Correlation analyses indicated significant negative correlations between BPH populations and T_min and RH₂ in early and normal transplanting, while positive correlations with SSH were observed. Validation showed satisfactory accuracy for early and normal transplanting (RMSE: 0.237-0.749), but lower accuracy for late transplanting (RMSE: 2.033-3.259).

Interpretation: These findings show the importance of weather parameters in predicting BPH infestations, with temperature, humidity, and sunshine hours playing significant roles. The study highlights the necessity for integration of pest management strategies time of planting and weather conditions to effectively mitigate BPH impacts on rice yields.

Key words: Multiple regression model, Nilaparvata lugens, Rice, Weather parameters

Copyright © 2025 Triveni Enterprises. All rights reserved. No part of the Journal can be reproduced in any form without prior permission. Responsibility regarding the authenticity of the data, and the acceptability of the conclusions enforced or derived, rest completely with the author(s).