Journal of Environmental Biology

Testing the accuracy of rainfall forecast for its usability in planning smart agricultural practices and its user response

A.N. Islam¹*, R.L. Deka², K. Medhi² and K. Sarmah²     

¹Department of Natural Resource Management, College of Horticulture and Farming System Research (Assam Agricultural University), Nalbari-781 336, India

²Department of Agricultural Meteorology, College of Agriculture (Assam Agricultural University), Jorhat-785 013, India

Received: 30 June 2025                   Revised: 22 February 2026                   Accepted: 16 April 2026

*Corresponding Author Email: athar.islam@aau.ac.in                  *ORCiD: https://orcid.org/0009-0001-7712-1682

Abstract

Aim: Compute the accuracy of district level rainfall forecast issued during last five years and compare its accuracy with the block level forecast and evaluate its usability by farmers.

Methodology: 5-years rainfall forecasts were verified qualitatively by using Ratio Score (RS) and quantitatively by using statistical technique like RMSE. The error structures for verification of quantitative precipitation were also followed to discriminate between probability of success and failure forecasts. 5-Day Block level forecasted rainfall was also verified for three blocks viz., Kaliapani, Jorhat and Titabor, and the accuracy was compared with the district level forecasted data for assessing its practical utility. Further, based on the agromet advisory bulletins (AABs) disseminated, feedback were collected from 700 respondents regarding the field level usability of the AABs and the weather forecasts issued during the study period.

Results: Qualitative verification of district level rainfall forecasts from 2018-19 to 2022-23 showed that, Ratio Score (RS) had a good acceptability with a mean score of 72% on annual basis and 80%, 76%, 67% and 66% during winter, post-monsoon, monsoon and pre-monsoon seasons, respectively. Quantitatively, 65% rainfall forecasts were found correct and usable on an annual basis with highest probability of success during winter (96%; i.e., highly accurate) and lowest during monsoon (30%). Probability of success along with the RS of the block level rainfall forecasts were somewhat fair when compared with the district level forecast. Further, feedback analysis from 700 respondents revealed that 86% of the farmers followed the agromet advisories and out of the total sample, 81% were satisfied with the rainfall forecast and 17% farmers were partially satisfied.

Interpretation: District level rainfall forecast had a good acceptability with fair quantitative and qualitative forecast accuracy. Though the qualitative acceptability of rainfall forecast during monsoon was satisfactory, there was a lag in forecast of the exact quantum of rainfall occurred, which resulted in reduced probability of success with high RMSE values for the season. Further, Block level forecast accuracy showed slightly better results in most of the cases and therefore it can be effectively utilized over district level rainfall forecast for formulation and dissemination of micro level smart weather-based advisories.

Key words: Accuracy, Agromet Advisory Bulletins, Block level, District level, Qualitative, Quantitative, Rainfall forecast

Copyright © 2026 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).