Co-Design of Climate Agrometeorological Prediction System to Aid Small Scale Farmers

Abstract

The Majority of farmers in Namibia are not able to access meteorological information and even where available processing requires human expertise which is rarely available. The purpose of this study is to develop an agrometeorological system to predict weather parameters that can aid small scale farmers to plan the type of crops to grow and the use of related pesticides. The aim of this study is to collect climate agrometeorological data from real-world environment using IoT sensors, pre-train the prediction model with the real-world environment dataset using IoT, to predict weather parameters such as temperature. In addition, an interactive Chatbot is developed for effective early warning of Agrometeorological conditions and the performance of the pilot prototype is evaluated. The mixed method research with the experimental approach is used for this study, both quantitative and qualitative data was collected through the use of interviews and questionnaires. An experiment is carried out to train and test the proposed model using real world dataset. Design Science Research Methodology (DSRM) was employed by designing and creating artifacts of the model and validating the results through pilot testing and the result of the study shows that predicting weather parameters such as humidity and temperature using IOT surpass other prediction methods with an accuracy rate of 98.2% with real-world environment dataset. The study further concludes that predicting weather parameter using IoT sensors provides fast, accurate and reliable real-time weather forecast. In addition, the study recommends translating the system in local languages to solve the communication barrier between the local users and the system.