Article Preview
Top1. Introduction
Agriculture is a field that is considered the foremost part of the monsoon and eventually donates to India's GDP and leads to the growth in the social economy. Generally, rainfall in India occurs in midyear. The prediction of climatic factors in India is attracted by the subcontinent and global communities (Vathsala & Koolagudi, 2017). Atmospheric conditions and cloud formation are most demanding to represent in models and may lead to many complications (Abishek, 2017). Nowadays, various meteorology observatory stations obtain different data types depending on weather characteristics taken from different sites. For instance, Guangdong's meteorological department adapts various sensors to gather the weather variables, like wind speed, dew point, temperature, and humidity (Qiu et al., 2017).
Numerous rainfall prediction techniques are adapted for predicting the rain at national and regional levels. Primarily, there are two techniques for predicting rainfall, which is dynamical and empirical techniques. The empirical methods are devised by analyzing the past data of weather and relation with different oceanic and atmospheric variables worldwide. In dynamic methods, the prediction is produced by physical models using an equation that forecast the growth of global climate concerning atmospheric conditions. The dynamical techniques are adapted with statistical weather forecasting strategies. The information on rainfall is essential for planning food production, managing water resources. The incidence of prolonged heavy rain or dry period at vital phases for crop growth and design may reduce crop production. Hence, rainfall prediction is considered an imperative factor in agricultural countries (Zaw & Naing, 2008). Flood is generally occurred due to extreme and continuous rainfall with more river discharge causing huge damages. The flooding causes devastation of ecological resources, massive economic losses, shortage of foods, and starvation. Hence, the major distress of different countries is to observe flooding to reduce the possessions of floods (Wu, et al., 2015).
Rainfall prediction is normally analyzed month-wise to determine the rainy season's beginning and ending duration. The values attained via monthly rainfall are more accurate than those obtained by seasonal values in rainfall distribution. Monthly rainfall predictions play an important role in hydrology as well as agriculture. Hence, if predicted correctly, the system can yield better decision-making. The prediction mechanism is also widely useful in the motivating the students in academic field (Nguyen, 2021), (Nguyen, 2020), (Thi, et al., 2021). The prediction of performance based on linear regression yields better accuracy (Rajalaxmi et al., 2019). Moreover, rainfall prediction models can be categorized as data-driven and physical. The physical techniques devise physical edict required for modeling the pertinent physical process to predict the rainfall, whereas data-driven models operate on previous data to predict the future (Bagirov, et al., 2017). It is more challenging for predicting effectual rainfall and thus, requires analysis of temperature and humidity. The traditional way to predict rainfall concentrates on the three factors, namely humidity, temperature, and received rainfall (Abishek, 2017). Rainfall is a natural phenomenon whose forecasting is considered a major issue. Precise rainfall information is important for managing and planning water resources and is imperative for flooding prevention and reservoir operation. In addition, rainfall poses a strong power on traffic and other human activities. Furthermore, rainfall is the most complicated and complex attribute of the hydrology cycle for understanding and modeling due to atmospheric pressures that produce rainfall. The significant ranges of variations over huge scales cause an increase in time and space. However, accurate rainfall prediction is a challenging task in operational hydrology, despite several forecasting weather advances (Nayak, et al., 2013), (Iliopoulou and Koutsoyiannis, 2020).