Agriculture is inextricably linked to the environment. Climate change has an effect directly on agricultural activities. India gets severely impacted if there is a loss of yields, which affects human lives. Hence, monitoring climate and its impact on the agricultural field is essential for a country like India. This chapter proposes a novel deep learning-based hybrid vector autoregressive–gated recurrent unit model (VAR-GRU model) for weather forecasting involving the four important weather parameters such as temperature, pressure, humidity, and wind speed for the cities of Bengaluru and temperature, pressure, dew point, and wind speed for the cities of Dongsi. The effectiveness of the proposed VAR-GRU model is proven by comparing its performance metrics (MAE, MSE, RMSE, and R2 Score) with that of other baseline models such as LSTM, VAR, GRU, and another hybrid VAR-LSTM model. The outcomes of this research work can help in increasing crop yields by utilizing the weather forecasting results in smart farming applications.
TopIntroduction
Agriculture is an important industry in the world. It plays a significant role in global economic activity and supplies the majority of food in the global food sector. However, this industry is completely dependent on the environment and its climate. Agriculture relies mostly on external environmental factors such as rainfall and sunlight for plant growth as well as development. The other climatology parameters, such as temperature, humidity, and wind, are also intricately tied to agriculture and have an impact on the growth of the crops (Vining, 1990). In recent years, Extreme Weather Events (EWE) such as heavy rainfall causing floods and higher temperatures causing droughts directly affect the agricultural sector (Cogato, Meggio, De Antoni Migliorati, & Marinello, 2019). This EWE results in a lack of food security, which puts human lives in a dangerous situation (Harvey, et al., 2018). Large populations in countries like India and China depend on a thriving agricultural sector to guarantee there is never a national food shortage. However, agricultural productivity declines when temperatures increase, resulting in unfavorable outcomes. Hence, it is very necessary to make a more accurate and exact prediction of the meteorological parameter.
Several methodologies such as numerical, and statistical methods that include Autoregressive Integrated Moving Average (ARIMA) (Shivhare, Rahul, Dwivedi, & Dikshit, 2019), Seasonal Autoregressive Integrated Moving Average (SARIMA) (Narasimha Murthy, Saravana, & Vijaya Kumar, 2018), Fb-Prophet (Toharudin, et al., 2020), Vector Moving Average (VMA), Vector Autoregressive Moving Average (VARMA) (Kadiyala & Kumar, 2014), etc., are adopted to predict the climatology parameter. However, these models are not able to predict long data sequences. Hence, researchers are more focused on machine learning (ML) models like support vector regression (SVR) (Hasan, Nath, & Rasel, 2015), xg-boost (Zheng & Wu, 2019), Support Vector Machine (SVM) (Zendehboudi, Baseer, & Saidur, 2018), Random Forest (RF) (Lahouar & Slama, 2017), Linear regression (LR) (Anusha, Chaithanya, & Reddy, 2019), etc. With the advancement of artificial intelligence and computational platforms, Deep Learning (DL) models such as Long Short Term Memory (LSTM) (Qing & Niu, 2018), Bi-directional LSTM (Bi-LSTM) (Peng, Zhang, Zhou, & Nazir, 2021), Convolutional Neural Network (CNN) (Lawal, Rehman, Alhems, & Alam, 2021), Neural Bias Expansion Analysis for Time Series (NBEATS) (Sabat, Nayak, Pati, & Das, 2022), Gated Recurrent Unit (GRU) (Darmawan, Yuliana, & Hadi, 2022), Deep Neural Network (DNN) (Ren, et al., 2021), Bidirectional GRU (Bi-GRU) (Liu, Gan, Chen, & Shu, 2023), became popular and widely used for predicting climatology parameters. The statistical approaches are appropriate for regression issues, but deep learning models are appropriate for complex data sets and long-term forecasting. Hence, this study is being examined and intended to create a hybrid model that includes both statistical and deep learning models. The climatic parameter is a multivariate parameter, which means one parameter is dependent on another. For example, temperature predictions are based on modeling historical time series data that includes not just temperatures but also humidity, wind, and other variables. Hence, a multivariate regressive model of ARIMA called Vector Auto-regressive (VAR) and a DL model GRU are combined to get more accurate and long-term forecasting results.