Two-Stage Spatiotemporal Time Series Modelling Approach for Rice Yield Prediction &amp; Advanced Agroecosystem Management

Santosha Rathod,Ponnuraj Jeyakumar,Rahul Patil,Arun Kumar Swarnaraj,Brajendra Parmar,Ponnuvel Senguttuvel,Amtul Waris,Dhumannatarao Venkata Krishna Nageswara Rao,Gabrijel Ondrasek,Nirmala Bandumula,Pitchiahpillai Muthuraman,Raman Meenakshi Sundaram,Channappa Gireesh,Amit Saha,Madhyavenkatapura Siddaiah Anantha,Shaik N Meera

doi:10.3390/agronomy11122502

Abstract

A robust forecast of rice yields is of great importance for medium-to-long-term planning and decision-making in cereal production, from regional to national level. Incorporation of spatially correlated adjacent effects in forecasting models in general, results in accurate forecast. The Space Time Autoregressive Moving Average (STARMA) is the most popular class of model in linear spatiotemporal time series modelling. However, STARMA cannot process nonlinear spatiotemporal relationships in datasets. Alternately, Time Delay Neural Network (TDNN) is a most popular machine learning algorithm to model the nonlinear pattern in data. To overcome these limitations, two-stage STARMA approach was developed to predict rice yield in some of the most intensive national rice agroecosystems in India. The Mean Absolute Percentage Errors value of proposed STARMA-II approach is lower compared to Autoregressive Moving Average (ARIMA) and STARMA model in all examined districts, while the Diebold-Mariano test confirmed that STARMA-II model is significantly different from classical approaches. The proposed STARMA-II approach is promising alternative to classical linear and nonlinear spatiotemporal time series models for estimating mixed linear and nonlinear patterns and can be advanced tool for mid-to-long-term sustainable planning and management of crop yields and patterns in agroecosystems, i.e., food supply and demand from local to regional levels.

Highlights

IntroductionThe development and implementation of successful adaptation policies and strategies that increase the resilience of agroecosystems to climate change are crucial for food supply and security, especially for the most widely grown and consumed commodities
Suitable candidate the Autoregressive Moving Average (ARIMA) models were fitted to rice yield data, beginning by examining the stationarity of the data using Augmented Dickey-Fuller (ADF) test which revealed a positive trend over time, indicating that the time series under consideration is non-stationary in nature, first differencing was done to make the series stationary wherever necessary
As explained in previous section, the Space Time Autoregressive Moving Average (STARMA) (∅10, ∅11, θ11 ) model was built for rice yield series in first stage, where by diagnostic checking of residuals it was found that probability value of multivariate Box-Pierce non-correlation test is significant

Summary

Introduction

The development and implementation of successful adaptation policies and strategies that increase the resilience of agroecosystems to climate change are crucial for food supply and security, especially for the most widely grown and consumed commodities. Spatiotemporal modelling of rice yield across different, but climatologically closely related (i.e., adjacent) agroecosystems could help to understand the trends in rice cultivation to make the future roadmaps, as well as to assess the impact on the rice supply and demand at both national and regional markets. With this importance, yield of rice, as a model crop has been chosen for this study

Methods

Results

Discussion

Conclusion