Forecasting of crop water stress indicators using machine learning algorithms

Authors: ANDRADE, MANUEL - University Of Nevada; O`Shaughnessy, Susan; Evett, Steven - Steve

Submitted to: Journal of the ASABE
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 1/12/2023
Publication Date: 2/8/2023
Citation: Andrade, M.A., O'Shaughnessy, S.A., Evett, S.R. 2023. Forecasting of crop water stress indicators using machine learning algorithms. Journal of the ASABE. 66(2):297-305. https://doi.org/10.13031/ja.15213.
DOI: https://doi.org/10.13031/ja.15213

Interpretive Summary: As water resources for agriculture become more limited, the efficiency of the conversion of water to crops needs to improve. One means by which crop water productivity can be improved is the use of sensors to monitor crop water stress and inform variable rate irrigation systems to apply water in a site-specific manner within a field. Scientists at ARS-Bushland and University of Nevada-Reno developed an irrigation supervisory control and data acquisition system (ISSCADAS) to simplify sensor-based irrigation scheduling. The system uses wireless sensor networks to collect canopy temperature, soil water content and weather data to automate the building of prescription maps. However, at times, data loss from the moving network of wireless canopy temperature sensors may occur, or data collection may be prevented when the irrigation system does not travel across the entire field. In this study, an artificial neural network (ANN) algorithm was trained based on historical canopy temperature, weather data, irrigation treatment level, days after planting and days from the last irrigation to predict canopy temperature data. The ANN model was able to reasonably estimate canopy temperature, allowing the ISSCADAS to build an accurate prescription map. This result is important as it adds redundancy to the ISSCADAS and enables crop status monitoring when there are gaps in measured canopy temperature data.

Technical Abstract: Recent advances can provide farmers with irrigation scheduling tools based on crop stress indicators to assist the management of Variable Rate Irrigation (VRI) center pivot systems. These tools were integrated into an Irrigation Scheduling Supervisory Control and Data Acquisition System (ISCCADAS) developed by scientists with the USDA-Agricultural Research Service (ARS). The ISSCADAS automates the collection of data from a network of wireless infrared thermometers (IRTs) distributed on a center pivot’s lateral and in the field irrigated by the center pivot, as well as data from a wireless soil water sensor network and a microclimate weather station. This study analyzes the use of Artificial Neural Networks, one of the most popular machine learning algorithms, for the forecasting of canopy temperatures obtained by a wireless network of IRTs mounted on a three-span VRI center pivot irrigating corn near Bushland, TX, during the summer of 2017. Two case studies were conducted for this purpose using data collected from periodic scans of the field performed during the growing season by running the pivot dry. In the first case, data from the first three scans were used to train an Artificial Neural Network (ANN) and canopy temperatures estimated using the ANN were then compared against canopy temperatures measured by the network of IRTs during the fourth scan. In the second case, data from the first six scans were used to train ANNs and canopy temperatures estimated using the ANN were then compared against canopy temperatures measured by the network of IRTs during the seventh scan. The Root of the Mean Squared Error (RMSE) of ANN predictions in the first case ranged from 1.04 °C to 2.49 ºC, whereas the RMSE of ANN predictions in the second case ranged from 2.14 °C to 2.77 °C. To assess the impact of ANN accuracy on irrigation management, estimated canopy temperatures were fed to a plant-stress based irrigation scheduling method and the resulting prescription maps were compared against prescription maps obtained by the same method using the canopy temperatures measured by the network of IRTs. In the first case no difference was found between both prescription maps. In the second case only one plot (out of 26) was assigned a different prescription. Results of this study suggest that machine learning techniques can be used to assist the ISSCADAS in situations where canopy temperatures cannot be measured by the network of IRTs due to poor visibility conditions, or because the center pivot cannot traverse the field within a reasonable amount of time.