A new probabilistic forecasting model for canopy temperature with consideration of periodicity and parameter variation

Abstract

Continuous measurement of canopy temperature is an important indicator of plant water status of crops and the ability to predict canopy temperature will assist in the implementation of this technology for guiding crop irrigation scheduling. By noting that canopy temperature is related to its environmental weather variables which change over time of the day and have different effect or contribution to canopy temperature, this paper presents a probabilistic model to predict canopy temperature by using weather variables which can be obtained from weather model predictions. Unlike the existing models which consider only the linear correlation, the proposed model allows the model parameters to vary according to a periodic function which is designed to capture the variation over the time of the day. The continuity of parameter changes is guaranteed by varying the model parameters periodically and smoothly. Two case studies using cotton experiments from Australia and the United States are conducted to compare the model performance with the existing published models. Using the predictions of canopy temperature an index of crop stress is also predicted in order to evaluate its influence in the irrigation scheduling. Results show that the proposed model is superior to the existing published models in its ability to predict canopy temperature into the future and has utility in assessing when crop stress will occur, to assist with irrigation scheduling. Further evaluations suggest that the air temperature is a dominate weather variable for forecasting canopy temperature but the inclusion of the other weather variables still improves the forecast.