Discipline: Technology and Engineering
Subcategory: Electrical Engineering
Diana Soto - California State University, Bakersfield
Recent drought conditions in California have forced the state to use water more efficiently. California has one of the most productive agricultural regions in the world and relies primarily on the water collected during the winter months, as snow, in the Sierra Nevada Mountains. Irrigation allows for production of agriculture during the summer months when precipitation is at its lowest. Given the persistency of drought conditions in California, improving irrigation planning is of great value. The most fundamental quantity in irrigation planning is evapotranspiration (ETo), whose forecast can be used for water demand estimation and reduction in over-irrigation. The focus of this research study is on ETo forecast for Kern County by means of time series analysis. ETo is the loss of water by either evaporation or transpiration (evaporation of water from the leaves of plants) into the atmosphere. We developed an autoregressive moving-average (ARMA) model on the collected ETo. We obtained accurate models that represent the changes in ETo in Kern County. The structure for each ARMA model is decided using Final Prediction Error criterion. ARMA models are trained using least squares regression and are capable of capturing the seasonal trends in ETo. These seasonal trends are then combined using fuzzy logic to achieve a generic model with year-long ETo forecasting capability. The results are promising and the proposed forecast offers higher irrigation efficiency. Our future research plans include taking weather quantities into account for ETo forecast and combining the proposed ETo forecast with the irrigation planning to develop a comprehensive irrigation scheme.
References: Luo, Yufeng, et al. ‘Medium range daily reference evapotranspiration forecasting by using ANN and public weather forecasts.’ Water Resources Management 29.10 (2015): 3863-3876.
Tian, Di, and Christopher J. Martinez. ‘The GEFS-based daily reference evapotranspiration (ETo) forecast and its implication for water management in the southeastern United States.’ Journal of Hydrometeorology 15.3 (2014): 1152-1165.
Egrioglu, Erol, Cagdas Hakan Aladag, and Ufuk Yolcu. ‘Fuzzy time series forecasting with a novel hybrid approach combining fuzzy c-means and neural networks.’ Expert Systems with Applications 40.3 (2013): 854-857.
Funder Acknowledgement(s): This material is based upon work supported by, or in part by, the U. S. Army Research Laboratory and the U. S. Army Research Office under contract/grant number W911NF1510498.
Faculty Advisor: Saeed Jafarzadeh, sjafarzadeh@csub.edu
Role: I worked on developing an autoregressive moving-average model on Kern County ETo. Also, I worked on selecting the best ARMA structure by using the Final Prediction Error criterion. Lastly, I worked on combining the seasonal trends using fuzzy logic to create a year long ETo forecast to plan irrigation efficiently.