Modeling monthly crop coefficients of maize based on limited meteorological data: A case study in Nile Delta, Egypt

Accurate estimation of crop evapotranspiration (ETc) is essential for water resources management, planning, and scheduling. This study focuses on estimating, predicting and modeling crop coefficients kc of maize which is an important parameter for calculating ETc by using artificial neural networks models (ANN) and fewer parameters. Altogether, four major maize producing Egyptian governorates (Ad Dakahliyah, Al Gharbiyah, Ash Sharqiyah, and Al Ismailiyah) were selected and the monthly data of minimum and maximum temperature, solar radiation, wind speed, and vapor pressure deficit were extracted from open access data (GIS-raster) over the period from 2006 to 2016. The analyzed datasets were divided into two segments from 2006 to 2014 for training and from 2015 to 2016 for testing. The results indicated that data combination of minimum temperature, maximum temperature, and solar radiation was the best artificial intelligence model for predicting kc in four sites with differing hidden layers. The hidden neuron layers were (7, 5), (8, 6), (9, 6), and (9, 7) for Ad Daqahliyah, Al Gharbiyah, Ash Sharqiyah, and Al Ismailiyah, respectively. There was a statistically significant consistency between the measured and modeled values in four locations, and the analysis showed the distributional differences between the actual FAO CROPWAT Model and modeled values were small. The accuracy of the best model and correlation coefficients for prediction kc are close to 1. Thus, the developed model was proven to produce high accuracy and it is recommended to predict the accurate value of kc with limited climatic factors. Also, this study help water users to create new kc database for each region and updated it yearly according to climatic conditions. © 2020 Elsevier B.V.