Agricultural Water Management 163 (2016) 363–379 Contents lists available at ScienceDirect Agricultural Water Management jou rn al hom ep age: www.elsevier.com/locate/agwat Deployment of artificial neural network for short-term forecasting of evapotranspiration using public weather forecast restricted messages Seydou Traore a,∗ , Yufeng Luo a,b,c , Guy Fipps a a Department of Biological and Agricultural Engineering, Texas A&M University, Scoates Hall, Room#322, College Station, TX 77843, USA b State Key Laboratory of Hydrology-Water Resources and Hydraulic Engineering, Hohai University, Nanjing, Jiangsu 210098, China c State Key Laboratory of Water Resources and Hydropower Engineering Science, Wuhan University, Wuhan, Hubei 430072, China a r t i c l e i n f o Article history: Received 24 October 2014 Received in revised form 29 September 2015 Accepted 11 October 2015 Available online 27 October 2015 Keywords: Neural networks Evapotranspiration forecast Near-future irrigation Restricted weather messages Crop water demand a b s t r a c t Near-future irrigation demand forecasting is important information for anticipating decisions on crop irrigation scheduling and planning water allocation in large irrigation command areas of Texas. The key determinant that is required for estimating irrigation demand in advance is toward the evapotranspira- tion forecast. Normally, in rich data environment, current reference evapotranspiration (ETo) is estimated by the well-known FAO56 PM method which requires bunch of observed climatic data. In poor data envi- ronment for either current or future estimation, this well-known method application is restricted. Indeed, the correctness of ETo forecast remains a challenging computational task, since inaccurate weather vari- ables can alter the forecast accuracy. Therefore, this study aims to employ artificial neural network (ANN) methodology for forecasting near future ETo values by using restricted climate information messages retrieved from public weather forecast source. Four ANNs learning algorithms including the Generalized Feedforward (GFF), Linear Regression (LR), Multilayer Perceptron (MLP) and Probabilistic Neural Network (PNN) are applied with three sets of inputs combination composed of minimum (Tmin) and maximum (Tmax) daily air temperatures, extraterrestrial radiation (Ra) and net solar radiation (Rs) to forecast ETo in Dallas. The coefficient of correlation (CC), mean square error (MSE), normalized mean square error (NMSE), mean absolute error (MAE) and mean square error skill score (MSESS) were used for the mod- els evaluation. Statistically, in comparison with FAO56 PM, the performances of ANNs models using only Tmax and Tmin predictors were inferior to those of Tmax, Tmin and Ra. With Tmax, Tmin and Rs input-sets, MLP yielded the highest accuracies (CC = 0.926; MSE = 0.770 mm/day, NMSE = 0.143 mm/day; MAE = 0.708 mm/day). Tmax is an important ETo forecast predictor, and the performance improvement relies mostly on Rs accuracy. With precise weather forecast information, ANN made ETo forecast possi- ble (Average CC = 0.860, MSESS = 0.738). These results can assist irrigation districts to accommodate in advance their crop water demand to near-future irrigation requirement. © 2015 Elsevier B.V. All rights reserved. 1. Introduction Irrigation water-use efficiency improvement is amongst the key strategies that can substantially save more water, and then re- allocate it to other users in Texas. The water resources in the State of Texas are shared by many users and for multiple uses. Based on existing literatures, the major water problems and challenges that Texas is facing today are the provision of sufficient water sup- ply for various users including irrigators and the maximization of water-use efficiency in a timely manner at irrigation district level. ∗ Corresponding author. E-mail address: se73traore@gmail.com (S. Traore). However, an effective water supply at irrigation field can be achieved by providing to the individual end-user or district a near- future irrigation demand forecasting tool. So, water users and planners can anticipate their irrigation water delivery and alloca- tion plan based on the public weather forecast information release. Irrigation demand forecast tool has not yet been deployed in the irrigation districts of the State of Texas. Many other sophisticated applications devices are currently being used in the State for irri- gation field data measurement in which include current soil water balance information, an essential component of irrigation schedul- ing and water management. These applications, even with their very wide spectrum of data requirement, are used for current irri- gation estimates. Moreover, reference evapotranspiration forecast for short- term near-future as opposed to current estimates, is central to http://dx.doi.org/10.1016/j.agwat.2015.10.009 0378-3774/© 2015 Elsevier B.V. All rights reserved.