Prognostication of energy consumption and greenhouse gas (GHG) emissions analysis of apple production in West Azarbayjan of Iran using Artificial Neural Network

The present study addresses the amount of input–output energy utilized in apple production in West Azarbayjan province of Iran. The environmental indices of greenhouse emissions during apple production were determined as another end of this investigation. Finally, the potential of a supervised Artificial Neural Network (ANN) approach was assessed to prognosticate the energy consumption and environmental indices of apple production in the studying location. The associated data for the production of apple were collected randomly from 100 orchardists by using a face to face questionnaire method. Energy inputs included human labor, machinery, diesel fuel, seeds, herbicide, pesticide, chemical fertilizers, manure, irrigation water and electricity. The total input and output energies of 77,064.24 MJ ha−1 and 802,695 MJ ha−1 were obtained for apple production in the study region where the value of total GHG emission was estimated at 1195.79 kg CO2eq ha−1. The results revealed that the total consumed energy input could be classified as direct energy (65.97%), and indirect energy (33.76%) or renewable energy (45.37%) and nonrenewable energy (46.97%). The modeling implementations indicated that the lowest RMSE and MAPE of 0.11 and 0.68 were obtained at 16 neurons. At this number of neurons, the best predicting model was achieved. The R2 values of 0.9879 and 0.9827 were obtained for energy input and environmental indices prediction, respectively. The promising ability of the developed ANN in this study indicates that ANN is powerful and robust tool to be served as a functional and dynamic field of studying interest in the realm of energy consumption modeling.