Environmental impact assessment of open field and greenhouse strawberry production

• Strawberry cultivation is modeled in Guilan province of Iran.
• Open field (OF) and greenhouse strawberry (GH) production systems are compared.
• Electricity and natural gas are the major input in GH that affects environmental impacts.
• Evaluations show that OF cultivation is more environmentally profitable.
• Results indicate that ANFIS can predict environmental indices with high accuracy.

A cradle-to-farm-gate life cycle assessment (LCA) study is conducted for open-field (OF) and greenhouse (GH) strawberry cultivation in Guilan province, Iran. The initial data including input materials and energy inputs exercised during production season was collected from 33 greenhouse holders and 77 farmers in the studied region. Data for the production of used inputs were taken from the EcoInvent®2.0 database, and SimaPro software was employed to analyze primary data. Ten impact categories including Abiotic Depletion potential (AD), Acidification potential (AC), Eutrophication potential (EU), Global Warming potential for time horizon 100 years (GW), Ozone Depletion potential (OD), Human Toxicity potential (HT), Freshwater and Marine Aquatic Ecotoxicity potential (FAET and MAET), Terrestrial Ecotoxicity potential (TE), and Photochemical Oxidation potential (PhO) were selected based on CML 2 baseline 2000 V2/world, 1990/characterization method. One ton of produced strawberries was chosen as the functional unit. The evaluations of the study indicated that OF production system in all selected impact categories was more environmentally profitable than GH production system except AC and EU. Based on the results electricity, natural gas and N-based fertilizers in both production stage and application in GH strawberry production produced the most environmental burdens. Moreover, artificial neural networks (ANNs) and adaptive neuro-fuzzy inference systems (ANFIS) were compared to model environmental indices of strawberry production on the basis of input materials. The results revealed that ANFIS models due to employing fuzzy rules are able to predict ten environmental indices with minimum error and the highest accuracy.