Document Type : Original Article
Authors
1
PHD Graduated, Water Resource Engineering, University of Birjand, Birjand, Iran
2
Professor of Water Engineering Department, University of Birjand, Iran
10.22077/jdcr.2026.12142.1286
Abstract
Considering the effects of climate change on crop yield, this study identified the climatic variables associated with saffron yield using an artificial neural network (ANN). After selecting the appropriate climate model for North Khorasan, Razavi Khorasan, and South Khorasan provinces, saffron yield in future periods was simulated under the CMIP6-based SSP1-2.6, SSP2-4.5, and SSP5-8.5 scenarios. Comparison of the results of the ACCESS-ESM1-5-5, HadGEM3-GC31-LL, and MRI-ESM2-0 climate models indicated that ACCESS-ESM1-5-5 was suitable for North Khorasan and South Khorasan, whereas MRI-ESM2-0 was suitable for Razavi Khorasan. After implementing the algorithm in the R environment, the results showed that the selected ANN models achieved R2 values of 0.93, 0.88, and 0.87, RMSE values of 0.288, 0.419, and 0.551, and MAE values of 0.232, 0.313, and 0.516 in North Khorasan, Razavi Khorasan, and South Khorasan, respectively, indicating satisfactory performance in predicting saffron yield. Analysis of the selected neural networks showed that rainfall during the vegetative period was selected in all three provinces. In addition, rainfall during the flowering period and minimum temperature during the flowering period were selected for North Khorasan; rainfall during the dormancy and flowering periods, maximum temperature during the dormancy and flowering periods, and minimum temperature during the flowering period were selected for Razavi Khorasan; and rainfall during the dormancy period and the number of hot days during the dormancy period were selected for South Khorasan. Future projections indicated that saffron yield responses to climate change scenarios differed among the provinces.
Keywords
Subjects
