Title : ( Determining Key Agronomic and Environmental Drivers of Saffron ( Crocus sativus L.) Yield and Quality Using LASSO Regression )
Authors: abdollah dorpoor , Mehdi Rastgoo , Ghorban Ali Asadi , Javad Feizy , Mohammadhasan Fallah ,Abstract
Saffron is a high-value crop of strategic and economic importance in Iran, particularly in semi-arid regions. Its productivity and quality are influenced by complex interactions among agronomic, environmental, and management factors. This study aimed to identify and analyze the key determinants of saffron yield (stigma yield per hectare and per kilogram of fresh flowers) and quality indices (crocin, picrocrocin, and safranal) using a data- driven approach. In 2023, data were collected from 99 saffron farms across eight counties in Razavi Khorasan Province, encompassing 75 variables that recorded climate, soil, management practices, and farmer demographics. Least Absolute Shrinkage and Selection Operator (LASSO) regression with 10-fold cross-validation was applied for variable selection and predictive analysis. Results indicated that stigma yield per hectare was primarily influenced by corm planting rate, organic fertilizer, corm weight, and field area, achieving R² = 0.63 and RMSE = 3.75 kg.ha⁻¹. For stigma yield per kilogram of fresh flowers, phosphorus fertilization, corm weight, and planting density were the strongest positive predictors, with R² = 0.70 and RMSE = 0.69 g.kg⁻¹. Moderate positive effects were observed for organic fertilizer and irrigation frequency, while quadratic effects suggested threshold responses for corm size and irrigation. For quality traits, phosphorus was the dominant positive predictor of crocin (β = 18.3) and picrocrocin (β = 3.97), whereas altitude and foliar spray frequency negatively affected picrocrocin and safranal. The effects of nitrogen and sulfur fertilizers were minor and nonlinear. Simplified models retained predictive accuracy (R² = 0.70), improving practical applicability. These findings highlight the importance of site- specific phosphorus management, corm quality monitoring, and optimized irrigation for enhancing saffron yield and quality. LASSO regression effectively identified influential variables, supporting precision agriculture and decision-support tools for sustainable saffron production under semi-arid conditions.
Keywords
, Climatic factors, Irrigation scheduling, Nonlinear effects, Nutrient management, Picrocrocin, Variable selection@article{paperid:1107315,
author = {Dorpoor, Abdollah and Rastgoo, Mehdi and Asadi, Ghorban Ali and جواد فیضی and محمدحسن فلاح},
title = {Determining Key Agronomic and Environmental Drivers of Saffron ( Crocus sativus L.) Yield and Quality Using LASSO Regression},
journal = {زراعت و فناوری زعفران},
year = {2026},
volume = {13},
number = {3},
month = {April},
issn = {2383-1529},
pages = {291--306},
numpages = {15},
keywords = {Climatic factors; Irrigation scheduling; Nonlinear effects; Nutrient management; Picrocrocin;Variable selection},
}
%0 Journal Article
%T Determining Key Agronomic and Environmental Drivers of Saffron ( Crocus sativus L.) Yield and Quality Using LASSO Regression
%A Dorpoor, Abdollah
%A Rastgoo, Mehdi
%A Asadi, Ghorban Ali
%A جواد فیضی
%A محمدحسن فلاح
%J زراعت و فناوری زعفران
%@ 2383-1529
%D 2026
