A Novel Hybrid MARS Algorithm-Particle Swarm Optimization Approach for Predicting and Optimizing K/Na, Ca/Na Ratios and Salt Tolerance Index in Sorghum Seedlings Under Salinity
JOURNAL OF SOIL SCIENCE AND PLANT NUTRITION, 2026 (SCI-Expanded, Scopus)
- Yayın Türü: Makale / Tam Makale
- Basım Tarihi: 2026
- Doi Numarası: 10.1007/s42729-026-03492-2
- Dergi Adı: JOURNAL OF SOIL SCIENCE AND PLANT NUTRITION
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Natural Science Collection (ProQuest), Biological Science Database (ProQuest)
- Ondokuz Mayıs Üniversitesi Adresli: Evet
Özet
Soil salinity severely limits crop establishment by disrupting ionic homeostasis and reducing biomass accumulation. Although potassium nitrate (KNO3) seed priming is a well-known strategy for enhancing stress tolerance, the nonlinear response surfaces governing ion regulation in sorghum seedlings under combined KNO3 and salinity treatments remain unexplored. With this aim, this pot experiment developed a hybrid modeling-optimization framework integrating multivariate adaptive regression splines (MARS) and particle swarm optimization (PSO) to predict and optimize Na+, K+, and Ca & sup2;(+) levels, K+/Na+ and Ca & sup2;(+)/Na+ ratios, and the salt tolerance index (STI) in sorghum seedlings under varying soil salinity levels and KNO3 priming doses. Increasing salinity severely disrupted ionic homeostasis, causing a 688.8% increase in shoot Na+ accumulation and 56.7% and 29.5% reductions in shoot K+ and Ca & sup2;(+) contents, respectively, under non-primed conditions at 14.02 dS m(-)& sup1;. However, priming with 25 and 50 mM KNO3 reduced Na+ accumulation by 68.3% and 52.5%, respectively, while increasing STI by 44.1% and 49.0%, respectively, at the same salinity level. Moreover, MARS models showed strong predictive accuracy (R & sup2;: 0.844-0.978; RMSE: 0.021-7.670; MAE: 0.017-6.181) and identified a critical salinity threshold of 9.55 dS m(-)& sup1;, beyond which ionic regulation deteriorated sharply. PSO-based multi-objective optimization identified 50 mM KNO3 as the most favorable priming dose for maximizing combined ionic balance and salt tolerance across salinity levels. Finally, the integration of MARS and PSO advances salinity research from descriptive assessment toward a more prescriptive, data-driven framework for optimizing crop performance under saline conditions.