Abstract
Potato (Solanum tuberosum L.) is a chlorine-sensitive crop. When soil Cl(-) concentrations exceed optimal thresholds, the yield and quality of potatoes are limited. Consequently, chloride-containing fertilizers are rarely used in actual agricultural production. Therefore, two years of field experiments under natural rainfall regimes with three chlorine application levels (37.5 kg ha(-1)/20 mg kg(-1), 75 kg ha(-1)/40 mg kg(-1), and 112.5 kg ha(-1)/60 mg kg(-1)) were conducted to investigate the leaching characteristics of Cl(-) in field soils with two typical textures for Northeast China (loam and sandy loam soils). In this study, the reliability of Cl(-) residual estimation models across different soil types was evaluated, providing critical references for safe chlorine-containing fertilizer application in rain-fed potato production systems in Northeast China. The results indicated that the leaching efficiency of Cl(-) was significantly positively correlated with both the rainfall amount and the chlorine application rate (p < 0.01). The Cl(-) migration rate in sandy loam soil was significantly greater than that in loam soil. However, the influence of soil texture on the Cl(-) leaching efficiency was only observed at lower rainfall levels. When the rainfall level exceeded 270 mm, the Cl(-) content in all the soil layers became independent of the rainfall amount, soil texture, and chlorine application rate. Under rain-fed conditions, KCl application at 80-250 kg ha(-1) did not induce Cl(-) accumulation in the primary potato root zone (15-30 cm), suggesting a low risk of toxicity. In loam soil, the safe application range for KCl was determined to be 115-164 kg ha(-1), while in sandy loam soil, the safe KCl application range was 214-237 kg ha(-1). Furthermore, a predictive model for estimating Cl(-) residuals in loam and sandy loam soils was validated on the basis of rainfall amount, soil clay content, and chlorine application rate. The model validation results demonstrated an exceptional goodness-of-fit between the predicted and measured values, with R(2) > 0.9 and NRMSE < 0.1, providing science-based recommendations for Cl-containing fertilizer application to chlorine-sensitive crops, supporting both agronomic performance and environmental sustainability in rain-fed systems.