Moisture Migration Behavior During Paddy Rice Drying Under Multi-Factor Stress and Its Effects on Starch, Protein, and Antioxidant Enzyme Activities

To investigate the effects of drying conditions on moisture migration and quality changes in paddy rice, freshly harvested grain from northeastern China in autumn 2025 was subjected to 30 thin-layer drying treatments at 30–50 °C, 20–30% relative humidity, and 3–4 m/s air velocity. Three-dimensional interpolation surfaces of mean water-loss rate were constructed, and apparent effective moisture diffusivity was analyzed to characterize moisture migration. Six models—Newton, Page, Henderson–Pabis, Logarithmic, Wang–Singh, and Midilli—were fitted to evaluate their potential for process applications. Changes in dry basis protein and amylose contents and in peroxidase (POD) and superoxide dismutase (SOD) activities before and after drying were also examined. The results showed that temperature was the domi nant factor affecting the drying process. When the temperature increased from 30 to 50 °C, the mean drying time decreased from 125.00 to 48.33 min, and the mean water-loss rate increased from 0.204 to 0.662 g/min. The water-loss-rate surfaces rose along the direction of increasing temperature, with small fluctuations associated with changes in relative humidity and air velocity. The mean bed-scale apparent effective moisture diffusivity (Deff*) increased from 1.572 × 10−8 to 3.843 × 10−8 m2/s. Among the six drying models, Midilli provided the best in-sample fit, with an R2 of 0.9988 and an RMSE of 0.0082. It also generally yielded lower Akaike information criterion (AIC) values, suggesting its potential for detailed characterization of drying kinetics and process-oriented applications. The Logarithmic model ranked second, while the Page model also offered practical advantages owing to its simplicity. Temperature had a significant main effect on protein content (p = 0.011), with the lowest mean at 50 °C (7.900 g/100 g). The main effect of air velocity on amylose content was significant, but it did not remain significant in the HC3 sensitivity analysis. None of the three drying factors had a significant main effect on POD or SOD activity. The Pearson correlations of protein and amylose with SOD were significant; after Holm correction, only the protein–SOD negative correlation remained significant. Higher temperatures improved drying efficiency within the tested range, whereas quality indicators responded differently. Based on the observed changes in drying efficiency and quality indicators, 35–40 °C can be regarded as a candidate temperature range that balances dehydration efficiency and quality preservation, providing a reference for subsequent process optimization. These findings provide physiological and biochemical information to support the coordinated evaluation of moisture migration and quality during thin-layer paddy drying and the future development of intelligent drying-control models.