Tianyang District is situated in the western region of Guangxi, within the Youjiang River Valley at coordinates 106°91′ E, 23°73′ N. The area experiences a southern subtropical monsoon climate characterized by a high average temperature, ample sunlight, plentiful rainfall, an extended frost-free period, and conducive conditions for year-round cultivation. The average annual temperature is 22°C, with an annual precipitation of 1100 mm. Heat peaks occur thrice annually in this region.

The experimental data were initially organized using Microsoft Excel 2020 software. IBM SPSS Statistics 22 software was employed for variance analysis, multiple comparisons, Pearson correlation analysis, and principal component analysis. Redundancy analysis was conducted using Canoco 5 software, and graphical representations were generated using Origin 2022.

As illustrated in Table 2 , the plant height reached its peak (123.00 cm) with the NPK fertilization treatment and was lowest in the control group. N deficiency treatment exhibited a substantial influence on the number of effective panicles per hectare. The highest number of effective panicles (14.92; tens of thousands) was observed in the complete fertilization treatment, while the lowest (11.38; tens of thousands) was recorded in the control group. Additionally, the grain-to-panicle ratio was notably lower in the complete fertilization treatment compared to the other experimental treatments. The grain yield in the complete fertilizer treatment reached 7680 kg/hectare, exhibiting a 25.60% increase compared to the control group. Furthermore, the highest straw yield was recorded in the complete fertilizer treatment, reaching 11098.20 kg/hectare, signifying a 32.43% increase relative to the control group.

The results of the principal component analysis conducted during the vegetative growth period are presented in Table 4 . Based on the extraction of eigenvalues greater than 1 for the principal components, a total of five principal components were extracted. The contribution rates of the five principal components were 17.883%, 16.413%, 14.114%, 11.483%, and 9.303%, respectively, resulting in a cumulative contribution rate of 69.195%. Principal component 1 was influenced by soil nitrate N, plant catalase, and chlorophyll; principal component 2 was influenced by soil TP and TN; principal component 3 was impacted by soil available K; principal component 4 was influenced by soil ammonium N and available P; and principal component 5 was affected by plant catalase and malondialdehyde. The results of the principal component analysis during the reproductive growth stage are presented in Table 5 . Following the extraction of eigenvalues greater than 1 for the principal components, a total of five principal components were extracted for this period. The contribution rates for these components were 20.289%, 18.620%, 14.042%, 12.361%, and 9.175%, resulting in a cumulative contribution rate of 74.487%. Principal component 1 was influenced by soil TP, total K, and plant catalase; principal component 2 was influenced by soil nitrate N and plant proline; principal component 3 was impacted by soil ammonium N, available P, and available K; principal component 4 was influenced by plant malondialdehyde and chlorophyll; and principal component 5 was affected by soil TN and plant peroxidase. The factors influencing the principal components extracted in the two periods varied, with the cumulative contribution rate in the reproductive growth period being higher than that in the vegetative growth period.

The scores of the five principal components and the overall scores of different treatments during the rice vegetative growth stage are presented in Table 6 . The comprehensive scores of the 12 indicators for each treatment were computed using F=0.258*F1 + 0.237*F2 + 0.204*F3 + 0.166*F4 + 0.134*F5. The ranking of the scoring results is as follows: the score for full fertilization was higher than that of the N deficiency treatment, followed by the K deficiency treatment, the P deficiency treatment, and finally, the blank treatment. The comprehensive scores of the five principal components for the various treatments at reproductive growth are presented in Table 6 . The comprehensive scores of the 12 indicators of each treatment were calculated using F=0.272*F1 + 0.250*F2 + 0.189*F3 + 0.166*F4 + 0.123*F5, and the order of the results is as follows: the score for the N deficiency treatment was higher than that of full fertilization, followed by the blank treatment, P deficiency, and K deficiency. The highest comprehensive scores were observed in the N deficiency treatment during the reproductive growth stage and the full fertilization treatment during the vegetative growth stage.

In Figure 6A , an exceptionally significant positive correlation was observed between chlorophyll and CAT (P< 0.01). Additionally, significant positive correlations were found between TP and TK, available K and proline, ammonium N and MDA, as well as TN and MDA (P< 0.05). Furthermore, a significant negative correlation was noted between POD and MDA (P< 0.05). No significant differences were observed among the other indicators. In Figure 6B , significant positive correlations were observed between TP and TK, as well as between TP and PRO. A significant negative correlation was identified between TP and CAT (P< 0.05), while a highly significant positive correlation was found between nitrate N and PRO (P< 0.01). Additionally, a significant positive correlation was noted between soil TK and TP across the two periods (P< 0.05).

In Figure 7A , the cumulative percentage of variance explained by the two ordination axes in the RDA result ranged from 29.23% to 38.39%. The ranking outcomes provide insights into the relationship between soil nutrients and rice enzyme activities. During the vegetative growth period, positive correlations were observed between PAC and PRO, as well as SOD, while negative correlations were noted between NAC and MDA, CAT, POD, and ROS. In contrast, during the reproductive growth period ( Figure 7B ), PRO, MDA, and ROS were negatively correlated with PAC and positively correlated with NAC.

Nutritional stress exerts a significant influence on the agronomic traits of rice. The findings of this study reveal that plant height, number of effective panicles, 1000-grain weight, and yield were reduced under conditions of N deficiency, P deficiency, and K deficiency compared to full fertilization. In conditions of N deficiency, the plant height and number of effective panicles in rice were observed to be the lowest among all treatments, highlighting the crucial role of N in these agronomic traits. Considering all yield factors collectively, it is evident that N, P, and K positively influenced the vegetative growth of rice, aligning with the findings reported by Zhang (Zhang et al., 2022). Research findings indicate that as the quantities of N, P, and K applied in the fertilization regimen were decreased, there was a corresponding decrease in rice yield (Wu and Zhang, 2023). The experimental results indicate that fertilization has the potential to enhance rice yield. N and P were identified as the primary limiting factors for rice yield, with N exerting the most significant influence on yield-related factors such as plant height, effective panicles, and 1000-grain weight. The constraint on growth imposed by nutritional stress may stem from the inadequate acquisition of essential and well-balanced nutrients by the organism.

Research findings suggest that the application of N fertilizer plays a crucial role in the vegetative growth and yield of rice (Djaman et al., 2016). As the rice plants developed, the activation coefficient of N progressively increased. The concurrent application of N and P enhanced the utilization efficiency of soil N. Similarly, the combined application of P and K improved the utilization efficiency of soil P and K. Numerous studies (Darmawan et al., 2009; Djaman et al., 2016) have demonstrated that P application can elevate the available P content in the soil, a finding that was corroborated by the results of the present study. P is a vital nutrient essential for crop metabolism, collaborating with N fertilizer to enhance rice growth. In this investigation, a highly significant negative correlation was observed between soil available P and ammonium N during the rice reproductive growth period, aligning with findings from prior research (Xiao et al., 2023). In cases of K deficiency, crops exhibit reduced absorption, underscoring the significance of K fertilizer application in augmenting the available K content in the soil (Nakamura et al., 2010). Consequently, fertilization enhances the soil NAC and KAC while diminishing the PAC.

With increasing stress intensity, there was an upward trend observed in POD activity. Specifically, during the rice vegetative growth period, the POD activity surpassed that of the full fertilization treatment in the K-deficient treatment, contrary to the findings reported by Zhu et al (Zhu et al., 2024). The analysis of CAT activity changes revealed a decreasing trend over time in the P deficiency treatment across the two periods, consistent with findings from prior studies (Huang et al., 2013; Hussain et al., 2017; Hussain et al., 2019). As a response mechanism to reactive oxygen stress, SOD exhibited a pattern similar to the changes in reactive oxygen concentration. There was no significant change observed between the two periods. However, under adverse stress conditions, SOD activity notably decreased (Wang et al., 2019). In the K deficiency treatment during the reproductive growth stage, there was a corresponding increase in SOD activity with the rise in ROS concentration, aligning with previous research findings. In an experiment involving foliar fertilizer application on rice, Peng Jiayuan (Peng et al., 2023) observed that the chlorophyll content of rice was elevated in the later growth stage, followed by a subsequent decrease until stabilization during the reproductive growth period. Stress was identified to exert an inhibitory effect on photosynthesis, a finding that is in line with the outcomes of the present study (Fariduddin et al., 2009).

In conclusion, CAT serves as the primary decomposer of H₂O₂ during the vegetative growth period, whereas POD plays a predominant role during the reproductive growth period. The alterations in SOD activity correspond with changes in reactive oxygen levels. These three enzymes collaborate to regulate the free radical content within the plant at a stable level, thereby preventing physiological and biochemical changes induced by free radicals. As the growth period advanced, N and P stress led to an accumulation of proline in the plants. The severity of membrane lipid peroxidation induced by N and K stress decreased, whereas N and P stress exacerbated membrane lipid peroxidation. To enhance the stress resistance of rice, it is recommended to reduce P fertilizer application and increase K fertilizer input during the reproductive growth period.

Soil nutrients have a direct impact on enzyme activities. In conditions of high N availability, plants can uptake a substantial amount, leading to an increase in proline and malondialdehyde contents in the plants. Concurrently, the activity of POD also increases accordingly. A study (Yan et al., 2021) has demonstrated that inadequate or excessive soil N supply resulted in reduced activities of SOD and POD, along with an accumulation of reactive oxygen species. Additionally, this condition led to elevated levels of the membrane lipid peroxidation product MDA and accelerated leaf senescence. In this experiment, POD, CAT, SOD, and MDA were all positively correlated with N, aligning with findings from a prior study (Meng, 2023). K, being a crucial constituent of plant enzymes, directly participates in substance synthesis. Elevated soil K content typically enhances plant enzyme activity. Nevertheless, in this experiment, the association between K and plant enzymes was not robust, contrary to the findings reported in previous studies (Meng, 2023). With prolonged stress exposure, PRO was notably influenced by K, while CAT was affected by P. The application of P and K could mitigate stress-induced damage and decrease the levels of oxidative substances. The relationship between plant photosynthesis, proline, and CAT suggests a synergistic effect, signifying that the efficient advancement of photosynthesis enhances enzyme activity. The potential explanation could be that organisms experiencing nutritional stress require extra energy and resources to address nutrient deficiencies, hindering their ability to efficiently mobilize defense mechanisms against other adverse factors. This weakened response severely impacts the plant’s survival and adaptation capabilities. Furthermore, under nutritional stress conditions, plants may enhance their capacity to uptake and transport specific nutrients. They can also improve their efficiency in acquiring limited nutrients from the soil by modifying their root structure and physiological functions in response to nutrient deficiencies.