Skip to main content
NCBI home page
Food Chemistry: X logoLink to Food Chemistry: X
. 2026 Mar 28;35:103797. doi: 10.1016/j.fochx.2026.103797

Multi-year field evidence for minimum nighttime temperature thresholds in rice fatty acid composition during grain filling

Wenjie Yu a,b,1, Qiuning Wang a,b,1, Qifang Sun c, Qiming Zhu a,b, Tao Liu a,b, Shengxiang Yi c, Gaowei Zhang c, Jie Zhu a,b, Jingze Cao a,b, Jinwang Li a,b, Yan Li d, Xiaoqing Fan d, Xiali Guo c, Yanling Song e,⁎, Liping Luo a,b,⁎,2
PMCID: PMC13068580  PMID: 41971403

Abstract

Fat content shapes rice eating and nutritional quality, yet its climate response remains unclear. In this study, we combined six years of multisite field observations across nine provinces in southern China with an interpretable random forest model to identify drivers of rice fat. The model showed that the minimum nighttime temperature during grain filling (TMIN) was identified as the primary factor. Fat content increased with TMIN, whereas the unsaturated fatty acid index (IUFA) exhibited three distinct accumulation patterns and generally declined at higher TMIN. This pattern indicates a physiological tradeoff potentially linked to changes in energy metabolism and membrane stability of rice. Patterns were independently validated in 2024 field trials at 11 sites. Furthermore, projections indicate higher fat under warming but declining IUFA, especially in early- and mid-season rice. These results support climate-resilient cultivation, sustainability, and quality management of rice.

Keywords: Rice, Climate change, Minimum temperature, Fat, Fatty acids

Highlights

  • •

    Interpretable RF predicts rice fat from climate using 6-year, 9-province field data

  • •

    TMIN is identified as the primary climate driver influencing rice fat content

  • •

    Three TMIN ranges and their fat and fatty acid accumulation patterns are revealed

1. Introduction

Climate change has increased the risk of the climate system reaching critical thresholds, primarily due to rising temperature, which manifests as rising sea levels, more frequent extreme weather events, and rapid sea ice melt (Forster et al., 2023). All regions of the world are facing unprecedented climate change. One of the most concerning findings is that its negative impacts have been more extensive and severe than anticipated; particularly in relation to food quality and safety (Halford et al., 2015). Recent studies have revealed that intense global warming and elevated CO2 levels will lead to widespread reductions in minerals; proteins; and vitamins in crops; exacerbating nutritional deficiencies for over 600 million people worldwide and intensifying global “hidden hunger” and obesity (Kong et al., 2024; Loladze, 2014; Zhu et al., 2018). Moreover, climate change has already significantly impacted the appearance, processing characteristics, and eating quality of crops, have already, posing major challenges to the stability and development of agricultural production, commercial supply chains, and socioeconomic systems (Custodio et al., 2019; Xu, Zhao, & Zhai, 2020). Global temperatures are projected to rise by 1.9 °C and 4.0 °C by 2040 and 2100, respectively, accompanied by an increase in the frequency of various extreme weather events (Meinshausen et al., 2022). Climate change will intensify changes in crop quality, posing significant challenges to the production of high-quality agricultural products and the assurance of food quality and safety in the future.

Rice (Oryza sativa L.) is a cornerstone of global food production, feeding more than half of the world's population (Wu et al., 2022). Consumer demand for rice quality has reached unprecedented levels as living standards improve and the pursuit of a higher quality of life increasingly intensifies (Cao et al., 2024; Gong et al., 2023). Fat, an important component after starch and protein in rice, makes up 1% to 3% of rice, playing a crucial role in its appearance, texture, and nutritional quality (Wang et al., 2006; Wu et al., 2020). Researchers have shown that rice with higher fat content has a more appealing and unique lustre, making it more favoured by consumers. Additionally, fat can reduce starch water absorption and swelling by forming stable helical complexes with amylose, raising the gelatinization temperature and significantly improving rice palatability of rice (Chen et al., 2012). Moreover, rice fat is the primary source of rice bran oil, a major by-product of the grain industry, which is widely regarded as one of the healthiest edible oils.

Rice fat is primarily composed of fatty acids such as C18:1 (oleic acid), C18:2 (linoleic acid), and C16:0 (palmitic acid), with smaller amounts of C18:3 (linolenic acid), C18:0 (stearic acid), and C20:0 (arachidic acid), whereas unsaturated fatty acids. These unsaturated fatty acids confer health benefits to rice bran oil, including protective effects against cardiovascular disease, thereby contributing to its nutritional value (Wu et al., 2020). Furthermore; studies have shown that unsaturated fatty acids in rice positively influence its palatability (Wu et al., 2020). The index of unsaturated fatty acids (IUFAs) is regarded as an important indicator for evaluating the degree of fat unsaturation. In summary, fat and fatty acids are integral to both grain quality and the value of its by-product, rice bran oil. However, there is a lack of research on the impact of climate change on rice fat and fatty acids. Investigating the mechanisms underlying the changes and accumulation of fat and fatty acids in rice is strategically crucial for developing high-quality rice and enhancing the stability of future food supply chains.

Fatty acids composition adjusts in response to environmental changes to maintain the stability of cell membranes and other critical biological functions (Holm et al., 2022). The content and composition of fat and fatty acids in plants; animals; and even plankton have been significantly affected by climate change (Holm et al., 2022; Ouellet et al., 2021; Wu et al., 2020). The grain-filling period is critical time for rice quality, with temperature, light, and precipitation significantly influencing rice fat content (Jianguo, 2008; Lu et al., 2012; Wang et al., 2023). Studies have shown that insufficient light reduces rice fat synthesis but increases the proportion of unsaturated fatty acids, while moderate water control promotes the accumulation of various fatty acids during the grain-filling period (Wu et al., 2020). Notably; increasing in both daytime and nighttime temperatures have differential impacts on rice fat accumulation: high daytime temperatures inhibit fat synthesis; while high nighttime temperatures promote fat accumulation (Cooper et al., 2008; Kitta et al., 2005; Zeng et al., 2011). Moreover, studies examining different temperature thresholds have yield conflicting results. For instance, both low and high temperatures reduce rice fat content (Song et al., 2011; Zeng et al., 2011), whereas Wang et al. (2023) suggested that high temperatures promote fat accumulation. These studies indicate that different levels of climate change may have varying impacts on rice fat and fatty acids (Li et al., 2024). Although some researchers have explored the effects of climate change on rice fat and fatty acids; most studies have been conducted under stress or near-stress conditions; limiting their applicability in real-world agricultural production; this could exaggerate or underestimaten climatic effects; leading to inconsistencies in the research findings (Huang et al., 2023; Wernberg et al., 2012). Furthermore, controlled environments such as growth chambers and greenhouses could alter climate factors like water availability, light conditions, air turbulence, and temperature, compared to field experiments, leading to discrepancies between study results (Li et al., 2024). These make climate drivers play a dominate role in changes to rice fat under complex climate conditions in agricultural production. However, how these changes affect the content and composition of fatty acids remains unclear. Therefore, research on the impact of climate change on rice fat and fatty acids composition under real agricultural conditions is urgently needed to develop targeted breeding and adaptation strategies for high-quality rice in a changing climate.

In recent years, machine learning has become a key approach for uncovering changes in crop quality under complex climate variability. AI can automatically extract latent patterns from multi-source, heterogeneous, and large-scale spatiotemporal data, learn complex nonlinear relationships, and generalize across conditions, which confers clear advantages (Yu et al., 2025). For example; Kheir et al. developed a hybrid learning framework that combined traditional algorithms with automated modeling; integrating random forests and neural networks to predict Egyptian wheat yield and grain protein; iron; and zinc concentrations with high accuracy across 2400 observations (R2 = 0.78); and they extended the framework to future quality projections under CMIP6 scenarios (Kheir et al., 2024). Zhou et al. combined XGBoost with SHAP to evaluate the effects of water and temperature on eight winter wheat quality indicators in the North China Plain; showing that water supply generally outweighed temperature in importance and that interactions between warm dry and warm wet conditions produced markedly different quality responses (Zhou et al., 2025). Taken together, machine learning has become a central tool for revealing and forecasting climate driven variation in crop quality.

China is one of the world's leading rice, accounting for approximately 30% of global production. Southern China, known for its warm climate, is particularly significant, enabling multiple rice crops annually and serving as the country's main rice-producing region. This study presents a seven-year field experiment of early-season, mid-season, and late-season rice planting patterns across nine provinces in southern China, with aims to investigate the sensitivity of rice fat content and fatty acid composition to climate change under real agricultural conditions. This region contributes nearly 20% of the world's rice, playing a crucial role in global food quality and security. The objectives of this study were to: (1) develop a comprehensive predictive model for rice fat based on multiple climate factors and identify the primary climate drivers fat accumulation; (2) elucidate the regulatory patterns of rice fat and fatty acids influenced by key climate factors; (3) forecast the spatiotemporal evolutionary trends of rice fat and fatty acids under future climate change scenarios. This research aims to provide robust experimental evidence to support the optimization of adaptive strategies in rice-growing regions, guide targeted breeding improvements, enhance agricultural adaptation, and promote high-quality rice production under climate change.

2. Materials and methods

2.1. Experimental stations

From 2018 to 2023, field experiments on staged rice sowing were performed at agricultural meteorological experimental stations in nine major rice-producing provinces in southern China (n = 244) (Table S1; Fig. S1). The experiments adopted a completely randomized block design, with four blocks and four sowings spaced 10 d apart within each sowing season. All selected experimental sites planted the predominant local varieties and followed consistent cultivation management and agronomic practices to ensure the comparability and reliability of the experimental results. Single-cropping (mid-season) rice was grown in Hubei, Jiangsu, Anhui, Zhejiang, and Fujian provinces, while double-cropping (early-season and late-season) rice was cultivated in Hunan, Jiangxi, Guangxi, and Guangdong provinces, following local cultivation habits and agronomic management practices (Pan et al., 2021). To validate the consistency and generalizability of findings from the 2018–2023 period; the same experimental protocol was implemented at these sites in 2024 (n = 44). The experimental sites spanned ecological regions from tropical to subtropical monsoon climates; covering major rice-producing provinces in China and accounting for approximately 67% of the country's total rice production (Xu, Yuan, & Man, 2020).

2.2. Chemicals and reagents

Analytical-grade petroleum ether, hydrochloric acid, methanol, sodium hydroxide, boron trifluoride, n-heptane, sodium chloride, and anhydrous sodium sulfate were purchased from Aladdin (Shanghai, China). Standard solutions of 12 fatty acids, including C12:0 (lauric acid), C14:0 (myristic acid), C16:0 (palmitic acid), C16:1 (palmitoleic acid), C18:0 (stearic acid), C18:1 (oleic acid), C18:2 (linoleic acid), C20:0 (arachidic acid), C20:1 (gadoleic acid), C20:3 (linolenic acid), C22:0 (behenic acid), and C24:0 (lignoceric acid), were obtained from Sigma (St. Louis, MO, USA).

2.3. Fat and fatty acid contents of rice

All harvested rice samples were dried, dehulled, ground, and sieved through a 200-mesh screen. The rice fat content was determined according to the China national standard GB 5009.6–2016. Two grams of rice powder was transferred to a filter paper thimble. Then, the thimble was placed in a Soxhlet extractor with petroleum ether for 8 h. After extraction, the receiving flask was heated to evaporate the petroleum ether until the remaining volume was reduced to 1–2 mL. Finally, the remaining solution was dried to a constant weight, and the final weight was recorded. The fat content in the sample was determined by calculating the difference between the initial and final weights of the receiving flask.

The composition and content of fatty acids in rice were determined according to the national standard GB 5009.168–2016. Two grams of rice powder were used to extract the fat using the Soxhlet extraction method. Then, the extracted fat was mixed with 2 mL of 2% (v:v) sodium hydroxide methanol solution in a water bath at 80 °C for 30 min. Then, 3 mL of 15% boron trifluoride methanol solution was added in the water bath at 80 °C for 30 min. After the phases were separated, the upper layer was dehydrated with anhydrous sodium sulfate and allowed to stand for 5 min before the upper layer was collected again. Finally, the solution was filtered for gas chromatography analysis.

A gas chromatograph-mass spectrometer (Thermo, Trace1310 ISQ) equipped with an HP-5MS capillary column (30 m × 0.25 mm × 0.25 μm) was used for the analysis of fatty acids in rice. The temperature program was as follows: the reaction was maintained at 80 °C for 1 min, then increased to 200 °C at a rate of 10 °C/min, followed by an increase to 250 °C at 5 °C/min, and finally to 270 °C at 2 °C/min and held for 3 min. Detection was performed in splitless mode, and the injector temperature and carrier gas flow rate were set at 290 °C and 1.2 mL/min, respectively. Both the ion source and transfer line temperatures were set to 280 °C. The solvent delay time was 5 min, with a scan range of m/z 30 to 400.

2.4. Meteorological data

The detailed climate data were collected from local agrometeorological stations at nine experimental sites in southern China from 2018 to 2024. The meteorological data obtained included daily average temperature, maximum temperature, minimum temperature, precipitation, and sunshine duration. A total of 105 climate factors during the rice grain-filling period were calculated using daily climate data to evaluate the differential impacts of climate change across different scales and thresholds (Ingvordsen et al., 2015). These factors primarily included temperature changes at different thresholds, accumulated temperatures at various levels, variations in precipitation and its accumulation, changes in sunshine duration and its accumulation, and the number of consecutive days without sunlight, with precipitation, and without precipitation (representing extreme weather events) (Table S2).

The BCC-CSM2-MR is a medium-resolution climate model developed by the National Climate Center of China, incorporating future climate scenarios under varying emission levels. It has been widely used in studies forecasting future climate in China (Yu et al., 2025). In this study; climate data was obtained from the BCC-CSM2-MR model under the Coupled Model Intercomparison Project 6 (CMIP6) using three different Shared Socioeconomic Pathways (SSPs) to predict future climate change: SSP126 (low emission scenario; with warming limited to below 2 °C by 2100); SSP585 (highest emission scenario; with warming exceeding 4 °C by 2100); and SSP245 (medium emission scenario; with warming between 2 °C and 3 °C by 2100). Two representative time periods; the 2040s (2030–2050) and the 2070s (2060–2080); were selected to represent climate change in different future decades (Zhang et al., 2023).

2.5. Data analysis

2.5.1. Explainable machine learning model

Feature engineering is the process of extracting, selecting, and constructing useful features from raw data in machine learning, with the goal of enhancing model prediction accuracy, reducing the risk of overfitting, and improving computational efficiency (Fan et al., 2019). In this study, all constant terms were excluded from the climate factors. Then, a significance analysis (p < 0.05) was performed to assess the relationship between the retained climate factors and rice fat content, further eliminating those climate factors that were not significantly associated with changes in rice fat content. Next, correlation analyses were performed among all retained climate factors and between the climate factors and rice fat content. Climate factors with a correlation coefficient ≥ 0.5 were grouped together. Within each group, only the climate factor with the highest correlation coefficient to fat content was retained, in order to eliminate multicollinearity and retain climate factors that showed a significant impact on the model.

Shapley Additive exPlanations (SHAP) analysis is a model-agnostic machine learning interpretation method based on game theory used to evaluate and explain the black-box nature of machine learning models. SHAP highlights the contribution of each feature to the model's predictions, both globally and locally, by calculating the Shapley values for each feature, thereby explaining the model's prediction outcomes (Eyring et al., 2024). Random Forest (RF) is an ensemble learning model based on decision trees; each of which splits input data based on parameters like tree structure. RF can perform various tasks; including classification; regression; clustering; and variable selection. RF is resistant to overfitting and outliers; provides effective model feature explanations; and demonstrates strong predictive performance in modeling various climate-driven changes in crop yield and quality (Balogun & Tella, 2022). In this study, an RF model was developed to predict rice fat content based on multiple climate factors, using SHAP values derived from the TreeExplainer, which leveraged the structural characteristics of tree models.

2.5.2. Statistical analysis

Microsoft Excel was used for the organization and summarization of raw data to prepare for further analysis. Normality tests, significance analysis, correlation analysis, and the construction, interpretation, and visualization of machine learning models were all performed using Python (version 3.12.3). The spatial distribution of future climate was visualized using ArcMap (version 10.8).

3. Results

3.1. Rice fat content across different experimental stations over the years

The fat content of the 244 rice samples ranged from 1.10% to 2.96% (average = 2.01 ± 0.35%). A normality test was performed based on the fat content data of all rice samples prior to further statistical analyses (Fig. S2). The results showed that the fat content levels exhibited typical characteristics of a normal distribution, meeting the criteria of the Kolmogorov-Smirnov (p = 0.3118), Shapiro-Wilk (p = 0.2244), and Jarque-Bera (p = 0.8954) statistical tests. Based on the results of these normality tests, parametric tests such as one-way ANOVA and Pearson correlation analysis were selected for further statistical analysis.

The results of statistical analysis showed that the rice fat content was varied significantly across different experimental stations (Fig. S3A—C). Higher fat content was detected in early-season rice in Guangdong (2.41 ± 0.24%) and Guangxi (2.28 ± 0.27%), while mid-season rice in Jiangsu showed the lowest fat content (1.58 ± 0.30%). It was noted that the fat content of rice from the same experimental station showed considerable variations across different years. For example, the fat content of early-season rice in Jiangxi was decreased by over 30% from 2018 to 2019, while the fat content of mid-season rice in Jiangsu was increased by nearly 75% from 2019 to 2020. Furthermore, the fat content of the three types of rice across different sowing periods was decreased with increasing latitude, showing a clear geographical distribution pattern (Fig. S3D—F). These results indicated that, even with fixed cultivars and management practices each year, the fat content of rice still varied significantly across different experimental stations and different years, highlighting the potentially important impact of climate change on rice fat accumulation.

3.2. Construction of explainable machine learning model

3.2.1. Construction and validation of RF model

The RF model was constructed based on SHAP to predict the impact of multiple climate factors on rice fat content (Fig. 1A). The training and testing sets were set at 70% and 30%, respectively. The scatter plot and fitted curve of the RF model showed high consistency between the actual and predicted values, with R2 values of 0.6246 and 0.6730 for the testing and training sets, respectively (Fig. 1B). Additionally, the RF model maintained low Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Mean Squared Error (MSE) in both the training and testing sets, with minimal differences revealed between the two sets. The residual plot showed a uniform distribution, indicating strong predictive performance and generalization capability without systematic bias (Fig. 1C, D). The learning curves for the training and testing sets also showed an upward pattern and convergence as the number of iterations was increased, suggesting low likelihood of overfitting and strong generalization ability (Fig. 1E).

Fig. 1.

Fig. 1

Open in a new tab

Interpretable random forest model construction process (A), scatterplot of actual versus predicted values (B), test set versus training set model prediction parameters (C), residual plot (D), and learning curve (E).

3.2.2. Screening of the main climate factors driving fat content

SHAP analysis was performed to provide the global interpretation of the RF model to explain the role of each climate factor and to screen for the main driving climate factors. The results indicated that temperature-related climate factors had the most significant impact on rice fat content, particularly the different thresholds of TMIN (i.e., daily mean minimum temperature during the rice grain-filling period). Changes in TMIN-related climate factors played the most significant role in predicting rice fat content (Fig. 2A, B). According to the TMIN-SHAP dependence plot, rice fat content was significantly increased as temperature rose (Fig. 2C). When TMIN exceeded approximately 25 °C, a more pronounced accumulation of rice fat was observed, suggesting that variations in TMIN across different ranges led to distinct accumulation trends in rice fat. Furthermore, the results of comparative analysis of rice fat accumulation at different TMIN thresholds (Fig. 2D) showed that rice fat content was increased more significantly as the elevation of TMIN threshold occurrences rose. Specifically, when TMIN exceeded 24 °C, the rice fat content was increased by more than 4.60%, compared to that under 18 °C, and the occurrence of extreme TMIN (30 °C) led to an increase in rice fat content of over 10%, compared to lower TMIN (18 °C).

Fig. 2.

Fig. 2

Open in a new tab

Importance ranking of climatic factors based on SHAP analysis (A), eigenvalues and classification of climatic factors (B), SHAP dependence curves of TMIN (C), and relative trends (95% confidence intervals) in fat content with TMIN exceeding different thresholds (the number of days of TMIN with different thresholds is given in parentheses) (D).

3.3. Variation patterns in fatty acids with TMIN

The rice fatty acid composition was further analyzed to elucidate the specific impact of TMIN changes on the variation in rice fat content (Table S3). A total of 12 fatty acids were detected in all rice samples with varied concentrations, including C12:0 (4.52 ± 1.63 mg/kg), C14:0 (129.77 ± 42.92 mg/kg), C16:0 (4345.26 ± 1333.88 mg/kg), C16:1 (34.43 ± 14.53 mg/kg), C18:0 (453.57 ± 217.76 mg/kg), C18:1 (7383.54 ± 2528.69 mg/kg), C18:2 (4753.59 ± 1411.81 mg/kg), C20:0 (90.29 ± 52.74 mg/kg), C20:1 (56.03 ± 20.25 mg/kg), C20:3 (69.57 ± 49.32 mg/kg), C22:0 (79.56 ± 29.63 mg/kg), and C24:0 (204.60 ± 84.43 mg/kg). Notably, C16:0, C18:1, and C18:2 were the dominant fatty acids in all rice samples, together accounting for more than 90% of total fatty acids.

Most fatty acids were significantly and positively correlated with each other, suggesting that they increased in parallel during fat accumulation. In addition, most fatty acids tended to increase with rising TMIN, with significant changes observed for C12:0, C14:0, C16:0, C18:0, and C18:1 (Fig. 3A). Surprisingly, although an increase in TMIN led to an overall increase in fat content, not all fatty acids exhibited a continuous accumulation trend, and the increases in the levels of fatty acids varied across different TMIN ranges. Within the range of 10 °C to 24 °C, an increase in TMIN led to an increase in the level of all fatty acids, ultimately contributing to fat accumulation. However, at TMIN above 24 °C, the increase in TMIN mainly resulted in the accumulation of specific fatty acids, such as C16:0, C18:0, C18:1, and C20:0 (Fig. 3B). These results indicated that the appearance of TMIN at different thresholds caused differential accumulation of fatty acids. Consistent with this shift, mean C16:0 increased from 3425.12 ± 867.79 mg/kg at TMIN≤24 °C to 4255.73 ± 1283.14 mg/kg at TMIN >24 °C, while C18:1 increased from 6464.96 ± 2555.60 to 7530.29 ± 2448.48 mg/kg. These two fatty acids largely accounted for lipid accumulation at higher TMIN.

Fig. 3.

Fig. 3

Open in a new tab

Characterization of rice fatty acids. (A) Correlation analysis of TMIN and the 12 fatty acids detected in rice. Positive and negative correlations are indicated in red and blue, respectively. Correlation coefficients are provided in the upper right triangular matrix. Levels of significance are represented by p < 0.05 (*), p < 0.01 (**), and p < 0.001 (***), respectively. (B) Variation patterns in rice fatty acid composition based on changes in TMIN. (C) Proportions of monounsaturated fatty acids and polyunsaturated fatty acids as well as variations in the index of unsaturated fatty acids (IUFAs) based on changes in TMIN. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

The totals of saturated fatty acids (SFAs), monounsaturated fatty acids (MUFAs), and polyunsaturated fatty acids (PUFAs) were largely driven by a small number of constituent fatty acids (Fig. S4). SFAs showed the strongest positive correlations with C16:0 and C18:0, and were also positively correlated with other saturated fatty acids, including C12:0, C14:0, C20:0, C22:0, and C24:0. MUFAs were most strongly correlated with C18:1, and were also positively correlated with other monounsaturated fatty acids, such as C16:1 and C20:1, underscoring the central contribution of C18:1. PUFAs were most strongly correlated with C18:2 (with the highest statistical significance). Among meteorological variables, TMIN was significantly positively correlated with MUFAs and also showed a positive association with SFAs, whereas its correlation with PUFAs was weaker or not significant.

Further analysis indicated that the proportions of monounsaturated and polyunsaturated fatty acids, as well as the IUFAs, also showed differential changes with variations in TMIN (Fig. 3C). At lower TMIN ranges, the proportion of polyunsaturated fatty acids was increased along with IUFAs. However, at TMIN above approximately 16 °C, the proportion of polyunsaturated fatty acids and IUFAs were significantly decreased. The major SFA C16:0 increased from 3044.34 ± 684.85 mg/kg at TMIN ≤16 °C to 3911.68 ± 1171.28 mg/kg at TMIN >16 °C. The major MUFA fraction increased from 6040.63 ± 2370.55 to 7081.52 ± 2561.66 mg/kg, whereas the major PUFA fraction increased more modestly, from 4719.71 ± 2135.74 to 5126.42 ± 2143.16 mg/kg.

In summary, as TMIN was increased, the rice fat content was frequently increased. However, with changes in TMIN at different thresholds, the rice fatty acids and IUFAs exhibited three distinct variation patterns: (1) at TMIN <16 °C, an increase in TMIN led to an increase in the levels of all fatty acids and an increase in IUFAs; (2) at 16 °C ≤ TMIN ≤24 °C, an increase in TMIN caused an increase in the levels of all fatty acids and a decrease in IUFAs; and (3) at TMIN >24 °C, an increase in TMIN led to an increase in the levels of specific fatty acids, such as C16:0, C18:0, C18:1, and C20:0, accompanied by a decrease in IUFAs.

3.4. Validation of the effects of TMIN on rice fat and fatty acid composition

To further validate the influence of TMIN on rice lipid content, fatty acid profiles, and the IUFAs, coordinated field trials were conducted in 2024 across 11 representative rice-growing regions in China. Post-harvest samples were analyzed and correlated with corresponding TMIN data during the grain-filling stage.

The results revealed a consistent positive association between TMIN and fat content, echoing the trends predicted by multi-year, multi-site modeling (Fig. 4A). A similar pattern was observed in fatty acid composition (Fig. 4B), with levels of C16:0, C18:0, C18:1, and C20:0 increasing steadily as TMIN rose. Notably, several other fatty acids exhibited inflection points near 24 °C, suggesting a nonlinear threshold response under elevated temperature conditions. Further analysis of IUFAs showed a unimodal response to TMIN, increasing initially and reaching a peak around 16 °C before declining (Fig. 4C),. This temperature-dependent pattern was highly consistent with findings from previous multi-year, multi-location experiments.

Fig. 4.

Fig. 4

Open in a new tab

Field experiments with sowing by stages were conducted at 11 locations in 2024 (n = 44). (A) LOWESS-smoothed curve showing the relationship between rice fat content and TMIN; (B) trends in standardized fatty acid composition and content in response to TMIN; (C) variations in the index of unsaturated fatty acids (IUFAs) with TMIN.

Taken together, the 2024 validation trials reaffirm the pivotal role of TMIN in shaping the lipid metabolic profile of rice grains. The identification of 16 °C and 24 °C as critical climatic thresholds for IUFAs and specific fatty acids, respectively, provides a robust physiological basis for climate-informed quality zoning and targeted cultivation strategies.

3.5. TMIN, fat content, and fatty acid composition of rice at different crop stages

Comparative analysis was performed based on the variations in TMIN, fat content, and IUFAs under three planting regimes (Fig. S5). As sowing time was delayed, rice TMIN showed a significant decreasing trend, with early-season rice exhibiting the highest TMIN and late-season rice the lowest. The rice fat content was also decreased with the decline in TMIN, suggesting the significant impact of TMIN on rice fat content. In contrast, IUFAs showed an opposite variation trend to that of TMIN, with late-season rice showing the highest IUFAs and early-season rice the lowest. No significant difference was detected in IUFAs between mid-season and late-season rice. Further comparative analysis of the fatty acid compositions of the three types of rice revealed similar fatty acid profiles in both mid-season and late-season rice, while significant differences were detected in early-season rice. Higher proportions of C16:0 and lower proportions of C18:2 in early-season rice were primarily responsible for the differences in fatty acids composition and the lowest IUFAs compared to mid-season and late-season rice.

3.6. Temporal and spatial distributions of TMIN under projected climate change

The BCC-CSM2-MR model based on CMIP6 was used to predict the spatiotemporal changes in TMIN for early-season rice, mid-season rice, and late-season rice under the SSP126, SSP245, and SSP585 scenarios for the 2040s and 2070s (Fig. 5).

Fig. 5.

Fig. 5

Open in a new tab

Temporal and spatial distributions of TMIN during irrigation of early-season, mid-season, and late-season rice at different representative stages under three emission scenarios (i.e., SSP126, SSP245 and SSP585) and different TMIN ranges and their fat and fatty acid accumulation patterns.

For early-season rice, most rice-growing regions, except for the far west, will experience TMIN exceeding 24 °C both currently and in the future under all scenarios. Under the SSP585 scenario in the 2070s, all early-season rice regions will experience TMIN above 24 °C. With the increase in fat content in early-season rice across different years and scenarios, the proportions of fatty acids, such as C16:0, C18:0, C18:1, and C20:0, will continue to rise, ultimatelyleading to a decrease in IUFAs. Mid-season rice exhibits three different patterns in the levels of fatty acids and IUFAs changes from west to east. As TMIN increases and fat accumulates, the range of the first fatty acid variation pattern will gradually decrease, eventually disappearing by the 2070s under the SSP585 scenario. Most mid-season rice regions follow the second fatty acid change pattern, where an increase in the level of fat content will be accompanied by a continuous decrease in IUFAs in the future. Late-season rice primarily follows the first fatty acid change pattern under the SSP126 scenario in the 2040s and 2070s and under the SSP245 scenario in the 2040s. However, with increased emissions (from SSP126 to SSP585), the fat accumulation pattern in late-season rice regions will undergo significant changes; in particular, most late-season rice regions will shift from an increase in IUFAs to a decrease. In summary, under the changes in TMIN across different emission scenarios, the fat accumulation patterns of early-season rice and mid-season rice remain relatively stable, while the fatty acid composition and IUFAs of late-season rice will undergo significant changes.

4. Discussion

Machine learning techniques have become powerful tools for predicting climate change and its impacts. Particularly, the emergence of interpretable machine learning models has greatly improved our ability to understand prediction results and reveal the underlying drivers of model changes through methods such as causal inference, global interpretation, and feature importance (Eyring et al., 2024). RF models have been widely applied in agricultural climate quality studies due to their high performance; strong resistance to overfitting; and effective model feature explanation among machine learning techniques (Balogun & Tella, 2022).

In this study, an RF-based machine learning model was developed to investigate the impact of climate change on rice fat content (Fig. 1). This model demonstrated excellent predictive performance, explaining over 60% of the variation in rice fat content, indicating that rice fat content is highly sensitive to climate change. Furthermore, SHAP analysis revealed that TMIN was the primary driver of changes in rice fat content, showing a greater impact than both the average temperature (TAVG) and maximum temperature (TMAX) during the grain-filling period (Fig. 2). The result showed that under agricultural conditions, rice fat content increases with rising TMIN, consistent with the findings of Cooper et al. (2008) (Cooper et al., 2008).

Globally, climate change has caused TMIN to rise faster than daytime temperatures, and this trend is expected to intensify in the future (Zhang et al., 2020). Daytime heat stress typically has transient; short-term effects on crop growth; whereas warmer nights could have a long-term cumulative impacts and have already shown a greater effect on rice quality (Zhang et al., 2020). Higher nighttime temperatures will increase respiration in rice; leading to greater production of acetyl-CoA; a key intermediate in glycolysis (McAusland et al., 2023; Wu et al., 2020). Acetyl-CoA is a key substrate for de novo fatty acid synthesis in rice (Fig. 6). The increased production of acetyl-CoA, driven by enhanced respiration from higher TMIN, will promote fatty acid synthesis through the citrate-pyruvate cycle in mitochondria and the catalytic actions of acetyl-CoA carboxylase and fatty acids synthase (Wu et al., 2020). Therefore; the increased respiration resulting from rising TMIN; which boosts acetyl-CoA production; could contribute to the higher rice fat content. This is consistent with our observation that C16:0 and C18:1 largely drove variation in total fatty acids and lipid content; together accounting for more than 50% of total fatty acids and matching the dominant fatty-acid profiles reported previously (Samaranayake et al., 2022). This may reflect their roles as major storage fat in rice; as C16:0; C18:1; and C18:2 provide energy for seed germination and early seedling growth (Li et al., 2023). Notably; C16:0 and C18:1 were also among the fatty acids that increased most strongly with TMIN; potentially linked to enhanced activities of key enzymes in de novo fatty-acid synthesis under warmer nights (Baud & Lepiniec, 2010). We further found that changes in TMIN resulted in differential fatty acid synthesis patterns in rice; which have multifaceted impacts on rice quality (Fig. 3). At TMIN below 24 °C; the synthesis of all fatty acids was increased with rising temperature; leading to fat accumulation. However; at TMIN above 24 °C; the increase in TMIN only enhanced the synthesis of specific fatty acids; such as C16:0; C18:0; C18:1; and C20:0; while the synthesis of other fatty acids plateaued. In our results; compared with TMIN <24 °C; C16:0 and C18:1 increased by ∼800 mg/kg and ∼ 1050 mg/kg; respectively; corresponding to an increase of approximately 15%. The 24 °C threshold marked the point at which the pattern of fat accumulation in rice changes. These results highlight the critical role of TMIN in regulating the synthesis of specific fatty acids and overall rice fat content. For example; previous studies showed that high temperatures during the grain-filling period significantly increased the levels of C16:0; C18:0; and C18:1 (Kitta et al., 2005). Notably; Shi (2023) also found that the contents of C12:0; C22:0; and C24:0 exhibited a quadratic relationship with the mean temperature during the heading-to-maturity period; with the highest level observed at a mean temperature of 24.5 °C under conditions of adequate temperature and light. Our study further revealed that; in addition to C16:0; C18:0; C18:1; and C20:0; the other eight fatty acids showed a quadratic relationship with TMIN; initially increasing and then decreasing. This was likely because; while higher temperatures enhanced respiration and provided substrates for fatty acid synthesis; they also affect the activity of fatty acid synthases and desaturases; thereby impacting the synthesis of longer-chain fatty acids or the desaturation process (Fig. 6) (Xia et al., 2010). Different fatty acids play distinct physiological roles in rice response to environmental changes. Unsaturated fatty acids; such as C18:1 and C18:2; are crucial for maintaining the fluidity and flexibility of lipid membranes; which helps seeds function normally under varying temperature (Wang et al., 2019). This is consistent with our observation that C18:1 showed the largest increase at TMIN >24 °C, exceeding 1000 mg/kg. Our studies revealed that when TMIN exceeded 24 °C, it led to significant changes in rice energy metabolism and membrane stability, resulting in substantial alterations in rice quality.

Fig. 6.

Fig. 6

Open in a new tab

Fatty acid biosynthesis in rice. Fatty acids that no longer increase and those that continue to increase after TMIN exceeds 24 °C are highlighted in red and blue boxes, respectively. Dashed boxes indicate precursors or upstream pathway of Fatty acid biosynthesis. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

As the major storage fatty acids in rice, C16:0, C18:0, and C18:1 can substantially influence the overall degree of fatty-acid saturation. Higher TMIN may promote carbon supply and de novo fatty-acid synthesis, favouring the accumulation of saturated fatty acids such as C16:0 and C18:0 and contributing to increased SFAs. In parallel, the significant positive correlations between TMIN and both C18:1 and MUFAs suggest that, under warmer nights, carbon flux may be retained at the C18:1 stage, whereas further conversion to more unsaturated fatty acids (e.g., C18:2 and downstream PUFAs) may be constrained by reduced desaturase expression at high temperature (Baud & Lepiniec, 2010). In addition; fatty acids are essential components of cell membranes; playing a key role in maintaining the membrane fluidity and function. Saturated fatty acids; such as C16:0 and C18:0; enhance membrane rigidity; while unsaturated fatty acids; such as C18:1 and C18:2; improve membrane fluidity (Wang et al., 2019). Biological membranes are among the first components affected when plants are exposed to low-temperature stress. Therefore; the fluidity and stability of biological membranes are closely linked to plant cold tolerance (Wang et al., 2019). Numerous studies have demonstrated that an increase in IUFAs is beneficial for maintaining the stability of biological membranes in animals; plants; and microorganisms (Holm et al., 2022; Ouellet et al., 2021; Wu et al., 2020). Higher levels of unsaturated fatty acids in membrane lipids lower the phase transition temperature, helping maintain the liquid crystalline state of the membrane and ensuring membrane fluidity and normal function under low temperatures (Wang et al., 2019). Moreover; unsaturated fatty acids also play a crucial role in the nutritional quality; taste; and aging processes of rice (Chen et al., 2012; Wu et al., 2020). Our results showed that as TMIN was increased, the IUFAs in rice exhibit differential changes, which was probably due to the variations in membrane lipid composition caused by temperature fluctuations. When TMIN was above 16 °C, the rice IUFAs in were decreased as the temperature was increased. In the lower TMIN range (10 °C to 16 °C), IUFAs were decreased as the temperature was decreased (Fig. 3). Rice is sensitive to low temperatures, and temperatures below 15 °C or 16 °C are considered critical for cold damage, causing physiological harm to the plants (Li, Huo, Wu, Zhu and Hu, 2014, Wang, Wang and Zhao, 2019). This critical threshold is near the temperature at which IUFAs exhibit differential changes; as observed in this study. This phenomenon could be attributed to the cell structural damage caused by excessively low temperatures; which affects photosynthesis; respiration; and normal physiological functions in rice. Consequently; the plants are unable to properly regulate lipid metabolism in response to cold stress (Zheng et al., 2016). The changes in IUFAs at different temperature ranges highlighted the complex physiological responses of rice to temperature variations and emphasized the importance of understanding these responses for improving rice quality and resilience under changing climatic conditions. In addition, at TMIN >16 °C increases in SFAs (e.g., C16:0) exceeded those in PUFAs, suggesting that warmer conditions preferentially promote saturated-fatty-acid accumulation, accompanied by higher MUFAs (e.g., C18:1). By comparison, the more modest increase in PUFAs (e.g., C18:2) may indicate that, within this temperature range, further desaturation of C18:1 to more unsaturated fatty acids was not enhanced to the same extent. In summary, the rice fat content was increased with TMIN, while both the fatty acid composition and IUFAs showed differential accumulation patterns, with 24 °C and 16 °C identified as the thresholds for differential changes in the levels of rice fatty acids and IUFAs, respectively.

China is one of the world's leading rice producers, accounting to 30% of global rice output, particularly in southern China, where the warm climate allows for multiple rice harvests per year, making it the country's primary rice-producing region. The nine southern provinces of China covered in this study produce nearly 20% of the world's rice and play a critical role in global food quality and security (Xu, Yuan, & Man, 2020; Yin et al., 2018). Therefore, we predicted the changes in TMIN under different future scenarios of southern China. Currently and in the future, the TMIN for most early-season rice regions will exceed 24 °C, especially under the scenario of rapid temperature increase (SSP585), a significant increase in the fat content is projected in early-season rice, which could improve its appearance and flavour but also make it more susceptible to aging (Wang et al., 2006; Wu et al., 2020). Additionally, the increase in the proportions of C16:0, C18:0, and C20:0 and the decrease in IUFAs in early-season rice will impact its nutritional value, as previously indicated (Wu et al., 2020). For mid-season rice; the TMIN in most rice-growing regions is currently within the range of 16 °C to 24 °C; and it is expected to remain within this range even with TMIN increases under different climate scenarios in the future. This temperature range helps mid-season rice maintain relatively high fat content while keeping the decrease in IUFAs moderate; ensuring relative stability in both fat content and fatty acid composition; thus balancing flavour and nutritional value (Wang et al., 2006; Wu et al., 2020). Late-season rice will be more significantly affected by temperature increases. If the emissions are effectively controlled (SSP126), then the rice fat content in Jiangxi and Hunan provinces is expected to increase along with an increase in IUFAs over the next few decades, leading to significant improvements in rice quality. However, if temperature increases are too high (SSP585), then the decrease in IUFAs could have significant adverse effects on rice quality.

In addition, sustained temperature increases will be a global phenomenon in the future (Meinshausen et al., 2022); which will have significant impacts on rice quality worldwide. Particularly in tropical rice-producing regions like Southeast Asia; South Asia; and Africa (Yao et al., 2022); higher production temperatures will cause a continuous decline in rice IUFAs; affecting its nutritional value. This trend may also compromise rice bran oil; a major by-product of rice; by weakening its characteristic enrichment in unsaturated fatty acids and potentially reducing its bioactivity. Notably; these regions; including China; are not only key areas for global rice production but also crucial pillars for ensuring global food security and stability; making them highly vulnerable to high temperatures (Yao et al., 2022). The global decline in rice IUFAs due to rising temperatures could threaten nutritional health and food security on a global scale.

In addition to affecting fat content and fatty acid composition, rising temperature can also lead to an increase in chalkiness, a reduction in protein content, and a decrease in the brown rice yield. These changes negatively impact the appearance, nutrition, and processing quality of rice (Su et al., 2023). Our study revealed that under agricultural conditions; the increase in rice fat content with rising temperatures may have positive effects on rice quality. These findings indicate that the impact of temperature changes on rice quality is complex and multifaceted; involving the interaction of multiple physiological and biochemical processes. Specifically; rice may respond to temperature changes differently across various temperature ranges (Song et al., 2011; Wang et al., 2023; Zeng et al., 2011). Notably, many studies model using excessive stress conditions, which could lead to distorted results. Researchers should be cautious in selecting temperature ranges and thresholds for their experiments. Future research and practice should focus on balancing changes in other quality indicators caused by climate change while the fat content is increased, to ensure the overall quality of rice. These findings are essential for developing effective agricultural management and climate adaptation strategies.

5. Conclusion

In this study, an interpretable RF model based on six years of multi-location field experiments was developed, which effectively predicted the impact of climate change on rice fat content. TMIN was identified as the primary driver of changes in rice fat content based on SHAP analysis. The rice fat content showed an increasing pattern with rising TMIN. Furthermore, the rice fatty acid composition and IUFAs exhibited differential accumulation patterns with increasing TMIN, with 24 °C and 16 °C identified as the thresholds for changes in fatty acid composition and IUFAs, respectively. These findings were further validated by large-scale field trials conducted across 11 sites in 2024. The TMIN thresholds inform sowing windows and targeted night-temperature management to protect fat content and quality, resource-efficient production. Moreover, in the future, under different emission scenarios, the fat accumulation pattern of early-season rice remains relatively stable, with certain fatty acids increasing and IUFAs decreasing. Mid-season rice is expected to maintain relatively high fat content, with a moderate decrease in IUFAs, ensuring relative stability in both fat content and fatty acid composition. Late-season rice will be more significantly affected by temperature increases, with IUFAs shifting from an increase to a decrease, potentially causing substantial adverse effects on its quality. Overall, future warming is projected to increase rice fat content while reducing IUFAs, particularly in southern China. These findings provide an important scientific basis for adapting to climate change and formulating effective agricultural management strategies in southern China's rice-producing areas, which are crucial for ensuring food security and improving rice quality.

CRediT authorship contribution statement

Wenjie Yu: Writing – review & editing, Writing – original draft, Visualization, Validation, Methodology. Qiuning Wang: Writing – review & editing, Visualization, Validation, Methodology, Investigation. Qifang Sun: Writing – review & editing, Visualization, Validation. Qiming Zhu: Writing – review & editing, Visualization, Validation. Tao Liu: Validation, Methodology, Investigation. Shengxiang Yi: Writing – review & editing, Methodology. Gaowei Zhang: Writing – review & editing, Visualization. Jie Zhu: Writing – review & editing, Investigation. Jingze Cao: Validation, Investigation. Jinwang Li: Resources. Yan Li: Resources. Xiaoqing Fan: Resources. Xiali Guo: Resources. Yanling Song: Visualization, Validation, Resources, Funding acquisition. Liping Luo: Writing – review & editing, Resources, Methodology, Investigation, Funding acquisition, Conceptualization.

Funding

This work was supported by National Natural Science Foundation of China (U2542219), Beijing Technology and Business University Talent Introduction Programs Initiation Project (19008022338), Ministry of Science and Technology of China (19000720670), Jiangxi Province National High-level Talents Innovation and Entrepreneurship Project (0210224605), and the Cross-Innovation Open Project of Food Flavour and Health, Beijing Technology & Business University (FFHCI-2025073). Beijing Technology and Business University 2025 Program for Fostering Outstanding Doctoral Dissertations (19008025082).

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Footnotes

Appendix A

Supplementary data to this article can be found online at https://doi.org/10.1016/j.fochx.2026.103797.

Contributor Information

Yanling Song, Email: songyl@cma.gov.cn.

Liping Luo, Email: lluo2@126.com.

Appendix A. Supplementary data

Supplementary material

mmc1.docx (1.3MB, docx)

Data availability

Data will be made available on request.

References

  1. Balogun A.-L., Tella A. Modelling and investigating the impacts of climatic variables on ozone concentration in Malaysia using correlation analysis with random forest, decision tree regression, linear regression, and support vector regression. Chemosphere. 2022;299 doi: 10.1016/j.chemosphere.2022.134250. [DOI] [PubMed] [Google Scholar]
  2. Baud S., Lepiniec L. Physiological and developmental regulation of seed oil production. Progress in Lipid Research. 2010;49(3):235–249. doi: 10.1016/j.plipres.2010.01.001. [DOI] [PubMed] [Google Scholar]
  3. Cao S., Liu B., Wang D., Rasheed A., Xie L., Xia X., He Z. Orchestrating seed storage protein and starch accumulation toward overcoming yield-quality trade-off in cereal crops. Journal of Integrative Plant Biology. 2024;66(3):468–483. doi: 10.1111/jipb.13633. [DOI] [PubMed] [Google Scholar]
  4. Chen Y., Ju X., Dong W., Ma G. Advances in lipids and storage characteristics of rice. Food. Science. 2012;33(13):320–323. ://CSCD:4608384. [Google Scholar]
  5. Cooper N.T.W., Siebenmorgen T.J., Counce P.A. Effects of nighttime temperature during kernel development on rice physicochemical properties. Cereal Chemistry. 2008;85(3):276–282. doi: 10.1094/cchem-85-3-0276. [DOI] [Google Scholar]
  6. Custodio M.C., Cuevas R.P., Ynion J., Laborte A.G., Velasco M.L., Demont M. Rice quality: How is it defined by consumers, industry, food scientists, and geneticists? Trends in Food Science & Technology. 2019;92:122–137. doi: 10.1016/j.tifs.2019.07.039. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Eyring V., Collins W.D., Gentine P., Barnes E.A., Barreiro M., Beucler T., Zanna L. Pushing the frontiers in climate modelling and analysis with machine learning. Nature Climate Change. 2024 doi: 10.1038/s41558-024-02095-y. [DOI] [Google Scholar]
  8. Fan C., Sun Y., Zhao Y., Song M., Wang J. Deep learning-based feature engineering methods for improved building energy prediction. Applied Energy. 2019;240:35–45. doi: 10.1016/j.apenergy.2019.02.052. [DOI] [Google Scholar]
  9. Forster P.M., Smith C.J., Walsh T., Lamb W.F., Lamboll R., Hauser M., Zhai P. Indicators of global climate change 2022: Annual update of large-scaleindicators of the state of the climate system and human influence. Earth System Science Data. 2023;15(6):2295–2327. doi: 10.5194/essd-15-2295-2023. [DOI] [Google Scholar]
  10. Gong D., Zhang X., He F., Chen Y., Li R., Yao J., Yu G. Genetic improvements in rice grain quality: A review of elite genes and their applications in molecular breeding. Agronomy-Basel. 2023;13(5) doi: 10.3390/agronomy13051375. [DOI] [Google Scholar]
  11. Halford N.G., Curtis T.Y., Chen Z., Huang J. Effects of abiotic stress and crop management on cereal grain composition: Implications for food quality and safety. Journal of Experimental Botany. 2015;66(5):1145–1156. doi: 10.1093/jxb/eru473. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Holm H.C., Fredricks H.F., Bent S.M., Lowenstein D.P., Ossolinski J.E., Becker K.W., Van Mooy B.A.S. Global Ocean lipidomes show a universal relationship between temperature and lipid unsaturation. Science. 2022;376(6600) doi: 10.1126/science.abn7455. 1487-+ [DOI] [PubMed] [Google Scholar]
  13. Huang Y., Tao B., Lal R., Lorenz K., Jacinthe P.-A., Shrestha R.K., Ren W. A global synthesis of biochar’s sustainability in climate-smart agriculture- evidence from field and laboratory experiments. Renewable & Sustainable Energy Reviews. 2023;172 doi: 10.1016/j.rser.2022.113042. [DOI] [Google Scholar]
  14. Ingvordsen C.H., Backes G., Lyngkjaer M.F., Peltonen-Sainio P., Jensen J.D., Jalli M., Jorgensen R.B. Significant decrease in yield under future climate conditions: Stability and production of 138 spring barley accessions. European Journal of Agronomy. 2015;63:105–113. doi: 10.1016/j.eja.2014.12.003. [DOI] [Google Scholar]
  15. Jianguo L.I. Effects of the temperature and humidity in filling and fruiting stages on the yield and quality of rice. Journal of Anhui. Agricultural Sciences. 2008;36(8):3160–3162. ://CSCD:3273781. [Google Scholar]
  16. Kheir A.M.S., Ali O.A.M., Shawon A.R., Elrys A.S., Ali M.G.M., Darwish M.A., Feike T. Impacts of climate change on spatial wheat yield and nutritional values using hybrid machine learning. Environmental Research Letters. 2024;19(10) doi: 10.1088/1748-9326/ad75ab. [DOI] [Google Scholar]
  17. Kitta K., Ebihara M., Iizuka T., Yoshikawa R., Isshiki K., Kawamoto S. Variations in lipid content and fatty acid composition of major non-glutinous rice cultivars in Japan. Journal of Food Composition and Analysis. 2005;18(4):269–278. doi: 10.1016/j.jfca.2004.10.001. [DOI] [Google Scholar]
  18. Kong X., Hou R., Yang G. Effects of climactic warming on the starch and protein content of winter wheat grain under conservation tillage in the North China plain. Soil & Tillage Research. 2024;238 doi: 10.1016/j.still.2023.105995. [DOI] [Google Scholar]
  19. Li J., Li X., Bai Y., Xie Y., Li L., Mu S., Gao J. Transcriptome analysis of energy supply process during seed germination in <i>Phyllostachys edulis</i>. Plant Molecular Biology Reporter. 2023;41(3):489–511. doi: 10.1007/s11105-023-01378-6. [DOI] [Google Scholar]
  20. Li J.-L., Huo Z.-G., Wu L.-J., Zhu Q.-H., Hu F. Effects of low temperature on grain yield of rice and its physiological mechanism at the booting stage. Chinese Journal of Rice Science. 2014;28(3):277–288. doi: 10.3969/j.issn.1001-7216.2014.03.007. [DOI] [Google Scholar]
  21. Li S., Montes C.M., Aspray E.K., Ainsworth E.A. How do drought and heat affect the response of soybean seed yield to elevated O<sub>3</sub>? An analysis of 15 seasons of free-air O<sub>3</sub> concentration enrichment studies. Global Change Biology. 2024;30(9) doi: 10.1111/gcb.17500. [DOI] [PubMed] [Google Scholar]
  22. Loladze I. Hidden shift of the ionome of plants exposed to elevated CO<sub>2</sub> depletes minerals at the base of human nutrition. Elife. 2014;3 doi: 10.7554/eLife.02245. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Lu Y., Guo X., Yin D., Zhu H., Li X., Zhao Y., Zheng G. Effects of water supply on rice quality during grain filling stage in cold region. Bulletin of Soil and Water Conservation. 2012;32(5) 191-198,223. ://CSCD:4694980. [Google Scholar]
  24. McAusland L., Acevedo-Siaca L.G., Pinto R.S., Pinto F., Molero G., Garatuza-Payan J., Yepez E.A. Night-time warming in the field reduces nocturnal stomatal conductance and grain yield but does not alter daytime physiological responses. New Phytologist. 2023;239(5):1622–1636. doi: 10.1111/nph.19075. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Meinshausen M., Lewis J., McGlade C., Guetschow J., Nicholls Z., Burdon R., Hackmann B. Realization of Paris agreement pledges may limit warming just below 2°c. Nature. 2022;604(7905):304–+. doi: 10.1038/s41586-022-04553-z. [DOI] [PubMed] [Google Scholar]
  26. Ouellet V., Grenier P., Santschi D.E., Cabrera V.E., Fadul-Pacheco L., Charbonneau E. Projected economic losses from milk performance detriments under heat stress in Quebec dairy herds. Canadian Journal of Animal Science. 2021;101(2):242–256. doi: 10.1139/cjas-2020-0069. [DOI] [Google Scholar]
  27. Pan B., Zheng Y., Shen R., Ye T., Zhao W., Dong J., Yuan W. High resolution distribution dataset of double-season paddy rice in China. Remote Sensing. 2021;13(22) doi: 10.3390/rs13224609. [DOI] [Google Scholar]
  28. Samaranayake M.D.W., Abeysekera W., Hewajulige I.G.N., Somasiri H., Mahanama K.R.R., Senanayake D., Premakumara G.A.S. Fatty acid profiles of selected traditional and new improved rice varieties of Sri Lanka. Journal of Food Composition and Analysis. 2022;112 doi: 10.1016/j.jfca.2022.104686. [DOI] [Google Scholar]
  29. Shi L. 2023. Study on temperature and light ecology and utilization in the formation of yield and quality of high-quality japonica rice in the taihu rice region. [doctoral dissertation, Yangzhou university] [Google Scholar]
  30. Song G., Sun Z., Wang X., Liu Y. Effect of low temperature on rice quality in different growth period. Chinese agricultural. Science Bulletin. 2011;27(18):174–179. ://CSCD:4385764. [Google Scholar]
  31. Su Q., Rohila J.S., Ranganathan S., Karthikeyan R. Rice yield and quality in response to daytime and nighttime temperature increase-a meta-analysis perspective. Science of the Total Environment. 2023;898 doi: 10.1016/j.scitotenv.2023.165256. [DOI] [PubMed] [Google Scholar]
  32. Wang F., Wang Q., Zhao X. Research progress of phenotype and physiological response mechanism of plants under low temperature stress. Molecular. Plant Breeding. 2019;17(15):5144–5153. ://CSCD:6556545. [Google Scholar]
  33. Wang H.L., Wan X.Y., Bi J.C., Wang J.K., Jiang L., Chen L.M., Wan J.M. Quantitative analysis of fat content in rice by near-infrared spectroscopy technique. Cereal Chemistry. 2006;83(4):402–406. doi: 10.1094/cc-83-0402. [DOI] [Google Scholar]
  34. Wang X., Shang B., Zhang Y., Liu J., Jiang Y. Simulation model of nutrient quality for one-season rice in jianghuai regions based on light and temperature function. Chinese Journal of Agrometeorology. 2023;44(7):624–632. ://CSCD:7525451. [Google Scholar]
  35. Wernberg T., Smale D.A., Thomsen M.S. A decade of climate change experiments on marine organisms: Procedures, patterns and problems. Global Change Biology. 2012;18(5):1491–1498. doi: 10.1111/j.1365-2486.2012.02656.x. [DOI] [Google Scholar]
  36. Wu B., Yun P., Zhou H., Xia D., Gu Y., Li P., He Y. Natural variation in WHITE-CORE RATE 1 regulates redox homeostasis in rice endosperm to affect grain quality. Plant Cell. 2022;34(5):1912–1932. doi: 10.1093/plcell/koac057. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Wu Y., Yuan J., Zhang C., Zhang C., Chen T., Gu R., Xu K. The relationship between lipid and quality and its regulation in rice. Jiangsu journal of. Agricultural Sciences. 2020;36(3):769–776. ://CSCD:6759654. [Google Scholar]
  38. Xia H., Wang X., Li M., Xiao H. Improving fatty acid composition and increasing triacylglycerol content in plants by gene engineering: A review. Sheng wu gong cheng xue bao = Chinese Journal of Biotechnology. 2010;26(6):735–743. ://MEDLINE:20815252. [PubMed] [Google Scholar]
  39. Xu L., Yuan S., Man J. Changes in rice yield and yield stability in China during the past six decades. Journal of the Science of Food and Agriculture. 2020;100(8):3560–3569. doi: 10.1002/jsfa.10385. [DOI] [PubMed] [Google Scholar]
  40. Xu Y., Zhao Y., Zhai P. Advances in scientific understanding on climate change and food security from IPCC special report SRCCL. Progressus Inquisitiones de Mutatione Climatis. 2020;16(1):37–49. <Go to ISI>://CSCD:6676865. [Google Scholar]
  41. Yao W., Tong Y., Liu Q., Zang H., Yang Y., Qi Z., Zeng Z. Spatiotemporal change characteristics and trade trend of global rice production. Journal of Southern Argiculture. 2022;53(6):1776–1784. <Go to ISI>://CSCD:7300249. [Google Scholar]
  42. Yin X., Huang M., Zou Y. Changes in rice yield stability in southern China from 1949 to 2015. Agricultural & Environmental Letters. 2018;3(1) doi: 10.2134/ael2017.11.0038. [DOI] [Google Scholar]
  43. Yu W., Wang Q., Zhu J., Cao J., Liu T., Yi S., Luo L. Quantifying climate change impacts on rice chalkiness in southern China: Future trends and spatiotemporal patterns. Computers and Electronics in Agriculture. 2025;231 doi: 10.1016/j.compag.2025.109988. [DOI] [Google Scholar]
  44. Zeng K., Ju W., Zhou Y., Guan J., Wang S., Zhang Q. Influences of high temperature induced maturity on the quality and yield characteristics of early rice. Chinese agricultural. Science Bulletin. 2011;27(30):120–125. ://CSCD:4428753. [Google Scholar]
  45. Zhang D., Liu J., Li D., Batchelor W.D., Wu D., Zhen X., Ju H. Future climate change impacts on wheat grain yield and protein in the North China region. Science of the Total Environment. 2023;902 doi: 10.1016/j.scitotenv.2023.166147. [DOI] [PubMed] [Google Scholar]
  46. Zhang Y., Wang J., Wang Y., Chen Y., Zhu D., Chen H., Cao W. Response of rice starch synthesis to night temperature changes. Chinese Journal of Rice Science. 2020;34(6):525–538. doi: 10.16819/j.1001-7216.2020.0701. [DOI] [Google Scholar]
  47. Zheng G., Li L., Li W. Glycerolipidome responses to freezing- and chilling-induced injuries: Examples in <i>Arabidopsis</i> and rice. BMC Plant Biology. 2016;16 doi: 10.1186/s12870-016-0758-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Zhou W.M., Ata-Ul-Karim S.T., Kato Y., Liu H.Y., Wang K.C. Exploring the climate signal in the variation of winter wheat quality records in the North China plain. Agricultural and Forest Meteorology. 2025;369 doi: 10.1016/j.agrformet.2025.110567. [DOI] [Google Scholar]
  49. Zhu C., Kobayashi K., Loladze I., Zhu J., Jiang Q., Xu X., Ziska L.H. Carbon dioxide (CO<sub>2</sub>) levels this century will alter the protein, micronutrients, and vitamin content of rice grains with potential health consequences for the poorest rice-dependent countries. Science Advances. 2018;4(5) doi: 10.1126/sciadv.aaq1012. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary material

mmc1.docx (1.3MB, docx)

Data Availability Statement

Data will be made available on request.

Articles from Food Chemistry: X are provided here courtesy of Elsevier

RESOURCES