BU Yuanbo, CAO Shuhe and LI Linze
College of Earth Sciences, Jilin University, Changchun 130061, China
Abstract: To assess the indicative function of the fingerprint characteristics of mineral elements for small agricultural producing areas, 20 sets of soil surface samples and corresponding rice samples were collected from Songyuan and Hunchun of Jilin Province, China. Forty-six mineral elements of brown rice, soil, and rice husk were examined by inductively coupled plasma mass spectrometry. Ten characteristic elements (Li,Ag, Y, Bi, U, Eu, Er, Rb, Mo and As) were identified via multivariate statistics (Variance importance value analysis and rank sum test in SPSS and SIMCA software packages). The correlation of mineral elements in brown rice, soil, and rice husks of the two sample areas was analyzed and regression analysis models of characteristic mineral elements in brown rice were developed. The results indicate that a correlation exists among brown rice, soil, and rice husks in the same area, and the correlation tests using selected elements showed that all correlation coefficients were 0.65 or above. Differences in brown rice were found between different regions. Consequently, brown rice producing areas can be distinguished by the fingerprint characteristics of mineral elements.
Keywords: Jilin Province; rice producing area; soil-crop system; mineral element fingerprint
Rice has a long history of cultivation and consumption, which is of great significance to the scientific research on human survival and genetic variation as well as socio-economic development (Zhang &He, 2015). The japonica rice cultivation area of Jilin Province, China, has a size of around 867 000 hectares, accounting for about 16% of the japonica rice cultivation area of Northeast China (Gao, 2020). This rice cultivation area holds a special strategic position and is of great importance for ensuring China’s food security (Qu & Lu, 2021). With increasing popularity of various brands, identifying the origin of crops for brand protection purposes has attracted widespread attention. Mineral element fingerprint technology is one of the important methods for identifying signs of such origins (Techane & Girma, 2013).
Mineral element fingerprinting is a technique that can analyze the composition and content of mineral elements in organisms of different origins.Then, mathematical and statistical methods (such as ANOVA, cluster analysis, and discriminant analysis)can be used to filter out valid indicators and establish discriminant models and databases to achieve food traceability and confirmation (Jiang, 2018). Zhao(2013) sampled the four major wheat producing areas of Shanxi, Hebei, Henan and Shandong, and determined 24 mineral elements in wheat samples using inductively coupled plasma mass spectrometry(ICP-MS). Combining One-way ANOVA and principal component analysis showed a rate of 90.8%of sample discriminant analysis and a discriminant model was established, which was compared with the overall correct discrimination rate. Shi et al.(2020) used ICP-MS to analyze the contents of 40 mineral elements in 180 samples of Songjiang rice and non-Songjiang rice. Combined with multivariate statistics, they analyzed the fingerprint characteristics of mineral elements and established a discriminant model to trace the origin of Songjiang and non-Songjiang rice samples. An overall correct discrimination rate of the traceability model established by the screened mineral element indexes for the origin of rice in the training set of 93.0% was achieved. Yasui and Shindoh (2000) used inductively coupled plasma atomic emission spectrometer (ICP-AES) and inductively coupled plasma high resolution mass spectrometry (ICP-HRMS) to determine the contents of 19 mineral elements in 34 rice samples from 27 different regions of Japan. Combined with principal component analysis, they were able to successfully discriminate the origin of rice samples. Kaoru (2012)analyzed the contents of Al and Fe in rice from Japan,Thailand, the USA, and China for cross-validation,and achieved a final discrimination rate of 97%. The above studies show that mineral element fingerprinting characteristics can be used as indicators when characterizing information on the origin of agricultural products. Most current Chinese and international studies focus on the relevance of rice origins between international and inter-provincial regions at a large spatial scale. However, many natural factors (such as rainfall and climate), as well as human factors (such as farming measures and amount of applied fertilizer) will impact the fingerprint of selected mineral elements. How to overcome the problems associated with such changeable factors and screen stable and effective mineral element indexes of producing areas,especially the element combination representing regional characteristics in a small area, remains a challenging problem (Zhang et al., 2016). In this study,mineral element fingerprinting was used to explore the distinction of rice origins at the provincial scale.
1.1 Overview of the study area
Hunchun City is located in the east of Jilin Province, China. About four fifths of Jilin Province is mountainous terrain, and the remaining one fifth mostly consists of basins and small plains as well as pre-mountain alluvial fans and river alluvial fans.Jilin Province has a moderate temperate near oceanic monsoon climate, with an average annual rainfall of 616.8 mm, and an annual evaporation of 1 301.2 mm.Songyuan City is located in the west of Jilin Province, in the south of the Songnen Plain, with a flat,open, undulating and gentle terrain. It is mainly composed of the Songnen Alluvial Plain and the Songliao Divide Plateau Plain. The climate type is moderate temperate continental monsoon climate, with an average annual rainfall of 400–500 mm and an annual evaporation of 1 360 mm. The two cities were selected as the study areas due to their relatively close geographical location and certain differences in natural geographical conditions, topography and climate.
1.2 Sample collection
Twenty sets of soil surface samples and corresponding rice samples were collected before the rice harvest (Fig.1). Soil surface samples were collected by scraping away surface debris from sampling points,collecting the soil vertically up to 20 cm above and below the surface. About 4–6 points within 50 m in the same way were chosen and the sampling was repeated. Then, samples were mixed evenly, using the quadrat method of sampling, leaving approximately 1 000 g of sample in a clean cloth bag. Rice samples were collected by randomly harvesting about 1 000 g of mature rice ears within the collection area of soil sample points. Soil and rice samples were naturally dried indoors, soil samples were ground through a 2 mm sieve and rice ears were processed with a hulling machine to separate seeds from husks. Of each,200 g was filled into a sample bag for experimental determination.
Fig.1 Sample distribution map
In this study, brown rice was chosen because only the husk was separated from the seeds and seeds were not de-browned.
1.3 Sample testing methods and results
Element in soil, rice seeds, and rice husks were tested by ICP-MS, and the sample analysis was conducted at the Key Laboratory for Mineral Resources Evaluation in Northeast Asia, Jilin University, China.The statistics of the test results are shown in Tables 1,2 and 3.
2.1 Characteristics of mineral element contents
The statistical results presented in Tables 1, 2 and 3 show certain differences in the contents of mineral elements, such as B, Co, Rb, Mo and Ag, in soil, rice husks, and brown rice samples between Songyuan and Hunchun. However, the contents of other elements, such as Cu, Cd, Sb and W, were more similar.Yet other elements, such as Ga, Zr, Ba and Hf, in brown rice, Tb, Ho, Lu, Tl and Bi in soil, and La, Ce,Pr, Nd and Sm in rice husk samples showed large coefficients of variability (>100%), even in the same area. This indicates that they vary considerably within the same area.
2.2 Screening for characteristic elements
It is inefficient and uneconomical to test and analyze 46 elements in each experiment. In addition,because the contents of many elements are similar,these cannot reflect the characteristic differences between the two sampled places, resulting in large errors. Therefore, the statistical methods of rank sum test and VIP value analysis were combined to identify the mineral elements that can fully reflect the differences of brown rice samples between the two places,which are termed characteristic elements.
The 46 elements measured in brown rice wereanalyzed and a load diagram was drawn (Fig.2). According to the relevant literature, elemental points closer to the center are those with where the difference in the contents of elements in the brown rice samples between the two locations is small (Liuet al., 2019). Elemental points farther away from the center are those where the difference in the content of brown rice between the two locations is large. First, a preliminary analysis of the differences in the contents of 46 elements between the two sites was conducted,and then, VIP analysis was carried out using SIMCA software (Table 4) to compare elements with large differences in the samples. Sixteen elements with significant differences (VIP value >1) were identified,namely Li, B, Sc, Co, Ga, As, Rb, Y, Mo, Ag, La,Sm, Eu, Er, Bi and U. After that, SPSS software was used to conduct a rank sum test (Table 5) to further filter the elements screened by VIP analysis. Finally,10 elements were identified as characteristic, namely Li, Ag, Y, Bi, U, Eu, Er, Rb, Mo and As.
Fig.2 Load graph of 46 elements in brown rice in Songyuan City and Hunchun City
Table 1 Statistical analysis of 46 elements in brown rice samples from Songyuan City and Hunchun City
Table 2 Statistical analysis of 46 elements in soil samples from Songyuan City and Hunchun City
Table 3 Statistical analysis of 46 element in husk samples from Songyuan City and Hunchun City
2.3 Correlation analysis
Correlation analysis was conducted for the characteristic mineral element contents determined for soil, rice husk, and brown rice in Songyuan and Hunchun. The relevance between soil, rice husk and brown rice in the same area was compared. Table 6 and Table 7 show the results of correlation analyses of brown rice, soil, and husk samples in the cities of Hunchun and Songyuan, respectively.
Table 4 Analysis of VIP values
Table 5 Results of rank-sum tests
Table 6 Relevance of brown rice, soil and rice husk samples in Hunchun City
Table 7 Relevance of brown rice, soil and rice husk samples in Songyuan City
The obtained correlation coefficients (Tables 6 and 7) are above 0.65 (P < 0.05) for the same region, which is large. This indicates that there is a correlation between soil, rice husks, and the mineral elements in brown rice in the same region, showing a geographical characteristic. This suggests that a correlation exists between the mineral elements in soil,rice husk, and brown rice in the same region, with a geographical characteristic.
2.4 Partial least squares regression analysis
Partial least squares regression analysis was conducted on the measured content of 10 mineral elements in brown rice samples from Songyuan(Pink points in Fig.3) and Hunchun (Blue points in Fig.3). After excluding individual anomalies, samples from the same region cluster, indicating that the intra-group variability is small. Samples from different regions are clearly separated, indicating that the differences in mineral element contents in brown rice from these two regions are large, i.e., there are apparent differences in these 10 mineral elements in brown rice from the different sampled regions.
Fig.3 Regression analysis of mineral element contents in brown rice samples of different areas
The measured and selected elements were classified into three categories: trace elements, rare earth elements, and radioactive elements (Table 8). The selected characteristic elements contained very little radioactive elements, which are therefore not discussed here. The high proportion of trace elements and rare earth elements indicates that the differences in the content of trace elements and rare earth elements between the two sampled cities are substantial. On the one hand, the important role trace elements play in the human body is increasingly recognized and emphasized. Their availability is particularly important for the growth and development of children, and trace element testing plays an important role in guiding nutrition and preventing the occurrence of diseases.On the other hand, in recent years, studies have continuously shown that rare earth elements can interfere with the function of the human immune system and cause potential harm to human health. Therefore, the contents of trace elements and rare earth elements in plant foods should be further investigated
Table 8 Element classification
The identified 46 elements were analyzed and the results showed that the coefficient of variation in the mineral element content of brown rice, rice husk,and soil from the same area was large, while the content of certain elements showed little difference between the two areas. Therefore, SPSS and SIMCA software were used to screen characteristic elements.Ultimately, 10 elements were identified as characteristic for distinguishing brown rice from these two cities and a relevant model was established to prove that these 10 elements can be used for mineral element fingerprinting.
Previous studies have shown that the mineral element content in rice is not only related to the soil,but also to many anthropogenic and natural factors such as rainfall, temperature, and fertilization in that year (Wanget al., 2019; Shiet al., 2020). Such experiments combining rice and soil can disclose the relationship. Because stable isotopes are likewise natural fingerprints of plants, stable isotope analysis techniques can also be used and both methods can be combined in the future to identify mineral element fingerprint combinations more accurately(Chunget al., 2016; Pilgrimet al., 2010).
The contents of 46 mineral elements in soil, rice husk, and brown rice samples gathered from Songyuan City and Hunchun City were examined by ICPMS. Ten characteristic elements were identified using rank sum test and VIP value analysis in SPSS and SIMCA software. Correlation analysis and partial least squares regression analysis were performed for these 10 elements, and a model was developed. These 10 elements can be used for mineral element fingerprinting for Songyuan and Hunchun brown rice. The mineral element content in brown rice and rice husk correlated with the mineral element content in local soil, and there were differences in the content of these 10 mineral elements between different regions. These results indicate that mineral elements within the soilcrop system are significantly correlated. Furthermore,geographical variation exists in mineral elements in crops from different regions, which can be expressed based on mineral element fingerprinting.