Introduction

Cinnamomum cassia Presl. is a member of the Lauraceae family and is present in many different countries, including China, India, Seychelles, Madagascar, Sri Lanka, and Vietnam1. In the tropics and subtropics, it is one of the most valuable plant resources for the economy. Essential oils contained in C. cassia are synthesized and accumulated in its seeds, calyxes, leaves, twigs, and barks, making them an integral part of the plant1. Terpenes and aromatic chemicals may be found in abundance in the oils extracted from cassia leaves. Tran-cinnamaldehyde is the major active constituent of cinnamon oils2. Like cassia bark, the oil extracted from cassia leaves is used for flavoring, medicinal, and most notably, cola-like beverages3. C. cassia is now mainly distributed in the provinces of Hainan, Guangdong, Fujian, Yunnan, and Guangxi in China. Approximately 80% of the global supply of C. cassia oil comes from Guangdong and Guangxi Provinces.

C. cassia has a very long medicinal history in Chinese literature, recorded as far back as the era of the Warring States (475 BC-221 BC) in the “Classic of Shanhai Nanshan”4. The Shennong Materia Medica Classic of the Eastern Han Dynasty (25 AD-220 AD) listed C. cassia as a first-class and precious Chinese medicine5. It is now also documented by various books, such as Pharmacopoeia of the People’s Republic of China (PPRC)6, etc. According to existing records, before the Qing Dynasty (1616 AD − 1912 AD), the C. cassia production area was mainly in Guangxi, with a scattered distribution. After Qing Dynasty, the producing area gradually expanded to Guangxi and Guangdong7. With the change of time, the distribution area of C. cassia is also changing.

Climate is the most important factor in determining the distribution range of species8. Global warming is increasingly affecting the distribution and buildup of herbal plants’ secondary metabolites9,10. For example, average temperature and precipitation affected the distribution and accumulation of active constituents of Zanthoxylum nitidum10. Continuous in-depth studies on the chemical constituents of cinnamon have revealed significant differences in the major chemical constituents and pharmacological activities of C. cassia from different regions11. The growth and development of C. cassia are intrinsically linked to climate changes, which in turn affects the cultivation and propagation of C. cassia medicinal materials12. Recently, predicting alterations in suitable habitats and increasing yields with the use of plant distribution models based on environmental variables has now become standard practice13. Since the MaxEnt model is so straightforward, it may be readily interpreted13. The MaxEnt model is based on the principle of maximum entropy, using known environmental data and species distribution points to find the maximum entropy of the specific probability distribution and then predict its distribution model14. The optimal distribution of high-value medicinal plants may be determined using these techniques15. However, little is known about how different climatic and environmental variables affect the distribution and habitat quality of C. cassia. There are few reports on the distribution pattern and ecological suitability of C. cassia. However, under the conditions of severe climate and environmental changes, it is particularly urgent to study the migration and diffusion of this species and its responses and adaptations to global warming.

Accordingly, C. cassia was the subject of this investigation, and samples collected from various suitable habitats were analyzed using chemometrics, Fourier transform infrared reflection (FTIR), and high-performance liquid chromatography (HPLC), according to the MaxEnt model and forecast prospective suitable growth regions and changes in C. cassia’s geographic distribution as the climate shifts. These are the dominant themes that the study intends to cover: (1) When considering past, present, and future climatic scenarios in China, what is the probable distribution and spatial pattern of C. cassia? (2) How do environmental variables affect the spatial distribution of C. cassia habitat? (3) In what ways can future climate changes alter the optimal distribution and spatial pattern of C. cassia? (4) How is the quality of C. cassia in different suitable areas? The answers to these questions may form the foundation for future efforts to protect C. cassia and build a reserve for the plant, as well as for the delineation of appropriate planting locations, the artificial introduction, and the culture of the species.

Materials and methods

Sources of species information and sample collection

C. cassia occurrence data were derived from the Chinese Virtual Herbarium (http://www.cvh.ac.cn/) and the National Specimen Information Infrastructure (http://www.nsii.org.cn/), as well as author collection C. cassia (93 locations). A total of 232 locations were obtained from different provinces. Subsequently, records with evident geocoding errors are removed, and records of occurrences with a straight-line distance of < 10 km are removed using the Arcmap analysis tool. A total of 158 valid occurrence records were used in the study (Fig. 1; Table S1). In 2022, 31 samples of C. cassia were gathered from southern China after a detailed assessment of the species’ suitable habitats (Table 1), of which 24 and 7 samples were respectively obtained from high suitable habitats, medium- and low-suitable habitats. Professor Rongjing Zhang positively recognized the plant specimen (South China Agricultural University, China). Specimen vouchers were deposited in the South China Agricultural University’s Department of Botany herbarium (SCAUB). For details, please refer to Table 1.

Figure 1
figure 1

(a) Spatial distribution pattern of C. cassia as determined by 158 valid occurrence records in China (Created using ArcGIS 10.4.1, https://www.esri.com/). (b) C. cassia in the field. (c) Leaf of C. cassia.

Full size image
Table 1 C. cassia collection sites.
Full size table

Environmental variable

The 19 bioclimatic variables data were obtained from the WorldClim website (https://www.worldclim.org/) (Table S2). Current climate data was derived from the baseline period of 1970–2000. Data from the past on the climate include the “Last Glacial Maximum” (LGM, which occurred around 22,000 years ago) and the “Mid-Holocene” (MH, around 6000 years ago). Two time intervals, 2041–2060 and 2081–2100, were selected for the predicted climatic data. We used the global climate model (BCC-CSM; Beijing Climate Center Climate System Model) from the sixth CMIP6 phase of the Coupled Model Intercomparison Project to predict future weather patterns, and we compared our results to two socioeconomic scenarios with shared pathways (SSP126 and SSP585). The most optimistic forecast for the future of greenhouse gas (GHG) emissions is shown by SSP126, while the most pessimistic forecast is depicted by SSP585 11,16. SSP126 and SSP585 have good significance with a combination of local realities17. The Harmonized World Soil Database (version 1.2) was used to collect the data for the 11 different soil variables (https://www.fao.org/soils-portal/data-hub/soil-maps-and-databases/harmonized-world-soil-database-v12/en/) (Table S2).

Using the MaxEnt model for multiple operation analysis, we ranked the importance of each environmental variable and analyzed their association with the aid of SPSS (v 20.0) and ArcGIS (v 10.4.1) (Table S2). The small contribution rate was omitted when the correlation coefficient measured by |spearman| between the two variables was > 0.80 (Fig. S1). In the end, a total of 13 environmental variables (bio_2, bio_4, bio_5, bio_11, bio_14, bio_15 and bio_18 for climatic variables and aw_class, s_oc, t_caco3, t_clay, t_oc and t_sand for soil variables), and 3 topographic variables (elevation, aspect, and slope; https://www.worldclim.org/) were included in model analysis.

Species distribution modeling

ArcGIS (v 10.4.1) and MaxEnt software (v 3.4.4) were used to analyze and assess the 158 C. cassia distribution point data across different climatic scenarios. Additionally, 75% of the distribution points were used as the training set for the prediction model, while the remaining was used to evaluate model performance after it was run ten times.

We used the area under the receiver operating characteristic (ROC) curve (AUC) to determine how reliable the MaxEnt model was. Permutation importance and percentage contribution were determined using the jackknife technique. An AUC value < 0.5 indicates unsuitability for the growth of the species16. The artificial break method was utilized to segregate the zones suitable for C. cassia. Four types of suitable habitats were established: highly suitable (0.5–1), moderately suitable (0.30–0.50), poorly suitable (MTSPS − 0.30), and unsuitable habitats (0 - MTSPS) (Table S3).

HPLC analysis

To accurately characterize the local plant populations, we collected fresh leaf samples at random from three different plants for each sample. Thereafter, the samples were dried in the air and then crushed into a powder (60 mesh) after being rinsed using tap water. Methanol was used in an ultrasonic extraction of the sample, with the power set at 350 W, the frequency set at 35 kHz, and the temperature set at 40 °C.

HPLC (1260 A, Agilent) was employed to analyze the samples for trans-cinnamaldehyde content. The Lichrospher-C18 column (4.6 mm ID x 250 mm, 5 μm) was used in the chromatographic analysis. The column was maintained at a constant temperature of 30 ° C. The determination was performed with a detection wavelength of 273 nm and acetonitrile/water (35:65 − 38:62 − 38:62 − 50:50) as a mobile phase at 1.00 mL/min flow rate. Six microliters of fluid were injected. Further details on sample preparation and HPLC application can be found in PPRC6.

To determine whether trans-cinnamaldehyde accumulation is influenced by habitat suitability, we employed the trans-cinnamaldehyde content as a quality evaluation standard. Calibration plots of the trans-cinnamaldehyde (CAS: 14371-10-9; TCI, Shanghai, China) standard were constructed from the peak areas (y) obtained for 5 different concentration solutions (x) to establish a standard curve (y = 306382x + 421.72, r² = 0.9991) (Fig. S2).

FT-IR analysis

For the experiment, we employed an FTIR spectrometer (a Nicolet 5700 from Thermo Nicolet Corp.) equipped with a deuterated triglycine sulfate (DTGS) detector. With a 4 cm−1 resolution, IR spectra were obtained between 400 and 4000 cm−1. A steady 25 °C temperature and a relative humidity of 30% were maintained in the room. Pieces of KBr tablets were prepared for infrared spectrometric investigation by combining dry KBr powder (100 mg) with C. cassia leaf powder (200 mesh, 2.0 mg). Three sets of measurements were collected using three scans each, yielding a total of nine spectra; the average chart was then used as the final spectrum. These spectra were processed by removing the effects of air, CO2 interference, and water vapor. After importing the spectrum data into OMNIC 8.0 for baseline correction and smoothing, we used Unscrambler 9.7 to standardize normal variations.

HCA and OPLS-DA analysis

In this research, we employed the Euclidean distance and the hierarchical cluster analysis (HCA). As the metrics for cluster analysis, we employed the common peak area in HPLC and the relative absorbance in FTIR to establish the distance matrix of the chromatographic area and samples. SPSS 20.0 (SPSS Inc., Chicago, USA) was used to generate the dendrogram and execute cluster analysis.

The HPLC peak area and FTIR relative absorbency of 31 batches of C. cassia samples were determined utilizing the SIMCA software and an orthogonal partial least squares discriminant analysis (OPLS-DA) model was constructed. The OPLS-DA and subsequent scatter plots were produced in SIMCA 14.1.

Results and discussion

Evaluation of model accuracy

The predictive accuracy of the MaxEnt model is expressed as an AUC value between 0 and 1. Accuracy is measured by how close the area AUC is to 1 and values closer to 1 indicate greater accuracy16. The results showed that all average training AUC values and test set (10 replications) of the model were between 0.98 -1(Current, Rcp26-2050s and Rcp85-2050s: 0.988; LGM, MH, Rcp26-2090s and Rcp85-2090s: 0.987) (Fig. S3). The findings demonstrate that the MaxEnt model calculation attained excellent accuracy and that the reliability of the experimental data was high.

Environmental variables analysis

Some surveys and research reports examined where C. cassia was distributed in China, but they didn’t consider environmental variables that may have affected its distribution Using the Spatial Analyst Tools module in ArcGIS, the main ecological niche parameters of main environmental variables for determining the optimal distribution area were extracted and analyzed (Table 2). The analysis results show that bio_4 (Temperature seasonality), bio_18 (Precipitation of warmest quarter), t_clay (Topsoil clay content), bio_11 (Mean temperature of coldest quarter), bio_14 (Precipitation of driest month), awc_class (Soil available water content), s_oc (Substrate soil organic carbon), bio_15 (Precipitation seasonality), and bio_2 (Mean diurnal range) were the primary influencing factors on the distribution of C. cassia. Their percent contribution cumulative values and permutation importance reached 97.8% and 87.2%, respectively. Additionally, bio_18, bio_4, t_clay, and bio_14 were important environmental variables (Table 2), with bio_18 and bio_4 having more weight in the analysis of only one variable (Fig. S4), indicating that precipitation, soil, and temperature were significant elements influencing the potential distribution of C. cassia.

Table 2 Percentage percent contribution and permutation importance of environment variables.
Full size table

Based on the response curve of four important environmental variables (bio_18, bio_4, t_clay, and bio_14) to the probability distribution of C. cassia distribution (Fig. S5), it can be observed that the environmental variables suitable for C. cassia growth are: precipitation of warmest quarter (Bio_18) of 367.20–1697.81 mm and the optimal value of 775.55 mm, temperature seasonality (Bio_4) of 402.85-870.67 and the optimal value of 564.36, topsoil clay content (t_clay) of 41.40-51.88% w and the optimal value of 47.44% weight, and precipitation of driest month (bio_14) of 17.92–60.76 mm (Table 3). Guangxi and Guangdong, especially the Xijiang River Basin, have suitable temperatures and abundant rainfall, making them authentic production areas for C. cassia. Similar to the research results of Panax notoginseng and Zanthoxylum nitidum, bioclimatic variables, especially temperature and precipitation have a profound effect on the geographical distribution of herbal plants10,16. In addition, soil factors (topsoil clay content and soil available water content) contribute significantly to C. cassia distribution. Different climatic conditions, CO2 concentration, acid rain and CO2 fixation by plants cause alterations in soil pH and organic matter content18. In addition, the topsoil clay is higher in the south than in the north of China, and the topsoil clay affects its fertility and drainage performance19. It is reported that the soil conditions and microorganisms were different in different regions, which affected the composition and content of secondary metabolites of medicinal Citrus20. Therefore, in future research, we should further study the relationship between C. cassia quality and soil microorganisms in different suitable habitats, to better reveal the localization mechanism of C. cassia.

Table 3 The suitable habitat range of key environmental variables affecting the potential distribution of C. cassia.
Full size table

Distribution prediction in the current environmental conditions

After the Qin Dynasty (about 1700 years ago), the cultivation method of C. cassia shifted from wild growth to artificial cultivation, resulting in an expansion of the production areas. Figure 2a, b demonstrates that C. cassia, under the present-day climatic scenario, is found primarily in southern China between 18 and 33 °N, especially between 20 and 26 °N. There was a total of 1,606, 718.01 km2 of suitable habitat (containing both highly and moderately suitable habitats) (Fig. 3; Table 4). The highly-suitable habitat area was 706,129.08 km2, mainly distributed in two provinces of Guangxi and Guangdong, as well as in the south of Guizhou and Chongqing, the east of Taiwan, the southwest of Hubei, and the northeast of Hainan and Fujian Province. The moderately suitable area was 900,588.93 km2 (Fig. 3; Table 4), mainly distributed in most areas of Hainan, Hunan, Jiangxi, Fujian, Guizhou, and Zhejiang, the south of Yunnan and Anhui, the east of Sichuan and Hubei, and the central region of Taiwan, as well as a few areas in the south of Tibet (the southern part of Motuo County) and Jiangsu Province (Fig. 2b). It can be observed that the northern part of Yunnan and the southwestern part of Sichuan are within the 30°N latitude, but they have a relatively low distribution of C. cassia. This may be because these areas are more than 1,000 m above sea level. C. cassia is most suitable for growth at an altitude of 300–350 m, and it is not suitable for C. cassia to grow above 1000 m21. This suggests that altitude is also a limiting factor for the distribution of C. cassia. It was reported that C. cassia is widely distributed in China, with production areas in Guangxi, Guangdong, Yunnan, Fujian, and other regions1,7. Throughout history, the preferred production areas were mainly located in the Guangxi-Guangdong region. Now the authentic production areas in Guangxi and Guangdong have officially formed a pattern, accounting for more than 80% of the total area of C. cassia distribution in China2. These areas mainly have a subtropical monsoon climate and abundant water and heat resources. The results also showed that precipitation in the driest and wettest quarters had significant effects on the distribution of cinnamon. The water requirement in the dry season is more than 26 mm, and the water in the wettest season is more than 449 mm, so in the daily field management, attention should be paid to ensuring adequate water. Good concordance between predicted and actually observed distributions suggests that the MaxEnt-simulated suitable area may be utilized as a guide in selecting cultivation locations for C. cassia planting planning.

Figure 2
figure 2

Climate change scenarios-related suitable habitats for the C. cassia. a, b: Current; c: Last glacial maximum (LGM); d: Middle holocene (MH); e: 2050 (average for 2041–2060), SSP126; f: 2050 (average for 2041–2060) SSP585; g: 2090 (average for 2081–2100), SSP126; h: 2090 (average for 2081–2100), SSP585 (Created using ArcGIS 10.4.1, https://www.esri.com/).

Full size image
Figure 3
figure 3

Changes in C. cassia’s suitable habitat (in km2) across various climatic scenarios under different periods.

Full size image
Table 4 The suitable habitat areas of C. cassia under different future climate scenarios.
Full size table

Distribution prediction in the past environmental conditions

The distribution of C. cassia under past climate scenarios is shown in Fig. 2c-d. Compared with the current situation, suitable habitats of C. cassia were limited during the Last Glacial Maximum (LGM) (Fig. 2c). The distribution of highly suitable habitats was decreased to 22–23 °N, mainly in the southern of Guangxi and a few areas in the southern part of Guangdong with only 58,423.00 km2 (Figs. 2c and 3; Table 4). This is probably due to the extremely cold climate during the LGM, in which the global average temperature dropped by approximately 5–10℃ compared to the present, and the precipitation levels also decreased16,22. According to previous research and records, wild C. cassia resources were abundant in ancient times. During the Han and Wei dynasties, most C. cassia was naturally distributed in Guangxi, with scattered distribution in Guangdong as well23, which agreed with the findings of this research, and further proved the reliability of our findings.

Compared to the LGM, the suitable habitat of the Mid-Holocene (MH) was extended to the surrounding areas (Fig. 2d). The areas of moderately (898,945.59 km2) and highly suitable habitats (521,486.20 km2) were both significantly increased (Fig. 3; Table 4). It covered almost the entire regions of Guangdong and Guangxi provinces, and the eastern regions of Fujian, Zhejiang, and Sichuan provinces. This was because of the dramatic temperature rise that occurred throughout the MH period24, which was suitable for the growth of C. cassia. Palaeoclimate research showed that the MH period was a relatively warm and humid period since the Last Glacial Maximum. The temperatures were approximately 1-1.5℃ higher than current levels, and the southern regions experienced 100–300 mm more rainfall compared to the present25,26. In particular, the average annual temperature in the Zhejiang and Fujian provinces, which are close to the East China Sea, will be about 2–3℃ higher than the current. This makes these regions more suitable for the growth of C. cassia27. In the late Holocene, there was a decreasing trend in temperature26. Thus, temperature and precipitation are important key variables affecting C. cassia’s habitat suitability dispersion.

Predictions of distribution in the future environmental conditions

Future C. cassia distribution as well as suitable habitat regions are depicted in Figs. 2e-h and 3. Both SSP126 and SSP585 future climate scenarios for the 2090s predict a significant reduction in the low-suitable habitat regions for C. cassia compared to the two 2050s future climate scenarios (Fig. 3; Table 5). The high-suitable and total habitat areas both exhibit an upward trend, particularly in the 2090s SSP585 scenario, and the newly expanded total area of suitable habitat is the highest (312,963.53 km2), and the distribution of total suitable habitat has expanded to 18–32 °N. (Fig. 2e-h; Table 5). As shown in Fig. 2h and Fig. S6, the highly suitable habitat areas of C. cassia have expanded to the south of the Yangtze River. Moreover, the highly suitable habitat areas had increased significantly in Fujian, Jiangxi, Hunan, Hubei, Zhejiang, Hainan, and Taiwan provinces, as well as in the south of Yunnan and Anhui provinces and the east of Sichuan. This demonstrates the potential for future scientific expansion of C. cassia planting areas. But it also means that the environment is changing and the temperatures are rising. According to reports, the yearly average surface temperature in southern China rose by 2.7 to 2.9℃, with warming increasing by 1.8 to 2.5℃. Additionally, the warming rose from south to north by > 2.5℃ with the annual precipitation increasing by 1.5–2.0%28,29. Changes in precipitation are closely related to temperature changes. Global climate warming has led to a general increase in atmospheric humidity and precipitation intensity30. This could potentially expand the range of suitable habitats for C. cassia. Changes in the environment will affect the range and distribution of many medicinal plants9,16,31. This suggests that temperature and precipitation may be key factors influencing the southward expansion of suitable habitat for C. cassia towards the Yangtze River region.

Table 5 Contracted and expanded area of suitable habitat for C. cassia under different future climate scenarios.
Full size table

Quality evaluation of C. cassia in different suitable habitats

HPLC fingerprint of C. cassia

To reveal the variations in the content of active constituents in natural medicinal materials from different suitable habitats, the trans-cinnamaldehyde content of 31 C. cassia samples collected from various suitable environments was analyzed by HPLC. According to the results, trans-cinnamaldehyde levels varied among 31 samples collected from various suitable habitats (Fig. 4A; green column). In general, the high-suitable habitats of Guangxi (GX1-GX16) and Guangdong (GD1-GD8) have relatively higher levels of trans-cinnamaldehyde, while the medium- and low-suitable habitats of Yunnan (YN1-YN5), Hunan (HN) and Hubei (HB) have relatively lower levels of trans-cinnamaldehyde. In high suitable habitats, the average level of trans-cinnamaldehyde (Fig. 4A; red columns, 0.85%) is notably higher than that in medium- and low- suitable habitats (Fig. 4A; blue columns, 0.30%). Previous studies have also shown that C. cassia produced in Guangxi and Guangdong is considered to be of the best quality23,32, this is related to the environmental parameters like sunlight, temperature, soil composition, and precipitation have a remarkable influence on plant secondary metabolite synthesis and accumulation2,33,34. It has been reported that the suitable growth temperature for C. cassia is between 22 °C and 26 °C. When temperatures exceed 20 °C, the content of trans-cinnamaldehyde increases; however, when temperatures exceed 30 °C, the growth rate of C. cassia slows down, and trans-cinnamaldehyde synthesis decreases. In regions with annual precipitation between 1200 mm and 1800 mm, the average content of trans-cinnamaldehyde reaches 5.2%35. Guangdong and Guangxi provinces are the primary high-suitability areas for C. cassia, where the climate is warm and suitable (with average temperatures in 2022 of 26.2 °C and 22.5 °C, respectively) and rainfall is abundant (with average precipitation in 2022 of 2114.3 mm and 1696.7 mm, respectively). These conditions make them ideal for C. cassia cultivation, and they are also the main production areas for C. cassia in China. The C. cassia grown in these regions has a high trans-cinnamaldehyde content, making it suitable for use in the pharmaceutical field. The moderate-suitability areas are mainly concentrated in the northern and western parts of the subtropics. Due to colder winter temperatures and insufficient rainfall, the trans-cinnamaldehyde content in places like Yunnan and Hubei is lower than that in Guangdong and Guangxi. Therefore, it is suggested that the medicinal materials produced in these regions can be used as food additives or spices. Our results further supported that the medicinal plants in suitable habitats contain higher levels of active constituents9,10,31.

Figure 4
figure 4

A. C. cassia’s trans-cinnamaldehyde content in various suitable habitats. HIG was the trans-cinnamaldehyde average content of high-suitable habitats (GX1-GX16 and GD1-GD8). MOD/LES was the trans-cinnamaldehyde average content of medium- and low-suitable habitats (YN1-YN5, HN, and HB). The columns having different uppercase letters are significantly different (p < 0.05). ** Very significant differences p < 0.01. B. OPLS-DA model of C. cassia from different geographical origins.

Full size image

To accurately and visually observe the differences between samples, an OPLS-DA model was constructed with 22 aroma components as the dependent variables and different geographical origins as the independent variables (Fig. 4B). In this analysis, the fixed index for the independent variables (R2x) was 0.513, the fit index for the dependent variables (R2y) was 0.714, and the model prediction index Q2 was 0.581. R2 and Q2 values greater than 0.5 indicate acceptable model-fitting results36. From Fig. 4B, it can be observed that the high-suitable habitats (green box) and the medium- and low-suitable habitats (blue box) samples of C. cassia clustered into distinct groups. GX13 (Chongzuo, Guangxi) and GX16 (Baise, Guangxi) are located in the western part of Guangxi, close to Yunnan, which makes them similar to the samples from the medium- and low-suitable habitats. OPLS-DA can effectively discriminate C. cassia samples from high-, medium- and low-suitable habitats.

Quality evaluation based on FTIR fingerprint of C. cassia

The power of FTIR as a quantitative tool lies in its ability to easily perform multicomponent analyses37. This technique is usually considered to be non-destructive, simple, and fast with good repeatability and accurate analysis of traditional Chinese medicine38. Figure 5a illustrates that the FTIR fingerprint spectra peaks of 31 batches of the C. cassia leaves had many similar peaks with characteristic infrared absorption spectra ranging between 1800 –600 cm−1. Several characteristics of the typical C. cassia spectrum seen in Fig. 5a were determined via analysis. The aldehyde carbonyl’s C = O stretching vibration shows up as the strongest peak at 1665 and 1628 cm−1, and the peak is strong and broad, which is consistent with the high content of cinnamaldehyde and other aldehydes in C. cassia leaves. The absorption peak at 1451 cm−1 during bending vibration absorption is highly indicative of a C-OH alcohol. Bands at 1250 cm−1 indicate the presence of the aromatic acid ester C-O-C symmetric expansion and the stretching vibration of the phenolic C-OH groups, whereas the frequency at 1070 cm−1 is associated with C-O, with stretching vibration mode. Absorption due to C-H bending vibration absorption is attributed to the peak at 780 cm−1. Finally, the band at 680 cm−1 is an olefin vibration absorption, and these vibrational characteristics are consistent with C. cassia volatile oil containing olefinic substances.

Figure 5
figure 5

IR spectra (a), a dendrogram of cluster analysis (b), and OPLS-DA score plot (c) from the IR spectra for 31 C. cassia samples.

Full size image

Because there is a significant overlap in the chemical composition fingerprints, distinguishing between the components of the samples is challenging. The IR spectra data of the 31 samples was chosen for HCA and OPLS-DA to more directly detect the variations in the components of the C. cassia leaves. The OPLS-DA model has a good fit with an R2x of 0.901 for the independent variables, an R2y of 0.575 for the dependent variables, and a Q2 of 0.961 for the model prediction, indicating the reliability of this model36. As illustrated in Fig. 5b, c, both HCA and OPLS-DA analyses show the same result. All samples could be clearly divided into two categories. Samples (GX1-GX12, GX14-GX15, and GD1-GD7) from the high-suitable habitats were distinguished as a cluster, and samples (YN1-YN5, HN, HB) from the medium- and low-suitable habitats were classified as a cluster. In these two methods, GX13 (Baise, Guangxi), GX16 (Chongzuo, Guangxi), and GD8 (Puyang, Guangdong) are classified together with the samples from the medium- and low-suitable habitats. C. cassia is a tropical and subtropical plant that thrives in warm environments. It has been mainly distributed in the warm Xijiang River Basin from ancient times to the present39. Baise, Chongzuo, and Puyang are far away from the Xijiang River Basin and have different climatic and soil environments, which may lead to large differences in chemical composition between the three samples and other samples from high-suitable habitats. It can be seen that FTIR, combined with HCA and OPLS-DA, is an effective and accurate method for identifying different suitable habitats of C. cassia, proving once again that the MaxEnt model used in this study is accurate.

Conclusion and prospects

In this work, we evaluated and predicted the distribution of C. Cassia suitable habitats across all periods using the parameter-optimized MaxEnt model. According to the findings, MaxEnt’s simulation of suitable regions and climatic suitability characteristics is basically in line with reality. (1) The most suitable environmental conditions for the growth of C. cassia were: precipitation of warmest quarter at 579.33–1165.20 mm, precipitation of driest quarter at 26.32–50.68 mm, Temperature seasonality range at 402.85–870.67, and topsoil clay content at 41.40-51.88% weight (2) In the current era, Guangxi and Guangdong provinces offer the most conducive environment for C. cassia. Under both climate change scenarios (SSP126 and SSP585), the suitable regional area for this species is projected to expand from the 2050s to the 2090s. Notably, the most significant increase is expected under the SSP585 scenario for the 2090s, where the largest suitable habitat area will be primarily concentrated in the southern region of the Yangtze River. (3) The different suitable habitats have a significant influence on the quality and characteristics of C. cassia, with the trans-cinnamaldehyde content being significantly higher in the high-suitable habitats than in the medium- and low-suitable habitats. (4) Furthermore, FTIR and HPLC techniques, in conjunction with HCA and OPLS-DA demonstrated that C. cassia samples collected from a variety of suitable habitats could be efficiently categorized, detected, and assessed according to the bioactive chemicals. The workflow used in this study provides a useful reference for determining C. cassia’s suitable planting area, identifies present and potential future high-suitability regions, and implements a methodical plan and thorough understanding of the relevant cultivation technology to guarantee a harvest of sufficient quantity and quality to satisfy market need, and enables actual production by providing a theoretical foundation.

In general, despite global warming and rising levels of carbon dioxide in the atmosphere, the suitable habitat area for C. cassia is expanding and C. cassia cultivation can be expanded accordingly. Furthermore, we must be conscious of the fact that C. cassia from various regions varies greatly in quality. The versatile cassia tree is a rare gift of nature. It has diverse uses ranging from ornamental and horticultural to medicinal and industrial. So it is recommended that in the highly suitable areas of Guangxi and Guangdong, C. cassia, which is rich in essential oils and trans-cinnamaldehyde, should be used as a medicinal raw material, while in other regions it can be used as a raw material for food, daily perfume, chemical production and multipurpose plantations, including timber production. In addition, many forests are experiencing a significant decline in resilience due to global climate variability40. C. cassia trees, with their strong sprouting and regenerative abilities, can contribute to restoring forest resilience. Expanding C. cassia cultivation can help maintain the ecological balance in the region and promote forest recovery, particularly by facilitating increased carbon sequestration and reduced carbon emissions in the future.