The Changes of Leaf Reectance Spectrum and Leaf Functional Traits of Osmanthus Fragrans are Related to the Parasitism of Cuscuta Japonica

Background: Studies on the inuence of parasitism on plants based on hyperspectral analysis have not been reported so far. To fully understand the variation characteristics and laws of leaf reectance spectrum and functional traits after the urban plant parasitized by Cuscuta japonica Choisy. Osmanthus fragrans (Thunb.) Lour. was taken as the research object to analyze the spectral reectance and functional traits characteristics at different parasitical stages. Results: Results showed that the spectral reectance was higher than the parasitic reectance in the visible light and near infrared. The spectral reectance in 750 ~ 1400 nm was the sensitive range of spectral response of host plants to parasitic infection, which is universal at different parasitic stages. We established a chlorophyll inversion model (y=-65913.323x+9.783, R 2 =0.6888) based on the reectance of red valley (minimum band reectance in the range of 640 ~ 700 nm), which can be used for chlorophyll content of the parasitic Osmanthus fragrans. There was a signicant correlation between spectral characteristic parameters and chlorophyll content index. Through the change of spectral parameters, we can predict the chlorophyll content of Osmanthus fragrans under different parasitism degrees. Conclusion: After the host plant was invaded by parasitic plants, its leaf functional traits are generally characterized by thick leaf, small leaf area, small specic leaf area, low relative chlorophyll content, high dry matter content and high leaf tissue density. These ndings indicate that the host plant have taken certain trade-off strategy to maintain their growth in the environment invaded by parasitic plants. Therefore, we suspect that there may be leaf economics spectrum in the parasitic environment, and there was a general trend toward "slow investment-return" in the global leaf economics spectrum.


Background
The prevalence of plant disease affects agriculture and forestry, reduces the quantity and quality of products, and poses a huge threat to agricultural safety and urban ecology. Parasitic plants are one of the special groups commonly existing in the global ecosystems [2,3]. The common parasitic plants include Taxillidae, Mistletoe and Cuscutaceae [4,5]. Among them, Cuscutaceae is one of the most common parasitic plant species in China, and Cuscuta japonica Choisy is widely distributed [6]. Cuscuta chinensis is seriously short of chlorophyll and other important substances to maintain its photosynthesis due to the degradation of its roots and leaves, commonly known as Caulis Sinomenii (6,7). It usually parasitizes the root and stem of the host plant through its special root absorption, and depends on absorbing carbohydrates, inorganic salts and water from the host to maintain its survival, growth and reproduction [8]. Studies have shown that the host range of Cuscuta Japonica Choisy is quite wide, and the vast majority of herbaceous dicotyledonous and monotyledonous plants may become parasitic objects of Cuscuta japonica Choisy. Cuscuta japonica Choisy usually grows in a winding way and spreads rapidly [9,10,11]. In addition, it can quickly form roots and continue to grow after the stem is broken. Moreover, when the damage is serious, the whole host plant is often covered with the stem strips of Cuscuta japonica Choisy, causing the host plant to grow poorly and even causing the whole host plant to die [11]. Therefore, parasitic plants are one of the important factors that endanger the urban greening plants and seriously threaten the urban environment. Researches on Cuscuta japonica Choisy parasitism mainly focus on its own biological and ecological characteristics and its effects on photosynthetic physiology and ecosystem of parasitic objects [11,12,13]. For example, Beifen Yang et al. study the effects of Cuscuta japonica Choisy parasitism on the growth and reproduction of Solidago canadensis L. Sumin Guo et al. study the growth trade-off mechanism of Alternanthera philoxeroides (Mart.) Griseb. on Cuscuta japonica Choisy parasitism [14,15]. All of these indicated that the host plants often change their growth defense strategies to maintain their own survival and reproduction after being subjected to Cuscuta japonica Choisy parasitism stress. In addition, many studies have shown that the in uence of parasitic plants on the biomass of the community in which they live have many uncertainties, which are often affected by the community's own characteristics, external environment and other factors, and mainly negative effects [16,17,18]. Osmanthus fragrans Lour., one of the most common tree species in China, plays a major role in the main ecological, cultural and landscape functions of the city. However, the parasitism of Cuscuta japonica Choisy seriously hinders the normal growth of Osmanthus fragrans. Diagnosis, monitoring and early warning of urban tree health have always been the focus of international urban forestry research. Therefore, how to monitor and obtain the growth status and the infringed status of the damaged vegetation is the key to effectively prevent and control the infringement.
In recent years, with the rapid development of hyperspectral technology, it has been widely used in forestry monitoring. Hyperspectrum has many advantages, such as high resolution, abundant information, simple data acquisition and so on [19,20]. Different plants have different re ectance spectral characteristics, and the same plant also has different re ectance spectral characteristics under different growth stages and different growth conditions [21,22]. Such spectral characteristics vary depending on the type of plant, the growth stage, the chlorophyll content, the cell water content, and the health status (whether or not it is affected by diseases, insect pests or parasitic plants) [21,23,24]. Plant functional traits refer to a series of internal physiological functions and external morphological characteristics gradually formed during the long-term interaction between plants and environmental factors, thus avoiding and reducing the adverse effects of the environment on them to the greatest extent [25]. In 2004, Wright et al put forward the concept of 1eaf economics spectrum (LES) for the rst time by analyzing the correlation between leaf functional traits. LES is the general internal relationship among functional traits of leaves [26]. The leaf economics spectrum, that is, the intrinsic relationship among leaf functional traits, re ects the resource allocation trade-off strategy among the functional traits [27]. Leaf functional traits can objectively re ect the impact of environmental changes on plants and the adaptability of plants to the environment, and predict plant characteristics [27,28,29]. Therefore, we suspect that leaf functional traits can also be used to diagnose the relationship between biological interactions, and applying imaging spectral remote sensing technology to study and utilize the change information of the spectral characteristics of affected plants can provide a reliable basis for large-scale monitoring of the occurrence of diseases and insect pests. Related researches based on forestry hyperspectral mainly focus on plant yield, crop seed vigor, plant diseases, and plant feature extraction [29,30,31,32,33,34]. However, there are few studies on the response of plant leaf functional traits and leaf spectral characteristics to parasitic plant invasion have not been reported.
As plants are endangered by parasitic plants, they will grow poorly or even die within a certain period of time [30,35,36]. Therefore, how to monitor and obtain the growth status and the invasion of the damaged vegetation is the key to effectively prevent and control parasitism. In order to fully understand the changing characteristics and laws of leaf spectrum and leaf functional traits of host plants after being parasitized, and to further explore the response mechanism of leaf re ection spectrum and plant traits to plant parasitic stress. In this study, Osmanthus fragrans Lour., a typical greening tree species in China, was taken as the research object. Spectral re ectance characteristics and leaf functional traits of Osmanthus fragrans leaves before and after being parasitized by Cuscuta japonica Choisy and different parasitic areas were analyzed, and sensitive bands of Osmanthus fragrans response to parasitic stress were obtained. The results provide a reference for the monitoring and early warning of the parasitism of Cuscuta japonica Choisy. At the same time, it provides a new experimental basis for different measures to control Cuscuta japonica Choisy and provides a theoretical basis for an in-depth understanding of the parasitic damage mechanism of Cuscuta japonica Choisy and its control strategies.

Changes in Leaf Functional Traits of Osmanthus Fragrans Leaves Parasitized by Cuscuta japonica Choisy
In this study, we selected six plant functional traits which are sensitive to environmental changes and external stress, including chlorophyll content, leaf area, leaf thickness, speci c leaf area, leaf dry matter content and leaf tissue density. As shown in Fig. 1, there are signi cant differences in leaf functional traits of Osmanthus fragrans between healthy leaves and the leaves being parasitic by Cuscuta japonica Choisy. The chlorophyll content index, leaf area and speci c leaf area of Osmanthus fragrans were signi cantly lower than those after parasitism, and with the increase of parasitism intensity, these indexes gradually decreased (CK > T1 > T2 > T3). On the contrary, the leaf thickness, dry matter content and leaf tissue density of Osmanthus fragrans were signi cantly higher than those after parasitism, and these indexes gradually increased with the increase of parasitism intensity (CK < T1 < T2 < T3). Studies show that chlorophyll is one of the important pigments in plant photosynthesis. As the increase of parasitic intensity, chlorophyll content index decreases gradually. The reasons may be that the poor growth of Cuscuta japonica Choisy due to its foraging of host nutrients, water, and nutrients after parasitization [37,38,39]. In addition, the host plant lacks su cient light due to the overgrowth of the parasitic plant Cuscuta japonica Choisy, which results in large area shading [40,41].
Spectral characteristics of Osmanthus fragrans before and after parasitizing by Cuscuta japonica Choisy As shown in Fig. 2, before and after Cuscuta japonica Choisy parasitism, the trend changes of spectral re ectance curves of Osmanthus fragrans leaves generally tend to be consistent. In the visible to the near-infrared band (350 ~ 1800 nm), the spectral re ectance is obviously different, which generally showing before parasitism (average value 0.0676 ~ 1.2633) and after parasitism (average value 0.0430 1.0061), but after parasitism, the spectral re ectance of Osmanthus fragrans is slightly higher than that of healthy Osmanthus fragrans in the band 350 ~ 750 nm. In addition, in the range of 350 ~ 1800 nm, there are four main re ection peaks and ve main absorption valleys in the spectral re ection curve of Osmanthus fragrans leaves to all treatments, and their positions are basically the same. The re ection peaks are respectively located at 560 nm, 1150 nm, 1300 nm and 1650 nm, and the absorption valleys are respectively located in the ranges of 350-560 nm, 600-700 nm, 950-1050 nm, 1150 ~ 1250 nm, and 1400 ~ 1500 nm. According to Table 1, chlorophyll content index of Osmanthus fragrans gradually decreased with the deepening of parasitism. Previous studies have shown that chlorophyll content can better characterize the light re ection curve of plant leaves [31,42]. Therefore, the spectral re ectance curve of the parasitized Osmanthus fragrans leaves is higher than that of the non-parasitized healthy Osmanthus fragrans leaves, which may be related to the decrease of chlorophyll content.

Spectral characteristics of Osmanthus Fragrans leaves under the different parasitic intensity of Cuscuta japonica Choisy
As can be seen from Fig. 3, under different parasitic intensities of Cuscuta japonica Choisy, the leaf surface spectral re ectance curves of Osmanthus fragrans are basically the same, but the spectral re ectance values are signi cantly different. Spectral re ectance values generally decreasing with the deepening of parasitic intensity, and the re ectance values are CK > T1 > T2 > T3. In the visible light to near-infrared 350 ~ 1800 nm band, the spectral re ectance of Osmanthus fragrans leaves under different parasitic intensities is the most easily distinguished in the range of 750 ~ 1400 nm, which indicated that this band is the sensitive range of host plants' spectral response to parasitic infection, and at the same time, this change characteristic is common under different parasitic conditions. In addition, we can also see that the spectral re ectance curve slope of Osmanthus fragrans leaves has a sharp increasing trend in the range of 700 ~ 780 nm. Studies have shown that the phenomenon of plants increasing suddenly in this waveband belongs to the typical "red edge effect" characteristic of plants. At the same time, the spectral re ectance of the leaves of the host plant (Osmanthus fragrans) with different relative chlorophyll contents has a higher re ection platform in the range of 750 ~ 1400 nm, which is wavy and may be affected by the cell structure of the leaves [43,44]. Among them, the sample with the lowest re ection coe cient are the sample with the lowest chlorophyll content index (the highest parasitic intensity). The re ectance of the sample with the highest chlorophyll content (without parasitism) is the highest at 1150 nm, which is 0.998. There is a signi cant valley at 1350 ~ 1800 nm, which may be closely related to light absorption by water [45,46,47,48].
Dynamic changes of spectral characteristic parameters of Osmanthus fragrans in different parasitic stages From Fig. 4 and Fig. 5, it can be seen that the general trend of the rst derivative spectrum of the leaf surface of the host plant under different parasitic intensities from visible light to the near-infrared band (350 ~ 1800 nm) is basically the same, but there are some differences in values. After Osmanthus fragrans was parasitized by Cuscuta japonica Choisy, there was an obvious "blue shift" in the red edge of its leaf surface spectral curve. With the deepening of parasitic intensity, the degree of "blue shift" also increased, indicating that with the increase of parasitic intensity, the in uence on the red edge of the leaf surface became more severe. In addition, the slope of the red edge of the host plant decreased obviously after parasitization (CK > T1 > T2 > T3). Many studies show that the red edge slope has a good indication of chlorophyll content [49]. Combined with Table 1, it can be seen that with the deepening of parasitic intensity, chlorophyll is decreasing. Therefore, we suspect that the cause of this phenomenon may be related to the in uence of Cuscuta japonica Choisy on the photosynthesis of host plants. Generally speaking, the red valley re ectance of healthy plants is the highest, but with the deepening of parasitic intensity, the spectral red valley re ectance of host plants shows a decreasing trend with T1 > T2 > T3.
Under different parasitic conditions, the position of the yellow edge is not affected, and it is all at 570 nm. However, with the deepening of parasitic intensity, the slope of the yellow edge and the re ectivity of the green peak gradually decreases, while the position of the green peak presents shifts to the long wave direction. At this time, the re ectivity of the water stress wave band increases gradually. Studies have shown that the spectral re ectance of vegetation in the range of 1550 ~ 1750 nm is usually closely related to the cell structure and water content of plants, which indicated the water absorption characteristics [50]. Therefore, with the deepening of the invasion degree of Cuscuta japonica Choisy, the cell structure of the leaves suffers certain damage, and the cell uid of the leaves gradually decreases, thus causing the absorption of light to decrease and the re ection to obviously increase.
Correlation between chlorophyll content and spectral characteristic parameters of host plants with different parasitic degree of Cuscuta japonica Choisy As shown in Fig. 1, CCI of the host plant (Osmanthus fragrans) gradually decreased with the deepening of the parasitic intensity of Cuscuta japonica Choisy. Previous studies generally believed that chlorophyll was an important parameter to determine the spectral re ectance curve characteristics of plant [51].
When the vegetation is in a healthy growth state and the chlorophyll content is high, the position of the red edge moves towards the long wave direction [51,52]. However, when vegetation is subjected to external environmental stress, such as drought stress, high temperature stress or pest damage, the red edge position tends to the short-wave direction [53]. Figure 6 and Table 1 showed the correlation between different spectral parameters and CCI and LT. The results of correlation analysis between plant functional traits and spectral parameters show that they show different correlations. It can be seen that there was an extremely signi cant correlation between spectral parameters and CCI. There was a signi cant correlation between RGP and LT. Among RRV, RGP, RES, RWSB and CCI were all highly correlated. The correlation between valley re ectance and chlorophyll content reaching the maximum (y=-65913.323x + 9.783, R 2 = 0.6888), which indicated that red edge characteristics were very sensitive to parasitic infestation and can be used to characterize changes in chlorophyll content of Osmanthus fragrans under different parasitic degrees. As shown in Fig. 7, we tested the chlorophyll inversion model of red valley re ectance, and found that the prediction accuracy of this model was high and stable (R 2 = 0.8811, RMSE = 0.0004). Table 1 Pearson correlation analysis between chlorophyll content index and spectral feature parameters. * indicates that the correlation reaches a signi cant level at the level of P < 0.05. and ** indicates a signi cant correlation between functional traits. As can be seen from Table 2, there was an interdependent relationship between the functional traits of the leaves. There was a signi cant positive correlation between LA and SLA. There is a signi cant negative correlation between SLA and LDMC and LTD. LA was signi cantly negatively correlated with LDMC and LTD. There was a signi cant negative correlation between LT and LTD. There was a very signi cant positive correlation between LDMC and LTD. There was a signi cant positive correlation between CCI and SLA. At the same time, LT has a negative correlation with SLA and LA, but the correlation has not reached a signi cant level. Studies have shown that leaf functional traits can re ect the adaptability of plants to the environment, but compared with a single leaf functional trait, continuous leaf economic spectrum can better re ect the growth strategy and adaptation mechanism of plants [54,55]. In this study, there was an obvious tradeoff relationship between the functional traits of plant leaves, which indicated that when plants are damaged by parasitic plants, host plants show certain ecological trade-off strategies in terms of functional traits for survive. SLA is closely related to the growth and survival strategy of plants, and can represent the ability of plants to adapt the environment and obtain resources [56]. In this study, after being invaded by parasitic plants, the reduction of SLA of the host plants makes the plants more adaptable to resource-poor environment. LDMC represents the of plants to maintain nutrients, while LTD re ects the bearing capacity and defense ability of plant leaves, which is closely related to the turnover growth rate of leaves [56,57]. In this study, LDMC and LTD increased gradually with the increase of parasitic intensity, and showed a very signi cant positive correlation. This indicates that the plants can improve the nutrient retention ability of leaves under the adverse environment of parasitic stress, thus making more effective use of limited resources. The increase of LTD is bene cial to enhance the defense ability of plants against biological factors. To sum up, after the host plant was invaded by parasitic plants, its leaf functional traits are generally characterized by large leaf thickness, small leaf area, small speci c leaf area, low chlorophyll content index, high dry matter content and high leaf tissue density. Therefore, we suspect that the leaf economics spectrum may also exist in the parasitic environment, and there was a general trend toward "slow investment-return" type in the global leaf economics spectrum (Fig. 8).

Conclusion
In this paper, Osmanthus fragrans is used as the research object to analyze the spectral characteristics and leaf functional traits of different parasitic periods after the natural infection of Cuscuta japonica Choisy, revealing the relationship between the spectral characteristic changes and invasive processes after host susceptibility, aiming to use the hyperspectral remote sensing technology for parasitization. In addition, by establishing a correlation between spectral characteristic parameters and chlorophyll content, the research results can provide theoretical support for the prediction of plant diseases in the early stage. At the same time, it can provide a reference for monitoring and early warning of infringement, and a new experimental basis for different measures to control Cuscuta japonica Choisy. Main conclusions are as follows.
(1) Before and after being parasitized by Cuscuta japonica Choisy, the trend of spectral re ectance curve of Osmanthus fragrans leaves tend to be consistent in general. The spectral re ectance is obviously different, and it is generally higher before parasitism than after parasitism. In the range of 350 ~ 1800 nm, there are four main re ection peaks and ve main absorption valleys in the spectral re ection curve of Osmanthus fragrans leaves.
(2) Visible light band was di cult to re ect the harm of host plants, while the near-infrared band (750 ~ 1400 nm) has the greatest degree of spectral re ectance discrimination. This band was the sensitive range of spectral response of host plants to parasitic infection. At the same time, such variation characteristics were universal under different parasitic degree conditions, and can better re ect the harm of host plants.
(3) The position of red edge, the slope of red edge, re ectance of a green peak, and re ectance of water stress band can well re ect the invasion status of Cuscuta japonica Choisy in different parasitic stages. After parasitism, the red edge position of the host plant spectrum shifted to shortwave direction. With the deepening of parasitic intensity, the in uence on the red edge position becomes more and more serious, and the degree of "short wave migration" also increases. The red edge slope decreased, and the re ectivity of water stress band increased gradually.
(4) With the increase of parasitic intensity, the relative content of chlorophyll in host plants gradually decreases, and the spectral characteristic parameters are signi cantly correlated with them. Chlorophyll inversion model based on red valley re ectance has the highest accuracy (y=-65913.323x + 9.783, R 2 = 0.6888), and can be used for chlorophyll content of parasitic Osmanthus fragrans.
(5) After the host plant was invaded by parasitic plants, its leaf functional traits are characterized by large leaf thickness, small leaf area, small speci c leaf area, low relative chlorophyll content, high dry matter content, and high leaf tissue density. Therefore, we suspect that there may be leaf economics spectrum in the parasitic environment, and there was a general trend toward "slow investment-return" in the global leaf economics spectrum.

Research area and sample collection
Nanning city is located in the southwest of Guangxi province, between 107°45 ′-108°51 ′ east longitude and 22°13 ′-23°32 ′ north latitude. It is a humid subtropical monsoon climate with abundant sunshine and rainfall throughout the year. The annual average temperature is about 21.6℃, the annual average rainfall is 1304.2 mm, and the average relative humidity is 79% (Quoted from https://baike.baidu.com). The sampling area is located on the campuses of Guangxi University, Guangxi Finance and Economics University, and Guangxi Nationalities University. The straight distance of the three locations is about 8 km, and they all belong to community-based environments, ensuring the relative consistency of atmosphere, planting and maintenance management conditions. According to the proportion of the parasitic area of Cuscuta japonica Choisy to the crown area of the host plant, it is divided into four parasitic degrees (CK-Without parasitic, T1-Initial parasitism: less than 50%, T2-Parasitic metaphase: 50% ~ 80%, T3-Late parasitism: more than 80%), 30 Osmanthus fragrans of 15 ~ 20 years old and healthy growth per treatment were selected, and the planting location was away from the in uence of tall buildings and tall trees. Leaf samples were collected from 10: 00 a.m. to 12: 00 a.m. on June 2019. Ten mature and healthy leaves were cut from each tree, placed in an icebox and immediately brought back to the laboratory for spectral determination. The time from leaves collection to spectral measurement is controlled within 15 min, thus ensuring the original growth activity of leaf samples. As shown in Fig. 9, Fig. 9(a) is a plant that is not parasitic and Fig. 9(b) is a plant that is parasitic by Cuscuta japonica Choisy. Professor Wei Jiguang from Agricultural College of Guangxi University identi ed the plants and plant diseases involved in this study.
Leaf re ectance spectrum collection and calculation method of leaf functional traits FieldSpec3 near-infrared spectrometer (ASD, Malvern Panalytical, USA) was used to collect spectral data.
The spectrum band acquired by this instrument ranges from 300 nm to 2500 nm. The nal output spectral re ectance curve is the average of 10 repetitions. The spectral measurement process is shown in Fig. 2. The light source is the solar light source during 12: 00 to 13: 00. Whiteboard is manufactured from a sintered polytetra uoroethylene (PTFE) based material. In order to reduce human interference, instrument operators wear cotton work clothes.
In July 2019, thirty Osmanthus fragrans with different degrees of damage were selected at each test site, and thirty mature and healthy leaves were randomly collected from each plant during 9:00-12:00 in ne weather. The relative chlorophyll content index (CCI) was determined by CCM-200 plus chlorophyll meter As shown in Table 1, the hyperspectral characteristic parameters selected in this study include the position of red edge (REP), the slope of red edge (RES), the re ectance of red valley (RRV), the re ectance of green peak (RGP), the position of green peak (GPP), the re ectance of water stress band (RWSB), the slope of yellow edge (YES), and the position of yellow edge [58,59,60].

RES
The maximum re ectance in the red band (680 ~ 750 nm).

REP
The wavelength position corresponding to the maximum re ectance in the wavelength band 680 ~ 750 nm.

RRV
Minimum band re ectance in the range of 640 ~ 700 nm.

RGP
Maximum band re ectance in the range of 510 ~ 580 nm.

GPP
The wavelength position corresponding to the green peak re ectance in the wavelength band 510 ~ 580 nm.

RWSB
Maximum band re ectance in the range of wavelengths from 1550 ~ 1750 nm.

YES
The maximum re ectance in the yellow band (550-582 nm).        Leaf samples of Osmanthus fragrans (Thunb.) Lour. with or without parasitism of Cuscuta japonica