Assessing sources of variation on leaves reflectance spectra in coastal saltmarshes and seagrasses

Assessing sources of variation on leaves reflectance spectra in coastal saltmarshes and seagrasses | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Assessing sources of variation on leaves reflectance spectra in coastal saltmarshes and seagrasses Andre C. Costa-Neves, Cristina Galván, Bárbara Ondiviela This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7591763/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 7 You are reading this latest preprint version Abstract There is an urgent need for effective large-scale biodiversity monitoring across ecosystems, given the recent tendency toward global biodiversity loss. The assessment of plant spectral diversity offers a promising approach as it is intrinsically linked to phylogenetic and functional diversity. This study investigates the relationship between taxonomic, functional, and spectral diversity in temperate saltmarsh and seagrass ecosystems in the Gulf of Biscay. Using hyperspectral leaf reflectance data and functional traits from 19 plant species across four estuaries, these three dimensions of biodiversity were compared. The predictive power of spectral data was assessed for estimating biochemical and anatomical traits and species identification. Results reveal significant correlations between functional and spectral diversity, with species sharing similar functional traits exhibiting similar spectral signatures. Spectral diversity is significantly influenced by taxonomic classification, with higher taxonomic levels (e.g., order, class) explaining substantial part of the spectral variation. Spectral regions of 720–770 nm and 1330–1380 nm were important for species discrimination, achieving 98% accuracy. Partial least squares regression models successfully estimated functional traits (e.g., water content, carbon, phosphorus) with high precision in these environments. These findings demonstrate that spectral data can effectively capture taxonomic and functional diversity, offering an effective tool for large-scale biodiversity monitoring in estuarine ecosystems. This study underscores the potential of remote sensing to track biodiversity and ecosystem health, providing a foundation for future applications in conservation and management. Biological sciences/Ecology Earth and environmental sciences/Ecology Earth and environmental sciences/Environmental sciences Earth and environmental sciences/Ocean sciences Optical traits functional assembly wetland vegetation leaf chemistry Cantabrian estuaries imaging spectroscopy Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Highlights • Key spectral regions (720–770 nm, 1330–1380 nm) enable 98% species discrimination. • Partial least squares regression models accurately retrieve functional traits (R²: 0.42–0.92) from leaf spectra. • Spectral diversity is mostly explained by taxonomic and functional diversity in temperate estuarine ecosystems. Introduction In light of the current biodiversity loss crisis, there is an urgent need for effective methods to assess biodiversity on a large scale across diverse ecosystems. The assessment of biodiversity in situ can be difficult to acquire and time-consuming, due to logistic issues and different methodologies among studies (Rocchini et al., 2010). A promising approach is through remote sensing of plant optical diversity, which is intrinsically related to other forms of biodiversity, such as phylogenetic and functional diversity (Jetz et al. 2016; Wang and Gamon 2019). Biodiversity encompasses multiple dimensions, many of which are interrelated; it consists of a complex process that varies across space and time, from the genetic to the ecosystem level (Kaennel 1998; Rocchini et al. 2004). Some of the main components included in this broad concept are the species richness and the functional diversity. The functional diversity is based on the principle of “functional traits”, which are the morphological, physiological and phenological attributes that directly influence individual’s fitness (Violle et al., 2007). These traits represent key biological processes (Asner et al. 2017), providing a viable and complementary approach to traditional species diversity assessments (Asner et al. 2011; Jetz et al. 2016; Schweiger et al. 2018; Ordway et al. 2022). The adoption of a trait-based approach is being increasingly used in ecological studies (Violle et al. 2007). Functional traits, in a variety of ecosystems, have been successfully estimated using spectral data derived from leaf and canopy reflectance (Asner et al. 2015; Schweiger et al. 2017; Wang et al. 2020). However, these models are usually site-specific, and applying existing models to new datasets can often be unfeasible (Burnett et al. 2021), requiring the calculation of the regression coefficients for the statistical models for each target vegetation. Similarly, the ability to assess species richness and other diversity metrics (such as alpha and beta diversity) based on spectral heterogeneity in vegetation communities has been demonstrated across many ecosystems (Rocchini et al. 2010; Feret and Asner 2013; Schweiger and Laliberté 2022). These estimations are based on the concept of the spectral diversity, which is a biodiversity metrics estimated as a function of the electromagnetic radiation reflected from plant communities. The fundamental is that various plant traits influence light absorption and scattering, leading to detectable variations in plant optical properties (Ustin et al. 2009; Asner and Martin 2016). The use of the spectral diversity approach is particularly useful to monitor sites with difficult access and extensive areas (Schweiger et al. 2018). While no single method is universally applicable across all ecosystems, investigating spectral diversity in different environments offers a standardised, consistent, and cost-effective means of biodiversity assessment (Cavender-Bares et al. 2020). Novel approaches for evaluating biodiversity through spectral diversity have been tested and developed in a range of ecosystems, from grasslands (Dalmayne et al., 2013; Schweiger et al., 2018; Warren et al., 2014; Zhao et al., 2021) to wetlands (Ozesmi and Bauer 2002; Adam et al. 2010; Klemas 2013), as well as in tropical rainforests (Asner et al., 2017). Many of these studies have concluded that spectral reflectance can effectively identify plant species and functional types and can also be correlated with species richness estimations obtained through remote sensing. To effectively relate traditional biodiversity metrics with spectral diversity, it is essential to adopt suitable spectral scales for each habitat (Bohrer et al. 2022). Hyperspectral data enables a detailed analysis and differentiation of vegetation types that sometimes is not possible with broader-band multispectral sensors (Adam et al. 2010). Given the strong scale-dependence of biodiversity estimations (Rocchini et al. 2016; Wang and Gamon 2019), spatial resolution has proven more critical than spectral resolution for accurate classification of salt marsh vegetation (Belluco et al. 2006a). Thus, the identification of plant species based on spectral data has been demonstrated promising in coastal wetlands (Schmidt and Skidmore 2003; Belluco et al. 2006; Valle et al. 2015; Curcio et al. 2023), particularly in the non-inundated zones (Adam et al. 2010; Klemas 2013). However, it essential to acquire reliable, repeatable and standardized in situ information to calibrate and validate the biodiversity estimations across diverse ecosystems (Bohrer et al. 2022). To support such efforts, a detailed investigation of the relationship between biodiversity dimensions in North Atlantic temperate coastal wetlands is required. As a first step, this involves assessing species diversity and biochemical uniqueness in situ and linking these attributes to spectral diversity (Adam et al. 2010). This study is based on the hypothesis that salt marsh and seagrass species exhibit distinct optical and functional traits, resulting in species-specific spectral signatures that can be predicted using hyperspectral data. The main objectives of this study are (1) to explore how the taxonomic and functional diversity are related to the spectral diversity in salt marshes and seagrasses ecosystems; and (2) to evaluate the predictive potential of leaf spectral reflectance to capture functional and taxonomic diversity. For this purpose, we aim to determine whether plant functional groups exhibit similar spectral traits; to what extent taxonomic diversity explains the spectral variability of the plant community; in which frames the intraspecific spectral variation is consistently lower allowing interpretations of interspecific spectral differences; and how accurately can functional traits be inferred from leaf spectral reflectance in salt marshes. Materials & Methods Study area The study was conducted in four estuarine systems along the Gulf of Biscay: Oyambre Estuary, Santander Bay, the Joyel Marshes, and the Santoña Marshes (Fig. 1). These estuaries were selected due to the presence of well-developed and healthy Atlantic temperate saltmarsh and seagrass habitats. They are representatives of the geomorphological and ecological diversity of estuarine vegetation found in northern Spain. These habitats are distributed along two main environmental gradients: salinity, which determines the presence of halophyte (e.g., cordgrass Spartina sp., Salicornia sp. and Limonium vulgare ) and subhalophyte (rushes Juncus maritimus , common reed Phragmites australis and scirpus Bolboschoenus maritimus ) species; and emersion, which drives vegetation zonation patterns, with seagrasses (i.e., Zostera marina and Zostera noltii ) colonizing the tidal flats and saltmarshes occupying the mid and high intertidal. A total of eight study sites were distributed across the four selected estuaries, covering the two environmental gradients to capture variation in species and populations (Fig. 1). Three study sites were located in seagrass habitats (2 in the Santander Bay and 1 in the Santoña Marshes) and five study sites in saltmarshes (2 in the Oyambre estuary, 2 in the Joyel Marshes and 1 in the Santander Bay). All sites are located within protected conservation areas, including Natura 2000 network and national parks. Plant collection A total of 19 coastal estuarine plant species were sampled across the study area (Table 1). The field campaigns were carried out in 2023 during flowering periods from mid-July to late September. Each species was sampled in three different patches within the same study site, separate patches were selected in an attempt to maximise variability within the same population (Fig. 2). Each species was collected between one and three study sites, depending on its local occurrence and the ecological integrity of the population. Table 1 List of species sampled, with their respective taxonomic classification (genus, family, order, and class). Also, showing the number of sites where these species were collected (nº sites). Class Order Family Genus Species nº sites Liliopsida Poales Cyperaceae Bolboschoenus maritimus 1 Poaceae Spartina alterniflora 1 anglica 1 maritima 2 Phragmites australis 3 Puccinellia maritima 2 Alismatales Zosteraceae Zostera noltii 3 marina 3 Juncaceae Juncus maritimus 2 Magnoliopsida Caryophyllales Amaranthaceae Salicornia ramosissima 1 Sarcocornia perennis 1 Halimione portulacoides 2 Plumbaginaceae Limonium humile 2 dodartii 1 vulgare 1 Amaranthaceae Suaeda maritima 2 Asterales Asteraceae Aster tripolium 1 Inula crithmoides 2 Lamiales Plantaginaceae Plantago maritima 1 Healthy adult individuals of the target species, with branches well exposed to sunlight, were identified in the field and collected (Asner et al. 2015). Samples (~200 grams of plant material) were maintained in ice coolers until brought to the laboratory (less than 2 hours) and stored in the refrigerator (4°C) in the dark until the fresh weight and the leaf reflectance were measured. Spectral and morphological measurements took less than 5 hours since collection. The remaining plant material was frozen in the dark for biochemical analysis (i.e., foliar nitrogen, phosphorus, and pigment content). Spectral data Hyperspectral reflectance of the leaves of each plant sample was measured with ASD Field Spec 4 (Analytical Spectral Devices, Boulder, USA) connected to a leaf probe equipped with an internal halogen light source under laboratory conditions. The instrument has a spectral range between 350 and 2500 nm with 1 nm spectral resolution. Before starting the measurements, the instrument was turned on to warm up for at least 30 minutes, dark current calibrations and white references (using the white panel from the leaf probe) were measured at the beginning and every 30 minutes to reduce eventual noise and to obtain relative reflectance (Schweiger 2020). Fully grown and healthy leaves were selected from the top of the branches, removed, and gently cleaned with absorbent paper before measurement. The leaves were placed aiming to fill the entire field of view. Given that some species had narrow leaves, they had more than one leaf to cover the field of view. Narrow leaves were sustained by a small sample platform composed of a square window made of a slightly rigid plastic sheet adapted from Noda et al. (2013). This sample mount allowed the placement of several blades in between two square frames (4 cm x 4 cm wide) and is fixed by the pressure of binder clips to handle the small leaves easily, serving as a platform to stabilise the leaves and completely fill the view of the fore optic (Support information – SI 6). Each spectral measurement was set to be the average of 30 internal readings taken by the instrument. The number of averaged internal readings has to be a balance between acquiring quality measurements and avoiding overheating the leaf by exposing it too long to the internal halogen lamp (Neuwirthová et al. 2017; Caturegli et al. 2020). Each species had three samples per study site, and each sample had three reflectance measurements recorded (one per leaf). Thus, each plant species had 9 measurements in each study site where they were collected (Fig. 2). The spectral data collected in the laboratory was pre-processed first by correcting to match sensors using the R package spectrolab (Meireles et al. 2017). The noisy regions in the beginning and at the end of the spectra were excluded by trimming the spectrum to the length of 400 to 2400 nm (Schweiger et al. 2018). Lastly, each spectrum was vector normalised by dividing the reflectance value of each band by the magnitude of the spectral vector, where the magnitude ∣v∣ is calculated as the square root of the sum of the squared reflectance values across all bands, with n being the total number of bands and x i the reflectance value of waveband i . Functional traits Eight relevant functional traits, divided into anatomical (leaf mass per unit of area and water content on leaves) and biochemical traits (total organic carbon, foliar nitrogen, phosphorus, carotenoids, chlorophyll a and b) were analysed. Anatomical traits After the reflectance measurements, the same leaves had their petiole removed. They were scanned to calculate the leaf area with the software ImageJ (Schneider et al. 2012). They were also weighed fresh and after they had been dried in the oven (60°C; 72 h), the trait leaf mass per area (g/m2) was calculated based on the dry weight (DW) divided by the leaf area. After removal from the oven, the samples were left at room temperature for ten minutes before being weighed. This step allows the plant material to cool down and prevents it from reabsorbing moisture from the air. The water content (g water/ g DW) was calculated by the difference between the fresh weight and the dry weight. Biochemical traits The same dried leaves used for the anatomical traits measurements were used in the analysis of total organic carbon (TOC). The determination of TOC in leaves (%) was through dry combustion (with an NDIR infrared detector) after correcting for inorganic carbon. A more detailed explanation of the laboratory procedures to measure these traits is provided in the Supplementary Material (SI 1). Additional leaves were randomly selected (roughly 100g of plant material) from the sample for chlorophyll and carotenoids (µg/g), nitrogen (%) and phosphorus analysis (%). The quantification of chlorophyll a, chlorophyll b, and total carotenoids was carried out following the equations proposed by Lichtenthaler & Buschmann (2001) and Lichtenthaler (1987), based on spectrophotometric determinations at specific wavelengths. These remaining leaves were slowly unfrozen before the analysis in a refrigerator (6ºC for 24 h). Nitrogen content in plant material was determined using the Kjeldahl method, a standard procedure for measuring organic and ammoniacal nitrogen (Bradstreet 1954). Phosphorus content in plant material was measured using UV-VIS spectrophotometry. Data analysis Relationship between types of biodiversity Functional and spectral diversity relationship To assess whether functionally similar species exhibit similar spectral reflectance, we applied the methodology adopted by Schweiger et al. (2018) using dissimilarity matrices. Initially, the median reflectance values of each band (1nm wide) from the spectral data were compiled into principal components (PCs) for each species. Manhattan distances were then calculated for each species pair to construct the spectral dissimilarity matrix. Functional distance was determined using the z-scores of the measured traits for each species. Pairwise comparisons were plotted, and a linear regression was fitted for each species. Additionally, an overall linear regression with a 95% confidence level was computed for all species combined to evaluate the general trend. The significance of the correlation between spectral and functional distance matrices was assessed using a Mantel test with 999 permutations. The Mantel statistic (r), calculated between the pairwise elements of the two matrices, was also used to quantify the strength of the association. All analyses were conducted in R (R Core Team, 2024). Taxonomic and spectral diversity relationship To quantify the variance in leaf spectral reflectance explained by different hierarchical taxonomic levels, we fitted linear mixed-effects models (LMMs) to the spectral data. This approach is well-suited for handling unbalanced data and nested structures. Initially, Principal Component Analysis (PCA) was performed on normalised hyperspectral reflectance data, with PCs retained as new variables representing the main axes of spectral variation. The analysis was conducted using the prcomp function in R. PCs were retained until reaching a cumulative explained variance threshold of ≥90%, and used as response variables in the LMMs. The selected PCs were graphically displayed to compare the distances between species and order. The models (one for each PC) included taxonomic levels (class, order, family, genus, and species) as predictor variables and site as random effects. Variables with sufficient variation were retained in the model, implemented using the lme4 package. Variance components were extracted from the final models to assess the contribution of each taxonomic level to spectral variability in each PC. The variance explained by each taxonomic level was then expressed as a percentage of the total variance for each PC. This allowed us to evaluate how spectral variation is partitioned across taxonomic hierarchies and how different PCs capture distinct biological information. Also, the intraspecific and interspecific variation in spectral diversity was quantified by comparing variance within and between species for the selected PCs. Intraspecific variance was calculated as the average variance within each species, while interspecific variance was derived from the variance of species means. The ratio of inter- to intraspecific variance provided a measure of how much spectral variation is observed at the species level. Predictive potential among types of biodiversity Functional traits quantification by spectral data Predictive models to retrieve functional traits based on hyperspectral data were developed for total organic carbon, foliar nitrogen and phosphorus, chlorophyll a and b, carotenoids, water content, and leaf mass per area. For this purpose, the database was divided into 80% used to calibrate the model, and the remaining 20% to validate. A PLSR model was fitted using the pls package (R), where the predictor matrix consisted of the spectral data, and the response variable was defined as each of the functional traits separately (Burnett et al. 2021). To avoid overfitting, the optimal number of components used in the regressions was determined through cross-validation with the "onesigma" rule. A lower number of components could be selected when model accuracy was maintained. The accuracy of the models was evaluated based on the R 2 and the root-mean-square error (RMSE). The use of partial least squares regressions (PLSR) is a common approach due to its capability to incorporate the full information from the leaf reflectance spectrum into a few uncorrelated latent factors (Doughty et al. 2017). These latent factors can accommodate the high collinearity among the predictors (wavelengths) and the predictor variables (functional traits). Taxonomic classification by spectral data To identify plant species based on hyperspectral reflectance, two fundamental steps are required: a feature selection followed by a discriminant analysis (Ferreira et al. 2013; Prospere et al. 2014). Features selection reduces dimensionality by identifying a subset of variables that enhance generalisation and reduce computational demands, while maintaining or improving classification accuracy (Hennessy et al. 2020). While the discriminant analysis is well-suited for maximising class separability while preserving the information contained in the selected variables. It is a commonly used approach for species classification based on spectral reflectance, often resulting in highly accurate models (Clark et al. 2005). Initially, curve smoothing was applied by calculating the simple average across blocks of five adjacent bands for each species (Yu et al. 1999; Pu 2009). Then, a Mann-Whitney U test was selected as the feature selection method. It identified the wavebands that provide the highest discriminatory power between species. A total of 171 different pairs among species were tested for each band, the wavebands that accumulated a higher number of significantly different pairs (α=0.01) were first selected (Schmidt and Skidmore 2003; Ferreira et al. 2013; Hennessy et al. 2020). Subsequently, the optimal number of wavebands to include in the model was determined using a support vector machine algorithm. This process employed 10-fold cross-validation, implemented through the R package e1071 (Version 1.7-16), in conjunction with the caret package to identify the feature subset that would yield the highest accuracy while minimising the risk of overfitting (Clark et al. 2005). To classify the species using the selected wavebands, Linear Discriminant Analysis (LDA) was chosen for the discriminant analysis. It has been proven effective classifier (Yu et al. 1999; Clark et al. 2005; Pu 2009), and applicable in wetland environments (Prospere et al. 2014; Hladik and Alber 2014). Firstly, the dataset was split into training and test sets (80% and 20% of the data, respectively). The partitioning was performed using the caret package to ensure balanced class representation in both sets. A LDA model was fitted to the training data using the lda function. After training, the model was used to predict the class labels of the test set. The accuracy and performance of the model were assessed based on the kappa statistic calculated from the predicted class labels to the true test labels ( i.e., species). Results Relationship between types of biodiversity Functional and spectral diversity relationship Species with greater functional similarity (similar biochemical and morphological traits) also exhibited higher spectral similarity (similar leaf reflectance patterns). Pairwise comparisons between species successfully highlighted moderate degrees of similarity based on both functional and spectral distance matrices (Fig. 3 ); the mean slope of the regressions and respective standard deviation was b = 13.7 (± σ = 7.8). The positive slope of nearly all regression lines indicates a trend that functionally similar species also tend to be spectrally similar. These slopes quantify the strength of the relationship, with stronger associations observed in Suaeda maritima (b = 24.25), Phragmites australis (b = 22.69), and Halimione portulacoides (b = 21.91). In contrast, weaker positive relationships were found for Zostera noltii (b = 8.26), Zostera marina (b = 9.39), and Spartina maritima (b = 10.03), while Juncus maritimus (b=-7.21) was the only species with a negative relationship. The Mantel statistic r of 0.46 (p = 0.001) indicates a moderate, significant positive correlation between spectral and functional diversity. The observed value exceeded all percentiles (90th, 95th, 97.5th, and 99th ) of the null distribution generated by 999 permutations, confirming that the correlation did not arise by chance. The Mantel test thus supports the trends observed in the linear regression slopes, reinforcing the evidence that spectral variation reflects underlying functional differences across species. Taxonomic and spectral diversity relationship Individuals from the same species tend to exhibit more similar leaf spectral reflectance patterns, and this similarity is consistent across different taxonomic levels (Fig. 4 a). Spectral variance within species was approximately five times lower than the variance observed among species, with intra- to interspecific variation ratios of 5.7 for PC1 and 4.8 for PC2. Additionally, the absolute variance estimated for PC1 (σ²=2.43) was notably higher than that for PC2 (σ²=0.38). The first two principal components captured a substantial proportion of the total spectral variation, with PC1 explaining 84.4% and PC2 12.5%. Plants from the same order tend to cluster together in the two-dimensional space of the PCA plot (Fig. 4 b). Likewise, a substantial portion of the variance in spectral response captured by the PCs (> 90%) is explained by different taxonomic levels (Fig. 4 a). Higher taxonomic groups, such as order (40.41%) and class (49.74%), explain the largest fraction of variation in PC1 and PC2, respectively. However, genus and family also contribute considerably to spectral differences in estuarine plants, with genus explaining 23.41% and 15.78%, while family accounts for 16.63% and 12.76% of the variance in PC1 and PC2, respectively. Additionally, the site effects play a notable role in spectral diversity, contributing 7.6% and 9% of the variation in PC1 and PC2, respectively. Predictive potential among types of biodiversity Functional traits quantification by spectral data The results show that it is possible to retrieve biochemical and anatomical traits of estuarine plant species in northern Spain from spectral data with reasonable precision (Table 2 ). The R² values ranged from 0.42 to 0.92, while the percentage root mean squared error of prediction (%RMSEP) varied between 8.56% and 21.08%. The number of components selected through cross-validation for model construction ranged from 4 (for carbon) to 12 (for nitrogen and LMA). Table 2 Summary of PLSR model performance for the eight traits analysed. The Root Mean Square Error of Prediction (RMSEP) is reported in the same units as the respective traits, alongside its percentage (%RMSEP), which accounts for differences in trait magnitudes. The number of components used to build the model (nº Components) and the sample size (N) to both train and validate the model is indicated Trait Unit R 2 RMSEP %RMSEP nº Components N Water g/g of dry mass 0.92 0.86 8.56 10 288 Organic Carbon % 0.87 2.76 12.38 4 41 Foliar Phosphorus % 0.81 0.3 12.34 11 183 Leaf Mass per Area g/m 2 0.76 54.8 10.8 12 288 Foliar Nitrogen % 0.66 3.31 17.51 12 183 Chlorophyll a µg/g 0.53 130.79 20.06 11 131 Carotenoids µg/g 0.44 46.57 19.71 9 131 Chlorophyll b µg/g 0.42 63.73 21.08 11 131 The validation R² values and the %RMSEP indicated that the models performed well for most traits, particularly water content, organic carbon, phosphorus, and LMA (Fig. 6 ). Notably, the model for organic carbon exhibited strong performance despite having the smallest sample size for training and validation (N = 41) and the lowest number of components (nº Comp = 4), achieving a high proportion of variability explained (R²=0.87). In contrast, models for traits related to pigment content showed moderate accuracy, as indicated by their lower R² values and higher %RMSEP (Table 2 ). The models achieved R² values of 0.92, 0.81, and 0.76 for water content, phosphorus, and LMA, respectively, indicating that a substantial proportion of the variability in these traits was explained by the spectral data (Fig. 5 ). Scatterplots of observed versus predicted values for both calibration and validation datasets further confirmed the accuracy of the models, with points closely aligned to the 1:1 line for water content, phosphorus, nitrogen, carbon, and LMA. For pigment-related traits, the relationship was less robust, although most predictions fell within the 95% confidence interval. Residual histograms for both calibration and validation datasets resemble a symmetric distribution around zero, indicating no systematic bias in the predictions. Taxonomic classification by spectral data Two spectral regions provided the best discrimination between the estuarine species at the leaf level, these are between 720–770 nm and 1330–1380 nm. These regions were located at the short borders of the NIR and SWIR ranges, specifically in the red edge shoulder and in the downward slope of the NIR reflectance plateau (Fig. 6 ). Within these regions, 20 spectral bands are sufficient to develop parsimonious models capable of distinguishing between the leaves of different species with accuracies of around 97%. Maximum accuracy (100%) is achieved when 45 or more bands are included (Supplementary Material – SI 2). Given that each band is 5 nm wide, the selection of 20 adjacent bands corresponds to a 100 nm spectral range, which is grouped into two distinct clusters. When expanding the analysis to include 45 bands (a 225 nm range) to achieve 100% accuracy. The discriminatory frame on the red edge expands further into the NIR range, reaching 900 nm. Although the top selected bands are within the VIS and NIR regions, there are relevant frames in the SWIR range (between 1840nm-1620nm) that could also help distinguish between salt marsh and seagrass species (Fig. 7 ). The results from the confusion matrix and LDA model demonstrate a high level of accuracy in classifying the plant species based on the provided spectral data (Supplementary Material - SI 3). The overall accuracy of the model is 97.87%, with a 95% confidence interval ranging from 88.71% to 99.95%. This high accuracy is further supported by a Kappa statistic of 0.977, indicating almost perfect agreement between the predicted and actual classifications. The accuracy being greater than the no-information rate is highly significant (p < 0.001), suggesting that the model performs substantially better than random chance. The confusion matrix reveals that the model correctly classified all instances of most species, with no misclassifications observed for the majority of classes (Supplementary material – SI 3). For example, species such as Zostera marina , Zostera noltii , Aster tripolium , Inula crithmoides , Suaeda maritima , and Halimione portulacoides were all predicted with perfect sensitivity and specificity (100%). Similarly, species like Limonium dodartii , Limonium humile , Phragmites australis , Juncus maritimus , Spartina maritima and Puccinellia maritima were also classified without error. The model struggled with Salicornia ramosissima , misclassified as Sarcocornia perennis , both belonging to the family Amaranthaceae. The LDA model's group provides insights into the spectral characteristics that distinguish each species. For instance, Salicornia ramosissima and Sarcocornia perennis exhibit higher reflectance values in certain spectral bands (e.g., 735, 725, and 730 nm) compared to other species. Discussion These results show that different biodiversity dimensions can be estimated from spectral features, and that spectral traits can be used to infer taxa and functional traits. These findings corroborate previous studies demonstrating the links among spectral, functional, and taxonomic diversity in terrestrial vegetation (Ustin and Gamon 2010 ; Schweiger et al. 2017 ; Laliberté et al. 2020 ; Bohrer et al. 2022 ). Furthermore, the present study extends this understanding by providing evidence that such relationships also persist within estuarine ecosystems. It has been shown that each species is composed of a set of anatomical, biochemical and spectral characteristics that are closely related, supporting the development of models to estimate functional traits and species based on the spectral data. Also, the spectral reflectance carries strong taxonomic influence across multiple hierarchical levels. This could mean that spectral data is likely driven by inherent physiological, structural, or biochemical traits and could be used as a powerful tool for species discrimination and functional trait quantification in temperate estuarine ecosystems. Relationship between types of biodiversity Functional and spectral diversity relationship Our results indicate that species functionally similar exhibit more similar spectral responses, reinforcing the strong relationship between functional traits and leaf reflectance observed in our dataset. This relationship arises because of the direct influence the biochemical and biophysical traits of leaves have on the spectral signatures (Kumar et al. 2001 ). Since leaf spectral behaviour results from an energy-matter interaction between the radiating light and the physicochemical structures in the leaves, it is expected that the biochemical content and the anatomical structures of plants have a direct effect on the way electromagnetic energy interacts with the structures present in these leaves (Asner and Martin 2009 ; Cavender-Bares et al. 2020 ). The moderate positive correlation relationship between spectral and functional traits supported by the Mantel test, indicates that functionally similar species tend to exhibit greater spectral similarity (Asner et al. 2017 ). These findings reinforce the potential for remotely and non-destructively monitoring functional diversity over large spatial scales, offering promising applications for biodiversity conservation and ecosystem management (Jetz et al. 2016 ; Schneider et al. 2017 ). The traits selected in this study are well-known predictors of ecosystem function, serving as proxies for key ecological processes (Homolová et al. 2013 ). The uniqueness of each species’ biochemical fingerprint (Support information – SI 5) suggests that functional diversity could be estimated remotely, even without prior knowledge of functional groups in the field. Although there is a natural variation in leaf biochemistry within the same species (Asner and Martin 2011 ), the complexity of spectral differentiation increases as more traits are considered in the analysis (Asner and Martin 2009 ). The ability to quantify functional traits using spectral data opens new possibilities for large-scale monitoring of functional diversity (Schweiger et al. 2018 ). Given that functional biodiversity is linked to other dimensions of biodiversity, this approach could also be leveraged for remote sensing-based biodiversity assessments at broader spatial scales. Taxonomic and spectral diversity relationship Our study suggests that spectral reflectance carries strong taxonomic signals across multiple hierarchical levels, with closely related taxa having similar spectral characteristics. Since these two components account for almost all the observed variation (~ 97%), analysing them provides robust insights into the spectral behaviour of the assessed species. The not inclusion of order in the PC2 model due to low variation suggests that variability at this level has already been captured by other taxonomic groups. This supports the idea that spectral variation is hierarchically structured across taxonomic levels, with some levels (e.g., class) capturing more distinct variation. Plant traits are shaped by a combination of factors, such as genetic material, environmental conditions, and an individual’s life history (Ustin and Gamon 2010 ). Therefore, plants from the same taxon have comparable genetic material and share traits that will shape the spectral reflectance, which in turn influences their spectral reflectance. These similarities likely account for the substantial proportion of spectral variance explained by taxonomic classification, with related species exhibiting spectral characteristics shaped by their evolutionary history and environmental adaptation. As shown in this work, spectral reflectance tends to be conserved at higher taxonomic groups such as genus, family and to a larger extent order and class (Fig. 4 a). Although the consistency of the spectral reflectance is observed to be more accentuated in higher biological organizations, the variation of the spectral reflectance observed within the same species has been around five times smaller than observed between species. This suggests a degree of intra-specific plasticity (Asner and Martin 2011 ) while maintaining sufficient stability to distinguish species and capture broader taxonomic relationships. These findings reinforce the potential of spectral data for species identification and biodiversity monitoring in temperate estuarine ecosystems (Wang and Gamon 2019 ). A substantial portion of spectral variance remains unexplained by taxonomy alone (< 20%), as indicated by the residuals in our models. This suggests that other factors contribute to spectral reflectance variability, including measurement inconsistencies at the leaf level (Neuwirthová et al. 2017 ), intraspecific variations (Asner and Martin 2011 ) and evolutionary or functional convergence (Ustin and Gamon 2010 ). Environmental factors also play a role in shaping spectral reflectance, as shown by the site variable explaining a considerable part of the variability. This suggests that local environmental conditions, such as soil composition, salinity, and nutrient availability, also contribute to spectral differences among individuals of the same species. Further investigation into these factors could provide a more comprehensive understanding of the drivers of spectral diversity in estuarine plant communities. Predictive potential among types of biodiversity Functional traits retrieval by spectral data This study successfully developed models capable of retrieving functional traits from spectral data with sufficient precision for plant species in Atlantic temperate estuaries. The high accuracy of models for water content, phosphorus, and LMA (R²=0.92, 0.81, and 0.76, respectively) aligns with previous studies showing that these traits are strongly linked to specific spectral regions, such as the NIR and SWIR for water content (Asner and Martin 2008 ; Homolová et al. 2013 ). The symmetric distribution of residuals around zero for both calibration and validation datasets further confirm the absence of systematic bias in the predictions, supporting the reliability of the models. Among the analysed traits, the strong performance of the organic carbon model (R²=0.87) despite its small sample size (N = 41) and low number of components (nº Comp = 4) highlights the robustness of spectral data for predicting this trait (Schweiger et al. 2018 ). Although total organic carbon has fewer observations to train and validate the model, the dataset used to calibrate the predictive model included a great part of the possible variation observed in this plant trait (Burnett et al. 2021 ). The predictions for phosphorus and LMA followed the same pattern, also performing well when compared to existing studies, demonstrating high R 2 and relatively low %RMSEP (Homolová et al. 2013 ; Wang et al. 2020 ). The anatomical trait LMA, often expressed as its inverse, specific leaf area (SLA), influences spectral data differently from biochemical traits by physically interacting with electromagnetic radiation. LMA is a reliable indicator of leaf thickness and density and is a key trait widely used to classify plant functional groups (Poorter et al. 2009 ). Foliar nitrogen content is widely recognised as a trait that can be retrieved with high accuracy using spectral data. Our nitrogen predictions fell within the expected range but did not exceed the performance of existing models (Homolová et al. 2013 ). Since there is no consensus on which part of the spectrum should be used to estimate nitrogen content from spectral data, we used the full spectrum (400-2400nm), identifying wavelengths in the visible range, particularly in the red edge, as the most influential in modelling nitrogen content (Supplementary material – SI 4). The possibility to predict leaf water content has broad applications, such as remotely assessing hydric stress during drought events (Martin et al. 2018 ). This is due to the strong predictive potential of water content in leaves, which was also evident in our study. Compared to previous research, our model for leaf water content demonstrated high precision (Asner and Martin 2008 ; Wang et al. 2020 ), achieving the best performance among the traits analysed (R² = 0.92). While some studies highlight the importance of the red edge in estimating water content (Filella and Penuelas 1994 ), our results indicate that, beyond the red edge, certain wavelengths in the NIR and SWIR regions play a critical role in the PLSR models for water content (Supplementary Material – SI 4). This is consistent with expectations, as water has a strong influence on specific regions of the NIR spectrum. In contrast, pigment content showed lower predictive performance than reported in other studies. Two factors may have contributed to this outcome. First, we used a colorimetric method to estimate pigment concentration, whereas other studies used more precise techniques such as high-performance liquid chromatography (HPLC). Second, while most studies report pigment content per unit leaf area (Schweiger et al. 2018 ), our values were expressed on mass per mass (µg/g), which may have caused the reduction of the performance of the model. Due to the complex interactions between pigments and their overlapping spectral signatures, photosynthetic pigments are relatively complicated to estimate, especially at canopy level (Ustin et al. 2009 ). The ability to accurately predict key functional traits from spectral data has significant implications for ecological monitoring and ecosystem management. These traits have been successfully retrieved via remote sensing across diverse biomes (Jetz et al. 2016 ), and they are essential for understanding plant physiological processes, nutrient cycling, and ecosystem functioning (Wright et al. 2004 ). However, caution is needed when applying leaf-level models to canopy spectral data, as factors such as multiple scattering and other optical effects can influence prediction accuracy (Al Makdessi et al. 2019 ). This study contributes to this field by providing openly available coefficients for trait quantification in the region of North Spain. Phylogenetic classification by spectral data The feature selection method successfully identified the wavebands where species exhibit the greatest differences from one another at the leaf level. As shown by other authors, the red edge has been widely recognized for its potential in species discrimination using hyperspectral data (Vaiphasa et al. 2007 ; Ferreira et al. 2013 ; Prospere et al. 2014 ). However, the selection of a spectral frame in the far NIR is less common, despite previous studies highlighting its relevance for identifying wetland vegetation (Schmidt and Skidmore 2003 ; Adam et al. 2010 ). Although the wavelengths selected in this study are in these two frames, it is also possible to build discrimination models with other bands (Fig. 7 ), depending on the spectral resolution of the available sensor. It is important to note that the optimal number and the position of the spectral frames for species discrimination are not universally fixed, but rather depend on the feature selection method, the specific dataset, and the biological characteristics of the vegetation (Hennessy et al. 2020 ). The entire spectrum produces more accurate species-level classification (Hennessy et al. 2020 ), although studies discriminating plant taxa based on leaf reflectance often focus on the visible spectrum (Artigas and Yang 2006 ; Gross and Heumann 2014 ). This is due to the visible region exhibiting less variation due to leaf stacking effects compared to the NIR and SWIR regions (Neuwirthová et al., 2017 ). Many leaf-level studies prioritize this spectral range because of its strong correlation with pigment content (Kumar et al. 2001 ) and the incorporation of the red edge. In contrast, the NIR and SWIR regions often show high intraspecific variation, primarily influenced by water loss during transport and the time elapsed between collection and measurement (Ferreira et al. 2013 ). In this study, the gap between field collection and laboratory measurement was relatively short (< 4 hours), suggesting that the observed variations in these spectral regions were likely due to the biological characteristics of the species, potentially reflecting natural changes in water content. Although the maximum accuracy (100%) was achieved with 45 bands, the model with 20 bands (kappa = 97%) effectively reduced data dimensionality while maintaining high accuracy. However, it is important to consider how this accuracy may change when scaled to the canopy level using the same number of bands (Clark et al. 2005 ). Additionally, the proximity of the selected SWIR range (1335–1380 nm) to the water vapor absorption region (1350–1480 nm) raises concerns about potential noise, highlighting the need for further validation with canopy-level spectral measurements. It has been pointed out that spectral variations detected among species at the leaf scale do not readily correspond to assessments conducted at the landscape scale. Individual inconsistencies are observed throughout the broad spectrum. That can be related to the intrinsic optical properties of the leaves or artefacts from the technique (Hennessy et al. 2020 ). Since manipulating narrow leaves can be challenging (Noda et al. 2013 ), the attempt to fill the field of view can slightly alter the geometry of the leaves being measured, affecting the reflected spectrum. Although we try to maintain consistency regarding the age of the leaves sampled and the part of the leaf where the reading was performed, different leaf ages (Clark et al. 2005 ) or leaf layers (Neuwirthová et al. 2017 ) could have affected the final spectrum measured of each sample. The use of a squared sampling port significantly facilitated the handling of these conspicuous leaves to completely fill the field of view of the spectroradiometer, minimizing changes in angle and shades within the probe. The vector normalize technique was also adopted to reduce individual inconsistencies. This process standardizes the magnitude of the spectral curve, allowing for better comparison across spectra and reducing effects from leaf stacking and luminosity (Neuwirthová et al., 2017 ). The high accuracy and robust performance of the discriminant model underscore its potential for applications in ecological monitoring and species identification using spectral data. The misclassification of Salicornia ramosissima highlights the need for further refinement, potentially through additional training data or feature engineering, to improve the model's ability to distinguish between spectrally similar species either at leaf and canopy level. Conclusions This study shows the close connection between taxonomic, functional, and spectral diversity in estuarine plant communities on the temperate Atlantic coast. It contributes to the foundation of research involving the spectral diversity of vascular plants in the estuaries in North Spain by calculating the coefficients to retrieve biochemical and anatomical traits from spectral data. In fact, functionally similar species exhibited comparable spectral signatures, reinforcing the idea that spectral data can serve as a proxy for functional diversity. It offers a promising indication for monitoring functional diversity at large scales using remote sensing data. Similarly, taxonomic classification played a significant role in shaping spectral diversity. However, the study revealed that intraspecific variability may contribute to spectral diversity, indicating a need for further research to account for these influences. It can be concluded that methodologies previously developed for identifying plant traits and species composition in other ecosystems are also suitable for estimating biodiversity in the temperate estuaries of the Gulf of Biscay. This success is not driven by a single spectral band, but rather by multiple spectral frames located at the edges of the NIR and SWIR ranges. These findings reinforce the importance of the red-edge shoulder in capturing remotely sensed plant traits and translating them into information about species composition and biochemical structure. Nonetheless, further efforts are needed to refine these models to accurately monitor species distribution and phenological changes across spatial and temporal scales, ultimately enabling the production of more precise species distribution maps. This study is an essential step to upscale the monitoring of vegetation types using remote sensing technologies. Also, it offers a punctual assessment of these three types of biodiversity in estuaries. It is expected that the species composition, abundance and dynamics change with time and it is imperative that we are able to track these changes cost-effectively through larger areas, particularly in ecologically sensitive and dynamic environments like coastal wetlands. Future work should focus on validating these findings at the canopy and landscape levels, as well as exploring the integration of remote sensing data with other biodiversity metrics to enhance conservation and management efforts. Dataset repository The dataset produced and used by this work is available open-access through the link https://apidies.ihcantabria.com/swagger/index.html Plant identification was conducted by MSc. Andre C. da Costa Neves under the supervision of the Coastal Ecosystems Group at IHCantabria. Specimens that could not be identified at species level were sent to specialized taxonomists and subjected to genetic identification. Reference individuals were preserved as dried exsiccata in the institutional herbarium of IHCantabria and are available upon request. The number and details of the reference specimens are provided in the Supplementary Information (SI 7). Declarations Competing Interests: 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. Ethical considerations All experimental research and field studies on plants reported in this manuscript complied with institutional, national, and international guidelines and legislation. Collection of wild plant material was authorized by the Dirección General de Montes y Biodiversidad of the regional government of Cantabria (Spain), and all procedures adhered to Spanish national regulations as well as international conventions on biodiversity and genetic resources. Only small amounts of leaves and branches were collected and brought to the laboratory for data collection, while the source plants were left alive in their natural habitats. Funding: This study forms part of the ThinkInAzul programme and was supported by the Ministry of Science and Innovation with funding from the European Union NextGeneration EU (PRTR-C17.I1), and by the autonomous community of Cantabria, and the project MarshA (Restoration beyond biodiversity: How to integrate estuarine ecosystem services into nature-based management) (TED2021-129973B-I00) funded by MICIU/AEI/ 10.13039/501100011033 and by the European Union NextGenerationEU/PRTR. This manuscript also associated with the MARBEFES project (MARine Biodiversity and Ecosystem Functioning leading to Ecosystem Services), funded by the European Union under the Horizon Europe Programme, “HORIZON-CL6-2021-BIODIV-01” Theme, Grant Agreement no. 101060937 (marbefes.eu). Author Contribution A. C. N. Conceptualization; Methodology; Fieldwork (data collection, sample processing); Formal analysis; Data curation; Visualization; Writing—original draft. C. G. Conceptualization; Funding acquisition; Resources; Supervision; Project administration; Writing—review & editing. B. O. Conceptualization; Funding acquisition; Resources; Supervision; Validation; Writing—review & editing. All authors contributed to the study conception, design, also read and approved the final manuscript. Acknowledgement The work was developed under the Complementary Plan for R+D+i in Marine Sciences (PCM) and within the workframe of the DIES and MarshA projects. The authors thank the Hydrobiology Laboratory at IHCantabria for support with laboratorial analyses, and the Centro de Investigación y Formación Agrarias (CIFA) of the Government of Cantabria for assistance with biochemical trait measurements. Data Availability The dataset produced and used by this work is available open-access through the link [https://apidies.ihcantabria.com/swagger/index.html](https:/apidies.ihcantabria.com/swagger/index.html) References Adam E, Mutanga O, Rugege D (2010) Multispectral and hyperspectral remote sensing for identification and mapping of wetland vegetation: a review. Wetl Ecol Manag 18:281–296. https://doi.org/10.1007/s11273-009-9169-z Al Makdessi N, Ecarnot M, Roumet P, Rabatel G (2019) A spectral correction method for multi-scattering effects in close range hyperspectral imagery of vegetation scenes: application to nitrogen content assessment in wheat. Precis Agric 20:237–259. https://doi.org/10.1007/s11119-018-09631-9 Artigas FJ, Yang J (2006) Spectral discrimination of marsh vegetation types in the New Jersey Meadowlands, USA. Wetlands 26:271–277. https://doi.org/10.1672/0277-5212(2006)26[271:SDOMVT]2.0.CO;2 Asner GP, Martin RE (2008) Spectral and chemical analysis of tropical forests: scaling from leaf to canopy levels. Remote Sens Environ 112:3958–3970. https://doi.org/10.1016/j.rse.2008.07.003 Asner GP, Martin RE (2009) Airborne spectranomics: mapping canopy chemical and taxonomic diversity in tropical forests. Front Ecol Environ 7:269–276. https://doi.org/10.1890/070152 Asner GP, Martin RE (2011) Canopy phylogenetic, chemical and spectral assembly in a lowland Amazonian forest. New Phytol 189:999–1012. https://doi.org/10.1111/j.1469-8137.2010.03549.x Asner GP, Martin RE (2016) Spectranomics: emerging science and conservation opportunities at the interface of biodiversity and remote sensing. Glob Ecol Conserv 8:212–219. https://doi.org/10.1016/j.gecco.2016.09.010 Asner GP, Martin RE, Anderson CB, Knapp DE (2015) Quantifying forest canopy traits: imaging spectroscopy versus field survey. Remote Sens Environ 158:15–27. https://doi.org/10.1016/j.rse.2014.11.011 Asner GP, Martin RE, Knapp DE et al (2011) Spectroscopy of canopy chemicals in humid tropical forests. Remote Sens Environ 115:3587–3598. https://doi.org/10.1016/j.rse.2011.08.020 Asner GP, Martin RE, Knapp DE et al (2017) Airborne laser-guided imaging spectroscopy to map forest trait diversity and guide conservation. Science 355:385–389. https://doi.org/10.1126/science.aaj1987 Belluco E, Camuffo M, Ferrari S et al (2006) Mapping salt-marsh vegetation by multispectral and hyperspectral remote sensing. Remote Sens Environ 105:54–67. https://doi.org/10.1016/j.rse.2006.06.006 Bohrer G, Cavender-Bares J, Chaplin-Kramer R et al (2022) NASA Biological Diversity and Ecological Forecasting: current state of knowledge and considerations for the next decade. NASA Report. https://doi.org/xxx Bradstreet RB (1954) Kjeldahl method for organic nitrogen. Anal Chem 26:185–187. https://doi.org/10.1021/ac60085a024 Burnett AC, Anderson J, Davidson KJ et al (2021) A best-practice guide to predicting plant traits from leaf-level hyperspectral data using partial least squares regression. J Exp Bot 72:6175–6189. https://doi.org/10.1093/jxb/erab295 Caturegli L, Matteoli S, Gaetani M et al (2020) Effects of water stress on spectral reflectance of bermudagrass. Sci Rep 10:12055. https://doi.org/10.1038/s41598-020-72006-6 Cavender-Bares J, Gamon JA, Townsend PA (2020) Remote sensing of plant biodiversity. Springer, Cham. https://doi.org/10.1007/978-3-030-33157-3 Clark ML, Roberts DA, Clark DB (2005) Hyperspectral discrimination of tropical rain forest tree species at leaf to crown scales. Remote Sens Environ 96:375–398. https://doi.org/10.1016/j.rse.2005.03.009 Consejería de Desarrollo Rural, Ganadería, Pesca y Biodiversidad (n.d.) Anejo VII. Documento técnico de los espacios acuáticos litorales Natura 2000 en Cantabria. C. LIC Dunas del Puntal y Estuario del Miera. https://rednatura2000cantabria.ihcantabria.com/ESPACIOS/5_LITORALES/PLANES%20DE%20CADA%20LIC/Anejo%20VII_PLAN%20GESTION%20LIC%20PUNTAL.pdf Curcio AC, Barbero L, Peralta G (2023) UAV-Hyperspectral Imaging to estimate species distribution in salt marshes: a case study in the Cadiz Bay (SW Spain). Remote Sens 15:1419. https://doi.org/10.3390/rs15051419 Doughty CE, Santos-Andrade PE, Goldsmith GR et al (2017) Can leaf spectroscopy predict leaf and forest traits along a Peruvian tropical forest elevation gradient? J Geophys Res Biogeosci 122:2952–2965. https://doi.org/10.1002/2017JG003883 Feret JB, Asner GP (2013) Tree species discrimination in tropical forests using airborne imaging spectroscopy. IEEE Trans Geosci Remote Sens 51:73–84. https://doi.org/10.1109/TGRS.2012.2199323 Ferreira MP, Grondona AEB, Rolim SBA, Shimabukuro YE (2013) Analyzing the spectral variability of tropical tree species using hyperspectral feature selection and leaf optical modeling. J Appl Remote Sens 7:073502. https://doi.org/10.1117/1.JRS.7.073502 Filella I, Penuelas J (1994) The red edge position and shape as indicators of plant chlorophyll content, biomass and hydric status. Int J Remote Sens 15:1459–1470. https://doi.org/10.1080/01431169408954177 Gross JW, Heumann BW (2014) Can flowers provide better spectral discrimination between herbaceous wetland species than leaves? Remote Sens Lett 5:892–901. https://doi.org/10.1080/2150704X.2014.973077 Guinea E (1953) Ammophiletea, Crithmo-Staticetea, Salicornieteaque santanderiensis. An Jard Bot Madr 11:545–568. Hennessy A, Clarke K, Lewis M (2020) Hyperspectral classification of plants: a review of waveband selection generalisability. Remote Sens 12:113. https://doi.org/10.3390/rs12010113 Hladik C, Alber M (2014) Classification of salt marsh vegetation using edaphic and remote sensing-derived variables. Estuar Coast Shelf Sci 141:47–57. https://doi.org/10.1016/j.ecss.2014.01.011 Homolová L, Malenovský Z, Clevers JGPW et al (2013) Review of optical-based remote sensing for plant trait mapping. Ecol Complex 15:1–16. https://doi.org/10.1016/j.ecocom.2013.07.003 IH Cantabria (2017) CONVIVE-LIFE: Deliverable Action A1. Cartografía de hábitats y formaciones vegetales/caracterización de las comunidades biológicas del estuario/cartografía de especies invasoras. http://convivelife.es/wp-content/uploads/2017/11/Annex_IIA1_Final1.pdf Jetz W, Cavender-Bares J, Pavlick R et al (2016) Monitoring plant functional diversity from space. Nat Plants 2:16024. https://doi.org/10.1038/nplants.2016.24 Kaennel M (1998) Biodiversity: a diversity in definition. In: Bachmann P, Köhl M, Päivinen R (eds) Assessment of biodiversity for improved forest planning. Springer, Dordrecht, pp 71–81. Klemas V (2013) Remote sensing of emergent and submerged wetlands: an overview. Int J Remote Sens 34:6286–6320. https://doi.org/10.1080/01431161.2013.800656 Kumar L, Schmidt KS, Dury S, Skidmore AK (2001) Review of hyperspectral remote sensing and vegetation science. Imaging Spectrom Basic Princ Prospect Appl:111–155. Laliberté E, Schweiger AK, Legendre P (2020) Partitioning plant spectral diversity into alpha and beta components. Ecol Lett 23:370–380. https://doi.org/10.1111/ele.13429 Lichtenthaler HK (1987) Chlorophyll fluorescence signatures of leaves during the autumnal chlorophyll breakdown. J Plant Physiol 131:101–110. https://doi.org/10.1016/S0176-1617(87)80271-7 Lichtenthaler HK, Buschmann C (2001) Extraction of photosynthetic tissues: chlorophylls and carotenoids. Curr Protoc Food Anal Chem 1:F4.2. https://doi.org/10.1002/0471142913.faf0402s01 Martin RE, Asner GP, Francis E et al (2018) Remote measurement of canopy water content in giant sequoias (Sequoiadendron giganteum) during drought. For Ecol Manag 419:279–290. https://doi.org/10.1016/j.foreco.2017.12.002 Meireles JE, Cavender-Bares J, Townsend PA et al (2017) SPECTROLAB – a R package for processing and analysing spectral data. Methods Ecol Evol 8:1215–1221. https://doi.org/10.1111/2041-210X.12765 Neuwirthová E, Lhotáková Z, Albrechtová J (2017) The effect of leaf stacking on leaf reflectance and vegetation indices measured by contact probe during the season. Sensors 17:1202. https://doi.org/10.3390/s17061202 Noda HM, Motohka T, Murakami K et al (2013) Accurate measurement of optical properties of narrow leaves and conifer needles with a typical integrating sphere and spectroradiometer. Plant Cell Environ 36:1903–1909. https://doi.org/10.1111/pce.12100 Ordway EM, Asner GP, Burslem DFRP et al (2022) Mapping tropical forest functional variation at satellite remote sensing resolutions depends on key traits. Commun Earth Environ 3:64. https://doi.org/10.1038/s43247-022-00564-w Ozesmi SL, Bauer ME (2002) Satellite remote sensing of wetlands. Wetl Ecol Manag 10:381–402. https://doi.org/10.1023/A:1020908432489 Poorter H, Niinemets Ü, Poorter L et al (2009) Causes and consequences of variation in leaf mass per area (LMA): a meta-analysis. New Phytol 182:565–588. https://doi.org/10.1111/j.1469-8137.2009.02830.x Prospere K, McLaren K, Wilson B (2014) Plant species discrimination in a tropical wetland using in situ hyperspectral data. Remote Sens 6:8494–8523. https://doi.org/10.3390/rs6098494 Pu R (2009) Broadleaf species recognition with in situ hyperspectral data. Int J Remote Sens 30:2759–2779. https://doi.org/10.1080/01431160802555820 Red Natura 2000 Cantabria (n.d.) LICs Litorales: Marismas de Santoña, Victoria y Joyel. https://rednatura2000cantabria.ihcantabria.com/lics/litorales/santona2/ Rocchini D, Balkenhol N, Carter GA et al (2010) Remotely sensed spectral heterogeneity as a proxy of species diversity: recent advances and open challenges. Ecol Inform 5:318–329. https://doi.org/10.1016/j.ecoinf.2010.06.001 Rocchini D, Boyd DS, Féret JB et al (2016) Satellite remote sensing to monitor species diversity: potential and pitfalls. Remote Sens Ecol Conserv 2:25–36. https://doi.org/10.1002/rse2.13 Rocchini D, Chiarucci A, Loiselle SA (2004) Testing the spectral variation hypothesis by using satellite multispectral images. Acta Oecol 26:117–120. https://doi.org/10.1016/j.actao.2004.03.008 Schmidt KS, Skidmore AK (2003) Spectral discrimination of vegetation types in a coastal wetland. Remote Sens Environ 85:92–108. https://doi.org/10.1016/S0034-4257(02)00196-7 Schneider CA, Rasband WS, Eliceiri KW (2012) NIH Image to ImageJ: 25 years of image analysis. Nat Methods 9:671–675. https://doi.org/10.1038/nmeth.2089 Schneider FD, Morsdorf F, Schmid B et al (2017) Mapping functional diversity from remotely sensed morphological and physiological forest traits. Nat Commun 8:16015. https://doi.org/10.1038/s41467-017-01530-3 Schweiger AK (2020) Spectral field campaigns: planning and data collection. In: Cavender-Bares J, Gamon JA, Townsend PA (eds) Remote sensing of plant biodiversity. Springer, Cham, pp 385–423. https://doi.org/10.1007/978-3-030-33157-3_15 Schweiger AK, Cavender-Bares J, Townsend PA et al (2018) Plant spectral diversity integrates functional and phylogenetic components of biodiversity and predicts ecosystem function. Nat Ecol Evol 2:976–982. https://doi.org/10.1038/s41559-018-0551-1 Schweiger AK, Laliberté E (2022) Plant beta-diversity across biomes captured by imaging spectroscopy. Nat Commun 13:2767. https://doi.org/10.1038/s41467-022-30369-6 Schweiger AK, Schütz M, Risch AC et al (2017) How to predict plant functional types using imaging spectroscopy: linking vegetation community traits, plant functional types and spectral response. Methods Ecol Evol 8:86–95. https://doi.org/10.1111/2041-210X.12642 Ustin SL, Gamon JA (2010) Remote sensing of plant functional types. New Phytol 186:795–816. https://doi.org/10.1111/j.1469-8137.2010.03284.x Ustin SL, Gitelson AA, Jacquemoud S et al (2009) Retrieval of foliar information about plant pigment systems from high resolution spectroscopy. Remote Sens Environ 113:S67–S77. https://doi.org/10.1016/j.rse.2008.10.019 Vaiphasa C, Skidmore AK, de Boer WF, Vaiphasa T (2007) A hyperspectral band selector for plant species discrimination. ISPRS J Photogramm Remote Sens 62:225–235. https://doi.org/10.1016/j.isprsjprs.2007.05.006 Valle M, Palà V, Lafon V et al (2015) Mapping estuarine habitats using airborne hyperspectral imagery, with special focus on seagrass meadows. Estuar Coast Shelf Sci 164:433–442. https://doi.org/10.1016/j.ecss.2015.07.034 Violle C, Navas ML, Vile D et al (2007) Let the concept of trait be functional! Oikos 116:882–892. https://doi.org/10.1111/j.2007.0030-1299.15559.x Wang R, Gamon JA (2019) Remote sensing of terrestrial plant biodiversity. Remote Sens Environ 231:111218. https://doi.org/10.1016/j.rse.2019.111218 Wang Z, Chlus A, Geygan R et al (2020) Foliar functional traits from imaging spectroscopy across biomes in eastern North America. New Phytol 228:494–511. https://doi.org/10.1111/nph.16711 Wright IJ, Reich PB, Westoby M et al (2004) The worldwide leaf economics spectrum. Nature 428:821–827. https://doi.org/10.1038/nature02403 Yu B, Ostland M, Gong P, Pu R (1999) Penalized discriminant analysis of in situ hyperspectral data for conifer species recognition. IEEE Trans Geosci Remote Sens 37:2569–2577. https://doi.org/10.1109/36.789651 Additional Declarations No competing interests reported. Supplementary Files CostaNevesetalSupplementaryInformation.docx Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 17 Oct, 2025 Reviewers agreed at journal 06 Oct, 2025 Reviewers invited by journal 01 Oct, 2025 Editor assigned by journal 29 Sep, 2025 Editor invited by journal 24 Sep, 2025 Submission checks completed at journal 23 Sep, 2025 First submitted to journal 23 Sep, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7591763","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":528090075,"identity":"3e2c770d-9d7f-4e8a-bb6b-922d72562f13","order_by":0,"name":"Andre C. Costa-Neves","email":"","orcid":"","institution":"Instituto de Hidráulica Ambiental de la Universidad de Cantabria","correspondingAuthor":false,"prefix":"","firstName":"Andre","middleName":"C.","lastName":"Costa-Neves","suffix":""},{"id":528090076,"identity":"d81b86b2-739e-4e50-b55a-e62f2f029209","order_by":1,"name":"Cristina Galván","email":"data:image/png;base64,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","orcid":"","institution":"Instituto de Hidráulica Ambiental de la Universidad de Cantabria","correspondingAuthor":true,"prefix":"","firstName":"Cristina","middleName":"","lastName":"Galván","suffix":""},{"id":528090077,"identity":"f2414136-0285-4ae8-8f90-79db3875ee3f","order_by":2,"name":"Bárbara Ondiviela","email":"","orcid":"","institution":"Instituto de Hidráulica Ambiental de la Universidad de Cantabria","correspondingAuthor":false,"prefix":"","firstName":"Bárbara","middleName":"","lastName":"Ondiviela","suffix":""}],"badges":[],"createdAt":"2025-09-11 12:08:40","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7591763/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7591763/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":93516950,"identity":"715a5041-74b0-403a-9d4a-89d7a2edf4de","added_by":"auto","created_at":"2025-10-14 16:51:28","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1796123,"visible":true,"origin":"","legend":"","description":"","filename":"CostaNevesetal2025BiodivConservOriginalResearchrevision.docx","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/418f77a1a131efc4ac39c7cc.docx"},{"id":93517525,"identity":"b7bb4f0c-b6dc-4f45-a504-a76eedf25183","added_by":"auto","created_at":"2025-10-14 16:59:28","extension":"xlsx","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":11029,"visible":true,"origin":"","legend":"","description":"","filename":"Table1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/f6984eff504b927e3ebb0737.xlsx"},{"id":93517902,"identity":"a77ed2c0-a95f-4c34-a4d0-1d0de186f0db","added_by":"auto","created_at":"2025-10-14 17:07:28","extension":"xlsx","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":11808,"visible":true,"origin":"","legend":"","description":"","filename":"Table2.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/4fcbc7086053f50636f88f92.xlsx"},{"id":93516946,"identity":"a6202f9a-2f01-4938-9234-39307438bc72","added_by":"auto","created_at":"2025-10-14 16:51:27","extension":"json","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6054,"visible":true,"origin":"","legend":"","description":"","filename":"19ccdf5c113a47529105e4d660655fdd.json","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/1a4fa012e56d6d5909d081d7.json"},{"id":93516951,"identity":"4b6429db-7358-4516-b501-468d9bbf93d9","added_by":"auto","created_at":"2025-10-14 16:51:28","extension":"docx","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":364731,"visible":true,"origin":"","legend":"","description":"","filename":"CostaNevesetalSupplementaryInformation.docx","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/0595240ec3e900dd4a00d200.docx"},{"id":93516954,"identity":"f318ac3b-cffd-46e1-b220-ce7e77d97336","added_by":"auto","created_at":"2025-10-14 16:51:28","extension":"xml","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":174201,"visible":true,"origin":"","legend":"","description":"","filename":"19ccdf5c113a47529105e4d660655fdd1enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/a26961dd78045337e8e76ed7.xml"},{"id":93516976,"identity":"392e5d7c-15a4-41ca-a382-454ced0a0a7e","added_by":"auto","created_at":"2025-10-14 16:51:28","extension":"eps","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":57324101,"visible":true,"origin":"","legend":"","description":"","filename":"Figure1.eps","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/1d2bed279b71cf607bb0c2fc.eps"},{"id":93516953,"identity":"b54c2051-8d11-4d3f-8ccb-03e3989ac737","added_by":"auto","created_at":"2025-10-14 16:51:28","extension":"eps","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":655955,"visible":true,"origin":"","legend":"","description":"","filename":"Figure2.eps","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/7a8fb5ed4af9ce635f135b87.eps"},{"id":93597742,"identity":"8f4d7ab1-cdbc-4535-bbfe-c817aac7d8f5","added_by":"auto","created_at":"2025-10-15 14:21:56","extension":"eps","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":393707,"visible":true,"origin":"","legend":"","description":"","filename":"Figure3.eps","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/05b4a1d053d444a505195596.eps"},{"id":93517527,"identity":"77598228-8587-4b93-aae3-6c9b7ea559c1","added_by":"auto","created_at":"2025-10-14 16:59:28","extension":"eps","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":140246,"visible":true,"origin":"","legend":"","description":"","filename":"Figure4.eps","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/6b7784aa1abb4878c82bb2ba.eps"},{"id":93517534,"identity":"71cbe1d3-1bfe-4bae-a54d-824df2088cbe","added_by":"auto","created_at":"2025-10-14 16:59:28","extension":"eps","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":709113,"visible":true,"origin":"","legend":"","description":"","filename":"Figure5.eps","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/ed33c8678f9bf6c1ed174ebc.eps"},{"id":93518957,"identity":"8ed6801f-684f-4ed4-8fa2-8ab33540766f","added_by":"auto","created_at":"2025-10-14 17:15:28","extension":"eps","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":106693,"visible":true,"origin":"","legend":"","description":"","filename":"Figure6.eps","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/a5f2fe271c9a31af14e14eea.eps"},{"id":93516961,"identity":"0ae89514-58e4-4c79-a02a-aa75abda0452","added_by":"auto","created_at":"2025-10-14 16:51:28","extension":"eps","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":176421,"visible":true,"origin":"","legend":"","description":"","filename":"Figure7.eps","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/c046dfe556d7ab8de90277e4.eps"},{"id":93517908,"identity":"2efd57a6-4ea7-40e9-a8bb-30912d9e3276","added_by":"auto","created_at":"2025-10-14 17:07:28","extension":"png","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":732403,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/302eebd0d21d50a0aa8630f1.png"},{"id":93516966,"identity":"ff7576ac-1788-4840-9480-b1625d661d84","added_by":"auto","created_at":"2025-10-14 16:51:28","extension":"png","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":253865,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/55593b0f4346983ad37241ab.png"},{"id":93517903,"identity":"847499ac-7503-4686-a010-69c84d9e2305","added_by":"auto","created_at":"2025-10-14 17:07:28","extension":"png","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":141015,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/38fffb76e77f52060af23b3e.png"},{"id":93516974,"identity":"0ba7aa67-d217-4c75-9ac2-82d019c98614","added_by":"auto","created_at":"2025-10-14 16:51:28","extension":"jpeg","order_by":23,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":333905,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/353fd10c3d097c0e0d44587b.jpeg"},{"id":93519438,"identity":"c3701b10-b4e2-40ca-8bde-46debabb3476","added_by":"auto","created_at":"2025-10-14 17:23:28","extension":"png","order_by":24,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":246713,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/a7cd8623ea14922ea8e3222d.png"},{"id":93597487,"identity":"7a8d6d0d-2c0a-499e-ad9d-a1e6d44a250e","added_by":"auto","created_at":"2025-10-15 14:14:57","extension":"png","order_by":25,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":63740,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/0f72add0050cf493b6bfd738.png"},{"id":93516969,"identity":"ef34a781-d15d-4d1c-ac8b-fc4e7f4cfd1e","added_by":"auto","created_at":"2025-10-14 16:51:28","extension":"png","order_by":26,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":144519,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/ea175413596d95b10681e2c8.png"},{"id":93518953,"identity":"b80bf688-ca1d-4379-96d2-2d81037d613f","added_by":"auto","created_at":"2025-10-14 17:15:28","extension":"png","order_by":27,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":115960,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/e9ca5b98dd8deaca874cf44a.png"},{"id":93516967,"identity":"dc48e8a1-b406-40a6-b0d0-08cc50a118a6","added_by":"auto","created_at":"2025-10-14 16:51:28","extension":"png","order_by":28,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":60106,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/6081c32526ecf67b5da4318f.png"},{"id":93516973,"identity":"00cf1beb-dcb1-4b77-b893-4e366f179bab","added_by":"auto","created_at":"2025-10-14 16:51:28","extension":"png","order_by":29,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":40458,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/9f41193ddc2ce7a5bfd71c25.png"},{"id":93516971,"identity":"0cb8983b-34d7-46aa-a95e-2b8d648acc11","added_by":"auto","created_at":"2025-10-14 16:51:28","extension":"png","order_by":30,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":54379,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/538ceea63296b204a8529e0d.png"},{"id":93516965,"identity":"a331942e-469d-4624-bc67-7881af7715f7","added_by":"auto","created_at":"2025-10-14 16:51:28","extension":"png","order_by":31,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":82341,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/896ebbf73fb6dd28daa8b2c9.png"},{"id":93516970,"identity":"e70b1b22-21a3-437c-91ed-e5f2fdce6d85","added_by":"auto","created_at":"2025-10-14 16:51:28","extension":"png","order_by":32,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":21942,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/578bd3bf4a3f38780bddd6bb.png"},{"id":93516968,"identity":"660cf60e-5a19-4fa4-8188-d9d493ca8e6d","added_by":"auto","created_at":"2025-10-14 16:51:28","extension":"png","order_by":33,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":45156,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/fae6fa919bf4bce98ff759ab.png"},{"id":93517537,"identity":"5aadba76-5234-4943-a23b-afb31e731fb1","added_by":"auto","created_at":"2025-10-14 16:59:28","extension":"xml","order_by":34,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":172200,"visible":true,"origin":"","legend":"","description":"","filename":"19ccdf5c113a47529105e4d660655fdd1structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/507062d9258e8cc808bcf01b.xml"},{"id":93516972,"identity":"9777da37-2085-441d-bb2f-f8543e0c4b49","added_by":"auto","created_at":"2025-10-14 16:51:28","extension":"html","order_by":35,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":187352,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/18d43db0464153adecc99d0b.html"},{"id":93517521,"identity":"43554092-06e3-4c5a-a9a1-ce7b2b679c09","added_by":"auto","created_at":"2025-10-14 16:59:27","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":732403,"visible":true,"origin":"","legend":"\u003cp\u003eStudy area in North Spain (top) showing the location of the four estuaries (Santander Bay, Oyambre estuary, Joyel marshes, and Santoña marshes). Red circles indicate the sampling sites within each estuary where plant material was collected\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/6502aa6e4ecd5d777cc7e5bb.png"},{"id":93516939,"identity":"7de6fd35-48e3-4faa-a40b-264d1578b992","added_by":"auto","created_at":"2025-10-14 16:51:27","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":253865,"visible":true,"origin":"","legend":"\u003cp\u003eExperimental design to build the databases (yellow box), from site selection to data collection. In total, each study site had 3 samples per species and 9 leaves measured\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/8e35756c97b739986299f4fd.png"},{"id":93516941,"identity":"39188a96-216f-443e-b4a8-429a370bf4a4","added_by":"auto","created_at":"2025-10-14 16:51:27","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":141015,"visible":true,"origin":"","legend":"\u003cp\u003eRelationship between spectral diversity and functional diversity. Each point in the graph represents a value of a pairwise comparison between all pairs of species. The colourful lines show the regression of the relationship of each species, and the black line shows the overall linear regression of all species with a 95% confidence level\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/28c9bd6444916ecbc356c62f.png"},{"id":93517522,"identity":"20cc5f9e-f09d-44ba-b5b2-5dce72068b2e","added_by":"auto","created_at":"2025-10-14 16:59:27","extension":"jpeg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":333905,"visible":true,"origin":"","legend":"\u003cp\u003e(a) Stacked bar graph showing the percentage of the variance of spectral response (here compiled as PC1 and PC2) explained by taxonomic groups and site. The percentage explained by each of the PCs is displayed above. (b) PCA of spectral diversity, illustrated by taxonomic order. Points represent individual plant readings, with colours indicating their respective order\u003c/p\u003e","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/fbf6df72d00dd9a630ff50e7.jpeg"},{"id":93517523,"identity":"cf670541-2751-4967-8fcb-f282859e0f34","added_by":"auto","created_at":"2025-10-14 16:59:27","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":246713,"visible":true,"origin":"","legend":"\u003cp\u003ePLSR validation scatterplot for the prediction of all traits. These are leaf phosphorus content (Phosphorus), leaf nitrogen content (Nitrogen), leaf mass per area (LMA), water content (Water), organic carbon (Carbon), carotenoids leaf content (Carotenoids), chlorophyll a (Chl-a), and chlorophyll b (Chl-b). These plots summarized 20% of the data used for validation per species, comparing the predicted values (x-axis) and observed values (y-axis) with error bars (95%) to assess the robustness of the PLSR\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/b414f058ec4644e6dc7f6ce4.png"},{"id":93516943,"identity":"97f0358c-2c31-4459-827a-09679a58bdfc","added_by":"auto","created_at":"2025-10-14 16:51:27","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":63740,"visible":true,"origin":"","legend":"\u003cp\u003eSelected bands with higher discriminative power. The graph shows the smoothed mean reflectance of all species (black line), with the respective standard deviation (dashed grey lines). The y-axis displays the relative normalised reflectance. The red and blue lines indicate the selected frames from the feature selection method to achieve 97% and 100% accuracy in the discriminant analysis, respectively\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/2e23b07e33ad9b633fc91647.png"},{"id":93517901,"identity":"94622ca2-e4fb-4ff1-9050-97647e8621b0","added_by":"auto","created_at":"2025-10-14 17:07:27","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":144519,"visible":true,"origin":"","legend":"\u003cp\u003ePolar chart showing the most informative wavelengths selected through feature selection for discriminating among species. Each bar represents the number of significant pairwise differences (p \u0026lt; 0.01) between species within a 20 nm wavelength bin\u003c/p\u003e","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/efeb5abbd66f19b147a2040e.png"},{"id":93598538,"identity":"1f6878d0-06e4-4afa-b6cb-7df7ba71190b","added_by":"auto","created_at":"2025-10-15 14:28:21","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2494325,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/95e626b8-5171-4738-91bb-593556abfff0.pdf"},{"id":93516947,"identity":"b09b012c-c212-47f3-8fb5-1d9559bf38e2","added_by":"auto","created_at":"2025-10-14 16:51:28","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":364731,"visible":true,"origin":"","legend":"","description":"","filename":"CostaNevesetalSupplementaryInformation.docx","url":"https://assets-eu.researchsquare.com/files/rs-7591763/v1/547958787496d4aa74377903.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Assessing sources of variation on leaves reflectance spectra in coastal saltmarshes and seagrasses","fulltext":[{"header":"Highlights","content":"\u003cp\u003e\u0026bull; Key spectral regions (720\u0026ndash;770 nm, 1330\u0026ndash;1380 nm) enable 98% species discrimination.\u003c/p\u003e\u003cp\u003e\u0026bull; Partial least squares regression models accurately retrieve functional traits (R\u0026sup2;: 0.42\u0026ndash;0.92) from leaf spectra.\u003c/p\u003e\u003cp\u003e\u0026bull; Spectral diversity is mostly explained by taxonomic and functional diversity in temperate estuarine ecosystems.\u003c/p\u003e"},{"header":"Introduction","content":"\u003cp\u003eIn light of the current biodiversity loss crisis, there is an urgent need for effective methods to assess biodiversity on a large scale across diverse ecosystems. The assessment of biodiversity in situ can be difficult to acquire and time-consuming, due to logistic issues and different methodologies among studies (Rocchini et al., 2010). A promising approach is through remote sensing of plant optical diversity, which is intrinsically related to other forms of biodiversity, such as phylogenetic and functional diversity (Jetz et al. 2016; Wang and Gamon 2019).\u003c/p\u003e\n\u003cp\u003eBiodiversity encompasses multiple dimensions, many of which are interrelated; it consists of a complex process that varies across space and time, from the genetic to the ecosystem level (Kaennel 1998; Rocchini et al. 2004). Some of the main components included in this broad concept are the species richness and the functional diversity. The functional diversity is based on the principle of “functional traits”, which are the morphological, physiological and phenological attributes that directly influence individual’s fitness (Violle et al., 2007). These traits represent key biological processes (Asner et al. 2017), providing a viable and complementary approach to traditional species diversity assessments (Asner et al. 2011; Jetz et al. 2016; Schweiger et al. 2018; Ordway et al. 2022).\u003c/p\u003e\n\u003cp\u003eThe adoption of a trait-based approach is being increasingly used in ecological studies (Violle et al. 2007). Functional traits, in a variety of ecosystems, have been successfully estimated using spectral data derived from leaf and canopy reflectance (Asner et al. 2015; Schweiger et al. 2017; Wang et al. 2020). However, these models are usually site-specific, and applying existing models to new datasets can often be unfeasible (Burnett et al. 2021), requiring the calculation of the regression coefficients for the statistical models for each target vegetation.\u003c/p\u003e\n\u003cp\u003eSimilarly, the ability to assess species richness and other diversity metrics (such as alpha and beta diversity) based on spectral heterogeneity in vegetation communities has been demonstrated across many ecosystems (Rocchini et al. 2010; Feret and Asner 2013; Schweiger and Laliberté 2022). These estimations are based on the concept of the spectral diversity, which is a biodiversity metrics estimated as a function of the electromagnetic radiation reflected from plant communities. The fundamental is that various plant traits influence light absorption and scattering, leading to detectable variations in plant optical properties (Ustin et al. 2009; Asner and Martin 2016). The use of the spectral diversity approach is particularly useful to monitor sites with difficult access and extensive areas (Schweiger et al. 2018). While no single method is universally applicable across all ecosystems, investigating spectral diversity in different environments offers a standardised, consistent, and cost-effective means of biodiversity assessment (Cavender-Bares et al. 2020).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eNovel approaches for evaluating biodiversity through spectral diversity have been tested and developed in a range of ecosystems, from grasslands (Dalmayne et al., 2013; Schweiger et al., 2018; Warren et al., 2014; Zhao et al., 2021) to wetlands (Ozesmi and Bauer 2002; Adam et al. 2010; Klemas 2013), as well as in tropical rainforests (Asner et al., 2017). Many of these studies have concluded that spectral reflectance can effectively identify plant species and functional types and can also be correlated with species richness estimations obtained through remote sensing.\u003c/p\u003e\n\u003cp\u003eTo effectively relate traditional biodiversity metrics with spectral diversity, it is essential to adopt suitable spectral scales for each habitat (Bohrer et al. 2022). Hyperspectral data enables a detailed analysis and differentiation of vegetation types that sometimes is not possible with broader-band multispectral sensors (Adam et al. 2010). Given the strong scale-dependence of biodiversity estimations (Rocchini et al. 2016; Wang and Gamon 2019), spatial resolution has proven more critical than spectral resolution for accurate classification of salt marsh vegetation (Belluco et al. 2006a). Thus, the identification of plant species based on spectral data has been demonstrated promising in coastal wetlands (Schmidt and Skidmore 2003; Belluco et al. 2006; Valle et al. 2015; Curcio et al. 2023), particularly in the non-inundated zones (Adam et al. 2010; Klemas 2013). However, it essential to acquire reliable, repeatable and standardized in situ information to calibrate and validate the biodiversity estimations across diverse ecosystems (Bohrer et al. 2022).\u003c/p\u003e\n\u003cp\u003eTo support such efforts, a detailed investigation of the relationship between biodiversity dimensions in North Atlantic temperate coastal wetlands is required. As a first step, this involves assessing species diversity and biochemical uniqueness in situ and linking these attributes to spectral diversity (Adam et al. 2010). This study is based on the hypothesis that salt marsh and seagrass species exhibit distinct optical and functional traits, resulting in species-specific spectral signatures that can be predicted using hyperspectral data.\u003c/p\u003e\n\u003cp\u003eThe main objectives of this study are (1) to explore how the taxonomic and functional diversity are related to the spectral diversity in salt marshes and seagrasses ecosystems; and (2) to evaluate the predictive potential of leaf spectral reflectance to capture functional and taxonomic diversity. For this purpose, we aim to determine whether plant functional groups exhibit similar spectral traits; to what extent taxonomic diversity explains the spectral variability of the plant community; in which frames the intraspecific spectral variation is consistently lower allowing interpretations of interspecific spectral differences; and how accurately can functional traits be inferred from leaf spectral reflectance in salt marshes.\u003c/p\u003e"},{"header":"Materials \u0026 Methods ","content":"\u003ch2\u003eStudy area\u003c/h2\u003e\n\u003cp\u003eThe study was conducted in four estuarine systems along the Gulf of Biscay: Oyambre Estuary, Santander Bay, the Joyel Marshes, and the Santo\u0026ntilde;a Marshes (Fig. 1). These estuaries were selected due to the presence of well-developed and healthy Atlantic temperate saltmarsh and seagrass habitats. They are representatives of the geomorphological and ecological diversity of estuarine vegetation found in northern Spain. These habitats are distributed along two main environmental gradients: salinity, which determines the presence of halophyte (e.g., cordgrass \u003cem\u003eSpartina\u003c/em\u003e sp., \u003cem\u003eSalicornia\u003c/em\u003e sp. and \u003cem\u003eLimonium vulgare\u003c/em\u003e) and subhalophyte (rushes \u003cem\u003eJuncus maritimus\u003c/em\u003e, common reed \u003cem\u003ePhragmites australis\u003c/em\u003e and scirpus \u003cem\u003eBolboschoenus maritimus\u003c/em\u003e) species; and emersion, which drives vegetation zonation patterns, with seagrasses (i.e., \u003cem\u003eZostera marina\u003c/em\u003e and \u003cem\u003eZostera noltii\u003c/em\u003e) colonizing the tidal flats and saltmarshes occupying the mid and high intertidal.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eA total of eight study sites were distributed across the four selected estuaries, covering the two environmental gradients to capture variation in species and populations (Fig. 1). Three study sites were located in seagrass habitats (2 in the Santander Bay and 1 in the Santo\u0026ntilde;a Marshes) and five study sites in saltmarshes (2 in the Oyambre estuary, 2 in the Joyel Marshes and 1 in the Santander Bay). All sites are located within protected conservation areas, including Natura 2000 network and national parks.\u003c/p\u003e\n\u003ch2\u003ePlant collection\u003c/h2\u003e\n\u003cp\u003eA total of 19 coastal estuarine plant species were sampled across the study area (Table 1). The field campaigns were carried out in 2023 during flowering periods from mid-July to late September. Each species was sampled in three different patches within the same study site, separate patches were selected in an attempt to maximise variability within the same population (Fig. 2). Each species was collected between one and three study sites, depending on its local occurrence and the ecological integrity of the population.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003e1\u003c/strong\u003e List of species sampled, with their respective taxonomic classification (genus, family, order, and class). Also, showing the number of sites where these species were collected (n\u0026ordm; sites).\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"632\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 16.7457%;\"\u003e\n \u003cp\u003eClass\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 16.5877%;\"\u003e\n \u003cp\u003eOrder\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.3254%;\"\u003e\n \u003cp\u003eFamily\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 16.9036%;\"\u003e\n \u003cp\u003eGenus\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 15.6398%;\"\u003e\n \u003cp\u003eSpecies\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 15.7978%;\"\u003e\n \u003cp\u003en\u0026ordm; sites\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.7457%;\"\u003e\n \u003cp\u003eLiliopsida\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5877%;\"\u003e\n \u003cp\u003ePoales\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.3254%;\"\u003e\n \u003cp\u003eCyperaceae\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003eBolboschoenus\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.6398%;\"\u003e\n \u003cp\u003e\u003cem\u003emaritimus\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 15.7978%;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.7457%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5877%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.3254%;\"\u003e\n \u003cp\u003ePoaceae\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003eSpartina\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.6398%;\"\u003e\n \u003cp\u003e\u003cem\u003ealterniflora\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 15.7978%;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.7457%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5877%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.3254%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9036%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.6398%;\"\u003e\n \u003cp\u003e\u003cem\u003eanglica\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 15.7978%;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.7457%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5877%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.3254%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9036%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.6398%;\"\u003e\n \u003cp\u003e\u003cem\u003emaritima\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 15.7978%;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.7457%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5877%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.3254%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003ePhragmites\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.6398%;\"\u003e\n \u003cp\u003e\u003cem\u003eaustralis\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 15.7978%;\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.7457%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5877%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.3254%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003ePuccinellia\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.6398%;\"\u003e\n \u003cp\u003e\u003cem\u003emaritima\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 15.7978%;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.7457%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5877%;\"\u003e\n \u003cp\u003eAlismatales\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.3254%;\"\u003e\n \u003cp\u003eZosteraceae\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003eZostera\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.6398%;\"\u003e\n \u003cp\u003e\u003cem\u003enoltii\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 15.7978%;\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.7457%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5877%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.3254%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9036%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.6398%;\"\u003e\n \u003cp\u003e\u003cem\u003emarina\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 15.7978%;\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.7457%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5877%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.3254%;\"\u003e\n \u003cp\u003eJuncaceae\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003eJuncus\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.6398%;\"\u003e\n \u003cp\u003e\u003cem\u003emaritimus\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 15.7978%;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.7457%;\"\u003e\n \u003cp\u003eMagnoliopsida\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5877%;\"\u003e\n \u003cp\u003eCaryophyllales\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.3254%;\"\u003e\n \u003cp\u003eAmaranthaceae\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003eSalicornia\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.6398%;\"\u003e\n \u003cp\u003e\u003cem\u003eramosissima\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 15.7978%;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.7457%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5877%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.3254%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003eSarcocornia\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.6398%;\"\u003e\n \u003cp\u003e\u003cem\u003eperennis\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 15.7978%;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.7457%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5877%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.3254%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003eHalimione\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.6398%;\"\u003e\n \u003cp\u003e\u003cem\u003eportulacoides\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 15.7978%;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.7457%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5877%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.3254%;\"\u003e\n \u003cp\u003ePlumbaginaceae\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003eLimonium\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.6398%;\"\u003e\n \u003cp\u003e\u003cem\u003ehumile\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 15.7978%;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.7457%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5877%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.3254%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9036%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.6398%;\"\u003e\n \u003cp\u003e\u003cem\u003edodartii\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 15.7978%;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.7457%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5877%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.3254%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9036%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.6398%;\"\u003e\n \u003cp\u003e\u003cem\u003evulgare\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 15.7978%;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.7457%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5877%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.3254%;\"\u003e\n \u003cp\u003eAmaranthaceae\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003eSuaeda\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.6398%;\"\u003e\n \u003cp\u003e\u003cem\u003emaritima\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 15.7978%;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.7457%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5877%;\"\u003e\n \u003cp\u003eAsterales\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.3254%;\"\u003e\n \u003cp\u003eAsteraceae\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003eAster\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.6398%;\"\u003e\n \u003cp\u003e\u003cem\u003etripolium\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 15.7978%;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.7457%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5877%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.3254%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003eInula\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.6398%;\"\u003e\n \u003cp\u003e\u003cem\u003ecrithmoides\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 15.7978%;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.7457%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.5877%;\"\u003e\n \u003cp\u003eLamiales\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.3254%;\"\u003e\n \u003cp\u003ePlantaginaceae\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003ePlantago\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.6398%;\"\u003e\n \u003cp\u003e\u003cem\u003emaritima\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 15.7978%;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eHealthy adult individuals of the target species, with branches well exposed to sunlight, were identified in the field and collected (Asner et al. 2015). \u0026nbsp;Samples (~200 grams of plant material) were maintained in ice coolers until brought to the laboratory (less than 2 hours) and stored in the refrigerator (4\u0026deg;C) in the dark until the fresh weight and the leaf reflectance were measured. Spectral and morphological measurements took less than 5 hours since collection. The remaining plant material was frozen in the dark for biochemical analysis \u003cem\u003e(i.e.,\u003c/em\u003e foliar nitrogen, phosphorus, and pigment content).\u0026nbsp;\u003c/p\u003e\n\u003ch2\u003eSpectral data\u003c/h2\u003e\n\u003cp\u003eHyperspectral reflectance of the leaves of each plant sample was measured with ASD Field Spec 4 (Analytical Spectral Devices, Boulder, USA) connected to a leaf probe equipped with an internal halogen light source under laboratory conditions. The instrument has a spectral range between 350 and 2500 nm with 1 nm spectral resolution. Before starting the measurements, the instrument was turned on to warm up for at least 30 minutes, dark current calibrations and white references (using the white panel from the leaf probe) were measured at the beginning and every 30 minutes to reduce eventual noise and to obtain relative reflectance (Schweiger 2020).\u003c/p\u003e\n\u003cp\u003eFully grown and healthy leaves were selected from the top of the branches, removed, and gently cleaned with absorbent paper before measurement. The leaves were placed aiming to fill the entire field of view. Given that some species had narrow leaves, they had more than one leaf to cover the field of view. Narrow leaves were sustained by a small sample platform composed of a square window made of a slightly rigid plastic sheet adapted from Noda et al. (2013). This sample mount allowed the placement of several blades in between two square frames (4 cm x 4 cm wide) and is fixed by the pressure of binder clips to handle the small leaves easily, serving as a platform to stabilise the leaves and completely fill the view of the fore optic (Support information \u0026ndash; SI 6). Each spectral measurement was set to be the average of 30 internal readings taken by the instrument. \u0026nbsp;The number of averaged internal readings has to be a balance between acquiring quality measurements and avoiding overheating the leaf by exposing it too long to the internal halogen lamp (Neuwirthov\u0026aacute; et al. 2017; Caturegli et al. 2020). Each species had three samples per study site, and each sample had three reflectance measurements recorded (one per leaf). Thus, each plant species had 9 measurements in each study site where they were collected (Fig. 2).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe spectral data collected in the laboratory was pre-processed first by correcting to match sensors using the R package \u003cem\u003espectrolab\u003c/em\u003e (Meireles et al. 2017). The noisy regions in the beginning and at the end of the spectra were excluded by trimming the spectrum to the length of 400 to 2400 nm (Schweiger et al. 2018). Lastly, each spectrum was vector normalised by dividing the reflectance value of each band by the magnitude of the spectral vector, \u003cimg src=\"data:image/png;base64,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\" width=\"70\" height=\"40\"\u003ewhere the magnitude ∣v∣ is calculated as the square root of the sum of the squared reflectance values across all bands, \u003cimg src=\"data:image/png;base64,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\" style=\"width: 118px; height: 45.0955px;\" width=\"118\" height=\"45.0955\"\u003e with \u003cem\u003en\u003c/em\u003e being the total number of bands and x\u003csub\u003ei\u003c/sub\u003e the reflectance value of waveband \u003cem\u003ei\u003c/em\u003e.\u0026nbsp;\u003c/p\u003e\n\u003ch2\u003eFunctional traits\u003c/h2\u003e\n\u003cp\u003eEight relevant functional traits, divided into anatomical (leaf mass per unit of area and water content on leaves) and biochemical traits (total organic carbon, foliar nitrogen, phosphorus, carotenoids, chlorophyll a and b) were analysed.\u003c/p\u003e\n\u003cp\u003eAnatomical traits\u003c/p\u003e\n\u003cp\u003eAfter the reflectance measurements, the same leaves had their petiole removed. They were scanned to calculate the leaf area with the software ImageJ (Schneider et al. 2012). \u0026nbsp;They were also weighed fresh and after they had been dried in the oven (60\u0026deg;C; 72 h), the trait leaf mass per area (g/m2) was calculated based on the dry weight (DW) divided by the leaf area. After removal from the oven, the samples were left at room temperature for ten minutes before being weighed. This step allows the plant material to cool down and prevents it from reabsorbing moisture from the air. The water content (g water/ g DW) was calculated by the difference between the fresh weight and the dry weight.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eBiochemical traits\u003c/p\u003e\n\u003cp\u003eThe same dried leaves used for the anatomical traits measurements were used in the analysis of total organic carbon (TOC). The determination of TOC in leaves (%) was through dry combustion (with an NDIR infrared detector) after correcting for inorganic carbon. A more detailed explanation of the laboratory procedures to measure these traits is provided in the Supplementary Material (SI 1).\u003c/p\u003e\n\u003cp\u003eAdditional leaves were randomly selected (roughly 100g of plant material) from the sample for chlorophyll and carotenoids (\u0026micro;g/g), nitrogen (%) and phosphorus analysis (%). The quantification of chlorophyll a, chlorophyll b, and total carotenoids was carried out following the equations proposed by Lichtenthaler \u0026amp; Buschmann (2001) and Lichtenthaler (1987), based on spectrophotometric determinations at specific wavelengths. These remaining leaves were slowly unfrozen before the analysis in a refrigerator (6\u0026ordm;C for 24 h). Nitrogen content in plant material was determined using the Kjeldahl method, a standard procedure for measuring organic and ammoniacal nitrogen (Bradstreet 1954). Phosphorus content in plant material was measured using UV-VIS spectrophotometry.\u003c/p\u003e\n\u003ch2\u003eData analysis\u003c/h2\u003e\n\u003ch3\u003eRelationship between types of biodiversity\u003c/h3\u003e\n\u003cp\u003eFunctional and spectral diversity relationship\u003c/p\u003e\n\u003cp\u003eTo assess whether functionally similar species exhibit similar spectral reflectance, we applied the methodology adopted by Schweiger et al. (2018) using dissimilarity matrices. Initially, the median reflectance values of each band (1nm wide) from the spectral data were compiled into principal components (PCs) for each species. Manhattan distances were then calculated for each species pair to construct the spectral dissimilarity matrix. Functional distance was determined using the z-scores of the measured traits for each species. Pairwise comparisons were plotted, and a linear regression was fitted for each species. Additionally, an overall linear regression with a 95% confidence level was computed for all species combined to evaluate the general trend.\u003c/p\u003e\n\u003cp\u003eThe significance of the correlation between spectral and functional distance matrices was assessed using a Mantel test with 999 permutations. The Mantel statistic (r), calculated between the pairwise elements of the two matrices, was also used to quantify the strength of the association. All analyses were conducted in R (R Core Team, 2024).\u003c/p\u003e\n\u003cp\u003eTaxonomic and spectral diversity relationship\u003c/p\u003e\n\u003cp\u003eTo quantify the variance in leaf spectral reflectance explained by different hierarchical taxonomic levels, we fitted linear mixed-effects models (LMMs) to the spectral data. This approach is well-suited for handling unbalanced data and nested structures. Initially, Principal Component Analysis (PCA) was performed on normalised hyperspectral reflectance data, with PCs retained as new variables representing the main axes of spectral variation. The analysis was conducted using the \u003cem\u003eprcomp\u003c/em\u003e function in R. PCs were retained until reaching a cumulative explained variance threshold of \u0026ge;90%, and used as response variables in the LMMs. The selected PCs were graphically displayed to compare the distances between species and order. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe models (one for each PC) included taxonomic levels (class, order, family, genus, and species) as predictor variables and site as random effects. Variables with sufficient variation were retained in the model, implemented using the \u003cem\u003elme4\u003c/em\u003e package. Variance components were extracted from the final models to assess the contribution of each taxonomic level to spectral variability in each PC. The variance explained by each taxonomic level was then expressed as a percentage of the total variance for each PC. This allowed us to evaluate how spectral variation is partitioned across taxonomic hierarchies and how different PCs capture distinct biological information.\u003c/p\u003e\n\u003cp\u003eAlso, the intraspecific and interspecific variation in spectral diversity was quantified by comparing variance within and between species for the selected PCs. Intraspecific variance was calculated as the average variance within each species, while interspecific variance was derived from the variance of species means. The ratio of inter- to intraspecific variance provided a measure of how much spectral variation is observed at the species level.\u003c/p\u003e\n\u003ch3\u003ePredictive potential among types of biodiversity\u003c/h3\u003e\n\u003cp\u003eFunctional traits quantification by spectral data\u003c/p\u003e\n\u003cp\u003ePredictive models to retrieve functional traits based on hyperspectral data were developed for total organic carbon, foliar nitrogen and phosphorus, chlorophyll a and b, carotenoids, water content, and leaf mass per area. For this purpose, the database was divided into 80% used to calibrate the model, and the remaining 20% to validate. A PLSR model was fitted using the \u003cem\u003epls\u003c/em\u003e package (R), where the predictor matrix consisted of the spectral data, and the response variable was defined as each of the functional traits separately (Burnett et al. 2021). To avoid overfitting, the optimal number of components used in the regressions was determined through cross-validation with the \u0026quot;onesigma\u0026quot; rule. A lower number of components could be selected when model accuracy was maintained. The accuracy of the models was evaluated based on the R\u003csup\u003e2\u003c/sup\u003e and the root-mean-square error (RMSE).\u003c/p\u003e\n\u003cp\u003eThe use of partial least squares regressions (PLSR) is a common approach due to its capability to incorporate the full information from the leaf reflectance spectrum into a few uncorrelated latent factors (Doughty et al. 2017). These latent factors can accommodate the high collinearity among the predictors (wavelengths) and the predictor variables (functional traits).\u003c/p\u003e\n\u003cp\u003eTaxonomic classification by spectral data\u003c/p\u003e\n\u003cp\u003eTo identify plant species based on hyperspectral reflectance, two fundamental steps are required: a feature selection followed by a discriminant analysis (Ferreira et al. 2013; Prospere et al. 2014). Features selection reduces dimensionality by identifying a subset of variables that enhance generalisation and reduce computational demands, while maintaining or improving classification accuracy (Hennessy et al. 2020). While the discriminant analysis is well-suited for maximising class separability while preserving the information contained in the selected variables. It is a commonly used approach for species classification based on spectral reflectance, often resulting in highly accurate models (Clark et al. 2005).\u003c/p\u003e\n\u003cp\u003eInitially, curve smoothing was applied by calculating the simple average across blocks of five adjacent bands for each species (Yu et al. 1999; Pu 2009). Then, a Mann-Whitney U test was selected as the feature selection method. It identified the wavebands that provide the highest discriminatory power between species. A total of 171 different pairs among species were tested for each band, the wavebands that accumulated a higher number of significantly different pairs (\u0026alpha;=0.01) were first selected (Schmidt and Skidmore 2003; Ferreira et al. 2013; Hennessy et al. 2020). Subsequently, the optimal number of wavebands to include in the model was determined using a support vector machine algorithm. This process employed 10-fold cross-validation, implemented through the R package \u003cem\u003ee1071\u003c/em\u003e (Version 1.7-16), in conjunction with the \u003cem\u003ecaret\u003c/em\u003e package to identify the feature subset that would yield the highest accuracy while minimising the risk of overfitting (Clark et al. 2005).\u003c/p\u003e\n\u003cp\u003eTo classify the species using the selected wavebands, Linear Discriminant Analysis (LDA) was chosen for the discriminant analysis. It has been proven effective classifier (Yu et al. 1999; Clark et al. 2005; Pu 2009), and applicable in wetland environments (Prospere et al. 2014; Hladik and Alber 2014). Firstly, the dataset was split into training and test sets (80% and 20% of the data, respectively). The partitioning was performed using the \u003cem\u003ecaret\u003c/em\u003e package to ensure balanced class representation in both sets. A LDA model was fitted to the training data using the \u003cem\u003elda\u003c/em\u003e function. After training, the model was used to predict the class labels of the test set. The accuracy and performance of the model were assessed based on the kappa statistic calculated from the predicted class labels to the true test labels (\u003cem\u003ei.e.,\u003c/em\u003e species).\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003eRelationship between types of biodiversity\u003c/p\u003e\u003cp\u003eFunctional and spectral diversity relationship\u003c/p\u003e\u003cp\u003eSpecies with greater functional similarity (similar biochemical and morphological traits) also exhibited higher spectral similarity (similar leaf reflectance patterns). Pairwise comparisons between species successfully highlighted moderate degrees of similarity based on both functional and spectral distance matrices (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e); the mean slope of the regressions and respective standard deviation was b\u0026thinsp;=\u0026thinsp;13.7 (\u0026plusmn; σ\u0026thinsp;=\u0026thinsp;7.8). The positive slope of nearly all regression lines indicates a trend that functionally similar species also tend to be spectrally similar. These slopes quantify the strength of the relationship, with stronger associations observed in \u003cem\u003eSuaeda maritima\u003c/em\u003e (b\u0026thinsp;=\u0026thinsp;24.25), \u003cem\u003ePhragmites australis\u003c/em\u003e (b\u0026thinsp;=\u0026thinsp;22.69), and \u003cem\u003eHalimione portulacoides\u003c/em\u003e (b\u0026thinsp;=\u0026thinsp;21.91). In contrast, weaker positive relationships were found for \u003cem\u003eZostera noltii\u003c/em\u003e (b\u0026thinsp;=\u0026thinsp;8.26), \u003cem\u003eZostera marina\u003c/em\u003e (b\u0026thinsp;=\u0026thinsp;9.39), and \u003cem\u003eSpartina maritima\u003c/em\u003e (b\u0026thinsp;=\u0026thinsp;10.03), while \u003cem\u003eJuncus maritimus\u003c/em\u003e (b=-7.21) was the only species with a negative relationship.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eThe Mantel statistic r of 0.46 (p\u0026thinsp;=\u0026thinsp;0.001) indicates a moderate, significant positive correlation between spectral and functional diversity. The observed value exceeded all percentiles (90th, 95th, 97.5th, and 99th ) of the null distribution generated by 999 permutations, confirming that the correlation did not arise by chance. The Mantel test thus supports the trends observed in the linear regression slopes, reinforcing the evidence that spectral variation reflects underlying functional differences across species.\u003c/p\u003e\u003cp\u003eTaxonomic and spectral diversity relationship\u003c/p\u003e\u003cp\u003eIndividuals from the same species tend to exhibit more similar leaf spectral reflectance patterns, and this similarity is consistent across different taxonomic levels (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea). Spectral variance within species was approximately five times lower than the variance observed among species, with intra- to interspecific variation ratios of 5.7 for PC1 and 4.8 for PC2. Additionally, the absolute variance estimated for PC1 (σ\u0026sup2;=2.43) was notably higher than that for PC2 (σ\u0026sup2;=0.38). The first two principal components captured a substantial proportion of the total spectral variation, with PC1 explaining 84.4% and PC2 12.5%.\u003c/p\u003e\u003cp\u003ePlants from the same order tend to cluster together in the two-dimensional space of the PCA plot (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb). Likewise, a substantial portion of the variance in spectral response captured by the PCs (\u0026gt;\u0026thinsp;90%) is explained by different taxonomic levels (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea). Higher taxonomic groups, such as order (40.41%) and class (49.74%), explain the largest fraction of variation in PC1 and PC2, respectively. However, genus and family also contribute considerably to spectral differences in estuarine plants, with genus explaining 23.41% and 15.78%, while family accounts for 16.63% and 12.76% of the variance in PC1 and PC2, respectively. Additionally, the site effects play a notable role in spectral diversity, contributing 7.6% and 9% of the variation in PC1 and PC2, respectively.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003ePredictive potential among types of biodiversity\u003c/p\u003e\u003cp\u003eFunctional traits quantification by spectral data\u003c/p\u003e\u003cp\u003eThe results show that it is possible to retrieve biochemical and anatomical traits of estuarine plant species in northern Spain from spectral data with reasonable precision (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). The R\u0026sup2; values ranged from 0.42 to 0.92, while the percentage root mean squared error of prediction (%RMSEP) varied between 8.56% and 21.08%. The number of components selected through cross-validation for model construction ranged from 4 (for carbon) to 12 (for nitrogen and LMA).\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eSummary of PLSR model performance for the eight traits analysed. The Root Mean Square Error of Prediction (RMSEP) is reported in the same units as the respective traits, alongside its percentage (%RMSEP), which accounts for differences in trait magnitudes. The number of components used to build the model (n\u0026ordm; Components) and the sample size (N) to both train and validate the model is indicated\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"7\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTrait\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eUnit\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eR\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eRMSEP\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003e%RMSEP\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003en\u0026ordm; Components\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWater\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eg/g of dry mass\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.92\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.86\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e8.56\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e10\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e288\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eOrganic Carbon\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e%\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.87\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e2.76\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e12.38\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e41\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eFoliar Phosphorus\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e%\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.81\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e12.34\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e11\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e183\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLeaf Mass per Area\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eg/m\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.76\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e54.8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e10.8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e12\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e288\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eFoliar Nitrogen\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e%\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.66\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e3.31\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e17.51\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e12\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e183\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eChlorophyll a\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e\u0026micro;g/g\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.53\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e130.79\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e20.06\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e11\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e131\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCarotenoids\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e\u0026micro;g/g\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.44\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e46.57\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e19.71\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e9\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e131\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eChlorophyll b\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e\u0026micro;g/g\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.42\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e63.73\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e21.08\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e11\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e131\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eThe validation R\u0026sup2; values and the %RMSEP indicated that the models performed well for most traits, particularly water content, organic carbon, phosphorus, and LMA (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e). Notably, the model for organic carbon exhibited strong performance despite having the smallest sample size for training and validation (N\u0026thinsp;=\u0026thinsp;41) and the lowest number of components (n\u0026ordm; Comp\u0026thinsp;=\u0026thinsp;4), achieving a high proportion of variability explained (R\u0026sup2;=0.87). In contrast, models for traits related to pigment content showed moderate accuracy, as indicated by their lower R\u0026sup2; values and higher %RMSEP (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). The models achieved R\u0026sup2; values of 0.92, 0.81, and 0.76 for water content, phosphorus, and LMA, respectively, indicating that a substantial proportion of the variability in these traits was explained by the spectral data (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e). Scatterplots of observed versus predicted values for both calibration and validation datasets further confirmed the accuracy of the models, with points closely aligned to the 1:1 line for water content, phosphorus, nitrogen, carbon, and LMA. For pigment-related traits, the relationship was less robust, although most predictions fell within the 95% confidence interval. Residual histograms for both calibration and validation datasets resemble a symmetric distribution around zero, indicating no systematic bias in the predictions.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eTaxonomic classification by spectral data\u003c/p\u003e\u003cp\u003eTwo spectral regions provided the best discrimination between the estuarine species at the leaf level, these are between 720\u0026ndash;770 nm and 1330\u0026ndash;1380 nm. These regions were located at the short borders of the NIR and SWIR ranges, specifically in the red edge shoulder and in the downward slope of the NIR reflectance plateau (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e). Within these regions, 20 spectral bands are sufficient to develop parsimonious models capable of distinguishing between the leaves of different species with accuracies of around 97%. Maximum accuracy (100%) is achieved when 45 or more bands are included (Supplementary Material \u0026ndash; SI 2). Given that each band is 5 nm wide, the selection of 20 adjacent bands corresponds to a 100 nm spectral range, which is grouped into two distinct clusters. When expanding the analysis to include 45 bands (a 225 nm range) to achieve 100% accuracy. The discriminatory frame on the red edge expands further into the NIR range, reaching 900 nm. Although the top selected bands are within the VIS and NIR regions, there are relevant frames in the SWIR range (between 1840nm-1620nm) that could also help distinguish between salt marsh and seagrass species (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eThe results from the confusion matrix and LDA model demonstrate a high level of accuracy in classifying the plant species based on the provided spectral data (Supplementary Material - SI 3). The overall accuracy of the model is 97.87%, with a 95% confidence interval ranging from 88.71% to 99.95%. This high accuracy is further supported by a Kappa statistic of 0.977, indicating almost perfect agreement between the predicted and actual classifications. The accuracy being greater than the no-information rate is highly significant (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), suggesting that the model performs substantially better than random chance.\u003c/p\u003e\u003cp\u003eThe confusion matrix reveals that the model correctly classified all instances of most species, with no misclassifications observed for the majority of classes (Supplementary material \u0026ndash; SI 3). For example, species such as \u003cem\u003eZostera marina\u003c/em\u003e, \u003cem\u003eZostera noltii\u003c/em\u003e, \u003cem\u003eAster tripolium\u003c/em\u003e, \u003cem\u003eInula crithmoides\u003c/em\u003e, \u003cem\u003eSuaeda maritima\u003c/em\u003e, and \u003cem\u003eHalimione portulacoides\u003c/em\u003e were all predicted with perfect sensitivity and specificity (100%). Similarly, species like \u003cem\u003eLimonium dodartii\u003c/em\u003e, \u003cem\u003eLimonium humile\u003c/em\u003e, \u003cem\u003ePhragmites australis\u003c/em\u003e, \u003cem\u003eJuncus maritimus\u003c/em\u003e, \u003cem\u003eSpartina maritima\u003c/em\u003e and \u003cem\u003ePuccinellia maritima\u003c/em\u003e were also classified without error. The model struggled with \u003cem\u003eSalicornia ramosissima\u003c/em\u003e, misclassified as \u003cem\u003eSarcocornia perennis\u003c/em\u003e, both belonging to the family Amaranthaceae. The LDA model's group provides insights into the spectral characteristics that distinguish each species. For instance, \u003cem\u003eSalicornia ramosissima\u003c/em\u003e and \u003cem\u003eSarcocornia perennis\u003c/em\u003e exhibit higher reflectance values in certain spectral bands (e.g., 735, 725, and 730 nm) compared to other species.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eThese results show that different biodiversity dimensions can be estimated from spectral features, and that spectral traits can be used to infer taxa and functional traits. These findings corroborate previous studies demonstrating the links among spectral, functional, and taxonomic diversity in terrestrial vegetation (Ustin and Gamon \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2010\u003c/span\u003e; Schweiger et al. \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Lalibert\u0026eacute; et al. \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Bohrer et al. \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Furthermore, the present study extends this understanding by providing evidence that such relationships also persist within estuarine ecosystems.\u003c/p\u003e\u003cp\u003eIt has been shown that each species is composed of a set of anatomical, biochemical and spectral characteristics that are closely related, supporting the development of models to estimate functional traits and species based on the spectral data. Also, the spectral reflectance carries strong taxonomic influence across multiple hierarchical levels. This could mean that spectral data is likely driven by inherent physiological, structural, or biochemical traits and could be used as a powerful tool for species discrimination and functional trait quantification in temperate estuarine ecosystems.\u003c/p\u003e\u003cp\u003eRelationship between types of biodiversity\u003c/p\u003e\u003cp\u003eFunctional and spectral diversity relationship\u003c/p\u003e\u003cp\u003eOur results indicate that species functionally similar exhibit more similar spectral responses, reinforcing the strong relationship between functional traits and leaf reflectance observed in our dataset. This relationship arises because of the direct influence the biochemical and biophysical traits of leaves have on the spectral signatures (Kumar et al. \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2001\u003c/span\u003e). Since leaf spectral behaviour results from an energy-matter interaction between the radiating light and the physicochemical structures in the leaves, it is expected that the biochemical content and the anatomical structures of plants have a direct effect on the way electromagnetic energy interacts with the structures present in these leaves (Asner and Martin \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2009\u003c/span\u003e; Cavender-Bares et al. \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe moderate positive correlation relationship between spectral and functional traits supported by the Mantel test, indicates that functionally similar species tend to exhibit greater spectral similarity (Asner et al. \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). These findings reinforce the potential for remotely and non-destructively monitoring functional diversity over large spatial scales, offering promising applications for biodiversity conservation and ecosystem management (Jetz et al. \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Schneider et al. \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2017\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe traits selected in this study are well-known predictors of ecosystem function, serving as proxies for key ecological processes (Homolov\u0026aacute; et al. \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). The uniqueness of each species\u0026rsquo; biochemical fingerprint (Support information \u0026ndash; SI 5) suggests that functional diversity could be estimated remotely, even without prior knowledge of functional groups in the field. Although there is a natural variation in leaf biochemistry within the same species (Asner and Martin \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2011\u003c/span\u003e), the complexity of spectral differentiation increases as more traits are considered in the analysis (Asner and Martin \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2009\u003c/span\u003e). The ability to quantify functional traits using spectral data opens new possibilities for large-scale monitoring of functional diversity (Schweiger et al. \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). Given that functional biodiversity is linked to other dimensions of biodiversity, this approach could also be leveraged for remote sensing-based biodiversity assessments at broader spatial scales.\u003c/p\u003e\u003cp\u003eTaxonomic and spectral diversity relationship\u003c/p\u003e\u003cp\u003eOur study suggests that spectral reflectance carries strong taxonomic signals across multiple hierarchical levels, with closely related taxa having similar spectral characteristics. Since these two components account for almost all the observed variation (~\u0026thinsp;97%), analysing them provides robust insights into the spectral behaviour of the assessed species. The not inclusion of order in the PC2 model due to low variation suggests that variability at this level has already been captured by other taxonomic groups. This supports the idea that spectral variation is hierarchically structured across taxonomic levels, with some levels (e.g., class) capturing more distinct variation.\u003c/p\u003e\u003cp\u003ePlant traits are shaped by a combination of factors, such as genetic material, environmental conditions, and an individual\u0026rsquo;s life history (Ustin and Gamon \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2010\u003c/span\u003e). Therefore, plants from the same taxon have comparable genetic material and share traits that will shape the spectral reflectance, which in turn influences their spectral reflectance. These similarities likely account for the substantial proportion of spectral variance explained by taxonomic classification, with related species exhibiting spectral characteristics shaped by their evolutionary history and environmental adaptation.\u003c/p\u003e\u003cp\u003eAs shown in this work, spectral reflectance tends to be conserved at higher taxonomic groups such as genus, family and to a larger extent order and class (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea). Although the consistency of the spectral reflectance is observed to be more accentuated in higher biological organizations, the variation of the spectral reflectance observed within the same species has been around five times smaller than observed between species. This suggests a degree of intra-specific plasticity (Asner and Martin \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2011\u003c/span\u003e) while maintaining sufficient stability to distinguish species and capture broader taxonomic relationships. These findings reinforce the potential of spectral data for species identification and biodiversity monitoring in temperate estuarine ecosystems (Wang and Gamon \u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e2019\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eA substantial portion of spectral variance remains unexplained by taxonomy alone (\u0026lt;\u0026thinsp;20%), as indicated by the residuals in our models. This suggests that other factors contribute to spectral reflectance variability, including measurement inconsistencies at the leaf level (Neuwirthov\u0026aacute; et al. \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2017\u003c/span\u003e), intraspecific variations (Asner and Martin \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2011\u003c/span\u003e) and evolutionary or functional convergence (Ustin and Gamon \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2010\u003c/span\u003e). Environmental factors also play a role in shaping spectral reflectance, as shown by the site variable explaining a considerable part of the variability. This suggests that local environmental conditions, such as soil composition, salinity, and nutrient availability, also contribute to spectral differences among individuals of the same species. Further investigation into these factors could provide a more comprehensive understanding of the drivers of spectral diversity in estuarine plant communities.\u003c/p\u003e\u003cp\u003ePredictive potential among types of biodiversity\u003c/p\u003e\u003cp\u003eFunctional traits retrieval by spectral data\u003c/p\u003e\u003cp\u003eThis study successfully developed models capable of retrieving functional traits from spectral data with sufficient precision for plant species in Atlantic temperate estuaries. The high accuracy of models for water content, phosphorus, and LMA (R\u0026sup2;=0.92, 0.81, and 0.76, respectively) aligns with previous studies showing that these traits are strongly linked to specific spectral regions, such as the NIR and SWIR for water content (Asner and Martin \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2008\u003c/span\u003e; Homolov\u0026aacute; et al. \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). The symmetric distribution of residuals around zero for both calibration and validation datasets further confirm the absence of systematic bias in the predictions, supporting the reliability of the models.\u003c/p\u003e\u003cp\u003eAmong the analysed traits, the strong performance of the organic carbon model (R\u0026sup2;=0.87) despite its small sample size (N\u0026thinsp;=\u0026thinsp;41) and low number of components (n\u0026ordm; Comp\u0026thinsp;=\u0026thinsp;4) highlights the robustness of spectral data for predicting this trait (Schweiger et al. \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). Although total organic carbon has fewer observations to train and validate the model, the dataset used to calibrate the predictive model included a great part of the possible variation observed in this plant trait (Burnett et al. \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe predictions for phosphorus and LMA followed the same pattern, also performing well when compared to existing studies, demonstrating high R\u003csup\u003e2\u003c/sup\u003e and relatively low %RMSEP (Homolov\u0026aacute; et al. \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; Wang et al. \u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). The anatomical trait LMA, often expressed as its inverse, specific leaf area (SLA), influences spectral data differently from biochemical traits by physically interacting with electromagnetic radiation. LMA is a reliable indicator of leaf thickness and density and is a key trait widely used to classify plant functional groups (Poorter et al. \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2009\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eFoliar nitrogen content is widely recognised as a trait that can be retrieved with high accuracy using spectral data. Our nitrogen predictions fell within the expected range but did not exceed the performance of existing models (Homolov\u0026aacute; et al. \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). Since there is no consensus on which part of the spectrum should be used to estimate nitrogen content from spectral data, we used the full spectrum (400-2400nm), identifying wavelengths in the visible range, particularly in the red edge, as the most influential in modelling nitrogen content (Supplementary material \u0026ndash; SI 4).\u003c/p\u003e\u003cp\u003eThe possibility to predict leaf water content has broad applications, such as remotely assessing hydric stress during drought events (Martin et al. \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). This is due to the strong predictive potential of water content in leaves, which was also evident in our study. Compared to previous research, our model for leaf water content demonstrated high precision (Asner and Martin \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2008\u003c/span\u003e; Wang et al. \u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), achieving the best performance among the traits analysed (R\u0026sup2; = 0.92). While some studies highlight the importance of the red edge in estimating water content (Filella and Penuelas \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e1994\u003c/span\u003e), our results indicate that, beyond the red edge, certain wavelengths in the NIR and SWIR regions play a critical role in the PLSR models for water content (Supplementary Material \u0026ndash; SI 4). This is consistent with expectations, as water has a strong influence on specific regions of the NIR spectrum.\u003c/p\u003e\u003cp\u003eIn contrast, pigment content showed lower predictive performance than reported in other studies. Two factors may have contributed to this outcome. First, we used a colorimetric method to estimate pigment concentration, whereas other studies used more precise techniques such as high-performance liquid chromatography (HPLC). Second, while most studies report pigment content per unit leaf area (Schweiger et al. \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2018\u003c/span\u003e), our values were expressed on mass per mass (\u0026micro;g/g), which may have caused the reduction of the performance of the model. Due to the complex interactions between pigments and their overlapping spectral signatures, photosynthetic pigments are relatively complicated to estimate, especially at canopy level (Ustin et al. \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2009\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe ability to accurately predict key functional traits from spectral data has significant implications for ecological monitoring and ecosystem management. These traits have been successfully retrieved via remote sensing across diverse biomes (Jetz et al. \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2016\u003c/span\u003e), and they are essential for understanding plant physiological processes, nutrient cycling, and ecosystem functioning (Wright et al. \u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e2004\u003c/span\u003e). However, caution is needed when applying leaf-level models to canopy spectral data, as factors such as multiple scattering and other optical effects can influence prediction accuracy (Al Makdessi et al. \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). This study contributes to this field by providing openly available coefficients for trait quantification in the region of North Spain.\u003c/p\u003e\u003cp\u003ePhylogenetic classification by spectral data\u003c/p\u003e\u003cp\u003eThe feature selection method successfully identified the wavebands where species exhibit the greatest differences from one another at the leaf level. As shown by other authors, the red edge has been widely recognized for its potential in species discrimination using hyperspectral data (Vaiphasa et al. \u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e2007\u003c/span\u003e; Ferreira et al. \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; Prospere et al. \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). However, the selection of a spectral frame in the far NIR is less common, despite previous studies highlighting its relevance for identifying wetland vegetation (Schmidt and Skidmore \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2003\u003c/span\u003e; Adam et al. \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2010\u003c/span\u003e). Although the wavelengths selected in this study are in these two frames, it is also possible to build discrimination models with other bands (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e), depending on the spectral resolution of the available sensor. It is important to note that the optimal number and the position of the spectral frames for species discrimination are not universally fixed, but rather depend on the feature selection method, the specific dataset, and the biological characteristics of the vegetation (Hennessy et al. \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe entire spectrum produces more accurate species-level classification (Hennessy et al. \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), although studies discriminating plant taxa based on leaf reflectance often focus on the visible spectrum (Artigas and Yang \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2006\u003c/span\u003e; Gross and Heumann \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). This is due to the visible region exhibiting less variation due to leaf stacking effects compared to the NIR and SWIR regions (Neuwirthov\u0026aacute; et al., \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). Many leaf-level studies prioritize this spectral range because of its strong correlation with pigment content (Kumar et al. \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2001\u003c/span\u003e) and the incorporation of the red edge. In contrast, the NIR and SWIR regions often show high intraspecific variation, primarily influenced by water loss during transport and the time elapsed between collection and measurement (Ferreira et al. \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). In this study, the gap between field collection and laboratory measurement was relatively short (\u0026lt;\u0026thinsp;4 hours), suggesting that the observed variations in these spectral regions were likely due to the biological characteristics of the species, potentially reflecting natural changes in water content.\u003c/p\u003e\u003cp\u003eAlthough the maximum accuracy (100%) was achieved with 45 bands, the model with 20 bands (kappa\u0026thinsp;=\u0026thinsp;97%) effectively reduced data dimensionality while maintaining high accuracy. However, it is important to consider how this accuracy may change when scaled to the canopy level using the same number of bands (Clark et al. \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2005\u003c/span\u003e). Additionally, the proximity of the selected SWIR range (1335\u0026ndash;1380 nm) to the water vapor absorption region (1350\u0026ndash;1480 nm) raises concerns about potential noise, highlighting the need for further validation with canopy-level spectral measurements. It has been pointed out that spectral variations detected among species at the leaf scale do not readily correspond to assessments conducted at the landscape scale.\u003c/p\u003e\u003cp\u003eIndividual inconsistencies are observed throughout the broad spectrum. That can be related to the intrinsic optical properties of the leaves or artefacts from the technique (Hennessy et al. \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Since manipulating narrow leaves can be challenging (Noda et al. \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2013\u003c/span\u003e), the attempt to fill the field of view can slightly alter the geometry of the leaves being measured, affecting the reflected spectrum. Although we try to maintain consistency regarding the age of the leaves sampled and the part of the leaf where the reading was performed, different leaf ages (Clark et al. \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2005\u003c/span\u003e) or leaf layers (Neuwirthov\u0026aacute; et al. \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2017\u003c/span\u003e) could have affected the final spectrum measured of each sample. The use of a squared sampling port significantly facilitated the handling of these conspicuous leaves to completely fill the field of view of the spectroradiometer, minimizing changes in angle and shades within the probe. The vector normalize technique was also adopted to reduce individual inconsistencies. This process standardizes the magnitude of the spectral curve, allowing for better comparison across spectra and reducing effects from leaf stacking and luminosity (Neuwirthov\u0026aacute; et al., \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2017\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe high accuracy and robust performance of the discriminant model underscore its potential for applications in ecological monitoring and species identification using spectral data. The misclassification of \u003cem\u003eSalicornia ramosissima\u003c/em\u003e highlights the need for further refinement, potentially through additional training data or feature engineering, to improve the model's ability to distinguish between spectrally similar species either at leaf and canopy level.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eThis study shows the close connection between taxonomic, functional, and spectral diversity in estuarine plant communities on the temperate Atlantic coast. It contributes to the foundation of research involving the spectral diversity of vascular plants in the estuaries in North Spain by calculating the coefficients to retrieve biochemical and anatomical traits from spectral data.\u003c/p\u003e\u003cp\u003eIn fact, functionally similar species exhibited comparable spectral signatures, reinforcing the idea that spectral data can serve as a proxy for functional diversity. It offers a promising indication for monitoring functional diversity at large scales using remote sensing data. Similarly, taxonomic classification played a significant role in shaping spectral diversity. However, the study revealed that intraspecific variability may contribute to spectral diversity, indicating a need for further research to account for these influences.\u003c/p\u003e\u003cp\u003eIt can be concluded that methodologies previously developed for identifying plant traits and species composition in other ecosystems are also suitable for estimating biodiversity in the temperate estuaries of the Gulf of Biscay. This success is not driven by a single spectral band, but rather by multiple spectral frames located at the edges of the NIR and SWIR ranges. These findings reinforce the importance of the red-edge shoulder in capturing remotely sensed plant traits and translating them into information about species composition and biochemical structure. Nonetheless, further efforts are needed to refine these models to accurately monitor species distribution and phenological changes across spatial and temporal scales, ultimately enabling the production of more precise species distribution maps.\u003c/p\u003e\u003cp\u003eThis study is an essential step to upscale the monitoring of vegetation types using remote sensing technologies. Also, it offers a punctual assessment of these three types of biodiversity in estuaries. It is expected that the species composition, abundance and dynamics change with time and it is imperative that we are able to track these changes cost-effectively through larger areas, particularly in ecologically sensitive and dynamic environments like coastal wetlands. Future work should focus on validating these findings at the canopy and landscape levels, as well as exploring the integration of remote sensing data with other biodiversity metrics to enhance conservation and management efforts.\u003c/p\u003e\u003cp\u003eDataset repository\u003c/p\u003e\u003cp\u003eThe dataset produced and used by this work is available open-access through the link \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://apidies.ihcantabria.com/swagger/index.html\u003c/span\u003e\u003cspan address=\"https://apidies.ihcantabria.com/swagger/index.html\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e\u003cp\u003ePlant identification was conducted by MSc. Andre C. da Costa Neves under the supervision of the Coastal Ecosystems Group at IHCantabria. Specimens that could not be identified at species level were sent to specialized taxonomists and subjected to genetic identification. Reference individuals were preserved as dried exsiccata in the institutional herbarium of IHCantabria and are available upon request. The number and details of the reference specimens are provided in the Supplementary Information (SI 7).\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003ch2\u003eCompeting Interests:\u003c/h2\u003e\u003cp\u003eThe 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.\u003c/p\u003e\u003c/p\u003e\u003cp\u003e\u003ch2\u003eEthical considerations\u003c/h2\u003e\u003cp\u003e All experimental research and field studies on plants reported in this manuscript complied with institutional, national, and international guidelines and legislation. Collection of wild plant material was authorized by the \u003cem\u003eDirecci\u0026oacute;n General de Montes y Biodiversidad\u003c/em\u003e of the regional government of Cantabria (Spain), and all procedures adhered to Spanish national regulations as well as international conventions on biodiversity and genetic resources. Only small amounts of leaves and branches were collected and brought to the laboratory for data collection, while the source plants were left alive in their natural habitats.\u003c/p\u003e\u003c/p\u003e\u003ch2\u003eFunding:\u003c/h2\u003e\u003cp\u003eThis study forms part of the ThinkInAzul programme and was supported by the Ministry of Science and Innovation with funding from the European Union NextGeneration EU (PRTR-C17.I1), and by the autonomous community of Cantabria, and the project MarshA (Restoration beyond biodiversity: How to integrate estuarine ecosystem services into nature-based management) (TED2021-129973B-I00) funded by MICIU/AEI/ \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.13039/501100011033\u003c/span\u003e\u003cspan address=\"10.13039/501100011033\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e and by the European Union NextGenerationEU/PRTR. This manuscript also associated with the MARBEFES project (MARine Biodiversity and Ecosystem Functioning leading to Ecosystem Services), funded by the European Union under the Horizon Europe Programme, \u0026ldquo;HORIZON-CL6-2021-BIODIV-01\u0026rdquo; Theme, Grant Agreement no. 101060937 (marbefes.eu).\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eA. C. N. Conceptualization; Methodology; Fieldwork (data collection, sample processing); Formal analysis; Data curation; Visualization; Writing\u0026mdash;original draft. C. G. Conceptualization; Funding acquisition; Resources; Supervision; Project administration; Writing\u0026mdash;review \u0026amp; editing. B. O. Conceptualization; Funding acquisition; Resources; Supervision; Validation; Writing\u0026mdash;review \u0026amp; editing. All authors contributed to the study conception, design, also read and approved the final manuscript.\u003c/p\u003e\u003ch2\u003eAcknowledgement\u003c/h2\u003e\u003cp\u003eThe work was developed under the Complementary Plan for R+D+i in Marine Sciences (PCM) and within the workframe of the DIES and MarshA projects. The authors thank the Hydrobiology Laboratory at IHCantabria for support with laboratorial analyses, and the Centro de Investigaci\u0026oacute;n y Formaci\u0026oacute;n Agrarias (CIFA) of the Government of Cantabria for assistance with biochemical trait measurements.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eThe dataset produced and used by this work is available open-access through the link [https://apidies.ihcantabria.com/swagger/index.html](https:/apidies.ihcantabria.com/swagger/index.html)\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAdam E, Mutanga O, Rugege D (2010) Multispectral and hyperspectral remote sensing for identification and mapping of wetland vegetation: a review. Wetl Ecol Manag 18:281\u0026ndash;296. https://doi.org/10.1007/s11273-009-9169-z\u003c/li\u003e\n\u003cli\u003eAl Makdessi N, Ecarnot M, Roumet P, Rabatel G (2019) A spectral correction method for multi-scattering effects in close range hyperspectral imagery of vegetation scenes: application to nitrogen content assessment in wheat. Precis Agric 20:237\u0026ndash;259. https://doi.org/10.1007/s11119-018-09631-9\u003c/li\u003e\n\u003cli\u003eArtigas FJ, Yang J (2006) Spectral discrimination of marsh vegetation types in the New Jersey Meadowlands, USA. Wetlands 26:271\u0026ndash;277. https://doi.org/10.1672/0277-5212(2006)26[271:SDOMVT]2.0.CO;2\u003c/li\u003e\n\u003cli\u003eAsner GP, Martin RE (2008) Spectral and chemical analysis of tropical forests: scaling from leaf to canopy levels. Remote Sens Environ 112:3958\u0026ndash;3970. https://doi.org/10.1016/j.rse.2008.07.003\u003c/li\u003e\n\u003cli\u003eAsner GP, Martin RE (2009) Airborne spectranomics: mapping canopy chemical and taxonomic diversity in tropical forests. Front Ecol Environ 7:269\u0026ndash;276. https://doi.org/10.1890/070152\u003c/li\u003e\n\u003cli\u003eAsner GP, Martin RE (2011) Canopy phylogenetic, chemical and spectral assembly in a lowland Amazonian forest. New Phytol 189:999\u0026ndash;1012. https://doi.org/10.1111/j.1469-8137.2010.03549.x\u003c/li\u003e\n\u003cli\u003eAsner GP, Martin RE (2016) Spectranomics: emerging science and conservation opportunities at the interface of biodiversity and remote sensing. Glob Ecol Conserv 8:212\u0026ndash;219. https://doi.org/10.1016/j.gecco.2016.09.010\u003c/li\u003e\n\u003cli\u003eAsner GP, Martin RE, Anderson CB, Knapp DE (2015) Quantifying forest canopy traits: imaging spectroscopy versus field survey. Remote Sens Environ 158:15\u0026ndash;27. https://doi.org/10.1016/j.rse.2014.11.011\u003c/li\u003e\n\u003cli\u003eAsner GP, Martin RE, Knapp DE et al (2011) Spectroscopy of canopy chemicals in humid tropical forests. Remote Sens Environ 115:3587\u0026ndash;3598. https://doi.org/10.1016/j.rse.2011.08.020\u003c/li\u003e\n\u003cli\u003eAsner GP, Martin RE, Knapp DE et al (2017) Airborne laser-guided imaging spectroscopy to map forest trait diversity and guide conservation. Science 355:385\u0026ndash;389. https://doi.org/10.1126/science.aaj1987\u003c/li\u003e\n\u003cli\u003eBelluco E, Camuffo M, Ferrari S et al (2006) Mapping salt-marsh vegetation by multispectral and hyperspectral remote sensing. Remote Sens Environ 105:54\u0026ndash;67. https://doi.org/10.1016/j.rse.2006.06.006\u003c/li\u003e\n\u003cli\u003eBohrer G, Cavender-Bares J, Chaplin-Kramer R et al (2022) NASA Biological Diversity and Ecological Forecasting: current state of knowledge and considerations for the next decade. NASA Report. https://doi.org/xxx\u003c/li\u003e\n\u003cli\u003eBradstreet RB (1954) Kjeldahl method for organic nitrogen. Anal Chem 26:185\u0026ndash;187. https://doi.org/10.1021/ac60085a024\u003c/li\u003e\n\u003cli\u003eBurnett AC, Anderson J, Davidson KJ et al (2021) A best-practice guide to predicting plant traits from leaf-level hyperspectral data using partial least squares regression. J Exp Bot 72:6175\u0026ndash;6189. https://doi.org/10.1093/jxb/erab295\u003c/li\u003e\n\u003cli\u003eCaturegli L, Matteoli S, Gaetani M et al (2020) Effects of water stress on spectral reflectance of bermudagrass. Sci Rep 10:12055. https://doi.org/10.1038/s41598-020-72006-6\u003c/li\u003e\n\u003cli\u003eCavender-Bares J, Gamon JA, Townsend PA (2020) Remote sensing of plant biodiversity. Springer, Cham. https://doi.org/10.1007/978-3-030-33157-3\u003c/li\u003e\n\u003cli\u003eClark ML, Roberts DA, Clark DB (2005) Hyperspectral discrimination of tropical rain forest tree species at leaf to crown scales. Remote Sens Environ 96:375\u0026ndash;398. https://doi.org/10.1016/j.rse.2005.03.009\u003c/li\u003e\n\u003cli\u003eConsejer\u0026iacute;a de Desarrollo Rural, Ganader\u0026iacute;a, Pesca y Biodiversidad (n.d.) Anejo VII. Documento t\u0026eacute;cnico de los espacios acu\u0026aacute;ticos litorales Natura 2000 en Cantabria. C. LIC Dunas del Puntal y Estuario del Miera. https://rednatura2000cantabria.ihcantabria.com/ESPACIOS/5_LITORALES/PLANES%20DE%20CADA%20LIC/Anejo%20VII_PLAN%20GESTION%20LIC%20PUNTAL.pdf\u003c/li\u003e\n\u003cli\u003eCurcio AC, Barbero L, Peralta G (2023) UAV-Hyperspectral Imaging to estimate species distribution in salt marshes: a case study in the Cadiz Bay (SW Spain). Remote Sens 15:1419. https://doi.org/10.3390/rs15051419\u003c/li\u003e\n\u003cli\u003eDoughty CE, Santos-Andrade PE, Goldsmith GR et al (2017) Can leaf spectroscopy predict leaf and forest traits along a Peruvian tropical forest elevation gradient? J Geophys Res Biogeosci 122:2952\u0026ndash;2965. https://doi.org/10.1002/2017JG003883\u003c/li\u003e\n\u003cli\u003eFeret JB, Asner GP (2013) Tree species discrimination in tropical forests using airborne imaging spectroscopy. IEEE Trans Geosci Remote Sens 51:73\u0026ndash;84. https://doi.org/10.1109/TGRS.2012.2199323\u003c/li\u003e\n\u003cli\u003eFerreira MP, Grondona AEB, Rolim SBA, Shimabukuro YE (2013) Analyzing the spectral variability of tropical tree species using hyperspectral feature selection and leaf optical modeling. J Appl Remote Sens 7:073502. https://doi.org/10.1117/1.JRS.7.073502\u003c/li\u003e\n\u003cli\u003eFilella I, Penuelas J (1994) The red edge position and shape as indicators of plant chlorophyll content, biomass and hydric status. Int J Remote Sens 15:1459\u0026ndash;1470. https://doi.org/10.1080/01431169408954177\u003c/li\u003e\n\u003cli\u003eGross JW, Heumann BW (2014) Can flowers provide better spectral discrimination between herbaceous wetland species than leaves? Remote Sens Lett 5:892\u0026ndash;901. https://doi.org/10.1080/2150704X.2014.973077\u003c/li\u003e\n\u003cli\u003eGuinea E (1953) Ammophiletea, Crithmo-Staticetea, Salicornieteaque santanderiensis. An Jard Bot Madr 11:545\u0026ndash;568.\u003c/li\u003e\n\u003cli\u003eHennessy A, Clarke K, Lewis M (2020) Hyperspectral classification of plants: a review of waveband selection generalisability. Remote Sens 12:113. https://doi.org/10.3390/rs12010113\u003c/li\u003e\n\u003cli\u003eHladik C, Alber M (2014) Classification of salt marsh vegetation using edaphic and remote sensing-derived variables. Estuar Coast Shelf Sci 141:47\u0026ndash;57. https://doi.org/10.1016/j.ecss.2014.01.011\u003c/li\u003e\n\u003cli\u003eHomolov\u0026aacute; L, Malenovsk\u0026yacute; Z, Clevers JGPW et al (2013) Review of optical-based remote sensing for plant trait mapping. Ecol Complex 15:1\u0026ndash;16. https://doi.org/10.1016/j.ecocom.2013.07.003\u003c/li\u003e\n\u003cli\u003eIH Cantabria (2017) CONVIVE-LIFE: Deliverable Action A1. Cartograf\u0026iacute;a de h\u0026aacute;bitats y formaciones vegetales/caracterizaci\u0026oacute;n de las comunidades biol\u0026oacute;gicas del estuario/cartograf\u0026iacute;a de especies invasoras. http://convivelife.es/wp-content/uploads/2017/11/Annex_IIA1_Final1.pdf\u003c/li\u003e\n\u003cli\u003eJetz W, Cavender-Bares J, Pavlick R et al (2016) Monitoring plant functional diversity from space. Nat Plants 2:16024. https://doi.org/10.1038/nplants.2016.24\u003c/li\u003e\n\u003cli\u003eKaennel M (1998) Biodiversity: a diversity in definition. In: Bachmann P, K\u0026ouml;hl M, P\u0026auml;ivinen R (eds) Assessment of biodiversity for improved forest planning. Springer, Dordrecht, pp 71\u0026ndash;81.\u003c/li\u003e\n\u003cli\u003eKlemas V (2013) Remote sensing of emergent and submerged wetlands: an overview. Int J Remote Sens 34:6286\u0026ndash;6320. https://doi.org/10.1080/01431161.2013.800656\u003c/li\u003e\n\u003cli\u003eKumar L, Schmidt KS, Dury S, Skidmore AK (2001) Review of hyperspectral remote sensing and vegetation science. Imaging Spectrom Basic Princ Prospect Appl:111\u0026ndash;155.\u003c/li\u003e\n\u003cli\u003eLalibert\u0026eacute; E, Schweiger AK, Legendre P (2020) Partitioning plant spectral diversity into alpha and beta components. Ecol Lett 23:370\u0026ndash;380. https://doi.org/10.1111/ele.13429\u003c/li\u003e\n\u003cli\u003eLichtenthaler HK (1987) Chlorophyll fluorescence signatures of leaves during the autumnal chlorophyll breakdown. J Plant Physiol 131:101\u0026ndash;110. https://doi.org/10.1016/S0176-1617(87)80271-7\u003c/li\u003e\n\u003cli\u003eLichtenthaler HK, Buschmann C (2001) Extraction of photosynthetic tissues: chlorophylls and carotenoids. Curr Protoc Food Anal Chem 1:F4.2. https://doi.org/10.1002/0471142913.faf0402s01\u003c/li\u003e\n\u003cli\u003eMartin RE, Asner GP, Francis E et al (2018) Remote measurement of canopy water content in giant sequoias (Sequoiadendron giganteum) during drought. For Ecol Manag 419:279\u0026ndash;290. https://doi.org/10.1016/j.foreco.2017.12.002\u003c/li\u003e\n\u003cli\u003eMeireles JE, Cavender-Bares J, Townsend PA et al (2017) SPECTROLAB \u0026ndash; a R package for processing and analysing spectral data. Methods Ecol Evol 8:1215\u0026ndash;1221. https://doi.org/10.1111/2041-210X.12765\u003c/li\u003e\n\u003cli\u003eNeuwirthov\u0026aacute; E, Lhot\u0026aacute;kov\u0026aacute; Z, Albrechtov\u0026aacute; J (2017) The effect of leaf stacking on leaf reflectance and vegetation indices measured by contact probe during the season. Sensors 17:1202. https://doi.org/10.3390/s17061202\u003c/li\u003e\n\u003cli\u003eNoda HM, Motohka T, Murakami K et al (2013) Accurate measurement of optical properties of narrow leaves and conifer needles with a typical integrating sphere and spectroradiometer. Plant Cell Environ 36:1903\u0026ndash;1909. https://doi.org/10.1111/pce.12100\u003c/li\u003e\n\u003cli\u003eOrdway EM, Asner GP, Burslem DFRP et al (2022) Mapping tropical forest functional variation at satellite remote sensing resolutions depends on key traits. Commun Earth Environ 3:64. https://doi.org/10.1038/s43247-022-00564-w\u003c/li\u003e\n\u003cli\u003eOzesmi SL, Bauer ME (2002) Satellite remote sensing of wetlands. Wetl Ecol Manag 10:381\u0026ndash;402. https://doi.org/10.1023/A:1020908432489\u003c/li\u003e\n\u003cli\u003ePoorter H, Niinemets \u0026Uuml;, Poorter L et al (2009) Causes and consequences of variation in leaf mass per area (LMA): a meta-analysis. New Phytol 182:565\u0026ndash;588. https://doi.org/10.1111/j.1469-8137.2009.02830.x\u003c/li\u003e\n\u003cli\u003eProspere K, McLaren K, Wilson B (2014) Plant species discrimination in a tropical wetland using in situ hyperspectral data. Remote Sens 6:8494\u0026ndash;8523. https://doi.org/10.3390/rs6098494\u003c/li\u003e\n\u003cli\u003ePu R (2009) Broadleaf species recognition with in situ hyperspectral data. Int J Remote Sens 30:2759\u0026ndash;2779. https://doi.org/10.1080/01431160802555820\u003c/li\u003e\n\u003cli\u003eRed Natura 2000 Cantabria (n.d.) LICs Litorales: Marismas de Santo\u0026ntilde;a, Victoria y Joyel. https://rednatura2000cantabria.ihcantabria.com/lics/litorales/santona2/\u003c/li\u003e\n\u003cli\u003eRocchini D, Balkenhol N, Carter GA et al (2010) Remotely sensed spectral heterogeneity as a proxy of species diversity: recent advances and open challenges. Ecol Inform 5:318\u0026ndash;329. https://doi.org/10.1016/j.ecoinf.2010.06.001\u003c/li\u003e\n\u003cli\u003eRocchini D, Boyd DS, F\u0026eacute;ret JB et al (2016) Satellite remote sensing to monitor species diversity: potential and pitfalls. Remote Sens Ecol Conserv 2:25\u0026ndash;36. https://doi.org/10.1002/rse2.13\u003c/li\u003e\n\u003cli\u003eRocchini D, Chiarucci A, Loiselle SA (2004) Testing the spectral variation hypothesis by using satellite multispectral images. Acta Oecol 26:117\u0026ndash;120. https://doi.org/10.1016/j.actao.2004.03.008\u003c/li\u003e\n\u003cli\u003eSchmidt KS, Skidmore AK (2003) Spectral discrimination of vegetation types in a coastal wetland. Remote Sens Environ 85:92\u0026ndash;108. https://doi.org/10.1016/S0034-4257(02)00196-7\u003c/li\u003e\n\u003cli\u003eSchneider CA, Rasband WS, Eliceiri KW (2012) NIH Image to ImageJ: 25 years of image analysis. Nat Methods 9:671\u0026ndash;675. https://doi.org/10.1038/nmeth.2089\u003c/li\u003e\n\u003cli\u003eSchneider FD, Morsdorf F, Schmid B et al (2017) Mapping functional diversity from remotely sensed morphological and physiological forest traits. Nat Commun 8:16015. https://doi.org/10.1038/s41467-017-01530-3\u003c/li\u003e\n\u003cli\u003eSchweiger AK (2020) Spectral field campaigns: planning and data collection. In: Cavender-Bares J, Gamon JA, Townsend PA (eds) Remote sensing of plant biodiversity. Springer, Cham, pp 385\u0026ndash;423. https://doi.org/10.1007/978-3-030-33157-3_15\u003c/li\u003e\n\u003cli\u003eSchweiger AK, Cavender-Bares J, Townsend PA et al (2018) Plant spectral diversity integrates functional and phylogenetic components of biodiversity and predicts ecosystem function. Nat Ecol Evol 2:976\u0026ndash;982. https://doi.org/10.1038/s41559-018-0551-1\u003c/li\u003e\n\u003cli\u003eSchweiger AK, Lalibert\u0026eacute; E (2022) Plant beta-diversity across biomes captured by imaging spectroscopy. Nat Commun 13:2767. https://doi.org/10.1038/s41467-022-30369-6\u003c/li\u003e\n\u003cli\u003eSchweiger AK, Sch\u0026uuml;tz M, Risch AC et al (2017) How to predict plant functional types using imaging spectroscopy: linking vegetation community traits, plant functional types and spectral response. Methods Ecol Evol 8:86\u0026ndash;95. https://doi.org/10.1111/2041-210X.12642\u003c/li\u003e\n\u003cli\u003eUstin SL, Gamon JA (2010) Remote sensing of plant functional types. New Phytol 186:795\u0026ndash;816. https://doi.org/10.1111/j.1469-8137.2010.03284.x\u003c/li\u003e\n\u003cli\u003eUstin SL, Gitelson AA, Jacquemoud S et al (2009) Retrieval of foliar information about plant pigment systems from high resolution spectroscopy. Remote Sens Environ 113:S67\u0026ndash;S77. https://doi.org/10.1016/j.rse.2008.10.019\u003c/li\u003e\n\u003cli\u003eVaiphasa C, Skidmore AK, de Boer WF, Vaiphasa T (2007) A hyperspectral band selector for plant species discrimination. ISPRS J Photogramm Remote Sens 62:225\u0026ndash;235. https://doi.org/10.1016/j.isprsjprs.2007.05.006\u003c/li\u003e\n\u003cli\u003eValle M, Pal\u0026agrave; V, Lafon V et al (2015) Mapping estuarine habitats using airborne hyperspectral imagery, with special focus on seagrass meadows. Estuar Coast Shelf Sci 164:433\u0026ndash;442. https://doi.org/10.1016/j.ecss.2015.07.034\u003c/li\u003e\n\u003cli\u003eViolle C, Navas ML, Vile D et al (2007) Let the concept of trait be functional! Oikos 116:882\u0026ndash;892. https://doi.org/10.1111/j.2007.0030-1299.15559.x\u003c/li\u003e\n\u003cli\u003eWang R, Gamon JA (2019) Remote sensing of terrestrial plant biodiversity. Remote Sens Environ 231:111218. https://doi.org/10.1016/j.rse.2019.111218\u003c/li\u003e\n\u003cli\u003eWang Z, Chlus A, Geygan R et al (2020) Foliar functional traits from imaging spectroscopy across biomes in eastern North America. New Phytol 228:494\u0026ndash;511. https://doi.org/10.1111/nph.16711\u003c/li\u003e\n\u003cli\u003eWright IJ, Reich PB, Westoby M et al (2004) The worldwide leaf economics spectrum. Nature 428:821\u0026ndash;827. https://doi.org/10.1038/nature02403\u003c/li\u003e\n\u003cli\u003eYu B, Ostland M, Gong P, Pu R (1999) Penalized discriminant analysis of in situ hyperspectral data for conifer species recognition. IEEE Trans Geosci Remote Sens 37:2569\u0026ndash;2577. https://doi.org/10.1109/36.789651\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Optical traits, functional assembly, wetland vegetation, leaf chemistry, Cantabrian estuaries, imaging spectroscopy","lastPublishedDoi":"10.21203/rs.3.rs-7591763/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7591763/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThere is an urgent need for effective large-scale biodiversity monitoring across ecosystems, given the recent tendency toward global biodiversity loss. The assessment of plant spectral diversity offers a promising approach as it is intrinsically linked to phylogenetic and functional diversity. This study investigates the relationship between taxonomic, functional, and spectral diversity in temperate saltmarsh and seagrass ecosystems in the Gulf of Biscay. Using hyperspectral leaf reflectance data and functional traits from 19 plant species across four estuaries, these three dimensions of biodiversity were compared. The predictive power of spectral data was assessed for estimating biochemical and anatomical traits and species identification. Results reveal significant correlations between functional and spectral diversity, with species sharing similar functional traits exhibiting similar spectral signatures. Spectral diversity is significantly influenced by taxonomic classification, with higher taxonomic levels (e.g., order, class) explaining substantial part of the spectral variation. Spectral regions of 720\u0026ndash;770 nm and 1330\u0026ndash;1380 nm were important for species discrimination, achieving 98% accuracy. Partial least squares regression models successfully estimated functional traits (e.g., water content, carbon, phosphorus) with high precision in these environments. These findings demonstrate that spectral data can effectively capture taxonomic and functional diversity, offering an effective tool for large-scale biodiversity monitoring in estuarine ecosystems. This study underscores the potential of remote sensing to track biodiversity and ecosystem health, providing a foundation for future applications in conservation and management.\u003c/p\u003e","manuscriptTitle":"Assessing sources of variation on leaves reflectance spectra in coastal saltmarshes and seagrasses","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-14 16:51:23","doi":"10.21203/rs.3.rs-7591763/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2025-10-17T09:51:59+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"304622313758556969418665847013220529303","date":"2025-10-06T13:36:30+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-10-01T13:14:11+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-09-29T13:35:33+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-09-24T14:50:46+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-09-23T16:58:48+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2025-09-23T14:26:05+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"dd7a81a3-5c50-4aa2-902f-76aaefc7ac90","owner":[],"postedDate":"October 14th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":56135712,"name":"Biological sciences/Ecology"},{"id":56135713,"name":"Earth and environmental sciences/Ecology"},{"id":56135714,"name":"Earth and environmental sciences/Environmental sciences"},{"id":56135715,"name":"Earth and environmental sciences/Ocean sciences"}],"tags":[],"updatedAt":"2025-10-14T16:51:23+00:00","versionOfRecord":[],"versionCreatedAt":"2025-10-14 16:51:23","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7591763","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7591763","identity":"rs-7591763","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}