Seasonal Spectral Separation of Western Snowberry and Wolfwillow in Grasslands with Field Spectroradiometer and Simulated Multispectral Bands

: Woody plant encroachment (WPE), the expansion of native and non-native trees and shrubs into grasslands, has led to degradation worldwide. In the Canadian prairies, western snowberry and wolfwillow shrubs are common encroachers, whose cover is currently unknown. As the use of remote sensing in grassland monitoring increases, opportunities to detect and map these woody species are enhanced. Therefore, the purpose of this study is to identify the optimal season for detection of the two shrubs, to determine the sensitive wavelengths and bands that allow for their separation, and to investigate differences in separability potential between a hyperspectral and broadband multispectral approach. We do this by using spring, summer, and fall ﬁeld-based spectra of both shrubs for the calculation of spectral separability metrics and for the simulation of broadband spectra. Our results show that the summer offers higher discrimination between the two species, especially when using the red and blue spectral regions and to a lesser extent the green region. The fall season fails to provide signiﬁcant spectral separation along the wavelength spectrum. Moreover, there is no signiﬁcant difference in the results from the hyperspectral or broadband approach. Nevertheless, cross-validation with satellite imagery is needed to conﬁrm the current results.


Introduction
Grasslands are among the largest ecosystems in the world, providing important ecologic and economic services [1]; however, they face multiple threats from climate change and human activity (e.g., conversion to cropland, biodiversity loss, expansion of invasive species), which can lead to their degradation [2].Woody plant encroachment (WPE) has become an important issue for grasslands in recent years.It is related to the expansion of native and non-native trees and shrubs into grasslands [3], and has been connected to changes in primary productivity, nutrient cycling, energy flow, the structure and function of the ecosystem [3]; these all lead to issues in rangeland management and livestock production.There exist various definitions of woody plant encroachment in the literature; except for the term "woody plant encroachment", the terms "woody plant invasion" [4], "woody thicketization" [5], "woody plant expansion" [6], "invasion of woody weed" [7], "xerification" [8], and "invasion of shrubs" [9] are also used.This is because WPE is a global phenomenon, and definitions depend on the precipitation gradient of the region.In particular, WPE occurs in the grasslands of the south-central and southwestern United States (mesquite and creosote brush) [10], North America (juniper) [11], South America (honey locust) [12], Southern Africa (Acacia and Grewia spp.) [13], Australia [14], Mongolia [15], Europe [16], and the Arctic (willow and Alnus spp.) [17].
WPE also takes place in the Canadian prairies, where tree encroachment (e.g., aspen, willow) has received more attention in the literature [18][19][20][21][22][23].For instance, trembling aspen (Populus tremuloides) is the dominant tree species encroaching on grasslands and pastures within the aspen parkland ecoregion in western Canada [24].Other species, such as willow (Salix spp.) and Balsam poplar (Populus balsamifera) are also encroachers, but to a lesser extent.Thorny buffaloberry is an encroaching species in Alberta [25].The most common encroachers that occur throughout several Canadian prairie ecoregions (i.e., aspen parkland, moist mixed grassland, mixed grassland) are western snowberry (Symphoricarpos occidentalis) [26] and wolfwillow (Elaeagnus commutate).Therefore, these two shrub species will be the main focus of this research, since they have been less studied.Moreover, the province of Saskatchewan will be our study area, since it includes the three previously mentioned ecoregions.An example of an encroaching shrub species in the rangelands of southern Saskatchewan is western snowberry, found in the commercial rangelands and provincial pastures of the Grand Coteau region and Weyburn.One can also find western snowberry and wolf willow in Burstall rangelands, the Northeast Swale of Saskatoon, Meewasin Valley, Kernen Prairie, and most of Saskatchewan's southern provincial parks (pers.comm.Mr. Merek Wigness, Dr. Eric Lamb, Dr. Thuan Chu, and pers.observ.).It is understood that shrub encroachment is either already an issue or might become an issue in most of southern Saskatchewan's rangelands.Nevertheless, the cover of these species within the prairies is currently unknown.
It is clear that maintaining grassland health is crucial, especially when food scarcity is estimated to rise, and sustainable management solutions are needed [27].This fits within the United Nations Sustainable Development Goal 15.3 on "Land degradation neutrality".Remote sensing can be used with success to fulfill this aim by mapping the spatiotemporal distribution of various encroaching species with the use of different methods and datasets [28,29]: for instance, to detect two Acacia species from hyperspectral imagery with the use of differences in their phenology in Namibia [30], to classify Prosopis and Vachellia spp. with an object-based approach in Kenya [31], to detect redberry juniper and honey mesquite in north central Texas with spectral contrast of a three-band aerial image [32], to classify three woody invasive species with spectral, textural, and structural features in Chile [33], and to detect six types of woody species with multispectral aerial imagery and LiDAR derived heights in the Netherlands [34].Overall, for species-specific detection, high spatial resolution is necessary.However, the use of high spectral and temporal resolution could compensate for the lack of spatial resolution, and is more preferable for regional and landscape scale mapping.Furthermore, when thinking about the phenological behavior of each woody species of interest, it might be necessary to define the optimal detection timeframe within the growing season for each one.We therefore focus our study on a seasonal spectral approach.Hyperspectral data have been used to detect WPE species due to their wide band range, which allow for the detection of finer spectral differences.In addition, field-based hyperspectral measurements offer the opportunity to fine-tune spaceborne and airborne sensors for larger-scale shrub species mapping by selecting appropriate spectral bands and regions with spectral separability metrics and statistics (e.g., InStability Index, Transformed Divergence, etc.).Afterwards, one can define remote sensing indices that use these bands and apply a broader land cover classification.
To our knowledge, no study has looked at seasonal hyperspectral and multispectral differences between western snowberry and wolfwillow for their potential detection with remotely sensed data, which can facilitate WPE management in the Canadian prairies.Therefore, the main purpose of this study is to derive the seasonal sensitive spectral regions for separation between western snowberry and wolfwillow shrubs in grasslands.Our main objectives are (1) to identify the optimal season for detection of the two shrub cover types, (2) to determine the sensitive wavelengths and bands that allow for their separation, and (3) to investigate differences in separability potential between a hyperspectral and broadband multispectral approach.

Study Area and Case Study Species
The study area is the University of Saskatchewan's Kernen Crop Research Farm (Saskatoon, Canada) in which WPE is an issue in its prairie stand.This area has a native remnant fescue prairie with common mixed prairie species which spans over 1.3 km 2 at about 8 km NE of Saskatoon in Saskatchewan (52 • 10 N, 106 • 33 W, 510 m mean elevation) [35,36] (Figure 1).More information about the study area can be found in [37].This site was chosen as representative of a grassland ecosystem and could be easily accessed during the pandemic restriction.The shrubs that are present consist of western snowberry (Symphoricarpos occidentalis Hook.) [38], wolf-willow (Elaeagnus commutata Bernh.Ex Rydb.), and wild prairie rose (Rosa arkansana) [39,40].In this study, we focus on western snowberry and wolf-willow, which are encroaching species in the site (Figure 2).Western snowberry (Symphoricarpos occidentalis) or otherwise western wolfberry, wolfberry, or buckbrush, occurs throughout most of the southwestern Canadian provinces and northern United States Great Plains [43].It is a deciduous rhizomatous short shrub (0.3-1.5 m) that forms dense colonies between 1 and 200 m [44].It is dominant in Saskatchewan along temporarily flooded shrublands of the northern Great Plains [45], next to water streams, at the base of steep slopes with runoff, and on north or east facing slopes [46].This shrub is common in the mixed-grass prairies.Specifically in Saskatchewan, it was found that western snowberry had lower density in areas with less water availability in comparison to sites with higher water availability [47].This species grows in continental climates with extreme temperatures and light to moderate rainfall [48], and it can survive moderate drought [49].Western snowberry grows on most soil types (e.g., silt, clay, fine sand, rocky substrates, and rich loams) apart from loose sands [50].Further, it is common on mild alkaline to slightly acidic soils [51].Western snowberry does well after disturbance, such as fire [38], and grazing [50].When it encroaches into grasslands, it leads to a decline in forage [52], as it shades out grasses [40], and facilitates the establishment of trees, such as trembling aspen [53].For the Northern Great Plains, fire cessation has led to the encroachment of western snowberry into the mixed-grass prairie [53].Generally, western snowberry can be an increaser in many productive range sites, or a decreaser on other ecosites [54].
Wolfwillow (Elaeagnus commutata) or silverberry belongs to the Elaeagnaceae family and, native to south Canada, it is a deciduous rhizomatous perennial shrub (1-4 m tall) [55].It forms thickets or loose colonies, and one of the ecosystems in which it occurs is the plains grasslands [56].This species can be found along streams, and near springs, while it can grow on different slopes, elevations, aspects, and soil conditions [49].It thrives in loamy soils, but is also found in dry, sandy, and gravel soils [57].Specifically, in the mixedgrass prairie, it is frequently found together with western wheatgrass, needlegrass, and rough fescue [58].In Saskatchewan, it is common on native fescue grasslands [59].In particular, wolfwillow had minor cover in the 1950s, in contrast to its currently wide distribution [60].Wolfwillow adapts well to areas that are disturbed.This is why it is increasing on rangelands that are overgrazed by cattle [58].Wolfwillow has the ability to spread fast through rhizomes [60], but it seems to not recover fast after burning [59].It is shade intolerant [61], justifying its common presence in open vegetation.This species is also resistant to drought, wind, and extreme cold temperatures up to −40 °C [57].Wolfwillow might Western snowberry (Symphoricarpos occidentalis) or otherwise western wolfberry, wolfberry, or buckbrush, occurs throughout most of the southwestern Canadian provinces and northern United States Great Plains [43].It is a deciduous rhizomatous short shrub (0.3-1.5 m) that forms dense colonies between 1 and 200 m [44].It is dominant in Saskatchewan along temporarily flooded shrublands of the northern Great Plains [45], next to water streams, at the base of steep slopes with runoff, and on north or east facing slopes [46].This shrub is common in the mixed-grass prairies.Specifically in Saskatchewan, it was found that western snowberry had lower density in areas with less water availability in comparison to sites with higher water availability [47].This species grows in continental climates with extreme temperatures and light to moderate rainfall [48], and it can survive moderate drought [49].Western snowberry grows on most soil types (e.g., silt, clay, fine sand, rocky substrates, and rich loams) apart from loose sands [50].Further, it is common on mild alkaline to slightly acidic soils [51].Western snowberry does well after disturbance, such as fire [38], and grazing [50].When it encroaches into grasslands, it leads to a decline in forage [52], as it shades out grasses [40], and facilitates the establishment of trees, such as trembling aspen [53].For the Northern Great Plains, fire cessation has led to the encroachment of western snowberry into the mixed-grass prairie [53].Generally, western snowberry can be an increaser in many productive range sites, or a decreaser on other ecosites [54].
Wolfwillow (Elaeagnus commutata) or silverberry belongs to the Elaeagnaceae family and, native to south Canada, it is a deciduous rhizomatous perennial shrub (1-4 m tall) [55].It forms thickets or loose colonies, and one of the ecosystems in which it occurs is the plains grasslands [56].This species can be found along streams, and near springs, while it can grow on different slopes, elevations, aspects, and soil conditions [49].It thrives in loamy soils, but is also found in dry, sandy, and gravel soils [57].Specifically, in the mixed-grass prairie, it is frequently found together with western wheatgrass, needlegrass, and rough fescue [58].In Saskatchewan, it is common on native fescue grasslands [59].In particular, wolfwillow had minor cover in the 1950s, in contrast to its currently wide distribution [60].Wolfwillow adapts well to areas that are disturbed.This is why it is increasing on rangelands that are overgrazed by cattle [58].Wolfwillow has the ability to spread fast through rhizomes [60], but it seems to not recover fast after burning [59].It is shade intolerant [61], justifying its common presence in open vegetation.This species is also resistant to drought, wind, and extreme cold temperatures up to −40 • C [57].Wolfwillow might also increase the available forage for cattle by fixing nitrogen, making it available to other surrounding species and plants [49].Hence, the complete eradication of wolfwillow should be avoided [62].Nevertheless, it seems that areas with wolfwillow are grazed less than half as much as neighboring grass areas that do not have this species [62].

Data Collection
We collected field hyperspectral data three times in the 2020 growing season (spring, summer, and fall).This was done with the use of a spectroradiometer (ASD field-portable FieldSpec Pro, Malvern Panalytical Inc., Boulder, CO, USA) between 10:00 and 14:00 during clear sky sunny weather conditions to maintain a stable ratio between diffuse and incoming solar radiation.The spectroradiometer collects between 350 and 2500 nm with a 1 nm band range.Reference measurements with a Spectralon panel were taken at least every 15 min.Measurements included the collection of shrub endmembers (i.e., wolfwillow and snowberry) that are encroaching in the study area.The spectroradiometer was placed close to the shrubs and at least 10 samples for each shrub species were measured to ensure the plant's spectral variation was captured (i.e., leaves, branches).This allows us to have a spectral signature for ~100% shrub cover of the existing species in the study area.

Data Processing 2.3.1. Calculation of Separability Metrics
For all collected spectral data, we removed the water absorption regions between 1350-1430 nm, 1750-1980 nm, and 2330-2500 nm, which caused noise in the data.Next, we calculated the spectral separability between wolfwillow and western snowberry for each season and wavelength.Various separability metrics calculate how separable two groups are.We used five univariate statistic methods that are provided in the "separability" function of the "spatialEco" package in R [63]: namely, the M-Statistic (M) [64], Bhattacharyya distance (B) [65], Jeffries-Matusita (JM) distance [66], Divergence (D) [67], and Transformed Divergence (TD) [68].These provide discrimination ability of each wavelength without taking into consideration their potential correlation [69].
Before the calculation of these separability metrics, a normality check to the wavelengths of each shrub per season was performed.For that purpose, we used the statistical Shapiro-Wilk test [70], which is considered more powerful than other statistical tests of normality [71], and has been used in similar studies [72].Nevertheless, we also used visual methods (i.e., quantile-quantile plot, density plot), since the test's power might be lower with a small sample size (e.g., below 30).For all seasons, western snowberry was normal for 87-95% of the whole wavelength spectrum (non-normality along the red-edge region and far-SWIR (Shortwave Infrared) in all seasons, and blue region during fall), and for wolfwillow, between 97-100% (non-normality for blue region in spring, and far-SWIR in fall).Even though some wavelengths were partially not normal, we did not consider this as an issue for the current spectral separability analysis, since these wavelengths will be aggregated during broadband simulation and some might not contribute to the spectral separability of the two shrubs.Furthermore, following a non-parametric separability approach for this small number of samples could result in larger biases than the slight deviation from normality for at most 13% of the current dataset.

Thresholding and Selection of Important Wavelength Regions
To select the final wavelength regions capable of separating western snowberry and wolfwillow, we had to identify cut-off thresholds for each of the separability metrics calculated.TD can have values between 0 and 2, with 2 providing maximum separability potential between groups.TD scales the divergence statistic, which looks at the difference between two distributions from their mean values of the log-likelihood ratio [73].Previous research shows that TD provides good separability when it has values above 1.8 [74] or 1.9 [75].Similarly, when TD has values between 1.5 and 1.8 or 1.9, two groups have moderate separation, whereas those with values below 1.5 have poor separation [74,75].
We therefore consider this classification from the literature for our own study (Table 1).JM is the scaled version of the B distance, which measures the divergence between two groups through the calculation of their cosine angle [65].Since JM also follows a scale from 0 to 2, we used the same threshold rules for this metric.Furthermore, when the M-statistic is >1, it is considered that there is good separation [64], so we used this approach, although for this statistic it is hard to define an intermediate separation level, because there is no upper limit.Similarly, one cannot define thresholds for the B and D statistics, since they continue to increase without an upper bound.Therefore, these statistics (B, D) only give a general idea of the important contributing wavelength regions towards the separability of the two shrubs.The final wavelength regions for which both the TD and JM have values above or equal to 1.8 were considered as having good separation.A similar ensemble approach was used for the moderate separability regions.We resampled the seasonal shrub spectra into the broadband Landsat 8, Sentinel-2A, and Sentinel-2B bands with the use of their spectral response functions.In particular, we performed the broadband simulation within the "hsdar" package in R with the use of the "spectralResampling" function [76].To determine if there was a significant difference between the two shrub groups in each season per simulated broadband, we performed multiple two-sampled t-tests assuming unequal variance per band.We used this test since the variances were unequal for some bands (based on results from a two-sided F-test) and due to the unequal sample size between the two shrub groups.We report the results of the t-tests in a table with two levels of adjusted p-value significance: below 0.05, and below 0.01.We further performed the same analysis as in Sections 2.3.1 and 2.3.2 with the broadband spectra.Before running these processes, we performed the Shapiro-Wilk test [70] to check if the assumptions of normality held.The results of the test showed that all simulated broadband for both shrubs had a normal distribution.

Seasonal Spectra of Shrub Species
One can see the average seasonal spectral reflectance for each shrub species in Figure 3a-c.The spectral signatures for both species are openly available through Figshare (https://doi.org/10.6084/m9.figshare.14541597.v1(accessed on 21 June 2021)).The reflectance of wolfwillow in the visible region (350-700 nm) is higher than that of western snowberry for all seasons, and so is the reflectance in the shortwave infrared (SWIR) region (1450-2350 nm).Rather noticeable is the lack of absorption for wolfwillow in the blue region (350-500 nm), which could be explained by the grey-blue appearance of its leaves (Figure 1).Further, the SWIR reflectance for western snowberry decreases slightly from spring to summer, and increases again during fall, whereas it increases throughout the seasons for wolfwillow.These patterns could be related to the seasonal leaf water content of each shrub.For the near infrared (NIR) region, western snowberry has higher reflectance than wolfwillow during spring, after which wolfwillow takes over for the summer and fall.Based on their different spectral signatures, it should be possible to detect each shrub.
The separability metrics can indicate how well each of the wavelengths contribute to this separation.

Hyperspectral Separability of Shrub Species
Seasonal separability between western snowberry and wolfwillow: When looking at the separability metrics for the two shrub species (Figure 4), we can see that separability is higher for the visible wavelengths during spring and summer, while it is higher for the far-SWIR region during fall.When looking at the whole wavelength spectrum, we see that the highest separability values correspond to the summer season, while the lowest to the fall season.In addition, TD and JM have similar results, with JM having lower values for most wavelength regions during spring and fall.Moreover, the M and B metrics show similar responses to the previous two; however, on a different scale, while the D metrics show some dissimilarity in the visible wavelength responses for spring and summer.The pattern similarity between all metrics provides additional reassurance towards trusting the thresholded results of the TD, JM, and M metrics.Based on the set thresholds (Table 1), the highest number of wavelengths that offer good separation are found in the summer (i.e., 17.57% for TD, 14.45% for JM), whereas only 0.12% of wavelengths offer moderate separation with the TD metric during fall (Table S1, Supplementary Materials).This is an initial indication towards the preferable selection of the summer season for the detection of both shrubs.Overall, the TD metric suggests a higher number of wavelengths that offer good separability for the two shrubs compared to the JM metric (13.25% vs. 0.00% in the spring, and 17.57% vs. 14.45% in the summer), whereas the M metric cannot distinguish moderate or good separation.
Wavelength regions sensitive to shrub species separation: To identify the wavelength regions that are important for the detection of each shrub species, we applied an ensemble method, in which the TD-and JM-defined wavelengths that are classified as moderate or good under both metrics are selected (Table 2).The reason for this is that both metrics allow for better interpretation and separation based on threshold establishment due to their upper limit (i.e., 2).The selected wavelength bands belong to certain spectral regions, and those that were below 10 nm wide were removed (e.g., some regions in the SWIR in the summer).The ensemble method could not be applied for the fall season, as none of the metrics included any wavelengths in the moderate or good category.In detail, the spring spectral regions in the blue (409-525 nm) and red-red edge (590-693 nm) offer good separation between the two species.For the summer season, the blue and red-red edge regions continue to offer good separation.In addition, the green region (532-577) is able to moderately separate the two species, since the reflectance of wolfwillow around the green peak is about 0.05 units higher than it was in spring.Furthermore, a narrow wavelength region in the far-SWIR (1981-1991 nm) offers moderate separation.During fall, both species are at the start of senescence, and although we do see some differences in their spectral signature and a few peaks in the separability metrics, these values are not high enough to allow for moderate or good separation.

Broadband Simulation and Shrub Species Spectral Band Difference
Broadband simulation: The mean values for each Landsat 8 and Sentinel-2A band per shrub species and season are presented in Table S2 (Supplementary Materials).The results for Sentinel-2B are very similar and are available in Table S3 (Supplementary Materials).
Shrub species spectral differences (two-sample t-tests): The two-sampled t-test p-values for each Landsat 8 and Sentinel-2A band per shrub species and season are presented in Table 3.Since the Sentinel-2B reflectance values are almost the same as those of Sentinel-2A, we did not perform t-tests on these.Several conclusions can be drawn from these results.First, we can see that the SWIR 1 region is not significantly different between the two species during spring.The same holds for the red edge (RE) 3 and RE 4 band of Sentinel-2A during spring, and the RE 1 band during fall.Season wise, we can see that all bands are significantly different during the summer season, whereas the blue, red and SWIR 1 bands

Discussion and Conclusions
Our results from the hyperspectral metrics, broadband metrics and two-sampled t-tests show that the summer season is the optimal one for the spectral separation of western snowberry and wolfwillow, as it has the highest number of significantly different spectral regions and bands.This is reasonable, since the summer is the peak of the growing season with the highest photosynthetic activity, during which differences between shrub species could become more obvious.For this reason, the summer season has been frequently selected for data acquisition when separating shrub species due to the higher vigor of vegetation in that season [31,33,34,77].Summer months have also shown better discrimination abilities compared to other months-even for separating an evergreen and a deciduous species [29].As for the optimal wavelength regions and bands, both blue and red are important, and more so in the summer.These two regions are influenced by stronger chlorophyll absorption for western snowberry compared to wolfwillow, based on their spectral signature.On the other hand, the green peak (around 550 nm) is similar for both shrubs, and is therefore not useful for classification in the spring.Nevertheless, this spectral region is moderately important during summer, where the reflectance of wolfwillow is significantly higher than that of western snowberry.Lastly, in the far-SWIR, there is moderate separation for a narrow hyperspectral region in spring, which is not represented in the broadband simulations.Although this region is significantly different in all seasons based on the two-sampled t-test, it is not strong enough to reflect its difference in the separability metrics.This region is most possibly related to the differences in water and moisture absorption between the two species.
Overall, when looking at the differences between the hyperspectral and broadband results for the separation of the two shrubs, we notice that the results are almost the same, except for a narrow region in the far-SWIR, which is not included in the broadband results.This leads us to the conclusion that hyperspectral data would not really improve the classification results for our specific study purposes, and that use of Landsat 8 or Sentinel-2 data would suffice.In addition, the increased number of spectral bands that Sentinel-2 data provide do not offer enhanced detection capabilities, since the NIR region that includes the red-edge bands is not one of the sensitive regions for western snowberry and wolfwillow classification throughout the seasons.
However, we must point out that our current simulated broadband results represent the leaf/branch scale and not the canopy scale.The reflectance properties of the two shrubs could be different at that scale due to canopy architecture, such as leaf angle distribution, density, biomass, and leaf area index, in which shadows and occlusions also play a role.In addition, these simulations do not represent satellite data conditions, which are strongly affected by the atmosphere, and which capture the land surface at a broader scale, in which topography also has a significant role.Furthermore, since Landsat and Sentinel-2 data capture the surface at a broader scale (10-30 m), each image pixel is usually a mixture of different land cover types (e.g., woody plants, grass, bare ground, rock).This is especially the case when WPE is at an early stage.Overall, grasslands can undergo different WPE stages (i.e., early, moderate, or advanced), resulting in different woody cover within an image pixel [78].A field-based study showed that the earliest WPE that could be identified was when it reached between 10% and 25% of an image pixel [37].However, more research with remotely sensed imagery is needed to verify this result.Nevertheless, even with mixed pixels, there exists a number of spectral unmixing techniques that could enhance WPE species specific mapping with coarse resolution pixels [79].With this technique, each pixel gets assigned to a fraction of its land covers, which are defined by endmembers.Two endmember classes that could be used for that purpose are the spectral signatures of western snowberry and wolfwillow that were used in this study.For the above reasons, the optimal season and bands detected in the current study for separation between the two woody shrubs might not coincide with their actual detection on the landscape.Therefore, the current results should be cross-validated with satellite-based remote sensing data, such as Landsat 8 and Sentinel-2.We plan to implement this in future research that will establish specific broadband multispectral indices optimally correlated with the two shrub species of this study, and with research that will investigate potential improvements in their detection with spectral unmixing techniques.
Author Contributions: Conceptualization, I.S. and X.G.; methodology, I.S.; writing-original draft preparation, I.S.; writing-review and editing, X.G.; visualization, I.S.; supervision, X.G.All authors have read and agreed to the published version of the manuscript.

Fi gure 1 .
Location of Kernen Prairie within the provincial boundaries of Saskatchewan (SK), Canada (upper figure), and on a Sentinel-2 image of 11 July 2020 (lower figure).Source of Canadian Provincial Boundaries: Statistics Canada (Open-Government License-Canada) [41], source of Sentinel-2 image: ESA ('Copernicus Service information 2020' for Copernicus Service Information) [42], source of Kernen Prairie boundary layer: Department of Plant Science, University of Saskatchewan (Dr.Eric Lamb).

Table 2 .
Seasonal wavelength bands and spectral regions that offer moderate and good separation between western snowberry and wolfwillow based on thresholds established in Table1(B = Blue, R = Red, RE = Red-Edge, SWIR = Shortwave infrared).