INTRONDUCTION
Rationale
Non-native species pose a significant threat to agricultural and native prairie ecosystems globally, as they negatively impact native species populations by outcompeting them for resources Defined as species that do not originate from a specific ecosystem and can reproduce independently, invasive species are known to cause environmental harm (Pejchar and Mooney, 2009) Additionally, climate change exacerbates the challenges posed by these non-native species, as predicted shifts in rainfall, temperature, nutrient levels, and soil disturbance may heighten habitat vulnerability to invasive plants (Hufnagel and Garamvửlgyi).
When non-native plants invade the environment, they are able to overcome native plants through direct or indirect competition (Robert et al.,
A study by Driscoll et al (2014) highlights that environmental weeds, which are non-native plants, invade natural areas and pose significant threats to native flora These invasive species can disrupt ecosystem functions and incur management costs that amount to billions of dollars annually.
Crested Wheatgrass (Agropyron cristatum) occupies 6 to 11 million hectares of grassland in the North American Great Plains, posing a significant threat to native ecosystems that are vital for biodiversity This invasive species adversely affects both animal and plant life, disrupting local habitats and impacting nutrition and energy sources within these ecosystems.
Non-native species pose a significant threat to agricultural and native prairie communities globally, as they negatively impact native species populations by outcompeting them Invasive species, defined as those that do not originate from a specific ecosystem and can self-propagate, have been shown to cause environmental harm (Pejchar and Mooney, 2009) Additionally, climate change exacerbates the effects of non-native species, with predicted shifts in rainfall, temperature, nutrient levels, and soil disturbance potentially increasing habitat vulnerability to these invasive plants (Hufnagel and Garamvửlgyi).
When non-native plants invade the environment, they are able to overcome native plants through direct or indirect competition (Robert et al.,
A study by Driscoll et al (2014) highlights that environmental weeds, which are non-native plants, invade natural areas and pose significant threats to native vegetation These invasive species disrupt ecosystem functions and incur management costs amounting to billions of dollars annually.
Crested Wheatgrass (Agropyron cristatum) occupies 6 to 11 million hectares of grassland in the North American Great Plains, posing a significant threat to native ecosystems that are crucial for biodiversity This invasive species adversely affects both plant and animal life, impacting their nutrition and energy sources.
Objectives
This study aims to evaluate the effectiveness of remote sensing in identifying non-native plants within prairie ecosystems To date, there has been no successful method developed for detecting these invasive species in mixed grassland environments using medium-resolution imagery.
This study aims to assist resource managers by investigating the spectral characteristics of crested wheatgrass using ground hyperspectral data at an affordable cost The key objectives include identifying the most effective vegetation indices for distinguishing crested wheatgrass and examining its behavior under various grazing regimes.
LITERATURE REVIEW
Remote sensing technique
Remote sensing involves gathering information about an object or phenomenon without direct physical contact, contrasting with local observation In contemporary applications, this term primarily refers to aerial sensor technologies that detect and classify objects on Earth, including those in the atmosphere and oceans, through signal transmission, such as electromagnetic radiation Remote sensing can be categorized into active remote sensing, where the initial signal is emitted from aircraft or satellites, and passive remote sensing, which relies on natural sources like sunlight to record information.
Figure 1 Remote sensing process (researchgate.net)
Recent advancements in image synthesis techniques enable the simultaneous detection of structural and functional properties of invasive plants using image spectrometry and LiDAR technology (Huang and Gregory, 2009) A study conducted by Ustin et al utilized high spectral resolution (224 bands at 10 nm) and spatial resolution (approximately 4 m) to create maps of various invasive species, such as eggplant, jubata, fennel, and giant reed, across diverse habitats at Camp Pendleton and Vandenberg Air Force Base in California, employing AVIRIS data from 2002.
The development of airborne and spaceborne imaging systems has enabled researchers to assess the physical and chemical properties of Earth's and planetary surfaces These technologies facilitate the monitoring of continental plant biomass, global water migration, surface heat flux, and the deformation of the Earth's surface caused by both human activities and natural processes, among various other applications (Beasley and Barnhart, 2017).
Zhou and Guo conducted a study in the northern mixed-grass prairie of Canada, utilizing SPOT-5 imagery to identify the invasion of Crested Wheatgrass Their findings indicated that a single-date SPOT-5 image with a 10-meter resolution effectively differentiates Crested Wheatgrass from native species in mixed grasslands (2007).
Using remote sensing can be a cost effective strategy to detect invasive weeds Historically, invasive species have often been identified by natural resource managers and volunteers who manually scouted (Shaw, 2005)
Hyperspectral data involves the detailed analysis of electromagnetic radiation's reflection, transmission, or absorption at high spectral resolution (Lukas et al., 2018) Characterized by large dimensions, mass, and numerous contiguous spectral channels, hyperspectral remote sensing images provide extensive information about the Earth's surface, including insights into space, radiation, and spectrum (Zebin et al., 2016) This rich data is invaluable for researchers in their efforts to analyze, process, and monitor information related to the Earth's surface.
Hyperspectral remote sensing captures ground surface reflection images across hundreds of wavelengths, ranging from 0.4 to 2.5 micrometers, with a spatial resolution of 10 to 30 meters This advanced technology is widely used in environmental monitoring, mineral exploration, and mining The key advantage of hyperspectral imagery is its ability to distinguish various objects and terrain types through their unique spectral signatures.
Hyperspectral imaging captures extensive spectral information, enabling enhanced characterization, identification, and classification of surfaces with greater accuracy (Wang, 2019) These images are structured as volumetric cubes composed of numerous spatial images, or spectral bands, each reflecting the response of ground objects at specific wavelengths (Cheung and Antonio, 2009) Due to the significant amount of autocorrelation data present, techniques like principal component analysis (PCA) are commonly employed to address this challenge (Bartold, 2008; Olesiuk and Zagajewski, 2008; Zagajewski, 2010).
Reflectance, defined as the ratio of energy reflected to the total energy absorbed by an object, is expressed as a percentage For instance, an object appears green because it reflects only the green wavelengths of the visible spectrum, while blue objects absorb all other light except for blue Spectral reflectance measures the reflected energy relative to incident radiation as a function of wavelength (Jain and Singh, 2003) Typically, reflectance is measured using a specular reflectometer The reflectance of solar mirrors significantly affects thermal efficiency, with a 1% decrease in reflectivity resulting in a nearly 1% drop in efficiency (Herrmann et al., 2017) Reflectance can be categorized based on the direction of reflection—specular, diffuse, or hemispherical—as well as by wavelength, such as spectral or solar-weighted reflectance (García et al., 2017).
Wavelengths play a crucial role in physical phenomena, while frequencies are associated with energy and radiation transmission (Anne, 2019) Visible wavelengths, ranging from 0.40 to 0.67 μm, interact with the outer electronic shells of transition metal ions in pigments, whereas thermal infrared wavelengths (4 – 14 μm) emitted from the Earth's surface engage with ions bound in crystalline lattices (Vincent, 2015) Additionally, the shortest wavelengths, found in the gamma, X-ray, and ultraviolet (UV) regions of the electromagnetic spectrum, highlight the diverse interactions within this spectrum (William and Adriano, 2017).
Vegetation indices (VIs) derived from canopy observations using remote sensing are effective algorithms for assessing plant cover, vitality, and growth dynamics These indices have been extensively utilized in remote sensing applications across various satellite and aerial platforms, with recent advancements in unmanned aerial vehicles (UAVs) enhancing their capabilities (Xue and Su, 2017) VIs are calculated from multispectral images (MSI) captured from the air and are commonly employed to quantify crop health, moisture, and nutrient content (Khan et al., 2018).
Crested Wheatgrass characteristics
Crested Wheatgrass is a resilient grass characterized by its numerous basic leaves, stems, and an extensive root system, making it highly resistant to cold, drought, and grazing It serves as a significant source of high-quality forage, thriving in soils from light sandy to medium clay types that facilitate rainfall absorption and possess good water retention However, this species tends to outcompete native vegetation, often forming nearly monotypic stands.
Crested Wheatgrass (Agropyron cristatum L.), a member of the Poaceae family, is a cold-season grass known for its extensive root system As the first perennial grass to emerge in spring, it offers high-quality forage during its early growth stage However, it exhibits lower below-ground biomass compared to native prairie species, leading to reduced root detritus and exudates that contribute to soil organic matter formation.
Crested Wheatgrass negatively impacts species diversity, leading to a decline in native plants in abandoned fields where it was once planted (Christian and Wilson, 1999) This invasive species can adversely affect local birds and wildlife (Reynold and Trost, 1980) Although early grazing of Crested Wheatgrass provides forage until native grasses are ready for grazing in early summer (Dwyer and Owen, 1984), it results in low biodiversity and altered ecosystem functions compared to native prairies (Christian and Wilson, 1999) Furthermore, Crested Wheatgrass is classified as a noxious weed in many regions worldwide and is recognized as a significant invasive threat in the United States and Canada (Diana, 2018).
Crested wheatgrass is more likely to invade disturbed areas compared to native prairie, with a relatively low invasion rate into native habitats (Coffin et al., 1996) As crested wheatgrass cover increases, the presence of native grasses decreases (Heidinga and Wilson, 2002) This species efficiently absorbs moisture at low temperatures and outcompetes native plants for phosphorus, hindering their establishment in crested sites (Lesica and DeLuca, 1996) Additionally, crested wheatgrass creates a ground layer with more exposed soil than native prairie, raising concerns about potential soil erosion (Lesica and DeLuca, 1996).
METHODS
Study area
My study area is Saskatchewan Landing Provincial Park (SLPP) (50°38′
Stewart Valley, situated approximately 50 kilometers north of Swift Current in southwest Saskatchewan, can be accessed by traveling north on Highway 4.
Figure 2 Location of Saskatchewan Landing Provincial Park
Figure 3 Saskatchewan Landing Provincial Park Grassland
Photo taken by: Dr Xulin Guo
Saskatchewan Landing Provincial Park spans 5,735 hectares at the western end of Lake Diefenbaker and features a mixed grassland ecology, representing the largest undisturbed primary prairie area in the provincial park system This park is significant to Saskatchewan for its anthropological, palaeontological, archaeological, and ecological contributions.
Data collection
Data collection utilized several instruments, including the Analytical Spectral Devices (ASD) FieldSpec 3 Pro System for measuring canopy reflectance and wavelength data during clear days from 9:30 to 14:30 GPS indices were recorded using a Garmin GPSMap 76S and cell phones for precise locational positioning at each site However, due to a malfunction of the Garmin GPSMap 76S, cell phones were used as a substitute for GPS identification.
Figure 4 Sampling design in each site
Field visits were conducted for this study to collect essential biophysical data, including the proportion of top layer cover, dominant species, and grazing activity Data collection took place from August 26th to August.
On the 28th, Dr Xulin Guo, Mr Yunpei Lu, and Ms Thuy Doan conducted a study at Saskatchewan Landing Provincial Park under cloudy and windy conditions They divided the grassland into ten sites, labeled from 9SL1 to 9SL10, with each site containing 20 quadrats Each quadrat was spaced 5 meters apart, measuring 50 x 50 cm for cover and 20 x 50 cm for biomass.
Quadrats were positioned vertically from the center and extended in four directions to effectively illustrate the variations in Crested Wheatgrass and the physiological data gathered for each standard plot A stratified random sampling method was employed to assess indicators of Crested Wheatgrass, Alfalfa, Juniper, Shrub, and Sage Alfalfa samples were collected in Alberta, while the other species were gathered from Saskatchewan Landing Provincial Park.
Figure 5 Designing the sampling area
Photo taken by: Dr Xulin Guo
Alfalfa is an introduced species, while Juniper, Sage, and Shrub are classified as non-grass species, and Crested Wheatgrass is identified as a non-native weed Data collection was conducted in a clockwise direction, focusing on areas where Crested Wheatgrass and the four non-grass species were predominant to gather relevant information.
Figure 6 Juniper in SLPP Figure 7 Alfafa in Alberta Photo taken by: Dr Xulin Guo Photo taken by: Xulin Guo
Figure 8 Sage in SLPP Figure 9 Shrub in SLPP Photo taken by: Dr Xulin Guo Photo taken by: Xulin Guo
Figure 10 Crested Wheatgrass and grazing activity in SLPP
Photo taken by: Dr Xulin Guo
In this study, ground hyperspectral data was collected using the Analytical Spectral Devices (ASD) FieldSpec 3 Pro System Research by Daniel highlights the use of spectroradiometers to measure plant reflectance, which aids in assessing tree stress levels Additionally, the reflection spectrum allows for effective monitoring of crop nitrogen and water levels The ASD FieldSpec Pro JR offers a comprehensive spectral range from 350 nm, enhancing its utility in agricultural assessments.
2500 nm It uses ASD’s FieldSpec Pro RS3 software (2003).
Pre – processing
Hyperspectral data analysis was conducted using Excel 2016 and IBM SPSS Statistics 22 to examine the spectral characteristics of crested wheatgrass This investigation required data from two Excel tables, with the first table containing wavelength values and reflectance measurements for each research site.
The "Spectral_SLPP_Data" table includes wavelength values and reflectance data for various species, including Alfalfa, Crested Wheatgrass, Juniper, Sage, and Shrub This data, referred to as the "Signature data" table, was originally organized into five separate Excel tables for Crested Wheatgrass and four non-grass species.
To streamline chart creation, I compiled a table that includes the Wavelength and Reflectance values for each species, such as Alfalfa, Crested Wheatgrass, Shrub, Sage, and Juniper Upon analyzing the distribution, I observed that certain sections lacked distinct differences among the grasses due to primary water vapor causing reflected noise To enhance chart quality, it is essential to eliminate three specific wavelength ranges that contribute to this noise: 1260 to 1560 nm, 1760 to 1960 nm, and 2450 to 2500 nm.
Data processing
This study primarily utilized remote sensing hyperspectral data for analysis According to Lukas et al (May, 2018), remote sensing hyperspectral is an effective tool applicable across various fields, including ecology, geology, analytical chemistry, and medical research.
For the "Spectral_SLPP_Data” table, I calculated the average for 20 quadrats at each wavelength value from 350 to 2500 of each research site Finally,
I have developed a new table of values that includes a wavelength column alongside 10 columns representing 10 research sites, labeled from 9SL1 to 9SL10 Each site contains average reflectance values ranging from wavelengths of 350 to 2500 nanometers This data has been utilized to create a graph illustrating the spectral properties across these 10 research sites Additionally, similar to the Signature data, there are three ranges of reflection noise caused by primary water vapor, specifically between 1260 to 1560 nanometers and 1760 nanometers.
1960 and 2450 - 2500) and I removed them The graphs I obtained are more convenient to compare the difference in spectral characteristics between sites
The study aimed to differentiate the distribution of Crested wheatgrass from four non-grass species using five key vegetation indices: NDVI, RDVI, SAVI, MSAVI, and PSRI These indices leverage the light spectrum for plant remote sensing, which includes the ultraviolet region (10-380 nm), visible spectra (blue: 450–495 nm, green: 495−570 nm, red: 620–750 nm), and near to mid-infrared bands (850–1700 nm) (Xue and Su, 2017).
Table 1 List of Hyperspectral Vegetation Indices have been used
R: original reflectance of red absorption region; 𝑅 800 : mean reflectance at 760- 900nm; 𝑅 670 : mean reflectance at 630-690nm; 𝑅 𝐺𝑟𝑒𝑒𝑛 : mean reflectance at wavelength 495-570nm
In the study, the author utilized R800, R670, and R500 to denote RNIR, RRed, and RGreen, respectively The analysis of the "Signature_Data" table revealed numerous reflectance values for R800, R670, and R500 across both the CW and four non-grass types Following the extraction of these reflectances, the average values for each wavelength—800, 670, and 500—were computed for each species Subsequently, the author applied the formulas from Table 1 to derive the Vegetation Indices.
Table 2 Average reflectances at wavelength 800, 670 and 500 of 5 grass types
Wheatgrass Juniper Alfafa Sage Shrub
I calculated the average reflectance of 20 quadrats across wavelengths ranging from 350 to 2500 nm for each site (9SL1 to 9SL10) and compiled the results into a single table Additionally, I utilized the equations from Table 1 to determine the Vegetation Indices for each study site.
The Vegetation Indices results obtained from the previous calculations are preliminary and primarily serve for visual observation To draw more accurate conclusions, a statistical analysis is necessary.
I did Hypothesis testing namely Tukey Post Hoc Test by IBM SPSS software
Hypothesis testing utilizes statistical methods to evaluate the likelihood of a hypothesis being true, providing a structured approach for decision-making based on sample evidence This technique helps assess the reliability of extrapolating findings from a sample to a larger population Among various methods, the Tukey test stands out for its effectiveness in detecting differences in pairwise comparisons, offering a less conservative approach To ensure accurate results, it is crucial to maintain an appropriate sample size, which minimizes standard errors and enhances the chances of rejecting the null hypothesis.
In this study, I utilized the Tukey Post Hoc Test to assess the predictions of Vegetation Indices across five species and ten research sites This method required me to import all reflectance values (R800, R670, and R500) into SPSS, as it does not rely on average reflectance values Additionally, I recalculated the Vegetation Indices for each reflectance value using Excel For the CW and four non-grass types, I assigned specific codes for data input in SPSS (Alfalfa: 1, Crested Wheatgrass: 2, Juniper: 3, Sage: 4, Shrub: 5) and followed the same procedure for the data sheet of the ten research sites.
RESULTS AND DISCUSSION
Spectral characteristics of Crested Wheatgrass
Figure 11 Spectral characteristic of CW and 4 non-grass types in SLPP
The graph in Figure 11 depicts the spectral characteristics of Crested Wheatgrass alongside four non-grass species in SLPP, with the x-axis representing wavelength (nm) and the y-axis indicating reflectance The study primarily focuses on Crested Wheatgrass, highlighted by gray curves, which reveals its spectral properties in comparison to other species Notably, the most significant differences among the five species occur in the near-infrared (NIR) region, although Crested Wheatgrass does not distinctly stand out within the wavelength range of 750nm to 1100nm In contrast, Crested Wheatgrass is clearly differentiated from the other species in the shortwave infrared (SWIR) region, particularly between 1550nm and 1770nm However, in the SWIR range of 2000nm to 2250nm, Crested Wheatgrass is nearly overshadowed by the other four species.
Detecting invasive species is complicated by spatial heterogeneity, particularly in mixed-grass prairies where Crested Wheatgrass typically invades areas with optimal conditions, avoiding saline or poor soils Factors such as shadows, soil reflectance, litter amount, bare soil, and moss presence significantly influence reflectance values Crested Wheatgrass exhibits higher levels of bare soil and standing dead material compared to native grasslands, which can obscure spectral signatures during classification Additionally, plant litter affects vegetation indices due to its visible and near-infrared spectroscopic responses Research by Zhou and Guo has shown that the spectral reflectance characteristics of Crested Wheatgrass and native grasslands reveal differences in their phenology and composition.
The best vegetation indices to differentiate Crested Wheatgrass
After calculating the average reflectance for R800, R670 and R500 of CW and
In the analysis of four non-grass vegetation types, the corresponding values were integrated into the Vegetation Indices equations outlined in Table 1 This method enabled visual predictions to determine which indicator most effectively differentiates between the various types of cover within the study area.
Table 3 Vegetation Indices of CW & 4 non-grass types
Indices Alfafa CW Juniper Sage Shrub
Figure 12 Five Vegetation Indices of CW & 4 non-grass types showed on graph
The graph (Figure 12) compares five Vegetation Indices, with the x-axis representing Crested Wheat and four non-grass types, while the y-axis displays the corresponding Vegetation Index values Notably, MSAVI and NDVI emerge as the most effective indices, demonstrating significant differences among the five species Additionally, SAVI and RDVI reveal differences among the grass types, although PSRI can be disregarded due to its minimal variance However, these observations are primarily visual predictions To accurately determine the optimal Vegetation Index for Crested Wheatgrass, the results of the Tukey Post Hoc Test conducted with SPSS software were utilized.
Ha: Not all group means are equal Significance level: 0.05
Sum of Squares df Mean Square F Sig
The One-way ANOVA analysis yielded a p-value of 0.000, leading to the rejection of the null hypothesis As noted in section 4.3, I designated Crested Wheatgrass as number 2, thus my focus is directed towards the row associated with this number in the "(I) Code" column.
Table 4 Comparisons of NDVI, RDVI, SAVI and MSAVI
(CW and 4 non-grass types) Vegetation
Alfafa CW, Juniper, Sage, Shrub
CW Alfafa, Juniper, Sage, Shrub Juniper Alfafa, CW, Sage, Shrub Sage Alfafa, CW, Juniper, Shrub Shrub Alfafa, CW, Juniper, Sage
Alfafa CW, Juniper, Sage, Shrub
CW Alfafa, Juniper, Shrub Sage
Juniper Alfafa, CW, Sage Shrub
Alfafa CW, Juniper, Sage, Shrub
CW Alfafa, Juniper, Shrub Sage
Juniper Alfafa, CW, Sage Shrub
Alfafa CW, Juniper, Sage, Shrub
CW Alfafa, Juniper, Shrub Sage
Juniper Alfafa, CW, Sage Shrub
A new table has been created to enhance observation of the data Table 4 demonstrates that the NDVI is the most effective index for distinguishing Crested Wheatgrass, as evidenced by all p-values being below 0.05 (see Appendix A) We are 95% confident that NDVI outperforms other Vegetation Indices in differentiating Crested Wheatgrass.
Table 5 Boxplots for 4 Vegetation Indices of CW and 4 non-grass types
A study by Zhou and Guo (2007) highlights the effectiveness of a new vegetation index, ExpNDMI, which enhances spectral separability between Crested Wheatgrass and native grasslands, leading to improved classification accuracy In my research, I utilized five vegetation indices to differentiate Crested Wheatgrass, while Rashmi et al focused on distinguishing the invasive grass Lantana from other vegetation classes This research demonstrates the potential of advanced vegetation indices in accurately identifying and classifying various grass species.
SAVI (Soil Adjusted Vegetation Index) is most suitable for distinguishing
In 2009, Lantana was identified as the leading species, followed closely by Perpendicular Vegetation Index-3, in the optimal bio-window Satellite-derived spectral vegetation indices (VIs) are widely utilized for estimating leaf chlorophyll content, although most have been developed and validated primarily on broadleaf species (Croft, Chen, and Zhang, 2013).
The behavior of Crested Wheatgrass in different Grazing regimes
Author did the same steps with hyperspectral data of 10 research sites I got the result below This result is also for visual prediction about the trend of
Vegetation Indices in 10 sites study
Table 6 Vegetation Indices for 10 sites study
Indices 9SL1 9SL2 9SL3 9SL4 9SL5 9SL6 9SL7 9SL8 9SL9 9SL10
NDVI 0.316164 0.336257 0.315769 0.377607 0.411943 0.424668 0.421518 0.483466 0.412266 0.351555 RDVI 0.169778 0.169995 0.180566 0.206328 0.21146 0.229145 0.213704 0.251381 0.211546 0.180104 SAVI 0.194834 0.170613 0.187282 0.211767 0.213256 0.234424 0.214676 0.254508 0.213318 0.181522 MSAVI 0.671209 0.643468 0.662951 0.68367 0.682503 0.704283 0.683314 0.721648 0.682544 0.653642 PSRI 0.18888 0.149886 0.176666 0.141778 0.131154 0.116871 0.125469 0.120778 0.134004 0.139605
Figure 13 Vegetation Indices of 10 sites study indicated on graph
The graph above (Figure 13) shows the Vegetation Indices trending for
The analysis of ten study sites reveals no significant differences in Vegetation Indices across these locations Consequently, I propose utilizing Vegetation Indices to examine the behavior of Crested Wheatgrass under varying grazing regimes To ensure accurate conclusions, I applied the Tukey Post Hoc Test, forming the basis for our hypotheses.
H0: àNDVI = àRDVI = àMSAVI = àSAVI = àPSRI
Ha: Not all group means are equal
The ANOVA table indicates a p-value of 0.000, leading to the rejection of the null hypothesis Analysis reveals that Crested Wheatgrass is predominant in both grazing area 9SL1 and non-grazing area 9SL5 The Tukey Post Hoc Test may demonstrate significant differences between these sites and others, suggesting that Vegetation Indices can effectively assess the behavior of Crested Wheatgrass under varying conditions Additionally, I prepared the necessary coding for data input into SPSS.
Table 7 Grazing & Dominant species information of 10 sites study in SLPP
9SL1 Y western wheatgrass, crested wheatgrass, northern wheatgrass, blue grama 9SL2 Y prairie sage, western wheatgrass, needle and thread
9SL3 Y western wheatgrass, blue grama, forb
9SL4 N western wheatgrass, blue grama, prairie sage, forb
9SL5 N needle and thread, prairie sage, crested wheatgrass, bluegrama, smooth brome 9SL6 Y western wheatgrass, blue grama, prairie sage, forb
9SL8 N western wheatgrass, kentucky blue grass, needle and thread
9SL9 N blue grama, western wheatgrass, prairie sage, sage brush
9SL10 N blue grama, western wheatgrass, smooth brome
The analysis of Multiple Comparisons Tables from a study across 10 sites indicates that there are no statistically significant differences in Vegetation Indices between 9SL1 and 9SL5, as several p-values exceed the significance threshold of 0.05 Consequently, it can be concluded that the five Vegetation Indices are insufficient for examining the behavior of Crested Wheatgrass under varying grazing regimes Additionally, boxplots illustrating the Vegetation Indices across the 10 research sites were created.
Table 8 Boxplots for 5 Vegetation Indices of 10 sites study
A study by Yang and Guo revealed significant differences in canopy height and the ratio of photosynthetically active vegetation cover to non-photosynthetically active vegetation cover (PV/NPV) between grazing and ungrazing areas, while canopy index (CI) did not indicate significant differences among grazing methods Research by Karnieli et al identified the Enhanced Vegetation Index (EVI) as the primary botanical indicator for examining the invasive mechanisms of non-native species in Mongolia's steppe biome, noting that grazing areas exhibited notably higher EVI values compared to ungrazing areas Furthermore, unattractive species have been found to invade grazing areas, displacing native grasses Vegetation indices (VI), calculated from reflectance data in red and near-infrared wavelengths, can be derived from biophysical parameters such as ground biomass and Leaf Area Index (LAI) Flynn et al (2008) demonstrated a strong correlation between the Normalized Difference Vegetation Index (NDVI), calculated using a ground-based sensor, and biomass.
CONCLUSION
Limitation and futher study
This study primarily relied on ground hyperspectral data and did not explore the depth of Crested Wheatgrass spectral characteristics, limiting the analysis to only five Vegetation Indices, with NDVI identified as the most effective for differentiation However, the analysis was simplistic and failed to assess the impact of Crested Wheatgrass in grazing versus non-grazing areas, suggesting that Vegetation Indices may not be the best method for examining its behavior under different grazing regimes Additionally, the Tukey Post Hoc Test may not have been appropriate for this research, and the study faced limitations regarding the location of study plots and accuracy evaluation points.
In the future, I aim to enhance this study by creating a detailed map of Crested Wheatgrass distribution at SLPP I propose utilizing Unmanned Aerial Vehicles (UAVs) to effectively detect invasive grasses in Canada's mixed-grass prairie Additionally, I will investigate various methods to assess the impact of Crested Wheatgrass in both grazing and non-grazing areas.
Conclusion
This study found that the spectral characteristics of Crested Wheatgrass can be distinguished from four non-grass types across three specific ranges: 750nm – 1100nm, 1550nm -1770nm, and 2000nm -2250nm While NDVI emerged as the most effective Vegetation Index for identifying Crested Wheatgrass, it is not the best method for assessing its behavior under various grazing regimes Limitations such as time constraints, data collection, and analysis were noted, suggesting that alternative Vegetation Indices or multi-date approaches may yield better insights and should be explored in future research Given the ongoing issue of exotic species invasions in many Canadian ecosystems, further investigation in this area is essential.
A Fernández-García et al., 2017 Mirrors The Performance of Concentrated Solar
Power (CSP) Systems, Pages 67-98 DOI: https://doi.org/10.1016/B978-0-
A Karnielia, Y Bayarjargala, M Bayasgalanb, B Mandakhc , Ch Dugarjavc , J
Burgheimera, S Khudulmurb, S N Bazhad, and P D Gunind, 2013 Do vegetation indices provide a reliable indication of vegetation degradation?
A case study in the Mongolian pastures International Journal of Remote Sensing Vol 34, No 17, 6243–6262 DOI: 10.1080/01431161.2013.793865
Anne M Hofmeister, 2019 Chapter 2 - Macroscopic Analysis of the Flow of
Energy Into and Through Matter From Spectroscopic Measurements and Electromagnetic Theory Measurements, Mechanisms, and Models of Heat
Transport Pages 35-73 DOI: https://doi.org/10.1016/B978-0-12-809981- 0.00002-4
Bartold M., 2008 Klasyfikacja drzewostanów na obrazie satelitarnym Hyperion
Beasley and Barnhart, 2017 Remote sensing The Leading Edge Vol.36(1), pp.11-
57 DOI: https://doi-org.cyber.usask.ca/10.1190/tle36010011.1
Cho-ying Huang and Gregory P Asner, 2009 Applications of Remote Sensing to
Alien Invasive Plant Studies Sensors (Basel) 9(6): 4869–4889 DOI:
Research by Christian and Wilson (1999) highlights the long-term ecosystem effects of an introduced grass species in the northern Great Plains, emphasizing its significant impact on local biodiversity and ecosystem dynamics Additionally, Coffin, Lauenroth, and Burke (1996) studied the recovery of vegetation in a semiarid grassland, finding that it took 53 years for the ecosystem to rebound after disturbance These studies underscore the importance of understanding both the introduction of non-native species and the resilience of native ecosystems over extended periods.
Curran, P.J., 1985 Principles of Remote Sensing, Longman Group Ltd,England D’Antonio, C M and Vitousek, P M 1992 Biological invasions by exotic grasses, the grass/fire cycle, and global change Annu Rev Ecol Syst 23:
Daniel Dallon, 2003 Comparison of the Analytical Spectral Devices FieldSpec
Pro JR and the Apogee/StellarNet Model SPEC-PAR/NIR Spectroradiometers Techniques and Instruments Source: https://digitalcommons.usu.edu/cgi/viewcontent.cgi?article04&context
Darcy Christopher Henderson, 2005 Ecology and Management of Crested
Wheatgrass Invasion Retrieved from: https://grasslandrestorationforum.ca/wp- content/uploads/2017/10/Henderson-Ecology-and-Management-of-
DeLuca and Lesica, 1996 Long-term harmful effects of crested wheatgrass on
The study by Dergousoff and Chilton (2016) investigates tick populations on small mammals in southern Saskatchewan, Canada, highlighting the presence of Rickettsia peacockii in Dermacentor andersoni nymphs Published in the Canadian Entomologist, this research contributes to the understanding of tick-borne diseases in Great Plains grassland ecosystems, emphasizing the ecological interactions within these environments.
Diana Quiroz, 2018 Agropyron cristatum (crested wheatgrass) DOI:
Driscoll et al., 2014 New pasture plants intensify invasive species risk
Proceedings of the National Academy of Sciences of the United States of America Vol.111(46), pp.16622-16627 DOI: 10.1073/pnas.1409347111
Dwyer, D D., and Owens, M E., 1984 Grazing crested wheatgrass range in the
Figure 1 Remote sensing process Source: https://www.researchgate.net/figure/Stages-in-a-remote-sensing- process_fig2_316473726
Figure 2 Location of Saskatchewan Landing Provincial Park Source: https://www.trailmeister.com/trails/saskatchewan-landing-provincial-park/ Flynn E.S., Dougherty C.T., Wendroth O., 2008 Assessment of pasture biomass sensors Agron J 100:114
Graham and Jennifer, 2016 Lake, hikes and history draw visitors to Saskatchewan
Graham and Jennifer, 2016 Lake, hikes and history draw visitors to Saskatchewan
H Croft, J.M Chen, Y Zhang, 2013 The applicability of empirical vegetation indices for determining leaf chlorophyll content over different leaf and canopy structures Ecological Complexity 17: 119–130
Heidinga, L and Wilson, S D 2002 The impact of an invading alien grass
(Agropyron cristatum) on species turnover in native prairie Divers Distrib 8: 249–258
Hofman L., Ries R E., Karn, J F and Frank A B, 1993 Comparison of seeded and native pastures grazed from mid-May through September 46: 251-254 Huete, A.R., 1988 “A soil adjusted vegetation index (SAVI).” Remote Sensing of
Environment, Vol 25(No 3): pp 295–309 DOI: 10.1016/0034-
The study by Hufnagel and Garamvửlgyi (2014) examines the effects of climate change on vegetation distribution, specifically focusing on the shifts in plant communities across the New World It highlights how changing climatic conditions are influencing the geographical distribution and composition of various plant species The research, published in *Applied Ecology and Environmental Research*, provides valuable insights into the ecological consequences of climate change, emphasizing the need for effective conservation strategies to address these shifts in vegetation patterns.
Iwaasa, A.D., Jefferson, P.G., and Birkedal, E.J 2014 Beef cattle grazing behavior differs among diploid and tetraploid crested wheatgrasses (Agropyron cristatum and A desertorum) Can J Plant Sci 94: 851-855
Jinru Xue, Baofeng Su, 2017 Significant Remote Sensing Vegetation Indices: A
Review of Developments and Applications Journal of Sensors DOI: https://doi.org/10.1155/2017/1353691
Joshi, C., de Leeuw, J., and van Duren, I C., 2004 Remote Sensing and GIS
Applications for Mapping and Spatial Modelling of Invasive Species 7:669-677
Khan, Z., Rahimi-Eichi, V., Haefele, S et al Estimation of vegetation indices for
Lesica P., and DeLuca T H., 1996 Long-Term Harmful Effects of Crested
Wheatgrass on Great Plains Grassland Ecosystems Journal of Soil and Water Conservation 14:136-138
Lucas Parra et al., 2000 Unmixing Hyperspectral Data
Lukas W Lehnert et al., 2018 Hyperspectral Data Analysis in R: The hsdar-
Package Journal of Statistical Software Volume VV, Issue II
Merzlyak, M.N., Gitelson, A.A., Chivkunova, O.B., and Rakitin, V.Y 1999 “Non- destructive optical detection of pigment changes during leaf senescence and fruit ripening.” Physiologia Plantarum, Vol 106(No 1): pp 135–141
In their 1989 overview on biological invasions, Mooney and Drake discuss the significant impacts of invasive species on ecosystems, highlighting their role in altering native biodiversity and ecosystem functions This work is part of a comprehensive volume edited by notable researchers, including Drake and Mooney, which presents a global perspective on the subject Additionally, Nagler et al (2000) contribute to the understanding of ecological interactions by examining the relationship between plant litter and soil, further emphasizing the importance of studying biological invasions and their effects on environmental health.
Reflectance Remote Sensing of Environment 71:207-215
National Post, 2016 Lake, hikes and history draw visitors to Saskatchewan
Landing Provincial Park Retrieved from: https://nationalpost.com/travel/lake-hikes-and-history-draw-visitors-to- saskatchewan-landing-provincial-park
Ngai-Man Cheung, Antonio Ortega, 2009 CHAPTER 10 - Distributed
Compression of Hyperspectral Imagery Distributed Source Coding, Pages 269-292 DOI: https://doi.org/10.1016/B978-0-12-374485-2.00015-9 Olesiuk D & Zagajewski B., 2008 Wykorzystanie obrazów hiperspektralnych do klasyfikacji pokrycia terenu zlewni Bystrzanki Teledetekcja Środowiska,
Park Management Services, 2014 Saskatchewan Landing Provincial Park
Management & Development Plan Retrieved from: https://pubsaskdev.blob.core.windows.net/pubsask-prod/97610/97610-
Pejchar L., Mooney H A., 2009 Invasive species, ecosystem services and human well-being Trends in Ecology and Evolution 24:497-504
Psomas A., Kneubühler M., Huber S et al., 2011 Hyperspectral remote sensing for estimating aboveground biomass and for exploring species richness patterns of grassland habitats Int J Remote Sens 32:9007–31
Qi, J., Chehbouni, A., Huete, A., Kerr, Y., and Sorooshian, S 1994 “A modified soil adjusted vegetation index.” Remote Sensing of Environment, Vol
R K Vincent, 2015 RADAR | Synthetic Aperture Radar (Land Surface
Applications) Encyclopedia of Atmospheric Sciences (Second Edition)
Pages 470-476 DOI: https://doi.org/10.1016/B978-0-12-382225-3.00331-
Rashmi Kandwal, C Jeganathan, Valentyn Tolpekin, S P S Kushwaha, 2009
Discriminating the invasive species, ‘Lantana’ using vegetation indices
Journal of the Indian Society of Remote Sensing 37: 275 DOI: https://doi.org/10.1007/s12524-009-0027-5
Reujean, J., and Breon, F 1995 “Estimating PAR absorbed by vegetation from bidirectional reflectance measurements.” Remote Sensing of Environment, Vol 51(No 3): pp 375–384 DOI: 10.1016/0034-4257(94)00114-3
Reynolds and Springfield, 1953 Reseeding southwestern range lands with crested wheatgrass Farmers’ bulletin (United States Dept of Agriculture) No
Reynolds and Springfield, 1953 Reseeding southwestern range lands with crested wheatgrass Farmers’ bulletin (United States Dept of Agriculture) No
Reynolds T D, Trost C H, 1980 The response of native vertebrate populations to crested wheatgrass planting and grazing by sheep Journal of Range Management 33: 122–125, DOI: http://dx.doi.org/10.2307/3898425
Robert A Tanner, Sonal Varia, René Eschen, Suzy Wood, Sean T Murphy, Alan
C Gange, 2013 Impacts of an Invasive Non-Native Annual Weed,
Communities in the United Kingdom DOI: 10.1371/journal.pone.0067271 Roger B Davis, Kenneth J Mukamal, 2006 Hypothesis Testing Circulation Vol
114 Issue 10 Page: 1078 - 1082 DOI: https://doi.org/10.1161/CIRCULATIONAHA.105.586461
In 1974, Rouse Jr., Haas, Schell, and Deering presented a paper at the NASA GSFC 3rd ERTS-1 Symposium, focusing on the use of Earth Resources Technology Satellites (ERTS) to monitor vegetation systems in the Great Plains The findings were published in NASA SP-351, highlighting the significance of satellite technology in environmental monitoring and management.
S K Jain, V P Singh, 2003 Chapter 3 - Emerging Techniques for Data
Acquisition and Systems Modeling Developments in Water Science
Volume 51, Pages 123-205 DOI: https://doi.org/10.1016/S0167- 5648(03)80057-6
S.L Ustin, D DiPietro, K Olmstead, E Underwood, G.J Scheer, 2002
Hyperspectral remote sensing for invasive species detection and mapping
IEEE International Geoscience and Remote Sensing Symposium DOI:
Sangseok Lee and Dong Kyu Lee, 2018 What is the proper way to apply the multiple comparison test? Korean J Anesthesiol 71(5): 353–360 DOI:
Shaw, D (2005) Translation of Remote Sensing Data Into Weed Management
Sutter, G C and Brigham, R M 1998 Avifaunal and habitat changes resulting from conversion of native prairie to crested wheat grass: Patterns at songbird community and species levels Can J Zool 76: 869–875
U Herrmann, D Kearney, M Rửger, C Prahl, 2017 System performance measurements The Performance of Concentrated Solar Power (CSP) Systems Modelling, Measurement and Assessment Pages 115-165 DOI: https://doi.org/10.1016/B978-0-08-100447-0.00005-5
Weidong Zhou and Xulin Guo Discriminating Invasive Crested Wheatgrass
(Agropyron Cristatum) in Northern Mixed Grass Prairie Using Remote
Saskatchewan, Saskatoon, SK, S7N 5A5 Retrieved from: https://harvest.usask.ca/handle/10388/9369
Weisstein, Eric W, 2004 "Hypothesis Testing." From Math World A Wolfram Web
Resource: http://mathworld.wolfram.com/HypothesisTesting.html
William Emery, Adriano Camps, 2017 Chapter 2 - Basic Electromagnetic
Concepts and Applications to Optical Sensors Introduction to Satellite Remote Sensing Pages 43-83 DOI: https://doi.org/10.1016/B978-0-12- 809254-5.00002-6
Yang, X., Guo, X Investigating vegetation biophysical and spectral parameters for detecting light to moderate grazing effects: a case study in mixed grass prairie Central European Journal of Geosciences 3, 336 (2011) doi:10.2478/s13533-011-0032-4
In their 2012 study, Yu et al developed a genetic linkage map for crested wheatgrass utilizing AFLP and RAPD markers, published in Genome (55: 327-335) Additionally, Zagajewski (2010) evaluated the effectiveness of neural networks and hyperspectral data for classifying vegetation in the High Tatras, as detailed in Teledetekcja Środowiska (43, 1-113).
Zebin Wu, Jinping Gu, Yonglong Li, Fu Xiao, Jin Sun, and Zhihui Wei, 2016
Distributed Parallel Endmember Extraction of Hyperspectral Data Based on Spark Scientific Programming Article ID 3252148 DOI: http://dx.doi.org/10.1155/2016/3252148
Zhangyang Wang, 2019 Chapter 2 - Bi-Level Sparse Coding: A Hyperspectral
Image Classification Example Deep Learning Through Sparse and Low-Rank Modeling Page 9 – 29 doi: https://doi.org/10.1016/B978-0-12-813659-1.00002-0
APPENDIX A Multiple Comparisons Tables (5 grass types)
Variable (I) Code (J) Code Mean Difference (I-J) Std Error Sig
* The mean difference is significant at the 0.05 level
Multiple Comparisons Table (10 sites study)
Dependent Variable (I) Code (J) Code Mean Difference (I-J) Std Error Sig
* The mean difference is significant at the 0.05 level.