Invasive species are one of the largest threats to wetlands biodiversity and ecosystem functioning. In the US, invasive species are estimated to cause $120 billion dollars per year in environmental damage and associated control costs: Lythrum salicaria (purple loosestrife) alone is estimated to cost $45 million per year as it spreads at a rate of 115,000 ha/yr across wetlands in the US [1]. With approximately 50% of wetlands destroyed or altered globally as a result of human activities [2], preserving and restoring the ecological integrity of the remaining wetlands has become a priority [3-5]. In order to meet future goals for aquatic ecosystem integrity, resource managers and decision makers need information and improved methods for identifying invasive plants and monitoring control measures.

Satellite remote sensing has been a useful tool in providing general information on wetlands types [6]; however, both spatial and spectral resolutions have limited the level of detail ultimately required for comprehensive wetland assessments. Recent advances in sensor technology and remote sensing science have promoted an interest in hyperspectral data for mapping wetlands at the species level [7-11]. Advanced spectroscopic systems possess capabilities to capture data at narrow spectral bandwidths on the order of three to ten nanometers, while contiguously covering large portions of the spectrum (e.g., 350-2500nm). This allows for small variations in plant/substrate absorptance and reflectance to be recorded. Incorporating such relatively high spectral detail makes it possible to explore species separability and precise process monitoring [10,12].

In the last few years several studies have utilized hyperspectral data for wetlands mapping. A primary goal in these investigations was to develop methods to utilize the increased level of data supplied via hyperspectral instruments. These studies can be grouped into methods to identify wavelengths of particular utility (processing techniques to extract and identify the most useful bands) and evaluating and improving classification techniques to map species of interest.

For example, Becker et al. [8] performed derivative analysis to identify unique points of inflection along spectra for wetland plants in a Great Lakes coastal wetland and identified eight bands as possessing the most utility for separation. The bands are located across the visible (VIS) and near-infrared (NIR) portions of the spectrum and are affiliated with domains that represent unique biophysical characteristics. The red-edge was highlighted as having particular strength in separation. Artigas and Yang [7] analyzed samples from a New Jersey coastal wetland and also identified the NIR region using a discrimination metric and derivative analysis. This research concluded that monotypic stands of Phragmites could be identified by using the unique NIR response. Schmidt and Skidmore [10] conducted MannWhitney ^-testing on field-level reflectance data from coastal salt marshes in the Netherlands and found wavelengths in the NIR to possess high abilities to statistically differentiate species.

Developing and improving classifications has also been a focus of hyperspectral wetland remote sensing. For example, Becker et al. [9] examined the optimal spectral and spatial resolutions for mapping Great Lakes coastal wetlands. A series of experiments tested different bands, band combinations, and pixels sizes in order to identify the most advantageous configurations to accurately map coastal wetlands. The results showed that narrow, strategically located bands were necessary to achieve acceptable resiliency levels when trying to limit the number of bands in order to maintain small pixel sizes. Li et al. [13] and Rosso et al. [11] used Spectral Mixture Analysis (SMA) and Multiple End-member (MESMA) techniques on airborne data to map coastal marshes in central California and found spectral similarity and increasing landscape heterogeneity to cause misclassification. Underwood et al. [14] found minimum noise fraction to outperform band ratio and continuum removal processing techniques when classifying varying densities of nonnative coastal plants of concern. Recently, Artigas and Yang [15] found that mapping Phragmites australis gradients in coastal New Jersey using airborne imagery was possible due to spectral differences associated with physiognomic characteristics. All these studies show the importance of utilizing unique spectral features to better differentiate/map wetland plant canopies.

A processing technique that allows for the extraction and modeling of individual spectral features (absorptance and reflectance) is continuum removal [13]. This technique is increasingly being implemented in hyperspectral vegetation investigations to isolate features of utility [10, 11, 17-19]. Used extensively in geological applications, continuum removal disregards albedo, and/or contributing background signal, to obtain individual features (absorption/reflectance) such as the precise location and depth of absorption features. By using continuum removal, Underwood et al. [14] achieved accurate maps of coastal invasives by applying continuum removal techniques to take advantage of unique water absorption features. Schmidt and Skidmore [10] found that applying continuum removal to salt-marsh vegetation spectra improved species separation in the visible spectrum, but decreased it in the near-infrared (NIR) and shortwave-infrared (SWIR) regions. They further suggest that if continuum removal eliminates noise from the soil background, moisture content, and canopy structure, then only the varying biogeochemical content of a species would determine separability levels. In this chapter, we explored the utility of continuum removal for identifying wetland invasive plant species.

One second well-established technique for characterizing spectra is derivative analysis. The methodology distinguishes the wavelength location where substantial inflection occurs [20-22]. In Becker et al. [8], second derivative approximations identified seven wavelengths (685, 731, 939, 514, 812, 835, 823, 560 nm) using contiguous data covering the visible and NIR regions from a Great Lakes coastal wetland community. In this chapter, we expanded upon the techniques applied in Becker et al. [8] to evaluate if invasives species of interest and different biological communities reproduce similar results.

The third identification approach utilized vegetation reflectance variation for distinguishing wetland invasive plant species. This approach, developed by Cochrane [23], uses a shape-filter representing the range of reflectance variation present in a species over the spectrum. Vegetation reflectance varies across the spectrum with the visible domain largely determined by the chlorophyll content, the NIR region a function of leaf structure and biomass volume, and the SWIR region largely determined by leaf water content and biomass volumes [23, 24]. Generally, differences in absorption determine the amount of variation and spectral overlap between species. The maximum and minimum reflectance for a given species creates its shape-space. If any of the comparison spectra (e.g., an invasive species vs. all others) fall outside the shape-space, then separation is possible. As absorption feature 'uniqueness' increases, the level of separability increases. As a byproduct, the shape-filtering technique identifies wavelengths that are useful for distinguishing between the species analyzed. The wavelengths identified might vary depending on species and uniqueness of absorption features. In this chapter we applied the shape filtering approach to identify invasives and wavelengths that were useful for doing so.

The overarching goal of this chapter was to investigate techniques for highlighting locations of unique spectral features for identifying wetland invasive plant species. This investigation utilizes three distinct processing methodologies (i.e., Derivatives, Continuum Removal, and Shape Filter) to explore their potential for delineating invasive species within the spectral domain of typical airborne and satellite hyperspectral sensors.

Was this article helpful?

0 0

Post a comment