Remote Sensing of Coastal Vegetation Phenology in a Cold Temperate Intertidal System: Implications for Classification of Coastal Habitats

Abstract

Intertidal vegetation provides important ecological functions, such as food and shelter for wildlife and ecological services with increased coastline protection from erosion. In cold temperate and subarctic environments, the short growing season has a significant impact on the phenological response of the different vegetation types, which must be considered for their mapping using satellite remote sensing technologies. This study focuses on the effect of the phenology of vegetation in the intertidal ecosystems on remote sensing outputs. The studied sites were dominated by eelgrass (Zostera marina L.), saltmarsh cordgrass (Spartina alterniflora), creeping saltbush (Atriplex prostrata), macroalgae (Ascophyllum nodosum, and Fucus vesiculosus) attached to scattered boulders. In situ data were collected on ten occasions from May through October 2019 and included biophysical properties (e.g., leaf area index) and hyperspectral reflectance spectra (Rrs(λ)). The results indicate that even when substantial vegetation growth is observed, the variation in Rrs(λ) is not significant at the beginning of the growing season, limiting the spectral separability using multispectral imagery. The spectral separability between vegetation types was maximum at the beginning of the season (early June) when the vegetation had not reached its maximum growth. Seasonal time series of the normalized difference vegetation index (NDVI) values were derived from multispectral sensors (Sentinel-2 multispectral instrument (MSI) and PlanetScope) and were validated using in situ-derived NDVI. The results indicate that the phenology of intertidal vegetation can be monitored by satellite if the number of observations obtained at a low tide is sufficient, which helps to discriminate plant species and, therefore, the mapping of vegetation. The optimal period for vegetation mapping was September for the study area.

Multi-Dimensional Low-Rank with Weighted Schatten p-Norm Minimization for Hyperspectral Anomaly Detection

“…Compared with conventional images such as RGB images, multispectral images, SAS images [1], and delay-Doppler images [2], hyperspectral images (HSIs) offer the advantage of capturing hundreds of contiguous spectral bands of the same scene. This unique characteristic of HSI proves to be beneficial for target detection and finds wide applications in various fields such as land cover classification [3][4][5], mineral survey [6][7][8], environmental protection [9][10][11], and other applications [12][13][14][15][16][17][18]. In hyperspectral target detection, when the target information is unknown, the unsupervised processing of the target detection is called anomaly detection.…”

Sensing Linked Cues for Ecosystem Risk and Decisions

“…Species sense the quality of the environment, and ecological data reflect the functioning of eco-environmental ties. There is no environment that is fully abiotic, and yet efforts to compile ecological data must be comprehensive of the flows of ecosystems over time; (2) The spatial connections among habitats (natural and self-emergent habitats and those of human-made design, which are reflected in geomorphological and infrastructural data, respectively) are the basis of any ecological function with strong climate feedback; thus, "climate neutral" efforts must consider the engineering of salient hydrologic flows and eco-geomorphological connections (broadly defined as ecological ties) whose scale-free organization is the optimal configuration of our ecosystem; (3) Networks of people's decisions, from the behavior of citizens to stakeholder development and management strategies, are critical for an ecosystem's function and intelligence, in which the latter is as much a conscious action as the reactions of species to information sensed in ecosystems. All these decisions are associated with ecological information (extracted by models as perceptrons) for which digitized information carries values and thresholds with respect to the functions of ecosystems to create forecasts, assess indicators and ecosystem states and define ecosystem services and controls (what is needed and/or desired, for which the definition of optimal trade-offs is essential).…”

“…In this Special Issue, many papers highlighted data and methods used to infer patterns across multiple scales and ecosystems, as well as to provide solutions, including predictive capabilities. For marine ecosystems, the delicate nature of the phytoplanktonenvironmental nexus was highlighted is in determining the extent and persistence of algal blooms [1], and the ways in which the phenology of coastal vegetation in a cold temperate intertidal system impacts remote sensing (and the subsequent classification of coastal habitats) was addressed [2]. Both studies actually emphasize how ecological conditions affect the information that can be gathered and yet add intrinsically uncontrollable (but measurable) uncertainty into monitoring technology; this is rather important and unappreciated since a large number of scientists and policy makers assume that all data are the undisputable, golden truth.…”