Incorporating the Plant Phenological Trajectory into Mangrove Species Mapping with Dense Time Series Sentinel-2 Imagery and the Google Earth Engine Platform

Abstract

Information on mangrove species composition and distribution is key to studying functions of mangrove ecosystems and securing sustainable mangrove conservation. Even though remote sensing technology is developing rapidly currently, mapping mangrove forests at the species level based on freely accessible images is still a great challenge. This study built a Sentinel-2 normalized difference vegetation index (NDVI) time series (from 2017-01-01 to 2018-12-31) to represent phenological trajectories of mangrove species and then demonstrated the feasibility of phenology-based mangrove species classification using the random forest algorithm in the Google Earth Engine platform. It was found that (i) in Zhangjiang estuary, the phenological trajectories (NDVI time series) of different mangrove species have great differences; (ii) the overall accuracy and Kappa confidence of the classification map is 84% and 0.84, respectively; and (iii) Months in late winter and early spring play critical roles in mangrove species mapping. This is the first study to use phonological signatures in discriminating mangrove species. The methodology presented can be used as a practical guideline for the mapping of mangrove or other vegetation species in other regions. However, future work should pay attention to various phenological trajectories of mangrove species in different locations. mangrove forests is essential for the accurate formulation of future studies and management of mangrove ecosystems [8][9][10].Remote sensing has served as a sustainable tool in the mapping and monitoring of mangrove forests for decades, primarily because of the logistical and practical difficulties involved in field surveys of the muddy environments [11,12]. Before the launch of the high resolution satellite sensors, it was impossible to accurately discriminate mangrove species with traditional medium resolution satellite data [13]. Recently, with the development of commercial sensors (high spatial resolution, hyperspectral, and active remote sensor), many studies have employed single or combined airborne or satellite imagery to map mangrove species [2,3,8]. However, so far, the mapping of mangrove species with freely accessible imagery still remains a challenge, as mangrove species often exhibit similar spectral signatures and spatial textures [10].The aforementioned issues indicate the need to explore more substantial features to improve the detection of mangrove species from remote sensing data. A phenological trajectory of plants can be acquired from the time series of remote sensing images through delineating the temporal variation in spectrum during the growing period [14]. To date, remotely sensed plant phenology has been widely used to conduct vegetation discrimination, but most of the studied vegetation has been crops or inland forests [15,16]. Although mangroves are evergreen plants, different mangrove species have different phenophase peaks [17]. However, to date, no studies have dealt with mangrove phenological trajectories in remote sensing-based ...

A Simple Phenology-Based Vegetation Index for Mapping Invasive Spartina Alterniflora Using Google Earth Engine

“…In recent years, intertidal mudflats of these natural reserves have been extensively invaded by S. alterniflora [34], [35].…”

“…Although not all images were available for each study site, they can help to visualize and analyze the seasonal change of vegetation greenness. The visual interpretation was also assisted by several previous studies including UAV images collected by Zhu et al [24], the reference samples by Liu et al [16] and Tian et al [25], and the resultant maps by Li et al [34] and Tian et al [25]. Once the samples were interpreted by one expert, they were reexamined by other experts and all experts discussed the ones with uncertainties.…”

“…For example, Tian et al [25] composed two 6-band Landsat imageries using cloud-free pixels that have maximum EVI value and using pixels that have maximum Plant Senescence Reflectance Index for S. alterniflora detection, respectively. Li et al [34] constructed time series Sentinel-2 imagery using harmonic analysis of time series algorithm (HANTS) at pixel level for Mangrove tree species mapping. Karasiak et al [49] constructed time series Formosat-2 imagery by filling cloudy pixels using temporal linear interpolation and conducted forest tree species classification.…”

Adaptive Threshold Model in Google Earth Engine: A Case Study of Ulva prolifera Extraction in the South Yellow Sea, China

“…The larger the value, the greater the possibility of clouds, so this paper set the value of the "cloud" band to 20. For the Sentinel-2/MSI images in the selected study area, this study used the QA60 band, which contains cloud information for cloud mask and image quality problems caused by cirrus clouds [34].…”

A Comparison of Gaofen-2 and Sentinel-2 Imagery for Mapping Mangrove Forests Using Object-Oriented Analysis and Random Forest

“…ANGROVE forests, as an important intermediate ecosystem, are extensively distributed in the intertidal zone along tropical and subtropical coasts [1]- [3]. These forests play a vital role in maintaining biodiversity, stabilizing shoreline, keeping water quality, preventing the storm and providing numerous societal services and goods [4], [5].…”