Quantifying the Uncertainty of Short-Term Vegetation Anomalies Detection Using Eo-Based Coarse-Resolution Vegetation Products

Conference paper (2022)

Authors

J. Zhou Central China Normal University, Key Laboratory for Geographical Process Analysis & Simulation of Hubei Province, Wuhan, Optical and Laser Remote Sensing - CEG
Xuan Liu Central China Normal University, Key Laboratory for Geographical Process Analysis & Simulation of Hubei Province, Wuhan
Xuqian Xiong Central China Normal University, Key Laboratory for Geographical Process Analysis & Simulation of Hubei Province, Wuhan
Li Jia Chinese Academy of Sciences
Jing Lu Chinese Academy of Sciences
Yilin Cui Key Laboratory for Geographical Process Analysis & Simulation of Hubei Province, Wuhan, Central China Normal University
Research Group
Optical and Laser Remote Sensing (CEG) (TU Delft)
Copyright: © 2022 J. Zhou, Xuan Liu, Xuqian Xiong, Li Jia, Jing Lu, Yilin Cui
DOI
help_outline
https://doi.org/10.1109/IGARSS46834.2022.9884555
Uncertainty EO-based vegetation products Vegetation anomalies Vegetation Condition Index
More Info
expand_more

Published Date

2022

Language

English

Reuse Rights

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Faculty

Civil Engineering & Geosciences

Department

Geoscience and Remote Sensing

Research Group

Optical and Laser Remote Sensing

Abstract

Satellite-based Earth Observation systems archived a variety of vegetation products during the last 50 years, which can reveal regional to global ecosystem dynamics across diverse spatiotemporal scales. The anomaly metrics such as Vegetation Condition Index (VCI) defined by comparing the current vegetation growth condition to historical average status based on long-term EO-based vegetation products were widely used to delineate abnormal vegetation variation exerted by either climatic or anthropogenic factors (e.g., droughts, wildfires). However, currently available long-term vegetation products may differ from each other in terms of sensors (observational platform or spectral bands), bio-physical definitions (e.g., NDVI, EVI, LAI, and VOD), spatiotemporal resolution, as well as the time-spans, which results in inconsistency across these vegetation products. Taking the VCI as an example, this study evaluated the uncertainty of vegetation anomalies detected based on different vegetation products over the middle reach of the Yangtze River by explicitly considering the effect of sensors, biophysical definitions, and time-spans. The preliminary results showed that VCI derived from NDVI products from different sensors (AVHRR vs. MODIS) induced significant inconsistent anomalies over most landscapes. The differences resulting from products with different biophysical definitions (NDVI vs. EVI, LAI, and VOD) are much lower than those from different sensors but still significant over specific areas. As for the time-spans, the 20-year NDVI based VCI presented a considerable reduction in variance over the study area on average compared to VCI calculated based on 5-year NDVI. In summary, caution should be taken when applying EO-based vegetation products for vegetation anomalies mapping, especially for quantitative assessment.