phenoC++: An open-source tool for retrieving vegetation phenology from satellite remote sensing data

Frontiers in Environmental Science

Environmental Science

Fengrui Jing¹  Qinchuan Xin²  Baozhen Ruan³  Xi Liao³  Xinchang Zhang³  Yongjian Ruan⁴ 

[1] Department of Geography, University of South Carolina, Columbia, SC, United States;School of Geography and Planning, Sun Yat-Sen University, Guangzhou, China;School of Geography and Remote Sensing, Guangzhou University, Guangzhou, China;School of Geography and Remote Sensing, Guangzhou University, Guangzhou, China;Key Laboratory of Urban Land Resources Monitoring and Simulation, Ministry of Natural Resources, Shenzhen, China;

关键词: vegetation phenology;    satellite remote sensing data;    PhenoCam;    C++ language;    the contiguous United States;   

DOI  :  10.3389/fenvs.2023.1097249

 received in 2022-11-13, accepted in 2023-02-07,  发布年份 2023

来源: Frontiers

【 摘 要 】

Satellite-retrieved vegetation phenology has great potential for application in characterizing seasonal and annual land surface dynamics. However, obtaining regional-scale vegetation phenology from satellite remote sensing data often requires extensive data processing and computation, which makes the accurate and rapid retrieval of regional-scale phenology a challenge. To retrieve vegetation phenology from satellite remote sensing data, we developed an open-source tool called phenoC++, which uses parallel technology in C++. phenoC++ includes six common algorithms: amplitude threshold (AT), first-order derivative (FOD), second-order derivative (SOD), third-order derivative (TOD), relative change rate (RCR), and curvature change rate (CCR). We implemented the proposed phenoC++ and evaluated its performance on a site scale with PhenoCam-observed phenology metrics. The result shows that SOS derived from MODIS images by phenoC++ with six methods (i.e., AT, FOD, SOD, RCR, TOD, and CCR) obtained r-values of 0.75, 0.76, 0.75, 0.76, 0.64, and 0.67, and RMSE values of 21.36, 20.41, 22.38, 19.11, 33.56, and 32.14, respectively. Satellite-retrieved EOS by phenoC++ with six methods obtained r-values of 0.58, 0.59, 0.57, 0.56, 0.36, and 0.40, and RMSE values of 52.43, 46.68, 55.13, 49.46, 71.13, and 69.34, respectively. Using PhenoCam-observed phenology as a baseline, SOS retrieved by phenoC++ was superior to MCD12Q2, while EOS retrieved by phenoC++ was slightly inferior to that of MCD12Q2. Moreover, compared with MCD12Q2 on a regional scale, phenoC++-retrieved vegetation phenology yields more effective pixels. The innovative features of phenoC++ are 1) integrating six algorithms for retrieving SOS and EOS; 2) quickly processing data on a large scale with simple input startup parameters; 3) outputting phenology metrics in GeoTIFF format image, which is more convenient to use with other geospatial data. phenoC++ could aid in investigating and addressing large-scale phenology problems of the ecological environment.

【 授权许可】

Unknown   
Copyright © 2023 Ruan, Ruan, Xin, Liao, Jing and Zhang.