Satellite-derived LAI products exhibit large discrepancies and can lead to substantial uncertainty in simulated carbon and water fluxes

REMOTE SENSING OF ENVIRONMENT

卷:206

Article

Liu, Yibo^1,2  Xiao, Jingfeng^2,3  Ju, Weimin⁴  Zhu, Gaolong⁵  Wu, Xiaocui⁶  Fan, Weiliang⁷  Li, Dengqiu⁷  Zhou, Yanlian⁸ 

[1] Nanjing Univ Informat Sci & Technol, Sch Appl Meteorol, Jiangsu Key Lab Agr Meteorol, Nanjing 210044, Jiangsu, Peoples R China

[2] Univ New Hampshire, Inst Study Earth Oceans & Space, Earth Syst Res Ctr, Durham, NH 03824 USA

[3] Nanjing Univ Informat Sci & Technol, Sch Appl Meteorol, Int Ctr Ecol Meteorol & Environm, Nanjing 210044, Jiangsu, Peoples R China

[4] Nanjing Univ, Int Inst Earth Syst Sci, Nanjing 210023, Jiangsu, Peoples R China

[5] Minjiang Univ, Dept Geog, Fuzhou 350108, Fujian, Peoples R China

[6] Univ Oklahoma, Dept Microbiol & Plant Biol, Ctr Spatial Anal, Norman, OK 73019 USA

[7] Zhejiang A&F Univ, Sch Environm & Resources Sci, Key Lab Carbon Cycling Forest Ecosyst & Carbon Se, Linan 311300, Peoples R China

[8] Nanjing Univ, Sch Geog & Oceanog Sci, Nanjing 210023, Jiangsu, Peoples R China

关键词: Leaf area index;    Gross primary productivity;    Evapotranspiration;    Uncertainty;    Ecosystem model;    Four-scale geometric optical model;    MODIS;   

DOI  :  10.1016/j.rse.2017.12.024

来源: Elsevier

【 摘 要 】

Understanding the terrestrial carbon and water cycles is crucial for mitigation and adaptation for climate change. Leaf area index (LAI) is a key biophysical parameter in process-based ecosystem models for simulating gross primary productivity (GPP) and evapotranspiration (ET). The uncertainty in satellite-derived LAI products and their effects on the simulation of carbon and water fluxes at regional scales remain unclear. We evaluated three satellite-derived LAI products - MODIS (MCD15), GLASS, and Four-Scale Geometric Optical Model based LAI (FSGOM) over the period 2003-2012 using fine-resolution (30 m) LAI data and field LAI measurements. GLASS had higher accuracy than FSGOM and MCD15 for forests, while FSGOM had higher accuracy than MCD15 and GLASS for grasslands. The three LAI products differed in magnitude, spatial patterns, and trends in LAI. We then examined the resulting discrepancies in simulated annual GPP and ET over China using a process-based, diagnostic terrestrial biosphere model. Mean annual total GPP for China's terrestrial ecosystems based on GLASS (6.32 Pg C yr(-1)) and FSGOM (6.15 Pg C yr(-5)) was 22.5% and 19.2% higher than that based on MCD15 (5.16 Pg C yr(-1)), respectively. Annual GPP based on GLASS and MCD15 increased over larger fractions of China's vegetated area (15.9% and 17.3%, respectively) than that based on FSGOM (12.6%). National annual ET based on GLASS (379.9 mm yr(-1)) and FSGOM (374.4 mm yr(-1)) was 7.9% and 6.3% higher than that based on MCD15 (352.1 mm yr(-1)), respectively.(-)Simulated ET increased in larger fractions of the vegetated area for GLASS (5.7%) and MCD15 (5.8%) than for FSGOM (3.9%). Our study shows that there were large discrepancies in LAI among satellite-derived LAI products and the biases of the LAI products could lead to substantial uncertainties in simulated carbon and water fluxes.

【 授权许可】

Free