Synthesis of long-term remote sensing LAI for applications

Jiafu Mao1*, Binyan Yan2, Xiaoying Shi1, Peter E. Thornton1, Forrest M. Hoffman3 and David M. Lawrence4

1Climate Change Science Institute/Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee, USA 2Jackson School of Geosciences, the University of Texas, Austin, TX, USA 3Climate Change Science Institute/Computer Science and Mathematics Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee, USA 4National Center for Atmospheric Research, Boulder, Colorado, USA

Overview: There is growing interest in employing large-‐scale long-‐term remote sensing Leaf Area Index (LAI) products to evaluate responses and feedbacks of terrestrial ecosystem dynamics to climate change. However, the mulCple LAI products were derived from different satellites, produced using different algorithms and global land cover classificaCons, developed at different spaCal and temporal resoluCons, and released with different data gaps and map projecCons. Understanding and resoluCon of the inter-‐LAI dataset differences are criCcal if these datasets are to be used to quanCfy the response of vegetaCon phenology and variability to climate, to force biogeophysical land models and to evaluate biogeochemical land models. Here, we demonstrate the homogenizaCon and intercomparison of four different satellite datasets at 0.5 degree spaCal resoluCon between 1982 and 2010. Also, we evaluate the Community Land Model (CLM) simulated LAI against these standardized products.

Sponsored by the U.S. Department of Energy, Office of Science, Biological and Environmental Research (BER) programs, and performed at Oak Ridge National Laboratory (ORNL). ORNL is managed by UT-Battelle, LLC, for the U.S. Department of Energy under contract DE-AC05-00OR22725. Please contact [email protected] for further information.

•  Magnitude and phenology of multi-year mean LAI are comparable among different products; •  Mismatches between the pre- and post-2000 period are found in the four products, and this

mismatch problem is particularly significant in the GLOBMAP and GLASS LAI; •  The GEOLAND2 LAI and the GIMMS LAI3g are relatively consistent in the temporal

changes and the increasing trends from 1982 to 2010;

•  Weak agreement in long-term trends among satellite-based LAI implies the importance of using multiple products to reduce the uncertainties spatially and temporally;

•  LAI trend and interannual variability are better simulated by CLM4 than the magnitude; •  More analysis and understandings are needed for further applications in the LSMs and

ESMs.

Synthesis of long-term remote sensing LAI for applications in Land Surface and Earth System Models

Intercomparisons among the four processed LAI products (1982-2010) Evaluations of the simulated LAI from different versions of CLM4 (1982-2005)

LAI Length Organiza2on Reference

GLASS

1982/01 ~ 2000/12 (AVHRR)

2001/01 ~ 2010/12 (MODIS)

Beijing Normal University, University of Maryland

Xiao et al.,

2013

GLOBMAP

1981/07 ~ 2000/02 (AVHRR)

2000/03 ~ 2011/12 (MODIS)

IGSNRR, Chinese Academy of Sciences

Liu et al., 2012

GIMMS LAI3g

1981 ~ 1999 (AVHRR)

2000 ~ 2011 (MODIS)

Boston University

Zhu et al., 2013

GEOLAND2

1982/01 ~ 2000/12 (AVHRR)

1999/01 ~ present (SPOT/VEGETATION,

MODIS, CYCLOPES)

Europe

Baret et al., 2013

Methods!Gap$filling$!  Short$gaps$(shorter$than$6$months):!spline'fi)ng'

(using'a'moving'window'method'to'find'a'local'op6mal'6me'series'for'spline'fi)ng)'

'

'!  Long$gaps$(all$others):$average'of'a'previous'

period'and'a'following'period'that'are'the'same'length'of'the'gap'

Gap!Filling!

Calculate!monthly!mean!

Spa4ally!Up6scaling!

QC!informa4on!

Four!306year!monthly!LAI!products!(0.05!to!0.5!degree)!

Four!LAI!datasets!

New!LAI!product!

An$example$of$filled$short$gap$

An$example$of$filled$long$gap$

Black:!original!LAI!Red:!filled!values!

² Gap filling • Short gaps shorter than 6 months: using a

moving window method to find a local optimal time series for spline fitting

• Long gaps: average of a previous period and a following period that are the same length of the gap

² QC information • Based on the percentage of gaps

• Similar QC information as MODIS LAI product

• Supplementary information of the status of each grid for users

² Spatially up-scaling • Spatial average weighted by grid areas

² LAI synthesis • According to different uncertainties

LAI data sets used in this study Homogenization methods

Mean&LAI&(1982U2010)&

IntroducJon|&Data&sets&and&Methods|&Findings&|&Summary&

Mean&peak&LAI&month&(1982U2010)&


ArcJc&

Boreal&

NT&

NE&

SE&

ST& GLASS GIMMS GLOBMAP GEOLAND2

Monthly&evoluJon&of&LAI&(1982U2010)&


GLASS GIMMS GLOBMAP GEOLAND2

1982Z2010' 1982Z1999' 2001Z2010'

LAI&trends&

Sig.'+'(1982Z2010)' Sig.'−'(1982Z2010)'


AVHRR' MODIS'

CLM45bgc_hrv_1deg4508&

CLM40cn_ornl&

CLM45bgc_1deg4081!

CLM40cn_1deg4502!

Climatological&mean&differences&(1982;2005)!

LAI&Gme&series&(1982;2005)!

Annual&LAI&Gme&series! Annual&LAI&anomalies!

Global!

NH!

SH!

Mean%of%four%satellite%LAI%Mean%of%GIMMS%and%GEOLAND2%CLM40cn_ornl%CLM40cn_1deg4502%CLM45bgc_hrv_1deg4508%CLM45bgc_1deg4081%!

GLASS GIMMS GLOBMAP GEOLAND2 CLM40cn_ornl%CLM40cn_1deg4502%CLM45bgc_hrv_1deg4508%CLM45bgc_1deg4081%!

Annual&cycle&of&LAI&trends&for&CLM&and&each&product&(1982;1999)&




Mean LAI Month of mean peak LAI

Monthly evolu2on of LAI LAI trends

Differences between the CLM4 LAI and each remote-‐sensing based LAI

Annual LAI 2me series Annual LAI anomalies Annual cycle of LAI trends



LAI&Gme&series&(1982;2005)!

Annual&LAI&Gme&series! Annual&LAI&anomalies!

Global!

NH!

SH!


GLASS GIMMS GLOBMAP GEOLAND2 CLM40cn_ornl%CLM40cn_1deg4502%CLM45bgc_hrv_1deg4508%CLM45bgc_1deg4081%!

Documents

Synthesis of long-term remote sensing LAI for applications