Regulated large-scale annual shutdown of Regulated large-scale annual shutdown of Amazonian isoprene emissions?Amazonian isoprene emissions?
New insight provided by satellite observations of formaldehyde (HCHO) and of New insight provided by satellite observations of formaldehyde (HCHO) and of vegetationvegetation
Michael Barkley & Paul Palmer, University of Edinburgh, UKIsabelle De Smedt & Michel van Roozendael, BIRA-IASB, BelguimThomas Karl & Alex Guenther, NCAR, USA
Michael Barkley, University of Edinburgh
Slide 2
Mapping isoprene emissions from space using HCHO
HCHO is significant product of isoprene oxidation
Isoprene emissions are typically main driver of variability in HCHO columns observed from space
August 1999
Michael Barkley, University of Edinburgh
Slide 3
Methodology
HCHO column data:♦ New 12 year data set from BIRA-IASB (De Smedt et al., ACP, 2008)♦ GOME: 1996-2002♦ SCIAMACHY: 2003-2007
Analysis:♦ EOF (Empirical Orthogonal Function) Analysis
♦ Take a set of time evolving data and separate the variability into standing oscillation patterns (EOFs) and principal components that show how each mode varies with time
♦ What does EOF analysis tell us?♦ Most important variability in HCHO columns & therefore of isoprene
emissions too Correlative data:
♦ GEOS-4 model temperature, PAR (diffuse & direct), precipitation ♦ ATSR / AATSR firecounts♦ MODIS Leaf Area Index (LAI) size of leaves♦ MODIS Enhanced Vegetation Index (EVI) colour of leaves♦ Multivariate ENSO Index (MEI) (Wolter and Timlin,
1998)
Michael Barkley, University of Edinburgh
Slide 4
Where When Importance
Michael Barkley, University of Edinburgh
Slide 5
Where When Importance
Michael Barkley, University of Edinburgh
Slide 6
Where When Importance
Michael Barkley, University of Edinburgh
Slide 7
HCHO over the Amazon
Michael Barkley, University of Edinburgh
Slide 8
Seasonal
Variability
Scenes contaminated by fire are excluded using firecounts & NO2 columns
Background HCHO, from the oxidation of non-isoprene species, is taken from a GEOS-Chem simulation for the year 2000
Satellite data consistent with surface observations of isoprene♦ ~30% higher
in dry season
Michael Barkley, University of Edinburgh
Slide 9
HCHO response to vegetation activity
LAI - Myneni et al., PNAS, 2007. ♦ Leaf flushing (new leaf growth) during onset of the dry season
EVI – Huete et al., GRL, 2006. ♦ Rainforests green-up with sunlight in dry season
HCHO – Barkley et al, GRL, submitted.♦ Majority of isoprene emitting species undergo leaf flushing prior to
dry season in anticipation of light-rich conditions
Michael Barkley, University of Edinburgh
Slide 10
HCHO response to vegetation activity
LAI - Myneni et al., PNAS, 2007. ♦ Leaf flushing (new leaf growth) during onset of the dry season
EVI – Huete et al., GRL, 2006. ♦ Rainforests green-up with sunlight in dry season
HCHO – Barkley et al, GRL, submitted.♦ Majority of isoprene emitting species undergo leaf flushing prior to
dry season in anticipation of light-rich conditions
Michael Barkley, University of Edinburgh
Slide 11
HCHO response to vegetation activity
LAI - Myneni et al., PNAS, 2007. ♦ Leaf flushing (new leaf growth) during onset of the dry season
EVI – Huete et al., GRL, 2006. ♦ Rainforests green-up with sunlight in dry season
HCHO – Barkley et al, GRL, submitted.♦ Majority of isoprene emitting species undergo leaf flushing prior to
dry season in anticipation of light-rich conditions
Michael Barkley, University of Edinburgh
Slide 12
Summary
HCHO columns provide key insight to Amazonian isoprene emissions:♦ Seasonal variations of HCHO consistent with satellite observations
of vegetation activity ♦ Data strongly suggest large-scale leaf flushing by majority of
isoprene emitters results in annual cessation of isoprene emissions ♦ Or, at the very least a severe reduction !
♦ Big question – Why ?
Towards the future:♦ Sustained long-term measurement program essential !♦ Integration of flux and concentration measurements with satellite
observations♦ HCHO from OMI & GOME-2 – higher spatial resolution ♦ Nested GEOS-Chem with improved VOC chemical
mechanism♦ CRI scheme (Jenkin et al, AE, 2008)
The end!The end!
Michael Barkley, University of Edinburgh
Slide 14
Higher EOF modes
Higher modes:♦ Show decreased contributions to the total variance♦ Contain retrieval artefacts (GOME diffuser plate)
Difficult to assign physical meaning discarded
Michael Barkley, University of Edinburgh
Slide 15
Top-down isoprene emissions
*** GOME HCHO from Harvard ***
Barkley et al., JGR, 2008