European Geosciences Union General Assembly 2006 Vienna, Austria, 02 07 April 2006

1
European Geosciences Union General Assembly 2006

Vienna, Austria, 02 07 April 2006
AERONET versus MODIS retrievals at different
spatial resolutions over south-east Italy
M. Santese, F. De Tomasi and M. R. Perrone Phys
ics Department, University of Lecce, Italy
(monica.santese_at_le.infn.it / Fax 39 0832
297505)

• Papers objectives
• 1. Contribute to the validation of MODIS aerosol
  products over
  
• south-east Italy and investigate the
  correlation dependence
  
• on spatial resolution and identify regional
  biases of Lecces
  
• AERONET data
• 2. Can MODIS help us to understand to what extent
  the Lecces
  
• AERONET site can be considered representative
  of a larger area
  
• and hence, locally-derived aerosol parameters
  can be of use in
  
• General Circulation and Chemical Transport
  Models?

Geographic location of the AERONET monitoring site
Aerosol optical thicknesses AOTs and Fine
Fraction parameters ? retrieved by AERONET
measurements from March 2003 to September 2004
at Lecces University, are compared to similar
MODIS_Terra data retrieved over ocean and
land-ocean at 550 nm and at different spatial
resolutions (50x50, 100x100, and 300x300km²)
co-located in space and time.
50x50 km²
100x100 km²
300x300 km²
Results
AERONET versus MODIS temporal evolution
AERONET and MODIS-ocean fine fraction parameters
AERONET versus MODIS AOTs
Over Ocean
Over land-ocean
50x50 km²
Over Ocean
Over land-ocean
50x50 km²
Comments The correlation factors R of linear reg
ressions span the 0.88-0.83 range.
MODIS AOTs meet expected uncertainties Over
ocean 70, 67, and 70 of data points of the
50x50 km², 100x100 km², and 300x300 km² window
size, respectively is within expected
uncertainties Over land-ocean 85, 88, and 82
of AOT values retrieved at 50x50 km², 100x100
km², and 300x300 km² window size, respectively
meets pre-specified accuracy conditions.
MODIS overestimates AOTs at low aerosol loadings
. This result can be due to the fact that the two
algorithms understimate the ground surface
reflectance. The slopes of the over land-ocean
regression lines is closer to unity the
land-ocean MODIS AOT values can better represent
the aerosol properties over south-east Italy.
50x50 km²
Comments MODIS AOTs follow the temporal evolutio
n of AERONET AOTs at all tested window sizes and
are characterized by a significant seasonal
dependence. The temporal evolutions of ocean and
land-ocean mean AOTs are not dependent on window
size.
100x100 km²
100x100 km²
300x300 km²
Comments The ?M temporal evolution is not affect
ed by the window size monthly average values of
the 50x50 km² window size are rather similar to
those of 300x300 km² window size.
?M monthly means depend on seasons and take va
lues in the 0.7- 0.8 and 0.4- 0.6 range in
spring-summer and autumn-winter, respectively.
?A monthly means span the 0.7- 0.8 range during
all year and are not significantly affected by
seasons. It is possible that the marked seasona
l evolution of ?M is mostly due to the
MODIS-ocean algorithm that underestimates the
fine fraction contribution on autumn-winter
months.
300x300 km²
300x300 km²
Fig.3(a),(b) Temporal plots (red dots) of
MODIS-ocean fine fraction ?M (c) temporal plot
(red dots) of the AERONET fine fraction parameter
?A. Black full dots and error bars represent
monthly average values and corresponding standard
deviations. Blue boxes show on each panel the
monthly distribution of data points.
Fig. 2 Scatter plots of MODIS AOT referring to
the 50x50, 100x100, and 300x300 km² window size
centered on Lecce versus AERONET AOT mean values
collocated in time. Solid red and black lines
represent the linear regression lines and the 11
lines, respectively, dashed lines are MODIS pre
launch expected uncertanties.
Fig.2 Temporal evolution of MODIS (blue dots)
and collocated in time AERONET (red dots) AOTs
referring to the 50x50, 100x100, and 300x300 km².
Open blue and red dots represent monthly averaged
values of MODIS and AERONET AOTs collocated in
time, respectively.
Conclusions
MODIS AOT meet expected uncertainties
Regression lines fitting ocean- and
land-ocean-MODIS AOT values indicate that MODIS
overestimates AOTs at low aerosol loadings
The slope of the regression lines fitting the
scatterplots with land-ocean-MODIS AOTs is
closer to unity the land-ocean-MODIS AOTs better
represent the aerosol properties over south-east
Italy.
The temporal evolution of the MODIS fine fraction
?M (fig. 3) depends on seasons, while the
AERONET fine fraction ?A doesnt vary during all
year MODIS-ocean algorithm underestimates the
fine fraction contribution on autumn-winter?
Despite previous investigations on the validation
of MODIS retrievals, the results of this study
refer to a single site on south-east Italy where
different aerosol types may converge during the
year and many aerosol types can superimpose
mainly in summer as a consequence of the weather
stability. Then, the area can be well suited to
test the performance of MODIS retrieval
algorithms.
All these results can allow inferring that
AERONET AOTs retrieved at Lecce can be considered
representative at least of a 300x300 km² area
centered on Lecce. Hence locally-derived aeros
ol parameters can be of use in General
Circulation and Chemical Transport Models.