Comparisons of ocean colour merged data sets

1
Comparisons of ocean colour merged data sets

• Stéphane Maritorena
• ICESS
• University of California, Santa Barbara

Special thanks to Odile, Antoine and the ACRI-ST
Team
2
OUTLINE

• Current Ocean Colour merged data sets and
  associated merging techniques and products.
• Coverage
• Time-series
• Matchups
• Frequency distributions
• Error estimates

3
Ocean Color Merged Data Sets

• REASoN (ICESS/UCSB)
• NASA OBPG
• GlobCOLOUR

4
MERGED DATA SETS
5
Daily coverage using SeaWiFS/Aqua and Meris (2003
data)
GlobCOLOUR
NASA REASoN
6
Ocean color sensorsaverage daily coverage (2003
data)
7
FREQUENCY OF COVERAGE (2005)
Days
Merged
SeaWiFS
Aqua
8
Relative contribution of each sensor to an 8-Day
composite
MERIS
AQUA
SeaWiFS
9
THE 30 ZONES
10
Time-series CHL - Monthly data (Contd)
11
Time-series CHL - Monthly data (Contd)
12
Time-series - CDM - Monthly data
13
Time-series BBP - Monthly data
14
Issues with the particulate backscattering product
Meris issue
Mostly a SeaWiFS issue

• Noise in the SeaWiFS Lwn(?) around gaps caused
  by clouds. In these areas, the Lwn(?) are
  sometimes higher than in nearby gap-free areas
  and this results in enhanced bbp (443) values.
• Problem does not exist in Aqua data
• MERIS shows some stripes of high BBP values on
  some swaths.

15
Matchups - CHL
In situ data set (NOMAD SeaBASS) with CHL
3100 stations CDM 700 stations BBP 180
stations Same day matchups, 3x3 box, 9 km data
(4.5 km for GlobCOLOUR).
16
BBP
CDM
GlobCOLOUR
REASoN
17
FREQUENCY DISTRIBUTION - CHL - 50N-50S Deep Water
18
FREQUENCY DISTRIBUTION - CDM - 50N-50S Deep Water
19
FREQUENCY DISTRIBUTION - BBP - 50N-50S Deep Water
20
Monthly Chlorophyll May 2006
21
(No Transcript)
22
Summary

• The GlobCOLOUR data set provides better daily
  coverage thanks to the use of the MERIS data.
• GlobCOLOUR also offers more products and several
  resolutions.
• CHL
• The three merged data sets are very consistent
  most of the time, except in coastal zones (Z lt
  1000 m).
• MERIS alone tends to produce higher CHL values
  than SeaWiFS or AQUA.
• AQUA alone tends to produce lower CHL values
  than SeaWiFS or MERIS

23
Summary (Contd)

• CHL (Contd)
• The REASoN and GlobCOLOUR (GSM) data contain less
  low CHL values than the NASA OBPG data. Also true
  for REASoN at high CHL values but better
  agreement in summer.
• CDM
• The agreement between the GlobCOLOUR and REASoN
  merged CDM products is excellent, always,
  everywhere.
• BBP
• SeaWiFS and MERIS BBP products are sometimes very
  (too) high. MODIS-AQUA BBP product appears more
  stable and reasonable.

24
Conclusion

• The three merged data sets look good and are in
  good agreement overall.
• Better agreement between the merged products than
  between the products from the individual sensors.
• Some issues exist (BBP mostly) that need to be
  looked at.