RS component Dimitry Van der Zande RBINS WP4 Intercalibration RS vs IN SITU WGS84 Seaborne measurements Total of 908 observations in the period 20032010 Water samples taken at depth of 05m 1m ID: 791974
Download The PPT/PDF document "WP4. Brainstorm Statistics" is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
WP4. Brainstorm Statistics-RS component-Dimitry Van der Zande (RBINS)
Slide2WP4.
Intercalibration
RS vs IN SITU
WGS84???
Seaborne measurements
Total of 908 observations in the period 2003-2010
Water samples taken at depth of 0.5m -1m
CHL determined using HPLC method (ISO certified measurements)
Most stations are located in the Belgian waters, but not all (North Sea, Mediterranean Sea).
Directly compared with MERIS dataset
Slide3WP4.
Intercalibration
RS vs IN SITU
WGS84???
MERIS
CHL dataset
ESA standard CHL1 and CHL2 products (MEGS 7.5)
Spatial resolution: 2km
Quality flagging
Actual
pixel value used
Maximal time difference allowed with in-situ observation: +/- 1h
Station ID
date IS
CHL IS
sensorID
date RS
CHL RS
230
4/22/03 10:53
6.31
2
4/22/2003 10:36
8.94
230
4/23/03 10:02
11.68
2
4/23/2003 10:0516.53MC166/16/03 10:185.5126/16/2003 10:075.722308/6/03 10:107.5428/6/2003 10:050.41MC48/6/03 10:506.4428/6/2003 10:050.091307/19/05 10:129.9827/19/2005 9:565.79BB56/7/2006 10:561.226/7/2006 10:380.86MH59/21/06 10:111.8929/21/2006 10:071.33W019/21/2006 10:301.429/21/2006 10:072.29ST193/17/2009 9:440.9423/17/2009 9:321.25ST513/23/2009 10:333.213/23/2009 9:431.52VC015/30/2009 10:590.5415/30/2009 10:461.28W046/16/2009 10:0611.7526/16/2009 10:1013.24MER4/20/2010 10:392.724/20/2010 10:316.02MH-44/21/2010 9:243.524/21/2010 10:003.46MER4/21/2010 10:062.424/21/2010 10:005.06ZBLR266/13/2010 10:490.5316/13/2010 10:340.67
Match-up results
28
matches with 1h time difference
Slide4WGS84???
WP4.
Intercalibration
RS vs IN SITU
Slide5WP4. Zone Delineation
CHL P90
(March-Oct 2003-2008)
A lot of stations seem to have the same name and are located close to each other
Slide6WP4. Zone Delineation
Time series extraction and processing tool (
TSproc
)
Slide7WP4. Zone Delineation
TSproc
: How to compare different CHL time series?
1) By comparing statistical descriptions:
CHLmean, CHLP90,
ABstart
CHLmean
CHLP90
ABstart
Slide8WP4. Zone Delineation
Data & metadata:
Remote Sensing data
TSproc
: How to compare different CHL time series?
1) By comparing statistical descriptions (
CHLmean
, CHLP90,
Abstart
) and look for clusters
First
results
do not show usable
clustering
Slide9WP4. Zone Delineation
Data & metadata:
Remote Sensing data
TSproc
: How to compare different CHL time series?
1) By directly comparing CHL time series of stations that are positioned within a predefined distance from each other.
CHL P90
(March-Oct 2003-2008)
Slide10WP4. Zone Delineation
Data & metadata:
Remote Sensing data
TSproc
: compare different CHL time series using simple regression analysis
Slide11WP4. Zone Delineation
Data & metadata:
Remote Sensing data
Slide12WP4. Zone Delineation
Data & metadata:
Remote Sensing data
-
determine
optimal
thresholds
for R² and
slope
-
Classify
BC zone
based
on the
actual
CHL profiles and
their
similarity
Slide13WP4. Zone Delineation
Data & metadata:
Remote Sensing data
TSproc
: compare different CHL time series using MAPE
MAPE=0.37
MAPE=0.29
MAPE=0.16
Slide14WP4. Zone Delineation
Data & metadata:
Remote Sensing data
TSproc
: compare different CHL time series using MAPE
MAPE
map
for station ‘DePauw9’
1
MAPE product showing how CHL time series in BCZ differ from station DePauw9
This product can be used to delineate zone of confidence (=zone with similar CHL patterns as station DePauw 9 in
considered time period
)
A classification product is possible to determine CHL zones in BCZ without taking the stations into account. Stations in the same zone can be considered similar