205
vides the means for fast data availability, however the accuracy of those algorithms still requires manual identification
by humans. We combined state of the art automatic image
classification by convolutional neural networks (deep learning) with a citizen science project to classify a large dataset
of ~3 million images from an Underwater Vision Profiler 5
(UVP5). On our website https://planktonid.geomar.de, citizen scientists can confirm or reject the automatic assignment
of UVP5 images to different plankton categories in a
memory- like game. Inbuilt quality controls and multiple
validations per image enable scientific analysis of the citizen
science data. We will present data on citizen scientist engagement, data quality and the distribution analysis of large protists (Rhizaria) in the Mauretanian, Benguela and Humboldt
Current upwelling systems.
3.2.3 Study Phytoplankton Dynamics
in the Arctic Using MERIS Sun-Induced
Fluorescence
J. R. El Kassar
1*
, R. Preusker
1
, J. Fischer
1
1
Institute of Space Sciences, Carl-Heinrich-Becker-Weg
6-10, 12165 Berlin, Germany
*corresponding author: jan.elkassar@mail.met.fu-berlin.
de
Keywords: Phytoplankton, Fluorescence, Chlorophyll,
MERIS, Arctic
Analyzing the physiology of phytoplankton with suninduced fluorescence (SIF) is a crucial aspect of ocean color.
In this study, however, SIF and chlorophyll content have been
used to study the phytoplankton dynamics in the Arctic.
Chlorophyll content is taken from MERIS’ “Algal Pigment
Index 1” (AL1) dataset, derived from a blue-green ratio. SIF
is estimated with the fluorescence line height (FLH) which is
processed from MERIS reflectances in the red spectrum. We
created a monthly climatology for the years 2003–2011 over
the Arctic (60°N–90°N) and conducted a regression analysis
between FLH and AL1. Monthly averages show that AL1
peaks in summer. FLH, however, peaks in spring and fall. The
regression analysis shows a unimodal relationship between
FLH and AL1 with high slopes in spring and autumn, whereas
the relationship splits into two separate modes in summer.
The upper mode shows high slopes similar to these in spring,
whereas the lower mode shows very low slopes indicating
weak or no correlation between AL1 and FLH. These two
regimes are also visible in the spatial distribution of the ratio
FLH/AL1. The results from the regression analysis correspond with the Arctic annual surface chlorophyll cycle.
Spatial patterns of FLH/AL1 also align with currents exporting fresh water from the Arctic and currents advecting warm,
salty water from the Atlantic which influence the stratification. This suggests that the results are related to seasonal and
regional stratification processes and the vertical distribution
of phytoplankton. The hypothesis is that FLH and AL1 contain information about phytoplankton from different depths
due to different signal depths in the red (FLH) and blue-green
(AL1) spectrum. Thus a combination of both parameters
might be suitable to analyze not only the physiology but also
the vertical distribution of phytoplankton.
3.2.4 Hyperspectral Simulation of Chl-a
Fluorescence in Optically Complex Waters
Therese Keck
1*
, Lena Kritten
1
, René Preusker
1
, Jürgen
Fischer
1
1
Institute for Space Sciences, Freie Universität Berlin,
Carl-Heinrich-Becker-Weg 6-10, 12165 Berlin, Germany
*corresponding author: therese.keck@wew.fu-berlin.de
Keywords: Ocean color, Fluorescence, Remote sensing
Phytoplankton is one of the main constituents in oceans,
coastal and inland waters. Observing phytoplankton from
space, mainly the pigments called chlorophyll are detected
due to a very characteristic spectral properties. After correcting the remote sensing signals for the atmosphere, several
techniques can retrieve the chlorophyll concentration. During
photosynthesis, the pigments convert a part of the incoming
visible light to photochemical energy for living. The other
part dissipates as heat and is emitted as chlorophyll-a fluorescence. Optically complex waters contain various constituents like colored dissolved organic matter (cDOM) or
sediments which can change the remotely sensed signal. The
chlorophyll-a fluorescence located close to 682 nm is found
to be relatively insensitive to other constituents. For phytoplankton populations close to the surface the fluorescence
line height (FLH) gives good results for chlorophyll-a concentrations. The simulation of radiance in and above water
enable us to understand how phytoplankton stratification and
additional constituents influence .the fluorescence peak. In
future there will be hyperspectral satellite sensors available
(e.g., EnMAP) which may be used for novel hyperspectral
fluorescence algorithms.
3.2.5 Characterization of CDOM and FDOM
in the Nordic Seas
Anna Raczkowska
1,2*
, Piotr Kowlaczuk
1
, Sławomir Sagan
1
,
Monika Zabłocka
1
, Mats A. Granskog
3
, Alexey K. Pavlov
3
,
Colin Stedmon
4
1
Institute of Oceanology, Polish Academy of Sciences, ul.
Powstańców Warszawy 55, 81-712 Sopot, Poland
2
Centre for Polar Studies, Leading National Research
Centre, 60 Będzińska Street, 41-200 Sosnowiec, Poland
3
Norwegian Polar Institute, Fram Centre, 9296 Tromsø,
Norway
4
National Institute for Aquatic Resources, Technical
University of Denmark, 2920 Charlottenlund, Denmark
*corresponding author: anraczkowska@gmail.com
Keywords: Absorption, Fluorescence, DOM, Polar
regions
Appendices
vides the means for fast data availability, however the accuracy of those algorithms still requires manual identification
by humans. We combined state of the art automatic image
classification by convolutional neural networks (deep learning) with a citizen science project to classify a large dataset
of ~3 million images from an Underwater Vision Profiler 5
(UVP5). On our website https://planktonid.geomar.de, citizen scientists can confirm or reject the automatic assignment
of UVP5 images to different plankton categories in a
memory- like game. Inbuilt quality controls and multiple
validations per image enable scientific analysis of the citizen
science data. We will present data on citizen scientist engagement, data quality and the distribution analysis of large protists (Rhizaria) in the Mauretanian, Benguela and Humboldt
Current upwelling systems.
3.2.3 Study Phytoplankton Dynamics
in the Arctic Using MERIS Sun-Induced
Fluorescence
J. R. El Kassar
1*
, R. Preusker
1
, J. Fischer
1
1
Institute of Space Sciences, Carl-Heinrich-Becker-Weg
6-10, 12165 Berlin, Germany
*corresponding author: jan.elkassar@mail.met.fu-berlin.
de
Keywords: Phytoplankton, Fluorescence, Chlorophyll,
MERIS, Arctic
Analyzing the physiology of phytoplankton with suninduced fluorescence (SIF) is a crucial aspect of ocean color.
In this study, however, SIF and chlorophyll content have been
used to study the phytoplankton dynamics in the Arctic.
Chlorophyll content is taken from MERIS’ “Algal Pigment
Index 1” (AL1) dataset, derived from a blue-green ratio. SIF
is estimated with the fluorescence line height (FLH) which is
processed from MERIS reflectances in the red spectrum. We
created a monthly climatology for the years 2003–2011 over
the Arctic (60°N–90°N) and conducted a regression analysis
between FLH and AL1. Monthly averages show that AL1
peaks in summer. FLH, however, peaks in spring and fall. The
regression analysis shows a unimodal relationship between
FLH and AL1 with high slopes in spring and autumn, whereas
the relationship splits into two separate modes in summer.
The upper mode shows high slopes similar to these in spring,
whereas the lower mode shows very low slopes indicating
weak or no correlation between AL1 and FLH. These two
regimes are also visible in the spatial distribution of the ratio
FLH/AL1. The results from the regression analysis correspond with the Arctic annual surface chlorophyll cycle.
Spatial patterns of FLH/AL1 also align with currents exporting fresh water from the Arctic and currents advecting warm,
salty water from the Atlantic which influence the stratification. This suggests that the results are related to seasonal and
regional stratification processes and the vertical distribution
of phytoplankton. The hypothesis is that FLH and AL1 contain information about phytoplankton from different depths
due to different signal depths in the red (FLH) and blue-green
(AL1) spectrum. Thus a combination of both parameters
might be suitable to analyze not only the physiology but also
the vertical distribution of phytoplankton.
3.2.4 Hyperspectral Simulation of Chl-a
Fluorescence in Optically Complex Waters
Therese Keck
1*
, Lena Kritten
1
, René Preusker
1
, Jürgen
Fischer
1
1
Institute for Space Sciences, Freie Universität Berlin,
Carl-Heinrich-Becker-Weg 6-10, 12165 Berlin, Germany
*corresponding author: therese.keck@wew.fu-berlin.de
Keywords: Ocean color, Fluorescence, Remote sensing
Phytoplankton is one of the main constituents in oceans,
coastal and inland waters. Observing phytoplankton from
space, mainly the pigments called chlorophyll are detected
due to a very characteristic spectral properties. After correcting the remote sensing signals for the atmosphere, several
techniques can retrieve the chlorophyll concentration. During
photosynthesis, the pigments convert a part of the incoming
visible light to photochemical energy for living. The other
part dissipates as heat and is emitted as chlorophyll-a fluorescence. Optically complex waters contain various constituents like colored dissolved organic matter (cDOM) or
sediments which can change the remotely sensed signal. The
chlorophyll-a fluorescence located close to 682 nm is found
to be relatively insensitive to other constituents. For phytoplankton populations close to the surface the fluorescence
line height (FLH) gives good results for chlorophyll-a concentrations. The simulation of radiance in and above water
enable us to understand how phytoplankton stratification and
additional constituents influence .the fluorescence peak. In
future there will be hyperspectral satellite sensors available
(e.g., EnMAP) which may be used for novel hyperspectral
fluorescence algorithms.
3.2.5 Characterization of CDOM and FDOM
in the Nordic Seas
Anna Raczkowska
1,2*
, Piotr Kowlaczuk
1
, Sławomir Sagan
1
,
Monika Zabłocka
1
, Mats A. Granskog
3
, Alexey K. Pavlov
3
,
Colin Stedmon
4
1
Institute of Oceanology, Polish Academy of Sciences, ul.
Powstańców Warszawy 55, 81-712 Sopot, Poland
2
Centre for Polar Studies, Leading National Research
Centre, 60 Będzińska Street, 41-200 Sosnowiec, Poland
3
Norwegian Polar Institute, Fram Centre, 9296 Tromsø,
Norway
4
National Institute for Aquatic Resources, Technical
University of Denmark, 2920 Charlottenlund, Denmark
*corresponding author: anraczkowska@gmail.com
Keywords: Absorption, Fluorescence, DOM, Polar
regions
Appendices