N OVERVIEW OF SPECTRAL IN VIVO FLUORESCENCE …

www.jerico-fp7.eu

AN OVERVIEW OF SPECTRAL IN VIVO FLUORESCENCEMETHODS FOR PHYTOPLANKTON TAXONOMY

JUKKA SEPPÄLÄ, SEPPO KAITALA, MIKA RAATEOJA, STEFAN SIMIS, PASI YLÖSTALO

FINNISH ENVIRONMENT INSTITUTE

5th FerryBox Workshop - Celebrating 20 Years of Alg@lineApril 24-25, 2013 Helsinki

Pigment taxonomy Spectral fluorescenceTaxonomic spectral groups

Phycobilin fluorescence Origins of Phycocyanin /Phycoerythrin fluorescenceInstrumentationExamples for Baltic Sea

Spectral fluorescence Measuring techniquesInstrumentationData analysisExamples


SPECTRAL FLUORESCENCE: two wavelengthdimensions - excitation and emission - allowingbetter separation of various substances.

Babichenko et al 2000

TAXONOMIC SPECTRAL GROUPS: In living phytoplankton,accessory pigments in PSII transfer energy to Chla but do not emitfluorescence, only Chla in PSII and phycobilin pigments emitfluorescence

Sepp

älä

2009

TAXONOMIC SPECTRAL GROUPS: Group-specificExcitation-Emission Matrix is the basis forfluorescence based taxonomy

Nodularia Aphanizomenon Green Synechococcus

Red Synechococcus Cryptomonad Green algae

Seppälä et al 2007

Babi

chen

ko,P

oryv

kina

,Kai

tala

1995

Sepp

älä

&Ba

lode

1998

PHYCOBILIN FLUORESCENCE, EARLY YEARS”We may not know exactly what we are measuring, but the patternsobserved are too strong to ignore” Cullen & Renger 1979.

Seppälä, Ylöstalo, Kuosa 2005

PHYCOCYANIN FLUORESCENCE, Operationaldetection of filamentous cyanobacteria started2005

PHYCOCYANIN FLUORESCENCE, Operationaldetection of filamentous cyanobacteria started2005

PHYCOERYTHRIN FLUORESCENCE, First tests in 2006with old Turner 10-AU

Niiranen 2008

PHYCOERYTHRIN FLUORESCENCE, New trials withLED fluorometer in 2012

Testing fluorometers for phycoerythrin (PE) detection:Chelsea Instruments PE Unilux in flow-through system during summer cruise

Phycoerythrin signal fromspectrofluorometry

exci

tatio

n

emission

PHYCOERYTHRIN FLUORESCENCE, New trials withLED fluorometer in 2012 2013 FerryBox.

FlowCAM to identify and count PE containinglarger cells (>5 µm)PE containing species: Mesodinium rubrum,Dinophysis norvegica, Cryptomonads, colonialcyanobacteria

Picocyanobacteria counts (microscopy, flowcytometry)

Photos: Baltic Sea Portal; nordicmicroalgae.org

PHYCOBILIN FLUORESCENCE: Current issues

• Instrument selection: Excitation/emission wavelengths vary betweeninstruments. Not all are strictly specific for Phycocyanin/Phycoerythrin


• Instrument selection: Excitation/emission wavelenghts vary betweeninstruments. Not all are strictly specific for Phycocyanin/Phycoerythrin

• Calibration: Solid secondary standards provide stable and traceable wayfor monitoring instrument performance, but chemical standards would beneeded for instrument comparisons and concentration measurements.




• Validation: More information is needed for biomass-fluorescencerelationship

Izydorczyk etal 2005Bowling etal 2010

Seppälä et al 2007


Attil

aet

al.u

npub

l.



• Validation: More information is needed for biomass-fluorescencerelationship

• Use of data: EO validation, Ecosystem model validation, visualization

SPECTRAL FLUORESCENCE: Measuring techniques

Several fluoroprobes Multichannel fluorometers Spectral fluorometers

1-4 k€/channelLimited taxonomicdiscriminationSensitive tobackground noiseOnly simple statisticspossibleSeveral manufacturers

>20 k€Limited detection ofphycobilinsCorrection forbackground noiseMultivariate statisticspossibleBbe-Moldaenke,JFEAdvantech

Lab instruments <20k€EEM measurementtime consumingBackground correctionMultivariate statisticspossibleNo commercial FBdevices

SPECTRAL FLUORESCENCE: How to retrievetaxonomic signal from fluorescence spectra

Ratio plots, clusteringSimilarity indicesPrincipal component analysisWaveletsLinear unmixing, multivariate regression

22 23 24

E longitude

Riga

LATVIA

Saaremaa

10 Nm

ESTONIAD

Mac

Inty

reet

al20

10

Sepp

älä

&Ba

lode

1998



Alex

ande

reta

l201

2

Ric

hard

son

etal

2010



Sepp

älä

2009

CHL a [µg L-1], observed

0 5 10 15 20 25

CH

La

[µg

L-1

]pr

edic

ted

0

5

10

15

20

25R2 = 0.98

CHL a in <2µm [µg L-1], observed

0 1 2 3

CH

La

in<2

µm[µ

gL

-1] ,

pred

icte

d

0

1

2

3R2 = 0.86

RMSECV = 0.23RMSECV = 0.69

Observed biomass [µg / L]10 100 1000

Pred

icte

dbi

omas

s[µ

g/L

]

10

100

1000R2 = 0.69

RMSECVFil. Cyano = 164Diatoms = 279Dinoflagellates = 459

SPECTRAL FLUORESCENCE: Identifying spatialstructures to locate patches and facilitate sampling

Alex

ande

reta

l201

2

PCA identifies 4spectral clusters,or species groups

Clusters arehorizontallyseparated (colorcoding)

…as well asvertically (colorcoding)

Study at Lake Victoria;

SPECTRAL FLUORESCENCE: Identifying seasonalstructures, Gulf of Finland

Chla a excitation spectra (380-700nm / 730 nm)PCA

%ex

plai

ned

# of PC

SPECTRAL FLUORESCENCE: Identifying seasonalstructures, Gulf of Finland

Chla a excitation spectra (380-700nm / 730 nm)PCA

3. PC

Season

Dep

th2. PC

1. PC

4. PC

5. PC

Wavelength

%ex

plai

ned

# of PC


What next?

Calibration / validation (HPLC, FlowCAM, FCM, EO)Collection of validation data (intelligence)Instrumentation (preferably spectral)New algorithms (multivariate)Use of data

N OVERVIEW OF SPECTRAL IN VIVO FLUORESCENCE …

Documents