Final Report Summary - PROTOOL (Productivity tools: Automated tools to measure primary productivity in European seas) Executive summary:primary production (PP), i.e. the process by plants and algae fix atmospheric CO2 using the sunlight as energy, forms the basis of all ecosystem processes. Understanding and management of marine ecosystems is thus impossible without accurate knowledge of how primary production responds to environmental change. In spite of this, primary productivity measurements are not part of the routine monitoring programs by most EU Member States, are not included in the Water or Marine Strategy Framework Directives (WFD and MSFD), and are seldom included in any European operational oceanography programs. The most likely reason is that the current methodology, based on the measurement of uptake of radioactive labeled CO2 is not suitable for routine applications. The PROTOOL project will now make this possible because it developed hardware, software and parameter estimation algorithms to measure primary production autonomously, with the aim of installing the equipment on ships of opportunity. Developing the tools to do this was the main aim of PROTOOL, and these goals have been realized so that implementation is the next phase.the hardware consisted of:1) an automated 'active' fluorometer to measure photosynthetic activity and phytoplankton biomass using the main pigment chlorophyll-a as a proxy,2) a spectral light reflectance setup allowing remote sensing at shipboard level (in order to obtain water quality parameters and3) an in-line absorption meter (PSICAM) to measure the light absorption by the algae and to obtain inherent optical parameters which can be used to derive information on the water constituents.project Context and Objectives:coastal seas, shelf seas and oceans are under severe anthropogenic pressure as a direct result of intensive fishing, pollution and eutrophication and indirectly via global climate change. Already about 50%, approximately 3 billion people, live near the coastline, a percentage which is expected to double by 2025, hence the anthropogenic pressure on coastal seas will rise. Such pressures have altered the ecosystem foodweb structure, and resulting 'regime shifts' in many marine ecosystems in the Pacific the Northern Atlantic Ocean, the North Sea, the Wadden Sea and the Baltic Sea have been documented. These regime shifts can have opposite effects on primary production and stimulate harmful/toxic algal blooms, causing deterioration in food quality of higher trophic levels and a decline in key ecosystem services, such as the fisheries. Of all factors, eutrophication appears to be the most common cause of regime shifts observed to date, especially in coastal seas, and has been a subject of international concern. Measures have been established to half the loads of phosphorous (P) and nitrogen (N), and these measures have been partially successful, at least in the southern North Sea: P but not N loadings have been substantially lowered leading to large changes in the N:P-ratio, an important structuring factor in phytoplankton communities, but also in increasing occurrences of coastal P-limitation. in order to accomplish our goals, we have divided the project in 5 subprojects (SPs):SP1: instrument developmentSP2: field testing and primary productivity measurementsSP3: conversion factorsSP4: IntercalibrationSP5: data centerSP1: instrument developmentthe objectives for SP1 are the following:- Develop a PROTOOL Fluorometer (WP2): this fluorometer will be developed in 3 stages: the first version (baseline model) will be put together from existing equipment, and with this module most of the field work will be done. This included the development of a PC controlled LED light panel by partner 8 (PSI). The performance of the baseline model should be input for the development of the prototype and the final version. The final version should be ready near the end of the project and will undergo only limited field testing.- Development of a PROTOOL absorption meter, the Point Source Integrating Cavity (PSICAM, WP3). The PSICAM should measure and quantify the spectral characteristics of the different water constituents (CDOM, suspended matter, algal absorption and algal biomass (as chla). From these measurements the rate of light absorption should be distinguished, a parameter necessary to quantify the rate of photosynthetic electron transport when using the bio-optical model.- Development of a PROTOOL reflectance module (WP4). This module performs hyperspectral remote sensing at shipboard level. The measurements should be used to obtain the important water quality (WQ) parameters chla and the light attenuation coefficient (Kd).hardware is not the only part of SP2. Some of the parameters are not measured directly but should be retrieved from the raw data using specially designed algorithms. This requires validation and for this reason field work has to be done.SP2, field testing has a dual role. The objectives of the subproject are:- To test the developed equipment under realistic field condition to check on its performance. Because there is a large diversity of water types in the EU the PROTOOL approach has to be able to work in all water types present in the EU. Therefore work was carried out in Estuaries (WP5), the brackish Baltic Sea with a large contribution of cyanobacteria, which are normally a challenge to active fluorometers, WP6), the North Sea (WP7, including work on the Celtic Sea and some Channel data) and the oligotrophic waters of the north east Atlantic (WP8). Originally we planned to measure in the Channel and Gulf of Biscay, but because the ferry used for this work was taken out of service, a number of cruises on the north east Atlantic were performed as an alternative.- To obtain data necessary for calibration and for validation of the algorithms developed for the different instruments in SP1. Due to the different types of water bodies we envisage an approach that might result in regional algorithm.SP3, conversion factorsPROTOOL adopted the Fast Repetition Rate fluorescence (FRRf) approach to measure marine primary productivity from seas and oceans. However, FRRf measurements suffer from two fundamental limitations: PP is measured (i) relative to the number of 'photochemical reaction centers' and not relative to the volume of water sampled; and (ii) in a 'currency' of electrons evolved (the so-called electron transfer rate, ETR) and not CO2 fixed. Therefore conversion factors must be applied to all FRRf-ETR measurements to yield measurements of primary production in units of C per unit time per unit area. Thus the fundamental aim of WP9 was to establish novel algorithms to enable implementation of these conversion factors.in order to convert the ETR to CO2 uptake knowledge of the 'quantum requirement for carbon fixation (QR) is required (assuming that one absorbed photon leads to the production on 1 electron finally extracted from water). Three parallel approaches were taken;- A meta-analysis was performed to compile existing QR data sets with environmental and taxonomic parameters to develop predictive algorithms- Collect new high resolution data to support past data sets during dedicated PROTOOL cruises, in particular for special marine regions (Baltic Sea and the Eastern Scheldt). In addition,- Perform additional laboratory culturing of phytoplankton to experimentally identity the role of key factors, e.g. nutrient availability upon the QR of different algal taxa.SP4, intercalibrationin order to facilitate the development of algorithms necessary for the different modules it is good to know if the measurements carried out during the fieldwork is comparable. For this reason a number of intercalibration exercises were carried out:- A pigment intercalibration exercise. This involved sending out pigment samples (both standards and field samples) collected by one partner to the other participating partners in the intercalibration exercise.- A primary production intercalibration. This intercalibration was aimed at testing the mathematical procedures followed to come to an estimate of estimates of daily primary production from a raw dataset. NIOZ mailed a dataset including irradiance data, data from a scintillation counter, added amount of radioactive bicarbonate, in situ and incubation temperatures, light attenuation coefficients and total dissolved inorganic carbon. The participants were asked to compute the photosynthetic characteristics, daily and integrated water column primary production.SP5, datacenter and disseminationproper data management is essential for the success of the project. Because of the large amount of data generated by the different PROTOOL modules and versions, automated fitting routines will be developed and special emphasis will be given to treatment of outliers, which will be manually checked. After quality control and quality assurance (QaQc), data will be added to the project database. We plan to host the database on the dataservers of the NIOZ and make the data available via a web interface.the NIOZ will develop two databases:- A C14 database, developed in MS Access. This database will hold the raw data obtained from the 14C-uptake experiments. It contains fitting routines, written in the R (see http://cran.r-project.org online) to fit the photosynthesis irradiance (P/I) data and obtain the parameters which describe the shape of the P/I-data (maximum rate of photosynthesis, photosynthetic affinity and light saturation parameter. It contains three different models to fit the P/I data. When the light attenuation coefficient is known and the hourly irradiance, it should be able to compute daily primary production in the water column and integrate the daily primary production estimates over a defined interval.- A MySQL database which will hold the data obtained from the different TriOS-Ramses sensors (data from the reflectance module, including GPS coordinates). There should be a software routine which flags unreliable data and after passing the filter criteria it should computes spectral reflectance and from this it the chla concentration and light attenuation coefficient according to 4 different algorithms (using an R-script). The database will also hold data from the PROTOOL fluorometer. From the FRRF data the RLC curve parameters will be calculated and from this it will compute daily water column primary productivity and integrate it over a set time.project Results:WP2: the PROTOOL fluorometerthe goal of the project was to develop and test the device for measuring and calculation of Primary production using variable fluorescence measurement and to implement the device with all other supportive devices into ferry box and other ships of opportunity to measure primary production in a simple and cheaper way with results comparable to present instruments (e.g. Fastracka device produced by Chelsea Instruments, Great Britain). During the project several generations of the instrument were constructed and tested: the first FRR prototype, the laboratory experimental version used for routine testing and finally the industrial flow-through version that was successfully incorporated into the ferry-box. Equipment proved to be functional and its parameters are in accordance with the expectations. The flow-through device is fully functional and is now under the implementation process. The design of the initial and the experimental versions of the FRRF fluorometers was derived from the standard fast fluorescence kinetics measuring devices produced by PSI company. hardware. The device runs under LINUX operation system and it can be remotely operated via internet connection. The Fast Repetition Rate fluorometer (FRRf) consists of 2 main parts: FFL-040 Fluorometer and Operator PC with web server. The FRRf instrument is built into the IP67 alluminium box case. It contains 2 holes for measuring tubes, power switch, light source intensity indicator and waterproof connectors. Left (measuring) port contains the measuring and actinic lights together. Right (adaptation) port contains actinic light only. This port is also used for circulation of the sample during the measurement. Actinic light intensity and light spectrum are the same for both ports. Temperature stabilization of the sample by the double sided tube is possible in the adaptation port. Circulating water in the outer tube maintains stable temperature of the sample. Measuring light sources (590nm and 460nm) are operated during the FRR induction sequence of the experiment. High power LED diodes controlled by high speed electronic driver circuit trigger measuring flash train of µs-scale light pulses (flashlets). Timing of these triggers may be controlled with 100ns accuracy. AD converter acquisition trigger is synchronized with light triggers also with 100ns accuracy. Both measuring light sources use 6 LEDs from each side of the tube. These LEDs are placed in perpendicular direction to the detector port. LED chips are placed close to each other for getting the best homogeneity of the light in the measuring tube. Double side excitation as well as optical parameters of the glass tube also improve the homogeneity of the light inside the measuring tube. LEDs illuminate 20mm of the sample in the middle of the detector port. Use of the high power LEDs must follow special design precautions for proper cooling of individual chips. Heating of the LEDs by current flow lowers the LED light intensity, especially in the amber measuring light source (see the LED documentation below). Actinic light sources (amber 590nm and blue 460nm) are operated during the rapid light curve (RLC) measurement. graphical User Interface. The FFL 040 fluorometer communicates with Operator PC via RS232 serial port. USB to serial converter is included for converting the data flow to USB bus standards. Data readout, data storage, pump and solenoid switching logic is controlled from the Operator PC. Specific communication protocol is implemented for this purpose, enabling to operate more FFL-040 fluorometers connected to one Operator PC. Operator PC reads out the calibration constants from all the connected FFL-040s.measurement with flow-through system runs under the control of the FRRF Operator PC. It is currently the standard HP laptop computer with Linux system installed. This software includes also the web server application (Java6 compatible) running for online network access. All data measured with the flow-through fluorometer FFL-040 are stored on Operator PC.the fluorometer GUI starts automatically after switching ON the FRR (Fast Repetition Rate) Linux controller Operator PC. Prior to the start of the automated measurements, few protocol parameters can be set to allow more user-defined variability. Automated setting of these parameters has not been implemented yet, but will be available in the commercial types of this device. The software plots actual data flow. FRR induction sequence fit is executed with the ST1 data. Results of this fitting procedure are presented. Different parameters are obtained for Blue and Amber colors. Standard deviation is calculated for all computed values. Statistical parameters and p-value inform about the quality of the fit. Fitting process is executed automatically by running the R-script file in the R language. WP3 Point Source Integrating Cavityimportant for the determination of primary production in the sea and for general water quality assessment is the exact measurement of the light absorption by various water constituents, as e. g. the absorption by phytoplankton itself. Goal of this work package was to build a compact, commercial instrument based on the point-source integrating-cavity absorption meter approach that can be used as an underwater in situ sensor as well as a flow-through bench instrument for ferry box implementation. The instrument consists of an integrating sphere, a special LED-light source and a miniature spectrometer. The design of the light source and the optical connection of the light source and the detector with the sphere were challenging. The first version of the instrument was tested in the field and the validity and reliability of the measuring concept and the new construction were proofed.the aims of this work package was to develop, first, a modular baseline version of the PSICAM absorption meter that is used to test the PSICAM principal in an automatic flow-through mode connected to a ferry box system and to develop and evaluate necessary calibration, cleaning, and maintenance procedure. Second, an autonomous version of the PSICAM shall be build as a stand-alone commercial instrument which can be part of the final PROTOOL-MD system.WP 4 Reflectance moduleaimsthe aim of this work package was to develop an independent module to obtain the spectral water-leaving reflectance (water colour) from ships-of-opportunity. The module supports integration with the other PROTOOL modules (WP2 and 3) through database connectivity. Three aspects define how reflectance complements the PROTOOL package:- Light absorption and scattering processes in the surface layer of the sea can be derived from the measured reflectance- Remote optical sensors (e.g. on satellites) are used to measure reflectance of large sea areas, allowing primary productivity measurements to be placed in a spatial context- Because the sensors are placed outside, the module records the actual intensity of sunlight available for photosynthesisthe development of the Reflectance module is described in detail in PROTOOL public report D4.15. The module provides support for the use of commercially available spectroradiometers (TriOS RAMSES), which are commonly used for in situ reflectance measurements.the development of a supporting platform for moving ships-of-opportunity included:- research into the optimization of the spatiotemporal coverage, notably avoiding capture of sun glint and minimizing the recording of reflected sky radiance.- development of an automated processing scheme to produce reflectance spectra from the spectroradiometric measurements- research into the performance of algorithms that interpret reflectance in terms of useful water quality parameters, such as the biomass of phytoplankton in terms of the pigment chlorophyll-athe algorithm development is described in detail in PROTOOL public report D4.16. Highlights of the developed hardware/software solutions as well as the algorithm results are provided below for each of the approaches listed above.optimizing spatiotemporal coveragereflectance is measured as the ratio of upwelling over down-welling light. In above-water measurements, a radiance sensor with a narrow field of vision is pointed at the water at an oblique angle. The signal measured with this sensor consists of light that has been reflected in the water column, but also a contribution of light that was reflected at the water surface. The latter does not carry information about light absorption and scattering processes in the water itself. To correct for this contribution, a separate sensor is pointed at the sky to record the spectral quality of the sky radiance. A third sensor captures all downwelling irradiance. Previous studies (Austin 1974, Mobley 1999) have shown that the best results are obtained when the water leaving radiance is captured at an angle away from the sun, to avoid sun glitter which is usually brighter than the radiance reflected in the water column.deriving reflectancefor every measurement, it is necessary to estimate the contribution of sky radiance reflected at the water surface to the water leaving radiance. After the viewing angle has been optimized to avoid direct sun glint, this contribution is still variable and depends on illumination conditions (clouds, sun angle) and the roughness (waves) of the sea surface (Doxaran et al. 2004, Toole et al 2000). The sky radiance contribution has previously been modelled after wind speed, as a proxy for wave slope. Unfortunately, this method can easily lead to inaccuracies, particularly in relatively clear waters where the water leaving radiance is weak compared to the reflected sky radiance. The existing methods also do not provide any means to judge, without expert involvement, whether the obtained reflectance is valid or not. Invalid measurements occur when reflectance is measured under broken cloud cover, in rain or fog, or when exposed ship heaving during storms.deriving water quality productstwo quantities that can be derived from reflectance are of central importance to estimate primary productivity: the absorption of light by photosynthetic pigments, and the number of photons available for light harvesting, in the mixed layer. Relatively little attention was given to the retrieval of the vertical diffuse attenuation coefficient that predicts light intensity at depth given the irradiance measured at the surface (which the reflectance module provides). This coefficient can be modelled after light absorption and backscattering and solar elevation (Lee et al. 2005), and relies primarily on the accurate inversion of reflectance to these bulk inherent optical properties. More complicated is the retrieval of partitioned absorption, such as the absorption of light by phytoplankton (often scaled to the concentration of the pigment chlorophyll-a). Quantifying this absorption from reflectance is one of the central themes in spaceborn remote sensing of the global ocean. In coastal seas, the optical composition of the water is complex, and reflectance algorithms must be designed to overcome this complexity.sub-Project 2 (SP2): Field testing and Primary Production mappingfor PROTOOL to succeed, fieldwork is necessary to 1) test the different developed PROTOOL equipment (baseline, prototype and final versions) under relevant field conditions, and 2) obtain the field data necessary for validation of the algorithms developed as part of sub-project 1 (instrument development). Most of the testing took place on research vessels, although reflectance modules have been used on the ferry Finnmaid on the Baltic Sea. In addition during this SP2 we obtained data from which the conversion factors (electron requirements for C-fixation) could be obtained. This data was part of the input for SP3, in which WP9 was devoted to evaluating, understanding and predicting the conversion factors needed to obtain rates of C-fixation from ETR.WP5 Field testing in the Dutch estuaries.aim and site descriptionestuaries are an important testing ground for the PROTOOL products. Estuaries carry out important ecosystem services (e.g. nutrient removal, nursery grounds for fish, sustain fisheries by virtue of primary production) but estuaries also face immense anthropogenic pressure because they concentrate the material released from the watershed in the rivers into a relatively small area. Hence, careful management of estuaries is a necessity, and this requires knowledge of primary production.reflectance resultsthe reflectance module was implemented by putting a system of 3 sensors on the bow of the ship and manually adjusting the viewing angle relative to the sun. The biweekly/monthly cruises delivered a large quantity of data and a series of filters were developed which were implemented in a MySQL database (see D11.24 for more information). Approximately 25% of the data do not pass the criteria for several reasons (instrument failure, sunglitter, measurement of different parts of the sky).primary productionthe PROTOOL FRR baseline fluorometer was used to map the parameter (Fv/Fm) which is the maximum photosynthetic quantum efficiency of PSII. This parameter is an indication of the algal condition, and a decrease in the maximum value might be related to nutrient limitation, photoinhibition or other stress parameters. Fv/Fm was low in winter in both estuaries, a possible effect of the low temperatures in winter. In the Western Scheldt the decrease in winter was more pronounced, especially in the brackish water region of the estuary, suggesting that salinity stress also added to the lowered Fv/Fm values. Lowered values during the growth season were mainly observed in the Eastern Scheldt, an indication that they might suffer from nutrient limitation.primary production form the FRRf measurements was calculated using 3 algorithms, including two new ones developed during the project and which remove the need for some assumptions (rate of light absorption in the Absorption algorithm and the concentration of PSII in the Sigma algorithm, see D9.22 and Oxborough et al 2012). All algorithms effectively captured the seasonal dynamics in daily primary production as measured with the 14C technique. Seasonal dynamics in the quantum requirements for C-fixation showed very similar behaviour in the two new algorithms with highest values outside the main growth season, although here there was a different pattern between both estuaries, where values were lower again in December and January in the Eastern Scheldt. conclusion: QR requirements should be determined during the growth season. If the QR are obtained the PROTOOL fluorometer is a valuable tool to measure (changes in) primary production and seems a monitoring suitable technique. In combination with high spatial resolution of water quality parameters obtained with the Reflectance module detailed maps of primary production can be obtained.WP 6 Field testing on the Baltic Seaaimsthe Baltic Sea is a brackish, shallow sea suffering from the effects of eutrophication and pollution combined with limited water exchange with the open ocean. The sea is considered one of the most polluted seas in the world and vulnerable to intensive shipping and fishing industries. Diatoms and dinoflagellates dominate the spring bloom while cyanobacteria often dominate in summer. Besides the occurrence of cyanobacteria blooms the sea has a unique optical composition due to high dissolved organic matter loading while the particle population of the open sea is dominated by weak light scatterers (phytoplankton rather than suspended sediments). These properties require dedicated techniques in the PROTOOL modules:- reflectance processing schemes optimized for clear coastal waters (WP4, D4.16 WP6)- unmixing of the dissolved organic matter absorption from the measured total absorption (WP3, WP6)- fluorescence excitation light sources targeting algal and cyanobacteria pigments (WP2, WP6)various stages of PROTOOL development were tested in the Baltic Sea during a series of dedicated bio-optical research cruises with R/V Aranda. A detailed overview of the testing activities is given in PROTOOL public report D6.tested modulesthe prototype absorption meter (PSICAM by TriOS, marketed as OSCAR, see WP3) was tested with discrete samples during two summer cruises and found to have sufficient sensitivity to characterize the phytoplankton component, albeit masked by strong dissolved organic matter absorption. A separate absorption unit for dissolved organic matter using cross-flow filtration of sea water, capillary waveguide techniques, and off-the-shelf spectrometers and light source was built, with dedicated software, and tested in the ship flow-through system. This system helps address the issue of unmixing the total absorption in the Baltic Sea. The final model PSICAM that was demonstrated in the public PROTOOL workshop in Helsinki was not field-tested in the Baltic Sea.supporting studiesthe optical requirements (excitation and emission configuration) for fluorescence induction instruments sensitive to both algae and cyanobacteria, was assessed using laboratory phytoplankton cultures and extensive simulation and statistical modelling. The resulting publication (Simis et al. 2012) describes the optimal configuration and pitfalls in instrument design, as well as some fluorescence characteristics in cyanobacteria that should be kept in mind when conducting fluorescence induction measurements. The PSI fluorometer conforms to the described specifications.implementation on ships of opportunitythe reflectance module was subjected to two tests on a ship-of-opportunity crossing the Baltic Sea. Improvements were implemented after the first test in summer 2011 and the module performed optimally during the second 2-month trial period in May-June 2012. The implementation can be achieved with only electricity supplied from the ship. An improved version of the deck box (with fewer electronic parts and embedded computer) is still being developed at SYKE.WP 7 Field testing on the North Seaintroduction and aimsthe production of organic carbon compounds from inorganic carbon dissolved in seawater is essential for supporting all food webs on Earth. In the sea, the majority of this conversion is driven by light and takes place in the surface sunlit layer known as the euphotic zone. The rate of carbon fixation or primary production varies according to several factors, but is primarily under the control of light and temperature as well as the availability of nutrients. Wind and tide-driven turbulence is also important in controlling the exposure of photosynthetic organisms to irradiance and nutrients. Most primary production is due to phytoplankton, but when sufficient light is available at the seabed then production by seaweeds, seagrasses and microphytobenthos may also be important. The aim of this WP of the PROTOOL project was to examine the photosynthetic physiology of North Sea phytoplankton, using automated optical techniques. Our knowledge of primary production in the North Sea is limited to a small number of in situ studies, with few reports of production measurements in the past decade, despite the importance of this region in supporting fisheries.PROTOOL Equipment testedseveral research cruise to the North Sea were used to test prototypes of the different modules (PSICAM, R-module, various types of fluorometers), to test the implementation in existing ferry-box system on the research vessel, and to test improvements for their operational use (like prevention of fouling, cleaning etc.).resultsA large set of optical and physiological data were collected on the ship cruise. One goal was to test whether the important state variables can be estimated from automated measuring system, like ferry-boxes that include online measurements of chlorophyll fluorescence, optical measurements, like absorption of water constituents (PSICAM), or remotely determined surface reflectance. All system can be used, e.g. to determine the concentration of chlorophyll a, some can be used to determine the factors that determines the light attenuation in water, like suspended and dissolved matter concentration, or their results can be used to estimate the light attenuation coefficient directly. conclusionthe results of this study showed that the proposed approach of PROTOOL to estimate primary production bear good prospects. They showed - that the mass-specific optical properties relevant for PP calculations are rather conservative in the North Sea, therefore,- that these properties and other relevant parameter like chl a and TSM concentration can be determined by automated continuous absorption measurement directly from the absorption values,- that, similarly, these properties can be remotely sensed by continuous reflectance measurements, these included also the attenuation coefficient in the water column,- that fluorescence blank value are a source for significant error in the determination of photosynthetic efficiencies when phytoplankton biomass is low and are not easy to estimate from other properties like temperature, salinity, etc. contamination (of optical surfaces) and bio-fouling had significant influence on the measurements and more work is need to avoid or correct for these effects in automated optical systemsWP 8 Field testing on the Atlantic OceanA fundamental aspect of the PROTOOL project was field testing of the different PROTOOL modules on research ships and ships of opportunity. In addition, it aimed to compile comprehensive data sets in order to produce maps indicative of biological activity over wide spatial and temporal resolution in coastal and open ocean waters. The collected data was then to be used to evaluate the conversion factors developed in SP3. Due to the limited availability of the PROTOOL modules during the early part of the project, mapping became the focus of WP8. The aim was to make measurement in clear waters where absorption by phytoplankton dominates the light attenuation with depth (ie Case 1), thus providing 'open ocean' field data for calibration and derivation of conversion factors. This was to augment the measurements made in other field regions (ie WPs 5, 6 and 7) where primary productivity is higher. Plant pigment, taxonomy and primary productivity measurements, together with simultaneously collected FFRF data, were the main focus of the work.conclusion. The scientific interpretation of the biological results is still at an early stage, but it is clear that throughout the open ocean North Atlantic blooms are patchy. During D365 phytoplankton activity was relatively high in the central regions of the Iceland Basin with chlorophyll a concentrations approaching 4 mg m-3. Other dominant pigments in the Iceland Basin eg fucoxanthin, 19-hexanoyloxyfucoxanthin and 19- butanoyloxyfucoxanthin with concentrations up to 0.8 mg m-3, 0.5 mg m-3 and 0.8 mg m-3 respectively suggest the presence of diatoms and coccolithophores, but peridinin is low suggesting an absence of dinoflagellates. WP9: conversion factors Science and Technology (S & T) results and foreground (WP9)A core goal of PROTOOL is to implement autonomous high-resolution measurements of marine primary productivity (MPP) from seas and oceans. Fast Repetition Rate fluorescence (FRRf) is a powerful means by which this goal can be accomplished; FRRf measurements and implementation into oceanographic programs has already revolutionized the way in which MPP is quantified but suffers from two fundamental limitations: MPP is measured (i) relative to the number of 'photochemical reaction centers' and not relative to the volume of water sampled; and (ii) in a 'currency' of electrons evolved (the so-called electron transfer rate, ETR) and not CO2 fixed. Therefore conversion factors must be applied to all FRRf ETR measurements to yield measurements that are meaningful to stakeholders, e.g. fisheries and climate scientists, who need to know the mount of CO2 fixed by phytoplankton into organic matter per unit volume of water. Therefore, the fundamental aim of WP9 was to establish novel algorithms to enable implementation of these conversion factors.to convert the ETR to CO2 uptake requires knowledge of the 'electron requirement for carbon fixation', which is the reciprocal of the Quantum Requirement for C-fixation (assuming 1 absorbed photon results in the production of 1 electron (ER(e,C), mol e-: mol CO2 uptake). Many studies in the past have attempted to reconcile independent measurements of ETR and CO2 uptake from the same water sample (and hence estimate ER(e,C)); from these exercises, values of ER(e,C) appear highly variable both within and between studies most likely as a result of taxonomic and environmental variability. Therefore, we implemented a parallel approach to (i) compile existing ER(e,C) data sets to develop predictive algorithms, (ii) collect new targeted and high resolution data to support past data sets, in particular for special marine regions (Baltic Sea and the Eastern Scheldt). In addition, support this approach with (iii) additional laboratory culturing of phytoplankton to experimentally identity the role of key factors, e.g. nutrient availability upon ER(e,C) to support interpretation of (i) and (ii).A meta-data approach was used to evaluate the dependency of ER(e,C) upon 'easy to measure environmental variables', such as light, temperature, nutrients etc. (to establish algorithms that could then predict ER(e,C) from these variables. Data from 14 studies comprised of ten research cruises and four time series studies covered a broad range of different geographical areas of the world's seas and oceans, including the temperate, tropical and subtropical Atlantic Ocean, Massachusetts Bay (USA), Bedford Basin (Canada), Ariake Bay (Japan), the Celtic and Irish Sea, the North Sea, Gulf of Finland, the Baltic Sea, UK and European shelf waters and the Pacific Ocean. In fact much of this data was from the northeast European shelf and therefore directly relevant to PROTOOL's immediate goals. Once the data sets were integrated and treated for uniformity, ca. 350 data points were examined by a stepwise multivariate approach. Firstly, a simple correlative exercise for each separate study, which demonstrated that the relationships between ER(e,C) and environment were dependent upon how data within and between data sets were grouped; secondly, a principal components analysis, in combination with cluster analysis, to form clusters of sites with similar environmental conditions; and thirdly, multiple stepwise regression to evaluate the relationship between each cluster and the environment (and hence generate the required algorithms).in producing these novel algorithms, we were able to subsequently develop a simple program (.xls macro) whereby researchers could input environmental data for their region of interest and extract appropriate values for FE:C. An important part of this 'tool' is that it is can be continually (iteratively) refined as new data becomes available; as such, we also produced (i) the database on line with which the original algorithms for FE:C were developed and (ii) an optimized protocol for collecting FE:C data so that new (but standarised) information can be uploaded for to the database data. All of htrse various components represent a major step change in the use of FRRf that is unprecedented in its 20 year history. Importantly, we acknowledge that presently our algorithms are restricted to examining surface waters. Even though the FE:C database is constructed from data collected across different depths, data does not unfortunately exist to determine whether FE:C will be very different (and predictable) with depth within any given body of water that has structure, e.g. a subsurface nutricline. A conceptual exercise was therefore performed to evaluate the influence of water structure on application of values of FE:C.ability to predict ER(e,C) from environmental data is in part dependent upon the size f the available data set and thus a component of PROTOOL was to collect additional data for specific 'regions of interest'. Two regions were focused upon and both yielded novel insight. Several campaigns in the Baltic Sea/Gulf of Finland (2010-2011) resulted in ca. 70 new data points for FE:C across large environmental and taxonomic gradients. Our previous analysis demonstrated that ca. 25-40% of all variance for ER(e,C) could be explained by environmental condition for this region; therefore, we examined whether taxonomy may comprise a significant source of variance. Indeed, a comparison of ER(e,C) against a bio-optical signature of cyanobacterial abundance demonstrated that a further ca. 45% of ER(e,C) variance could be explained. This study was extremely important in demonstrating how current algorithms MAY need to build in taxonomic information to become more accurate. Monthly data was also collected throughout the project for the Eastern Schedlt since no data for this important region was previously available. in order to address whether the concentration of photochemical reaction centres per unit volume of seawater (=[RCII]) could be estimated, we took a two-step approach. Unfortunately, [RCII] is exceptionally difficult to measure since standard techniques are slow and require material in concentrations that vastly exceed what is found in nature (by several orders of magnitude); therefore past FRRf studies have assumed constant values for [RCII] for specific groups of phytoplankton. Measurements of [RCII] have shown these assumed constants to be highly erroneous. Initially, past measures of [RCII], largely derived from phytoplankton cultures were compiled into a database to determine whether [RCII] could be predicted from 'easy to measure' environmental variables (e.g. light, temperature), i.e. an analogous approach as for FE:C. However, [RCII] could not be confidently correlated with any variables (either in isolation or combination), except that higher values of [RCII] could be negatively correlated with cell size. In part this inability to 'predict' [RCII] likely reflects that few samples of direct [RCII] measurements exist. Therefore we secondly adopted an entirely new approach to attempt to derive [RCII] from FRRf induction parameters via first principles. This approach proved extremely successful:A novel algorithm was generated that could predict [RCII] from measurements of the maximum and minimum fluorescnece yields and the effective absorption cross-section standardized to an instrument specific constant. Two past data sets of [RCII} measurements (one laboratory-based with high taxonomic diversity but constant biomass and one field based with high biomass diversity but constant taxonomic composition) and both demonstrated that this algorithm could explain 90% of variation of [RCII] measured conventionally. Importantly, the instrument-specific constant must be determined by calibrating the fluorescnece signals against conventional measures of [RCII] (as was performed for some of the PROTOOL instruments); once one instrument is calibrated, other instruments can be cross-calibrated based on factory settings of the excitation optics. Given that [RCII] can now be derived from FRRf, it was possible to incorporate the approach directly into existing algorithms to determine ETR (to result in a modified ETR that is per unit volume of seawater and not per [RCII]).WP10. Intercalibrationthis WP had the main task to carry out pigment calibration and a 'calibration' of the primary production measurements (it was not possible to send out a set of radioactive filters to also include a comparison of the radioactive counting efficiency as disintegrations per minute, but we expect this to an insigficant source of error. Because of the development of the new algorithms to obtain absolute rates of electron transport from the PROTOOL fluorometer a workshop was organised in June 2012, followed by a repeat of the exercise in August 2012. These data are not analysed yet.pigment calibration: rationaleover the last half-century methodology for plant pigment analysis has moved from simple spectroscopic determinations to High Performance Liquid Chromatography (HPLC) coupled to diode array or mass spectrophotometric detectors. In spite of continuous developments in instrumentation, the analysis of phytoplankton pigments is still a challenge for HPCL techniques. The use of different chromatography columns eg C8 and C18 reversed-phase stationary phases, the method of pigment extraction from the plant material and the way in which the extract is clarified poses difficulties. In general, C18 columns exhibit shape selectivity whereas monomeric C8 columns show special selectivity towards compounds with subtle differences in polarity. Evidence now suggests that methanol or ethanol are better solvents than acetone and that extraction time should be at least 16 hours with ultracentrifugation used for clarification. Even when such parameters are the same there appears to be subtle differences in methodologies that result in variability in reported concentrations. In view of this it was felt important that PROTOOL had a good understanding of the deficiencies in pigment data from multiple sources. firstly, to document the pigment methods used by the partners using a simple questionnaire. Of particular interest was information regarding extraction solvent, filter disruption method and timing, ways of clarification and also the type of column and gradients used for HPLC analysis, all of which have been the subject of some debate. Secondly, to analyse known standard concentrations and field samples prepared and distributed by the NERC partner. resultsresults from the questionnaire showed that although all partners were using similar HPLC methodology for the collection and harvesting of pigments, there were some slight variations with storage and solvent extraction and a number of differences in instrumentation. Generally, all partners were processing pigment samples by collecting water samples that were immediately filtered through GF/F filters and subsequently flash frozen with liquid nitrogen followed by storage at -80 C or alternatively storing directly at -80 C. The reason for this difference depended on the availability of liquid nitrogen at the field stations. Solvent extraction was either by 100% methanol or 100% acetone with the occasional use of 96% ethanol by 1 partner. Extraction times showed some variance. Four of the five partners left samples extracting for between 18 - 24 hours, while one partner extracted for just one minute with cell disruption. Three of the four partners who used longer extraction times also used sonification to disrupt the cells. Probably the most important difference was the use of C18 or C8 columns as this causes differences in separation efficiency and the possible misidentification of individual pigments and their allomers.initially, just one cross calibration of known standards and field samples for a range of plant pigments was envisaged and this was done in late 2010/early 2011. The standard mixtures were prepared from DHI standards by pipetting known volumes of each of the individual standards into a vessel and making this up to 12 ml. An aliquot of this mixture was then diluted to make an intermediate standard and then further diluted to make one at a low level. From these three homogenous mixtures 10, 1 ml aliquots were prepared and each of the 5 partners received duplicate samples of the three known concentrations. In this way it was possible to ensure that all partners received identical samples. The pigments chosen represented those that are commonly measured and/or those that are difficult to separate or have interferences.conclusions pigment calibrationthe conclusion from this work package is that the partners are able to measure chlorophylls a and b at the 50 ug L-1 level to a reasonable accuracy in known standard solutions. This is not true however for concentrations 1 and 2 orders of magnitude lower and this maybe due to extrapolation errors. All the PROTOOL partners produce standard calibration curves against which samples are compared and since inaccuracies occur at lower concentrations it is possible that non-linear calibration might be the cause. What is interesting is that in the second round errors do not appear to come form the use of C8 and C18 columns. Three of the five PROTOOL partners use C8 (although only two report data here) in order to separating chlorophyll-a and divinyl chlorophyll-a and, where this is not achieved, it is to be expected that the reported chlorophyll-a data would be higher, but this does not seem to be the case. However, the finding could also reflect the generally lower values reported by the partner using the C18 column. The spread of the data from the analysis of the field samples was better in the second inter-calibration exercise. Generally speaking values for the individual pigments were within the same range but no conclusions can be drawn that one laboratory had consistently higher levels for all pigments than another. During both exercises the data for fucoxanthin was less than ideal and this needs some investigation.calibration of primary production measurementsthis exercise was reported as deliverable D10.1a (and can be retrieved from the PROTOOL website). From this primary production exercise it could be concluded that the differences in calculation of daily primary production by the different participants were very small. Two major reason explain these (small) differences: the mathematical procedure used to fit the data (the algorithm used by the statistical packages used to fit the PI-data) and the fact that most partners did not correct for surface reflection of incident light. It is therefore suggested that all partners will adopt the same calculation procedure, and it has been included in the C14 database/fitting package.calibration of fluorometersthe development of the new Sigma and Absorption algorithms prompted us to organize a calibration workshop in June 2012 at the NIOZ. The main aim of this workshop was to obtain the instrument specific calibration factor KR (KR=[RCII]sigma(PSII)/FoE(LED) where [RCII] is the concentration of the photosynthetic reaction centre units of PSII, Fo the minimal fluorescence, sigma(PSII) the functional absorption cross section and E(LED) the intensity of the measuring beam of the LEDs. When KR is known the absolute rate of ETR can be calculated without assuming a PSII concentration per mg of chla. More info is given in D9.22 and D5.17. However, because the workshop was so recent, not all information is processed yet, but we expect this to result in 1-2 scientific publications. WP11: Data Centre and disseminationthe Data Centre had the main task to develop de website and the databases to store the results of the primary production measurements (14C and FRR fluorometer) and the Reflectance data. In addition dissemination was coordinated by this WP.websitedesign and implementation of the website is completed, the website is online since February, 2010 and can be found at http://www.protool-project.eu. The restricted part is used by the partners for sharing documents concerning the PROTOOL project. We added a link to the PROTOOL blog on the home page, which contains the newsletters mailed to a selected party of possible interested persons.databasesthe NIOZ developed two databases from which several fitting routines (R scripts) can be accessed.the 14C-database, developed in Microsoft Access is a database to deal with the raw data obtained from the 14C-measurements. It can use the scintillation counter data together with the necessary fieldparameters (temperature, chla concentration, total DIC, radioactive DIC added) and use these data to obtain the parameters which describe the shape of the photosynthesis irradiance (PI) curve (e.g the maximum rate of photosynthesis, the slope of the PI-curve, the light saturation parameter Ek). 4 different fitting routines are implemented (Eilers and Peeters 1988; Platt and Gallegos 1980; Platt and Jassby 1976; Webb et al. 2007). the FRRF/Reflectance database is developed in MySQL, but has a MS Access user interface. It contains both the TriOS sensor data to calculate the reflectance as the FRRF-data. The current input routine is targeted to the PROTOOL baseline fluorometer format, but we will change this to the other formats used as well.the TriOS sensor data are first going through a set of filter criteria and those measurements which do not pass the filter are flagged. For the Dutch estuaries this is approx. 25% of the data. After passing through the data the reflectance is calculated and from the reflectance spectra the chla concentrations is calculated according to 4 different algorithm (the open ocean SeaWIFS algorithm (O'reilly et al. 1998), the Modis algorithm (Carder et al. 2003), the Gons algorithm (Gons et al. 2002) and the adaptive two-band algorithm from Ruddick (Ruddick et al. 2001). Only the Ruddick algorithm proved satisfactorily in the Dutch estuarine waters. We tried a Kd algorithm (Gons et al. 1998), but we were not satisfied with the result, so new ones are in development but not ready at the time of writing.dissemination. The website gives information about the project, and on the results tab the reports and other outreach can be downloaded. PROTOOL also has its own twitter account (@protoolproject), although this operated at low key. In addition PROTOOL made a weblog for the newsletters, and these can be accessed via a link on the homepage http://www.protool-project.eu/blog/.potential Impact:this chapter of the report discusses the potential benefits from the PROTOOL project, and has hence a science fiction character. Despite the fact that we cannot see in the future a few things are clear, which will be described below.from previous contacts it was clear that the PROTOOL approach is welcomed by the HELCOM countries, and this interest was strengthened during the final end-user workshop organized by SYKE and NIOZ in Helsinki. Nevertheless it is uncertain at what rate our product will be implemented in future monitoring within the framework of the Marine Strategy Framework Directive, but one thing is clear: PROTOOL fulfilled its promise and developed a new method to measure primary production in an automated and autonomous was.implementation of the PROTOOL ideas has taken place in at least two projects:- Within the frame work of the Sea and Coastal Research (ZKO) program of the Netherland Organization for Scientific Research (NOW) the PROTOOL approaches are adopted (slightly modified) by the IN-PLACE project which develops a coastal monitoring station/strategy.- The Waterdienst from RWS asked the NIOZ to evaluate the PROTOOL approach as an alternative method to the standard way of measuring primary production in the monitoring program MONEOS, which is studying the changes in the Westerschelde as a result of dredging activities.economic aspectsPROTOOL developed new equipment. Really new are the OSCAR developed by TriOS optical sensors and the FFL040 fluorometer developed by Photon System Instruments (PSI).the CEO of TriOS writes:the sales perspectives for the new developed PSICAM (sold as OSCAR) have to be seen as very positive. The current design and functionality has an outstanding position in the world market, as there is no other commercial instrument available yet, which follows the PSICAM principle in a hyperspectral design.based on the fact that already more than 10 units are ordered from customers, the annual sales is expect to increase to 40-50 units a year within 2-3 years from now.the key-market at the moment is scientific institutes and organizations, as well as marine monitoring applications.TriOS is a leading supplier of instruments for this applications, the market entry of a new product is well known. The basic way is via conferences, workshops and mouth-to-mouth propaganda of satisfied customers. In addition the instrument will be presented in various exhibitions (like Oceanology International). Finally a wide sales network of international distributors will contribute to the sales as well.in a further step, the new instrument will be made available also for industrial applications, like in chemical or pharmaceutical research and production. If these applications can be made successfully, the number of sold units can properly be doubled.on a long-term run the successful marketing of the new PSICAM has positive impacts on the economic situation of TriOS and will lead into securing and generation of new qualified jobs inside the company but also at various suppliers.the CEO of PSI writes the following about the newly developed PROTOOL fluorometer:the PSI flow through fluorometer system FFL-2012 starts a new production line of monitoring fluorometers intended for the Gross Primary Production measurements. During the project a new design with additional functionality was developed and tested, which hadn't been yet on the market. There is no device on the market now, which offers such an FRRF measurement capability (multicolor) and online data availability (web server included as the one developed during the PROTOOL project). We aim to acquire 25% of the European market share of ferry box flow-through instrumentation within the next 5 years (200,000 EUROS/year). We expect an increase of 2 employees during next year in our production unit and 1 research staff for the further development. The device will be also presented on conferences worldwide to interest new customers all around the world.there is a potentially large market for a miniaturized version of the FRRF fluorometer for installation on buoys, gliders and monitoring stations. We want to apply a registered design intellectual property right on our design and software.these two statements clearly show that PROTOOL created economic success, and a future PROTOOL-2 is expected to do the same as the consortium has already plans for a follow-up and like to work on miniaturization of the PROTOOL equipment and develop similar equipment for the platforms mentioned by PSI (buoys, gliders and fixed monitoring stations).the third piece of equipment developed, the Reflectance module was less innovative and dealt with a reconfiguration of existing equipment, although SUKY developed a positioning sensor platform, but at the moment no plans exist to market this. The hardware description (excluding sensors) and software for shipborne reflectance measurements are open source resources hosted at http://sourceforge.net/projects/rflex/. However, the R-module is a very valuable addition to our monitoring equipment, and when widely adopted it can be a the reason for a surge in sales of the TriOS RAMSES sensors which form the hard of the R-module.we expect that the science produced by PROTOOL will have socio-economic impact: The generation of a new algorithm to quantify the concentration of photochemical reaction centres, [RCII] (and so in turn an '[RCII]-free ETR algorithm') as well as region-specific algorithms to convert ETRs into CO2 uptake rates will radically transform the nature and scale of FRR flourometry in aquatic sciences. These developments represent THE long awaited breakthroughs for the field and enable ETRs to be measured free from the practical constraints that have for so long limited widespread implementation of FRR fluorometers as 'productivity sensors'. Benefits include generation of higher resolution (and hence more accurate) marine primary productivity (MPP) estimates that can be used to improve the accuracy of forecasts of environmental change upon MPP though ecosystem models. More accurate MPP estimates also promise to improve remote sensing algorithms that provide real-time broad scale estimates of MPP.such improvements in capacity to measure and predict MPP have broad societal implications but importantly, we see such improvements absolutely fundamental in driving the future market of FRRf technology for MPP-based studies. The limitation is no longer on the inherent assumptions to the algorithm but on the technology. Variants of FRR fluorometers have already been developed (e.g. the PSI fluorometer as part of PROTOOL) to enable quantification of the contribution of different functional groups to be determined. Discussions are also now underway to develop this technology to examine MPP from benthic systems. Ultimately, such developments and the widening of the FRRf market(s) will require that instruments become smaller, more power efficient and able to transmit data wirelessly.wider implications with regard to the development realized by the PROTOOL projectWP10 contributes to the international debate on pigment analysis. There has been some concern about the inter-comparibilty of pigment data over recent years, in particular the need for a robust data set for in situ calibration of ocean colour data. The compilation of information from further inter-calibration exercise will help analysts unravel the inconsistencies in reported pigment data and help develop single methodologies for individual pigments and water bodies.WP8 provides comprehensive pigment and taxonomy data for the Bay of Biscay and the North Atlantic. Data for these regions is relatively sparse and this high resolution data over three years provides information on seasonal and annual variability. The data will be available to the international community and will be useful for scientific studies of changes in phytoplankton diversity and phenology and also for the in situ calibration of ocean colour satellite observations.FRRF methodology in fresh water research: A Czech perspective from a Czech member of the PROTOOL consortiumfor the technological development of the monitoring systems in The Czech Republic the PROTOOL approach was the main missing part in the information about hydrobiology of the water resources and standing water bodies in Czech Republic. The C14 primary production data are only scarce and done on specific places for scientific purpose only. The PROTOOL approach would make it easy to perform primary production measurement for routine monitoring, and because it can operate autonomously it does not require great man power to obtain and store the data.PROTOOL in a global contextnot only monitoring agencies can use the PROTOOL approach to get the necessary information needed for evaluating the ecological status of their seas. Phytoplankton plays a crucial role in the global C-cycle. Current estimates state that phytoplankton is responsible for about 50 Gton of CO2 net primary production: about 50% of the total global fixed CO2 is fixed by photosynthetic organisms, despite the fact that this is realized by less than 1% of the living C on Earth. This implies that phytoplankton is not a good indicator of primary production because of the rapid turnover times, a fact established by several studies. This fact also causes a large uncertainty in the estimates of net phytoplankton primary production (NPP). The estimates of NPP by phytoplankton vary twofold from ~35-75 GtonC/year. Part of this uncertainty lies in the fact that the phytoplankton biomass estimates from optical remote sensing (which use chlorophyll-a as proxy) is still difficult, especially in the more turbid case-2 waters. But more important is that no primary production measurements are available. Only two permanent monitoring stations exist: the HOT (Hawaiian Ocean Time Series) in the Pacific ocean in the vicinity of Hawai and BATS (the Bermuda Atlantic Time Series).the potential to use ships of opportunity is also recognized by the Scientific Committee on Oceanic Research (SCOR), a global NGO of scientists working on the oceans (see http://www.scor-int.org/ online). They developed the program OceanScope in which they set out a roadmap to use ships of opportunity to sampling platforms. The complete report can be found at:http://www.scor-int.org/Publications/OceanScope_Final_report.pdflist of Websites:http://www.protool-project.eu Related documents Final Report - PROTOOL (Productivity tools: Automated tools to measure primary productivity in European seas)
