Annex C-5 Phytoplankton primary production
Primary production is the only regular rate measurement in the Baltic Monitoring Programme. From these measurements it is possible to calculate the amount of organic material formed from light, carbon-dioxide and nutrients. Primary production has important links to eutrophication and sedimentation and, consequently, to deep water oxygen concentrations.
The measurement of primary production in water is carried out for, inter alia, the following purposes:
- to measure the ecophysiological response on different nutrient availability;
- to describe temporal trends in primary production.
Primary production should be measured with the "P/E - method", in an incubator. With this method the uptake rate of carbon is measured at a range of irradiance levels in order to get a relationship between photosynthesis and light. Pmax (maximum photosynthetic rate) and (initial slope of the P-E relationship) and Ek (the light saturation irradiance) can be calculated using this method.
The advantage of the method is that ecophysiological information of the phytoplankton assemblage can be derived from the P/E - relationship. It is also possible to calculate the daily production per m3 from these measurements after calculation and incorporation of the vertical attenuation coefficient and solar irradiance, and with some assumptions the annual primary production.
Sampling and analytical procedure
WORKING MANUAL AND SUPPORTING PAPERS ON THE USE OF A STANDARDIZED INCUBATOR-TECHNIQUE IN PRIMARY PRODUCTION MEASUREMENTS
(Version 22nd, February 1999, filename MANUAL4.doc)
Editors: Franciscus Colijn1 and Lars Edler2
1 Research- and Technology Centre Westcoast, Hafentörn, 25761 Büsum, Germany
2 Swedish Meteorological and Hydrological Institute, Oceanographic Services, Nya Varvet 31, SE-426 71 V. Frölunda, Sweden
Appendix Description of hose sampling method for phytoplankton measurements
Annex 1 Design and tests of a novel P-I incubator to be used for measuring the phytoplankton primary production in ICES monitoring studies
Annex 2 Light measurements and intercalibration of standard ICES incubators
Annex 3 Irradiance Differentiation and Control in the ICES incubator
The 14C method for the measurement of primary production in the sea has been used for more than 45 years. The database is considerable and it seems, as the method will continue to be an important tool in the monitoring of the status of the marine pelagic ecosystem. A major problem in the comparison of productivity data is, however, the use of different measuring methods. The differences stem from both conceptual and practical reasons. Within ICES long discussions have been held to create a database on primary production. However, the fear that the data were not comparable resulted in a workshop, where the methods applied by different laboratories were intercompared. Very significant differences in results were found between laboratories (Richardson, 1991). During its meeting in 1988 the ICES Working Group on Primary Production found that there was a need for a standardized method for primary production measurements to be used in monitoring studies of which the data were to be stored in the ICES data bank. It was decided to make a strict protocol for primary production measurements performed in an incubator. The intention was to make the incubator inexpensive and the method with as few steps as possible. Over the years that have passed since this decision, there have been long detailed discussions, but also fruitful tests of the incubator developed by Colijn et al (Annex 1, Annex 2). This manual with supporting papers is meant to serve as the protocol for future monitoring of primary production in the ICES area, and hopefully far beyond.
Although the initially intended simplicity has been left due to the wish to be able to measure full P-E relations, we still have given emphasis to obtain a concise and strict protocol, which does not leave much room for alternatives. Sometimes we have given alternatives where these do not affect standardization. However, in order to produce comparable data for a data bank we were obliged to keep the alternatives to a minimum and enable a rigorous quality assurance.
In summary, the purpose of this manual is to provide a strict protocol of the monitoring of Primary Production. Following this manual will ensure comparable data in the ICES database.
The P-E curve method should be used (for terminology we refer to Sakshaug et al., 1997). With this method the 14C uptake is measured at a range of irradiance levels in the incubator, in order to get an estimate of the photosynthesis rate versus irradiance. This can then be parameterized and give values of Pmax (maximum photosynthesis), a (maximum light utilization coefficient measured as the slope of the linear increase of photosynthesis against irradiance), Ek (the saturating irradiance) and, after calculation and incorporation of vertical attenuation and solar irradiance, the daily primary production per m2. Together with data on chlorophyll-a, Pmax can be normalized to obtain assimilation numbers.
The method for estimating primary production by the "ICES Incubator" given in this manual (cf. Annex 1) is intended for monitoring purposes. Measurements should be possible from small, as well as from large vessels. Because of this, simplifications from what could be considered to be the ideal method have been introduced. It should be pointed out that the "ICES Incubator" method is not meant as a replacement of other "P-E techniques". It has been designed to provide a reliable measurement of primary production, using a simple incubator and a standard protocol.
The incubator is a rectangular perspex tank (33 x 33 x 9 cm) with a turning wheel on which a maximum of 12 experimental bottles can be clamped. 10 fluorescent tubes (TLD 8W J8, no 33) illuminate it. The full description of the incubator is given in Annex 1. The standard protocol is presented here. The incubator is manufactured by HYDROBIOS, Kiel, Germany (see List of Manufacturers).
Standardization of the method involves strong reduction of the number of alternatives. However, a few are indicated (see text in italics) but the standard method is to be used to obtain quality assured data for the ICES data bank.
Mixed water columns
In areas where the euphotic zone is mixed and the phytoplankton community is uniformly distributed, one representative sample, obtained at 5 m depth is sufficient. It is important, however, to make sure that the water layer is mixed. This is easiest done with a CTD and fluorescens profile.
As an alternative an integrated sample can be taken with a hose (0-10 m) (Lindahl, 1986 and Annex 3). Mixed discrete samples from 0 to 10 m depth can also be used.
Stratified water columns
In stratified waters, where the phytoplankton community is not homogeneously distributed, a water sample should be obtained with a hose (see Appendix), covering the water column of interest. This single sample is treated as a mixed water sample.
If preferred, samples from different depths can be taken and incubated separately at temperatures similar to temperatures from the sampling depths. In that case more incubators may be needed, or subsequent incubations are to be made.
The hose sampling method can also be used as an alternative to sampling with water bottles, as the complete sample can easily be divided by depth for individual incubations by using clamps (Lindahl, 1986).
In conclusion, measurements of primary production in stratified water bodies are more complicated and will normally fall beyond "simple" monitoring strategies.
Experts to add figures
1. Placement of the incubator.
The incubator must be placed so that light conditions outside the incubator do not disturb the light climate inside the incubator. The incubator needs to be thermostatically controlled, to give the same temperature as the water sample. For samples from stratified waters differing in temperature two separate incubators should be used, or two consecutive incubations should be performed. The second water sample(s) should be kept dark and at the original temperature during the first incubation.
2. Incubation flasks.
Tissue culture flasks (see also 3.) of about 50 ml should be used. These flasks also work as paddles for the water-jet driven rotation of the flask-wheel. After each incubation, the flasks and the caps should be rinsed with diluted HCl (10%) and then several times with distilled water to avoid contamination. The flasks should be dried at 70 °C.
3. Irradiance levels in the incubator (for details see Annex 3)
A set of incubation flasks with different transmission levels, from 0 to 100% should be used (for manufacturer of special prepared bottles see Appendix). It is important that there should be enough measuring points to obtain a good measurement of Pmax and a. With the special prepared 12 bottles this is not a problem and after some experimentation with the incubator normally a series of 6 bottles will suffice to measure a reliable P-E relationship. The exact irradiance in each bottle should be measured, despite the transmission the manufacturer gives percentages. This can be done with a small sensor, which can be introduced into the bottles (a manufacturer of this calibrated sensor can be found in the Appendix). To obtain irradiance saturated photosynthetic rates (Pmax) a minimal irradiance of 500 µE.m-2.s-1 should be available. This is achieved by using 10 fluorescent lamps (TLD 8W J8, no 33). In case 500 µE.m-2.s-1 is not reached, a mirror behind the lamps and possibly on the other side of the tank will increase the irradiance flux.
4. 14C solution.
Dilution of commercially available 14C solution should be avoided due to the risk of contamination. The standard activity of every batch of 14C solution should be controlled by the liquid scintillation technique (see point 11). It is recommended to use ampoules which contain the whole amount of 14C needed for one incubation series. This reduces the number of measurements of the added 14C activity.
In case 14C solutions are prepared 'home-made' high grade chemicals and UHQ water must be used.
The final carbonate concentration of the solution should agree with the average carbonate concentration of the sea area being studied and the pH of the solution should be in the range of 9.5 - 10.0.
5. Accompanying field measurements.
In order to obtain a representative sample of phytoplankton it is important to have knowledge of the vertical distribution of the algae. This is easiest accomplished by a CTD-cast combined with an in situ chlorophyll-fluorescence cast. Measurements of the under-water irradiance in at least 5 different depths, in order to calculate the vertical irradiance attenuation coefficient are also necessary. If the daily production is going to be calculated, the daily surface irradiance must also be measured in hourly intervals.
Non-transparent and non-toxic sampling devices must be used. Sampling should take place in daylight, to avoid strong interference of inequality due to diel rhythms of the phytoplankton (Annex 1, Gargas et al., 1979).
After sampling but before incubation subsamples are taken for chlorophyll (Fig.1, Step 1) and TCO2 analysis (Fig. 1, Step 2).
The incubation should start as soon as possible, preferably within half an hour after sampling. All transfers of water samples should take place in subdued light, in order to avoid light-shock of the phytoplankton.
7. Total CO2 concentration. (Fig. 1, Step 2)
Total CO2 concentration should be calculated according to other standard methods, using titration of carbonate (Strickland and Parsons, 1972). In brackish waters, such as the Baltic Sea, the CO2 concentration can be calculated by the formulas of Buch (1945). In both cases temperature, salinity and pH must be measured.
8. Addition of 14C (Fig. 1, Step 3).
The 14C solution is added to the whole volume of sample needed to fill all the flasks. After thorough mixing, the flasks are filled. This procedure minimizes errors compared to pipetting the radioactive tracer to every individual incubation bottle.
Incubation bottles are filled with a measuring cylinder. The flasks should be filled up to the neck, leaving an air bubble in the flask. One dark flask from each original sample should be incubated.
The 14C solution should be added to the sample in such concentrations that statistically sufficient counts of the radioactivity in the phytoplankton can be obtained. A triplicate measurement of the added activity is needed (Fig.1, Step 4). These samples should be counted immediately to avoid loss of activity. Therefore in case direct counting is impossible the inorganic 14C should be mixed with ethanol-amine by pipetting 0.25 ml of sample with added activity together with 0.25 ml of ethanolamine. Scintillation cocktail can be added later and radioactivity determined.
As an alternative incubation flasks are first filled and then the 14C solution is added to every flask. It is important that the added volume is small and that a precise, calibrated micropipette is used.
9. Incubation. (Fig.1, Step 5)
The incubation time should be about 2 hours and the rotation speed should be approximately 10 rpm. Start and end of the incubation should be given in the protocol so that the precise incubation period (in minutes) can be used for the calculation. To achieve an unhampered rotation of the samples all positions on the wheel need to be filled (e.g. by using flasks filled with water).
10. End of incubation. (Fig. 1, Step 6)
After incubation the flask contents are filtered immediately. In case of high algal biomass or high sedimentation load it might be needed to filter a subsample. A defined portion should be taken and filtered.
Glass-fibre filters (GF/F, Ø 25 mm) should be used, since these filters are cheap, become opaque and are known not to disturb the counting procedure of the radiotracer. To avoid any contamination of the filter edges, prewetted filters should be used. The suction pressure should not exceed 30 kPa during filtration. The filters should be rinsed once with a small volume (5 ml) of filtered seawater from the original sample (use filtrate of the chlorophyll-a measurement!).
After filtration the filters should be placed in scintillation vials and dried at room temperature for 24 hours. Following addition of scintillation liquid, the samples should be kept in dark for at least 3 hours to reduce chemiluminescence.
As an alternative to filtration many scientists use the bubbling method to obtain the total (dissolved and particulate) primary production.
From each incubated sample a sub-sample of 10 ml (exactly) is pipetted into a scintillation vial and 0.2 ml of 80 % HCl is immediately added. In a ventilated cupboard, the vials are then bubbled with a fine jet of air bubbles for 20 minutes, or are left open for 24 hours. 10 ml of scintillation cocktail is added and the vials are shaken by hand for some seconds before scintillation counting.
The liquid scintillation technique should be used when counting the uptake of 14C. In order to get a statistically accurate measurement, 40 000 DPM, or counting for 10 minutes is needed to get a result with 1-% accuracy. Quench curves for different amounts of chlorophyll should be established and adding an internal standard, e.g. 14C -hexadecane or toluol, should check the measuring efficiency of the liquid scintillation counter. Normal counting efficiency calculation is done by using the channel ratio method. Modern scintillation counters are equipped with programs to facilitate efficiency calculations. The user is referred to the instructions of the manufacturer.
12. Calculation of carbon uptake (Fig. 1, Step 9).
The total carbon uptake is calculated from the equation:
(a) = Sample activity (minus back-ground), dpm
(b) = Total activity added to the sample (minus back-ground), dpm
(c) = Total concentration of 12CO2 in the sample water, µmol/L (or µM)
(d) = The atomic weight of carbon
(e) = A correction for the effect of 14C discrimination
k1 = subsampling factor (e.g. sample 50 ml, subsample 10 ml: k1=subsample factor 50/10= 5)
k2 = time factor (e.g. incubation time 125 minutes: k2= 60/125= 0.48)
The results will be given as µg C·L-1·h-1 per irradiance level and as well as the photosynthesis at light saturation (Pmax), the maximum light utilization coefficient (a), and light saturation parameter Ek , from the P-E curve (see below).
13. CALCULATION OF DAILY PRIMARY PRODUCTION
In order to calculate the daily primary production a number of parameters are needed. These include:
1. Vertical attenuation (extinction) coefficient, in m-1, measured with a calibrated irradiance meter.
In case no attenuation can be measured, Secchi disc values can be used by conversion. The attenuation coefficient is calculated as
Att. Coef. = x / Secchi depth (m)
where x is 1.7 - 2.3 (1.7 (Raymont, 1967), 2.3 (Aertebjerg and Bresta, 1984), 1.84 (Edler, 1997)). This factor changes with sea area. In principle, it increases with decreasing salinity.
2. Insolation (Hourly measurements of incoming radiation between 400 and 700 nm (PAR), in E.m-2.s-1
3. Pmax, Ek, and a.
The transformation of the hourly production corrected for dark uptake into daily production, which is the ultimate ecological goal, should follow the protocol outlined in Appendix 2.
A simple computer program for the calculation of the daily production is available (see list of manufacturer). After giving the raw data to the protocol, the software will calculate the daily production and combine the data in a database for the ICES data bank.
14. QUALITY ASSURANCE
In order to produce specified and confident data on primary production the performance of the measurement and the analytical procedures must follow a high quality system and operate in a state of statistical control. The method used shall be validated to meet the required specifications related to the use of the results. The validation includes selectivity, sensitivity, range, limit of detection and accuracy.
A high quality is maintained by using experienced and well trained personnel. Ring tests and intercomparisons ought to be conducted regularly.
The 14C tracer method is used to measure the incorporation of the added isotope in the form of NaH14CO3 as an estimate of the photoautotrophic growth, measured as photosynthesis of phytoplankton. The method is highly selective but nonphotosynthetic incorporation of 14C and non-biological fixation takes place simultaneously. This is measured as the dark uptake. It can not, however be used as respiration value as in the oxygen method. In the scintillation counting procedure interference may occur from the background values. They are, however, always subtracted from the uptake values.
Apart from the high selectivity the method is also very sensitive. Uptake rates of 0.05 µg C.L-1.hr-1 can easily be measured. There is no actual upper and lower limit of the method. The sensitivity can be improved by adding more 14C to the samples and/or by counting the incorporated radioactivity of the phytoplankton over a longer time which will improve the counting statistics.
The detection limit is set by the background radiation, the use of a zero time blank and dark incubations. The lower limit of detection is a sample should be defined as having activities at least three times the background values.
As mentioned above a range virtually does not exist. Uptake rates between 0.05 and 250 µg C.L-1.hr-1 can easily be measured. At very high uptake rates an increase in pH may occur. This would affect the distribution and availability of the bicarbonate ion. Under such conditions the incubation period should be reduced. Under natural marine and brackish water conditions this does not happen.
Random as well as systematic errors occur in this method. Random errors should be kept to a minimum by adopting the standard procedure with only few experimental steps in the whole process from sampling to scintillation analysis. Systematic errors may occur with the light source, irradiance levels, filtration technique and the scintillation counting. These errors should be kept to a minimum. Regular control of handling and function of the instruments used, as well as calibration of the instruments are necessary tools to control the errors. The use of an independent analytical method to measure the systematic error is not possible, since there is no better independent analytical method available as an alternative for the 14C tracer method. Certified reference material (CRM) exists for the calibration of the scintillation counter, or for the calibration of the counting procedure with the original samples by using the internal standard method (adding standard to a sample). Participation in intercomparison exercises is one of the possibilities to test the comparability and therefore the precision and error propagation in this method. In general, however, most variability of this method will be caused by the biological nature of the material. Therefore strict procedures for the measuring protocol are needed to obtain the best possible results.
The quality assurance should ensure that the data are fit for the purpose for which they have been collected, i.e. that they satisfy the detection limits and levels of accuracy compatible with the objectives of the monitoring program.
15. DATA DELIVERY:
In order to have the possibility to check and recalculate daily productivity data it is important that all laboratories deliver their data in the same format and that this includes the fixation rates for all irradiances. Data should be delivered in the ICES, Biological Reporting Format.
To be included
Literature and further reading:
Aertebjerg, G. and Bresta, A.M. (ed), 1984. Guidelines for the Measurement of Phytoplankton Primary Production. Baltic Marine Biologists Publication No. 1. 2nd edition.
Buch, K., 1945. Kolsyrejämvikten i Baltiska Havet. Fennia, 68(5):.
Colijn, F., Kraay, G.W., Duin, R.N.M., Tillmann, U. and Veldhuis, M.J.W., 1996. Design and test of a novel Pmax incubator to be used for measuring the primary production in ICES monitoring studies. ICES CM 1996/L3. (Annex 1 is a modified version of the original paper)
de Keyzer, J.. 1994. Irradiance Differentiation and Control in the ICES incubator. Zemoko, Maritiem (unpubl. report)
Edler, L., 1997. In: Report of the ICES/HELCOM Workshop on Quality Assurance of pelagic biological measurements in the Baltic Sea. ICES CM 1997/E:5
Gargas, E. and Hare, I., 1976. User´s Manual for estimating the Daily Phytoplankton Production measured in Incubator. Contribution from the Water Quality Institute. Danish Academy of Sciences. No. 2.
Gargas, E., Hare, I., Martens, P. and Edler, L., 1979. Diel changes in phytoplankton photosynthetic efficiency in brackish waters. Marine Biology 52: 113-122
HELCOM, 1988. Guidelines for the Baltic Monitoring Programme for the third Stage; Part D. Biological Determinands. Baltic Sea Environment Proceedings No 27 D. Baltic Marine Environment Protection Commission. Helsinki Commission.
Li, W.K.W and Maestrini, S.Y. (ed.). 1993. Measurement of Primary Production from the Molecular to the Global Scale. ICES Marine Science Symposia. Volume 197. ISSN 0906-060X.
Lindahl, O. 1986. A dividable hose for phytoplankton sampling. In Report of the ICES Working Group on Exeptional Algal Blooms, Hirtshals, Denmark 17-19 March, 1986. ICES, C.M. 1986/L:26.
O´Reilly, J.E. and Thomas, J.P., 1983. A Manual for the Measurement of Total Daily Primary Productivity. Biomass Handbook No. 10. SCAR, SCOR, IABO, ACMRR.
Raymont, J.E.G., 1967. Plankton and Productivity in the Oceans. Pergamon Press, Oxford.
Richardson, K. (ed), 1987. Primary Production: Guidelines for measurement by 14C incorporation. ICES.Techniques in Marine Environmental Sciences No. 5
Richardson, K. 1991. The ICES 14C Primary Production measurement intercomparison exercise. Mar. Ecol. Prog. Ser. 72: 189-201.
Sakshaug, E., Bricaud,A., Dandonneau, Y., Falkowski, P.G., Kiefer, D.A., Legendre, L., Morel, A., Parslow,J. and M. Takahashi, 1997. Parameters of photosynthesis: definitions, theory and interpretation of results. J. Plankt. Res. 19: 1637-1670.
Strickland J.D.H. and Parsons T.R. 1972. A practical handbook of seawater analysis. Fish. Res. Bd. Canada. Bull. 167. Ottawa. 310 pp.
List of Manufacturers
HYDROBIOS, c/o H. Fischer, Am Jägersberg 5-7, 24161 KIEL, Germany, Tel. +49-431-3696011, Fax: +49-431-3696021, E-mail: firstname.lastname@example.org
Incubation flasks and light sensor:
ZEMOKO, c/o, ing. Jan de Keyzer, Dorpsplein 40, 4371 AC Koudekerke, the Netherlands, Tel/Fax: +31-118-551182
Working Manual ICES Incubator
Irradiance Differentiation and Control in the ICES incubator.
SMHI, Oceanographic Services, Nya Varvet 31, SE-426 71 V. Frölunda.
Tel. +46 11 495 80 00, Fax. +46 31 751 8980, E-mail: email@example.com
Appendix: Description of hose sampling method for phytoplankton measurements
Manual for Marine Monitoring in the COMBINE Programme of HELCOM, Part C, Annex C-6. (Manual version 1.0 - Revised June 1998, HELCOM EC 8/97, updated by EC MON 3/98.
Lars Edler, Oceanographic Services, SMHI, SWEDEN
In order to get a true integrated sample of phytoplankton the mixing of discrete water bottle samples is not adequate. To overcome this the hose method should be used (Lindahl, 1986).
An armoured PVC hose is suitable. The inner diameter of the hose should be c. 20 mm, giving a sampled volume of c. 3 L with a 10 meter hose. The length of the hose should cover the 10 upper meters of the sea to be sampled, as well as the distance from the sea surface to the boat deck, from where it is operated.
In the end of the hose a PVC tube should be placed and secured with a hose clamp. A snap hook should be fastened at the hose clamp. Ten meters up, corresponding to the surface of the sea, a valve should be placed and secured with hose clamps. At the end of the tap a line should be connected. It must be possible to operate the line from the boat deck.
When sampling, the lower end of the hose is connected to the hydrographic wire with the snap hook. The end of the hose should be below the wire weight in order to avoid contamination. The valve must be open when lowering the hose. The hose is lowered slowly until the valve has reached the sea surface. The valve is closed by pulling the string connected to the tap. The hydrographic wire is then elevated. The valve is then opened and the content of the hose is filled into a bucket and mixed. Subsampling bottles are filled from the bucket.
After sampling the hose is rinsed carefully with fresh water and stoppers are put in both ends in order not to pollute the inside of the hose. A thorough cleaning with diluted HCl or other detergents should be made at the end of a cruise and the hose should be dried.
Lindahl, O., 1986. A dividable hose for phytoplankton sampling. ICES; C.M. 1986/L:26, Annex 3.
Primary production protocol
Annex 1 Colijn, F., Kraay, G.W., Duin, R.N.M., Tillmann, U. and Veldhuis, M.J.W., 1996. Design and test of a novel Pmax incubator to be used for measuring the primary production in ICES monitoring studies. ICES CM 1996/L3. (Annex 1 is a modified version of the original paper)
Annex 2 L.P.M.J. Wetsteyn1, L. Edler2, M.M. Steendijk1, G.W. Kraay3, F. Colijn4 & R.N.M. Duin5. Light measurements and intercalibration of standard ICES incubators
Annex 3 de Keyzer, J.. 1994. Irradiance Differentiation and Control in the ICES incubator. Zemoko, Maritiem (unpubl. report)
(revised version 4.December 1997)
ANNEX 1 Design and tests of a novel P-I incubator to be used for measuring the phytoplankton primary production in ICES monitoring studies
F. Colijn1, G.W. Kraay2, R.N.M. Duin3 , U. Tillman1 & M.J.W. Veldhuis2
1 Forschungs- und Technologiezentrum der Christian-Albrechts Universität Kiel, Hafentörn, 25761 Büsum, Germany
2 Netherlands Institute of Sea Research (NIOZ), P.O.Box 59, 1790 AB Den Burg, The Netherlands
3 National Institute for Coastal and Marine Management (RIKZ), P.O.Box 20907, 2500 EX The Hague, The Netherlands,
This paper has been developed during a period of several annual meetings of the ICES Working Group on Phytoplankton Ecology and its predecessors. Field and laboratory studies were made in successive years. We would like to acknowledge many colleagues who gave the possibility to test the method on board their ships or in their labs, like the meeting in Helsinki, Finland in 1989; the additional measurements made in the Indian Ocean in 1992-1993; and in the German Wadden Sea in 1995.Contents
An inexpensive and simple incubator for primary production measurements is presented along with a protocol (see Working Manual) to achieve strictly comparable and reliable 14C-fixation rates of phytoplankton. The incubator, based on Steemann-Nielsen and Aabye Jensen (1957), is comprised of incubation bottles revolving in a temperature controlled water bath at a fixed irradiance. The recommended protocol and incubator have been tested in different water types, such as Dutch , German and Finnish coastal waters, in the North Sea and in the Indian Ocean, and give reliable estimates of the photosynthetic rate at the fixed irradiance used. Coefficients of variation were between 0.6 and 7.6 in incubation experiments with three and five samples. No difference between Pmax measured in the Baltic incubator and the ICES incubator was observed.
The incubator has been used as a P-I incubator during cruises in the Indian Ocean by providing a series of bottles with different transmittance levels. These experiments show that actual P-I relations can be measured with a good fit of the P-I curve parameters, like initial slope ±, Ik, Iopt and Pmax values.
A series of measurements were performed for a period of one year at a monitoring station in the German Wadden Sea. These measurements showed the typical characteristics of P-I incubations with almost stable alpha values and temperature controlled Pmax levels. Correlations between chlorophyll and primary production were high.
Daily primary production values have been calculated based on the P-I relations after integration over time and depth on selected series of data and compared with a simple empirical equation based on Pmax, attenuation coefficient, daylength and daily insolation. The agreement between both methods was rather poor, and variable. Dependent on the calculation mode all values were roughly 1.5 to 2 times too high as compared to the integrated values based on one of the fitted P-I curve parameters. Further work needs to be done to improve this empirical formulation. The three equations used to calculate the daily primary production were comparable. Calculations not based on a sinoidal light function but on a rectangular mean irradiance level were 5-20 % higher.
Based on the measurements in the German Wadden Sea daily primary production has been calculated according to a strict format. This format will be proposed to standardize calculations.
Application of a known irradiance in the incubation bottles is still one of the most difficult parameters in this method, whereas the application of the tracer method itself is easy and straightforward.
Results of the Hirtshals intercalibration (c.f. Richardson, 1991) were discussed during the workshop of the ICES Working Group on Primary Production in Copenhagen (June 1988). The meeting adopted the following recommendation: "... that there is a need for a standardized primary production method to be used in monitoring studies with special coded data in the ICES data bank". The authors have accepted to comply with the request by building a simple and inexpensive incubator and proposing an appropriate protocol.
At present several procedures are available to measure daily depth-integrated primary production (mgC.m-2.d-1). Most of these methods are based on measurement of P (photosynthesis) vs. I (irradiance) relationships, of vertical attenuation coefficients, and solar irradiance (Aertebjerg Nielsen & Bresta, 1984; Gargas & Hare, 1976; Richardson, 1987).
The results of the Hirtshals intercalibration workshop (Anonymous, 1989; Richardson, 1991, 1993) have shown that calculation of integral daily primary production may contain a whole series of errors or assumptions which cause large differences in the final result. Substantial errors arise from handling of samples, incubation time, incubation handling, liquid scintillation counting, and calculation methods, but the main difference was due to the different types of incubators used (measurement of irradiance, differences in light quality etc.). Therefore, data offered to the ICES data bank are not comparable and therefore were never stored. This paper describes the use of an incubator and develops a strict protocol with as few steps as possible, and contains recommendations about the use of materials, to reach highest comparability of results.
Originally, our task, however, has been limited to this specific point and therefore no attempt has been made to propose a method to calculate integral daily production from single Pmax (mgC.m-3.h-1) measurements, assuming that the incubator has the possibility to measure Pmax at light saturation within a large range of irradiances. Several other assumptions have to be made to calculate daily primary production, including a vertically homogeneous distribution of algal biomass, similar photosynthetic characteristics of the phytoplankton and the same species composition throughout the water column. Also data on vertical attenuation and daily irradiance should be available. As shown by Riegman and Colijn (1991) calculations based on surface samples alone can underestimate areal primary production by 17%. As pointed out by Platt and Sathyendranath (1988) oceanic primary production might be well estimated from an irradiance model based on measurements of Pmax and ±, and a remotely sensed biomass field. Such estimates might be possible for the North Sea within the near future if both Pmax and ± are known.
Stimulated by the discussions and comments in the ICES WG we finally have attempted to use the ICES-incubator as a P-I incubator and to compare daily primary production values measured in the ICES-incubator with fully integrated values over time and depth, using P-I relations.
The incubator strongly resembles the one originally used by Steemann Nielsen & Aabye Jensen (1957), (cf. Postma & Rommets, 1970; Cadée & Hegeman, 1974). It is constructed of a rectangular perspex tank (h x b x w= 33 x 33 x 9 cm) with a turning wheel (max. 12 rpm, 18 cm in diameter) on which experimental bottles ( max. 12) are clamped. Illumination is provided by 10 Philips 8 W fluorescent tubes (TLD 8W J8, no.33) which can be switched off/on separately (Figure 1). Irradiance should in all cases be measured with an appropriate light sensor (e.g. LICOR, µE. m-2.s-1 or W.m-2) or the special sensor developed by de Keijzer (1994) (Annex 3). Our experimental set up gave a mean irradiance of 360 µE.m-2.s-1, providing a saturating 14C fixation rate (see results section). However, the light field is not homogeneous but ranged from 140 to 530 µE.m-2.s-1 depending on position of the flasks during revolution. The homogeneity of the light field can be easily improved by using a backscattering white polystyrene foam layer opposite to the fluorescent tubes. These irradiance measurements were done with a 2À-sensor and therefore are substantially lower than the earlier measurements in the incubator during the Indian Ocean cruise with a spherical sensor: with 10, 8, 6, 4 and 2 tubes and this polysterene layer we measured 1100, 850, 650, 300 and 250 µE.m-2.s-1, respectively as maximum irradiances. A full description of the irradiance distribution is given in (Annex 2) where several options for illumination were tested.
During one of the meetings it was discussed whether this incubator could be used to measure P-I relations. Indeed, this can be done by covering the incubation bottles with neutral density filters (e.g. Flash Light Lee), available in several transmission classes. An alternative is painting the bottles in different black intensities. Such tests have been performed during cruises in the Indian Ocean in 1993. However, this procedure did not fall into our primary goal as stated above in the recommendations of the 1988 meeting. Thus the incubator now no longer acts as a simple incubator again introducing several of the "old" uncertainties and errors, especially as far as irradiance levels in the bottles is concerned. During a later stage the problem how to obtain different irradiance levels in the incubation bottles has been solved by using a epoxy-resin layer of different attenuation (Annex 2).
Incubations are carried out in disposable tissue (ultraclean) culture flasks (e.g. Greiner, tissue culture flasks, 690160) containing 50 ml of sample. These flasks can be used several times without deterioriation of the vessel walls and are suited to adhere the epoxy-resin layers.
Temperature is controlled to within ± 0.1 °C by a suitable thermostat with enough capacity (Lauda, Colora). Water is recycled within the bath by an extra pump which also causes the revolution of the wheel, with the flasks acting as paddles. If only a few samples are incubated the open positions should be filled with flasks containing water to attain a constant turning of the wheel. A running seawater system on board the ship could be used instead of the thermostated water bath. The complete system is built by Hydrobios (Kiel, Germany), whereas the calibrated incubation bottles are sold by ZEMOKO (the Netherlands) (full adresses of both companies are given at the end). The cost per unit can be reduced if several incubators are built/ordered simultaneously.
Several tests by independent workers have been conducted with the apparatus in its former and improved form.
3.1. Test at the Netherlands Institute of Sea Research (NIOZ).
During the typical spring bloom of phytoplankton in Dutch coastal waters (plankton dominated by the diatoms Biddulphiaaurita, B. sinensis, Coscinodiscus concinnus, Skeletonemacostatum and colonies of Phaeocystis sp.), an incubation experiment was performed, according to the protocol (see Appendix). Incubation periods of 1 and 2 hours were tested, along with two filter types: Whatman GF/F (approximate pore size 0.7 µm, 47 mm) and Sartorius cellulose acetate 11106 (pore size 0.45 µm, 47 mm).
After filling the experimental bottles, 0.1 mL NaH14CO3 (Amersham) from a stock solution prepared with superclean distilled water containing one pellet of Ultrapure NaOH (pH =9), was added. Ampoules have been cleaned with 6N HCl. Total activity added, to be determined for each experiment, was 11.46 .106 dpm/ 0.1 ml. Precautions should be taken to use a pure 14C-bicarbonate solution, especially when release of extracellular dissolved organic carbon has to be measured (Bresta et al., 1987).
After incubation, samples were filtered within a few minutes through the two filter types. After fuming over concentrated HCl for 5 min in a desiccator, samples were counted in 10 ml Instagel in 20 ml glass scintillation vials. Cells on the filters were disrupted in a Bransom Ultrasonic device during 15 min. Without this disruption, counts can be up to 50% lower. Cpm's were converted into dpm's with a quench curve and the external standard channels ratio method. Results of the first experiment are compiled in Table I.
The results show a good reproducibility of the 14C fixation rates, an almost linear uptake over the 2h period, and a lower recovery and a higher variability of 14C on Sartorius cellulose acetate filters compared with GF/F filters (cf. Hilner & Bate, 1989). Dark values were about 2% of the light values.
3.2. Test at the Finnish Institute of Marine Research (Helsinki)
During an ICES workshop, the new incubator was tested on board the research vessel Aranda by making a direct comparison between the ICES incubator and the Baltic Sea incubator on July 6, 1989. A surface water sample containing cyanobacteria and several other species without dominance of a particular one was taken from the Baltic and divided into 14 bottles. To each bottle 0.1 ml of 2 mCi NaH14CO3 was added. Samples were incubated 2 h 25 min and filtered onto GF/F filters, and fumed over concentrated HCl for 10 min. Filters were disrupted by sonification and counted as above. Five samples were incubated in the ICES incubator, 5 in the Baltic Sea incubator at full light ( 400 µE.m-2.s-1), and another four samples were incubated at 50%, 25%, 10% and 5% of full light, respectively. Reduction of irradiance was obtained with neutral density filters.
Results are given in Table II. The full light samples in both incubators showed the highest fixation rates. The reproducibility was very high in both incubators. The single point measurements at the attenuated irradiances showed a good linearity, indicating that in this case four measurements suffice to estimate the photosynthetic efficiency a. Despite the difference in maximum irradiance in the two incubators, the same maximum fixation rate was measured, suggesting that photosynthesis was saturated at an irradiance of about 300 µE.m-2.s-1.
3.3. Tests in the North Sea by the National Institute of Coastal and Marine Management (RIKZ), formerly Tidal Waters Division at Middelburg (NL)
A similar but completely independent set of experiments was conducted during one of our regular sampling surveys of the North Sea within the EUZOUT (Eutrophication of the North Sea) project. Samples were taken at different stations in the North Sea (Fig. 2), covering both coastal and offshore waters, up to 370 km from the Dutch coast during a cruise from 25 to 27 July, 1989. Surface, thermocline and subthermocline samples were also incubated at the stratified stations. To 50 ml samples 10 µCi in 0.1 ml was added. In this case the results are also compared with Pmax values calculated from P-I measurements on the same samples incubated simultaneously but in another incubator (Peeters et al., 1991; Klein & van Buuren, 1992). Two comparisons of short (2 h) versus long (6 h) incubation times were made. All samples were filtered onto Whatman GF/F filters; after addition of 10 ml HCl, samples were bubbled with air for 20 min and counted as described in Peeters et al. (1991).
The results are given in Table III. Depending on the station a wide range of photosynthetic activities was observed. Coastal eutrophied stations showed rates up to 40 times higher than in the oligotrophic central part of the North Sea. Vertical profiles showed high rates in the thermocline or subthermocline layers. The long-term incubations showed an almost linear uptake over the 6 h period. Duplicate incubations generally showed a maximum difference of 10%.
Comparison of the Pmax in the ICES incubator with the Pmax in the P-I incubator shows that the ICES Pmax is somewhat higher than the latter Pmax. This confirms our findings in Helsinki which also showed that the ICES incubator measures a value close to Pmax. However, samples in the P-I incubator were run for about 6 h instead of 2 h in the ICES incubator.
3.4. Tests during Indian Ocean cruises (JGOFS) in 1992-1993 by NIOZ (Texel) east of African coast off Somalia and Kenya
During these cruises of which the results will be presented elsewhere a series of experiments were performed with bottles painted black with different degrees of transmittance resulting in a range of c. 4% to 100%. Irradiance in all individual bottles however had to be measured. Thus the incubator has now been used as a real P-I incubator. To increase the irradiance levels the backside of the incubator was covered with white polystyrene foam which gave a range of 40 to 1100 µE.m-2.s-1 in the bottles. A large sample of about 10 l has been taken from the surface during an evening cast at 18 h. LT. From this sample a P-I relation has been measured for 2h, including chlorophyll-a concentrations. Part (about 7 l) of the sample has been stored overnight in a dark cool box and incubated in a similar way the next morning at 6.00 h LT. The results of three of such series are given in Fig. 3 and Table IV. The P-I curves were analysed according to equations given by Jassby & Platt (1976), Platt et al. (1980) and Eilers & Peeters (1988). The first two equations showed comparable results whereas the third one showed higher Pmax values for both incubations. The former Pmax values were within 5% difference. Calculation of daily production also showed good agreement for the former two equations. However the Pmax and daily production values showed large differences between the two incubations (evening vs. morning) , mainly due to the higher Pmax values of the morning incubation due to a circadian rhythm (chlorophyll had slightly increased during the storage period) whereas also the initial slope a increased by 25%. More data of the Indian Ocean cruises are available but will be published elsewhere (Veldhuis and Kraay, in prep.).
3.5. Tests at the Station Büsum, along the German Wadden Sea in 1995
Within the framework of our monitoring studies in the German Wadden Sea, weekly incubations were made using the standard incubators, kindly provided by Mr. Bert Wetsteijn of the RIKZ in Middelburg. Contrary to the standard procedure, we used a direct cooling of the incubator in the lab by a Lauda cooler instead of the closed circuit with the copper tubing. This was done to be able to obtain very low incubation temperatures during winter time and does not have any further consequences for the measurements. The samples were illuminated from both sides to obtain sufficiently high irradiances up to 800 µE.m-2.s-1 for proper Pmax determination. Throughout these measurements we used the new incubation bottles and the improved irradiance setup as described in Wetsteijn et al. (1996). As a standard incubation time 2 hours were used, but in winter during low activities up to 4 hours were used. TL tubes were arranged to perform a homogeneous light field. Irradiance was measured inside the incubation bottles with the same equipment as developed by Wetsteijn et al.(1996). Mean irradiance values were based on twelve measuring points during one revolution of the wheel. The special incubation bottles prepared by ZEMOKO (see Wetsteijn et al., 1996) were used throughout the measurements. For one P-I measurement 8 bottles including one dark were used. Dark values were low but always subtracted from the light values. Added activity ranged from 0.5 to 3 µC in winter (volume 50 to 300 µl). Samples were filtered over 0.45 µm membrane filters (not GF/F) under reduced suction pressure (200 mm Hg), washed with 10 ml 'cold' filtered seawater and dried. Counting took place in Filter-count (Packard). Added activity was counted after dilution in 55 ml of sample and pipetting 50 µl of the mixture in counting vials. Calibration occurred according to the external standard ratio procedure of the liquid scintillation counter.
Primary production values were normalised to chlorophyll-a measured spectrophotometrically according to Lorenzen (1967).
The results are presented in figures 4 to 6. In Fig. 4 four representative examples of P/I curves are shown from different seasons. Curve fitting and calculation of P/I parameters was made according to the equation of Platt and Gallegos (198 ). The seasonal variations in P-I parameters is shown in Fig. 5. Chlorophyll specific maximum photosynthetic rates (Pbmax) ranged from 2.0 to 9.9 µG C./ µg Chlor/ h-1 and showed a large variation over the year and was highly significant correlated with water temperature (Fig. 6). In contrast, the slope of the P/I curves ranged from 0.0150 to 0.0375 µg C/ µg Chlor. h-1/ µE. m-2.s-1 (Fig. 5) and proved to be less variable and irrespective of water temperature. During the whole year no strong light inhibition at high irradiances could be observed. Ik values, used as a parameter of light adaptation, were relatively high throughout the year varying between 81 and 453 µE. m-2.s-1 (Fig. 5). Thus in spite of the low light conditions in the Wadden Sea due to high turbidity, no signs of low light adaptation of the phytoplankton could be detected. Further we conclude that based on the measured high Pbmax values and the natural mean low light levels in the Wadden Sea , the phytoplankton of the turbid inner parts is light limited and not nutrient limited throughout the year.
The results of these P-I measurements will be used, in combination with irradiance and attenuation measurements to calculate the daily and annual primary production at station Büsum (Tillmann et al., in prep.).
To a great extent the task accepted during the 1988 ICES meeting in Copenhagen has been fulfilled: a simple and inexpensive incubator has been built and tested. The tests so far show that the incubator works well, that it is simple to use, and that it also has the potential to measure P-I curves. However, it is not recommended as a P-I incubator, due to the fact that these already exist in a wide variety with more sophisticated irradiance regulation. Reproducibility and linearity of uptake rates are within the expected limits. Problems arising from different photosynthetic characteristics like a daily disparity or circadian rhythm in the same sample can not be solved. Because such differences can be quite large there is no simple solution except to incubate samples several times during the day. To reduce this kind of variability a practical and pragmatic solution could be to incubate all samples around noon.
The series measured at Station Büsum during 1995 show the consistent results which can be obtained with the incubator. Apart from minor changes such as the cooling device at low temperatures, we followed the protocol as described for the continuously mixed water mass. The series will be used to calculate the annual primary production, whereas we intend to continue the measurements to get a series for several years to see whether nutrient reductions influence the primary production in this part of the Wadden Sea. At the moment light limitation is the most important regulating factor.
Apart from the results obtained so far, there is a need for concurrent work with two types of incubators: an ICES type of incubator for monitoring studies and a more sophisticated type where P-I relations can be measured for physiological studies. Comparisons between this simple and maybe more complex types of incubator should be made by the individual scientists as part of an intercalibration study. Nevertheless, the limited amount of methodological steps is of great advantage and reduces several of the common errors. If the Working Manual is followed, data obtained in this way are directly comparable.
Discussions both in the ICES working group and with several colleagues have shown that there is a need for a further standardization step leading to the calculation of values per m2 from these Pmax measurements. As a first approach, empirical formulations like the one used by Cadée & Hegeman (1974) and DiToro et al.(1971) are useful. In Helsinki we decided that such a formulation should be derived, which then could be used to calculate a value per m2. A first attempt has been made to use such an empirical equation by comparing daily primary production calculated by integration and based on P-I parameters with this empirical estimate of daily primary production. The results (not given) showed that daily primary production calculated according to the equations given by Eilers and Peeters, Jassby and Platt, and Platt and Gallegos and in all cases with a sinusoidal irradiance give almost equal results. If daily primary production is calculated with a rectangular light distribution using one mean irradiance level the daily values are about 5 to 20% higher. If we use the empirical equation of Ditoro et al. (1971) we obtain values up to 1.5 to 2 times as high. Probably the calculation is not yet very realistic and we have further evaluated this procedure and finally present a calculation mode with the Working Manual. The calculation is also available on disk through SMHI in Sweden.
One should, however, realize that in all cases this value is only an estimate, due to physiological characteristics of phytoplankton (Neale & Marra, 1985; Savage, 1988; Vandevelde et al., 1989), and to an uneven vertical distribution of phytoplankton in the sea (Riegman & Colijn, 1991). Calculation of primary production under such circumstances can only be achieved if samples from different depths are incubated and their light-, temperature- and time-dependant fixation rates are known.
Based on a larger data set comprising Pmax data and simultaneous P-I measurements, we have calculated the daily primary production in Büsum as an example. The same calculation has been suggested to ICES for the calculation of primary production per m2 in different areas. A further step in modelling primary production could be the incorporation of time-dependent adaptation responses as described by Neale and Marra (1985). However, this was not the primary goal of the working group and therefore falls beyond the scope of this paper.
A recent paper of McBride (1992) also compiles several equations to calculate daily photosynthesis, one of which may be adopted by ICES as an alternative to the standard. The present method to calculate daily primary production is based on an numerical integration over time and depth which is very rapid and simple with modern PC's.
A problem which is not solved sofar is the irradiance needed to measure Pmax. In our opinion a procedure should be developed to relate the saturating irradiance for Pmax to the geographical latitude and the time of the year. Then a standardized incubation irradiance could be prescribed. A moment there is uncertainty because we have not tested it.
The authors would like to thank Juha-Markku Leppänen for the opportunity to work on board the R.V. Aranda during the ICES workshop in Helsinki and Louis Peperzak for his help during the EUZOUT cruise in July, 1989. Useful comments on the manuscript were given by several members of the working group, and by C. Peeters.Steve Bates is acknowledged for his efforts to improve the english version of the manuscript and Ken Jones for his long lasting stimulus to complete this manuscript. We would also like to thank Katherine Richardson for her initial involvement in this study and the comments she made on a earlier draft.
Aertebjerg Nielsen, G. & A-M. Bresta, 1984. Guidelines for the measurement of phytoplankton primary production. The Baltic Marine Biologists, 1.
Anonymous, 1987. ICES 14C primary production intercomparison exercise preliminary data report. CM1987/L:27.
Bresta, A-M., C. Ursin & L. Moller Jensen, 1987. Intercomparison of 14C-labelled bicarbonate solutions prepared by different institutes for measurement of primary productivity in natural waters and monoalgal cultures. J. Plankton Res. 9: 317-325.
Cadée, G.C. & J. Hegeman, 1974. Primary production of phytoplankton in the Dutch Wadden Sea. Neth. J. Sea Res. 8: 240-259.
DiToro, D.H., D.J. O'Connor & R.V. Thomann, 1971. A dynamic model of the phytoplankton population in the Sacramento-San Joaquin Delta. Avent. in Chem. 106: 131-180.
Eilers, P.H.C. & J.C.H. Peeters, 1988. A model for the relationship between light intensity and the rate of photosynthesis in phytoplankton. Ecol. Modelling 42: 199-215.
Gargas, E. & J. Hare, 1976. User's manual for estimating the daily phytoplankton production measured in an incubator, Water Quality Institute (Denmark) 2: 1-75.
Hilner, Th. & G.C. Bate, 1989. Filtertypes, filtration and post-filtration treatment in phytoplankton production studies. J. Plankton Res. 11(1): 49-63.
Jassby, A.D. & T. Platt, 1976. Mathematical formulation of the relationship between photosynthesis and light for phytoplankton. Limnol. Oceanogr. 21: 540-547.
Klein, A.W.O. & J.T. van Buuren, 1992. Eutrophication of the North Sea in the Dutch coastal zone, 1976-1990. Tidal Waters Division, Report 92.003: 1-70.
McBride, G.B., 1992. Simple calculation of daily photosynthesis by means of five photosynthesis-light equations. Limnol. Oceanogr. 37(8): 1796-1808.
Neale, P.J. & J. Marra, 1985. Short-term variation of Pmax under natural irradiance conditions: a model and its implications. Mar. Ecol. Prog. Ser. 26: 113-124.
Peeters, J.C.H., H. Haas & L. Peperzak, 1991. Eutrofiëring, primaire produktie en zuurstofhuishouding in de Noordzee (Eutrophication, primary production and oxygen management in the North Sea), Tidal Waters Division, Report GWAO 91.083 (in Dutch).
Platt, T. & S. Sathyendranath, 1988. Oceanic primary production: estimation by remote sensing at local and regional scale. Science 241: 1613-1620.
Platt,T., C.L. Gallegos & W.G. Harrison, 1980. Photoinhibition of photosynthesis in natural assemblages of marine phytoplankton. J. Mar. Res. 38: 687-701.
Postma, H. & J.W. Rommets, 1970. Primary production in the Wadden Sea. Neth.J. Sea Res. 4: 470-493.
Richardson, K. (ed.), 1987. Primary production: guidelines for the measurement by 14C incorporation. ICES, Techniques in marine environmental sciences, 5.
Richardson, K., 1991. Comparison of 14C primary production determinations made by different laboratories. Mar. Ecol. Prog. Ser. 72: 189-201.
Riegman, R. and F. Colijn, 1991. Evaluation of measurements and calculation of primary production in the Dogger Bank area (North Sea) in summer 1988. Mar. Ecol. Prog. Ser. 69: 125-132.
Savidge, G., 1988. Influence of inter- and intra-daily light-field variability on photosynthesis by coastal phytoplankton. Mar. Biol. 100: 127-133.
Steemann Nielsen, E. & E. Aabye Jensen, 1957. Primary oceanic production, the autotrophic production of organic matter in the oceans. Galathea Report, Vol. 1: 49-136.
Vandevelde, T., L. Legendre, S. Demers & J.C. Therriault, 1989. Circadian variations in photosynthetic assimilation and estimation of daily phytoplankton production. Mar. Biol. 100: 525-531.
Zimmerman, R.C., J. Beeler Soohoo, J.N. Kremer & D.Z. D'Argenio, 1987. Evaluation of variance approximation techniques for non-linear photosynthesis-irradiance models. Mar. Biol. 95: 209-215.
Table I. Fixation rates of sample from the Marsdiep tidal inlet (Wadden Sea, cf. Cadée & Hegeman, 1974).
|Filter type||Incubation time (h)||DPM||x ± sd cv|
|GF/F||1||45163||43724 ± 1757 4.0|
|GF/F||2||78384||79183 ± 2469 3.1|
|Sartorius||2||71228||67348 ± 5154 7.6|
|GF/F||2 ( in dark)||1142||1424|
|Sartorius||2 ( in dark)||1706||1424|
Table II. Samples from the inlet to the Helsinki harbour.
|ICES Incubator||Baltic Incubator|
|CPM/h x ± sd cv||CPM/h x ± sd cv|
|2518 2514 ± 17 0.6||2602 2553 ± 78 3.1|
Mean irradiance in ICES incubator: 297 mE.m-2.s-1;
Full irradiance in Baltic incubator: 400 mE.m-2.s-1.
Table III. Results from the North Sea cruise (25-27 July 1989); for location of stations see Fig.2. s=short term(c. 2 h), l=long term(c. 6 h) incubation; sur=surface, ther=thermocline, subther=subthermocline sample; Pmax derived from P-I measurements based on 6 h incubations
|DPM/2 h||DPM/2 h|
|NW100 sur||3385||4656||2242||TS100 sur||4590||5282||3982|
|NW70 sur||6951||6777||5532||TS100 ther||7541||7293||6213|
|TS370 sur/s||3293||2970||2699||TS100 subther||1478||1415||1085|
|TS370 sur/l||2755||---||---||TS10 sur||6583||6646||3454|
|TS275 sur/s||1741||1624||---||TS4 sur||59062||56027||53403|
|TS275 sur/l||1897||1870||1503||NW20 sur||17984||20844||15411|
* samples showed strong photoinhibition
Table IV. Example of results of experiments conducted in the Indian Ocean, location off Kenya and Somalia (Veldhuis & Kraay, in prep.) to show daily inequality.
Same sample was used for both incubations; parameters estimated by the equation of Platt et al. (1980). Calculation of daily primary production is based on ke = 0.1, daylength = 12 hrs., and mean surface irradiance = 1000 mE.m-2.s-1. SSE is the error sum of squares of the fitted model.
|Evening Incubation||Morning Incubation||Unit|
Legends to figures
Figure 1. Photograph of ICES incubator (see text).
Figure 2. Map showing location of sampling stations during the July cruise in the North Sea (Peeters et al., 1991).
Figure 3. P-I curves for two incubations on the same sample; a) in the evening, b) in the morning. Fitted curve is equation of Platt et al. (1980).
Figure 4. Examples of P-I curves measured at Station büsum; all curves were normalised to chlorophyll-a; fits were made with the equation of Platt (1980)
Figure 5. Seasonal course of P-I parameters at Station Büsum in 1995; all parameter calculations based on Platt et al. (1980)
Figure 6. Relation between assimilation number (Pbmax) with temperature for the measurements conducted in Büsum in 1995
Figure 7. Calculation of daily and annual primary production based on daily insolation , vertical attenuation and P-I curves from the ICES incubator
Experts to add missing figures and to number the above figures
L.P.M.J. Wetsteyn1, L. Edler2, M.M. Steendijk1, G.W. Kraay3, F. Colijn4 & R.N.M. Duin5
1 National Institute of Marine and Coastal Management (RIKZ), P.O. Box 8039, 4330 EA Middelburg, The Netherlands.
2 Swedish Meteorological and Hydrological Institute (SMHI), Doktorsgatan 9D, S-26252 Ängelholm, Sweden.
3 Netherlands Institute of Sea Research (NIOZ), P.O. Box 59, 1790 AB Den Burg, Texel, The Netherlands.
4 Forschungs- und Technologiezentrum Westküste, Hafentörn, D-25761, Büsum, Germany.
5 National Institute of Marine and Coastal Management (RIKZ), P.O. Box 20907, 2500 EX Den Haag, The Netherlands.
(Results from earlier performed light measurements in standard ICES incubators and from a workshop held on 9-11 March 1994 in Middelburg, presented at the meeting of the ICES WG on Phytoplankton Ecology in Copenhagen, 23-26 March 1994; additional revisions made after the meetings in Copenhagen, 23-26 March 1994 and in The Hague, 29-31 March 1995)
Since 1987 some of us have worked in a changing configuration on the construction and experimental performance including a standard protocol of a newly designed 'simple' and inexpensive incubator for primary production measurements. The original term of reference was to develop a simple and inexpensive incubator for use in monitoring studies.
During one of the meetings of the former ICES WG on Phytoplankton and the Management of their Effects, the original set-up was criticized because no P-I relations were measured. Therefore the design was adapted enabling the measurement of P-I relations at a range of 12 (including dark) irradiance levels. The incubator has been used as a P-I incubator during Indian Ocean cruises in 1992-1993 by NIOZ-workers (some results were presented in Colijn et al., 1993).
In the last report of the WG on Phytoplankton and the Management of their Effects (C.M.1993/ENV:7 Ref.:L) it was stated that the Dutch workers would be asked to explore the possibility of convening an evaluation workshop in The Netherlands. One of the objectives of this workshop would be to evaluate the reproducibility of measurements using the standard incubator and protocol in the hands of different users. At the end of 1993 funding for the manufacturing of four incubators, four filter/flask series (each with an irradiance gradient), some irradiance sensors and the execution of light measurements by an optical expert became possible, giving the opportunity to perform a reproducibility experiment before the next meeting.
In this report we will present 1) information on the used epoxy resin coating, 2) information on the used irradiance sensor, 3) some results from earlier performed extensive light measurements in the standard incubators and 4) the results from an intercalibration experiment with four incubators to check the comparability of identical incubators and the variability due to manipulation of the samples by different users. Information with respect to 1), 2) and 3) was taken from ZEMOKO (1994).
Incubators and incubation bottles
A short description of the incubator has been taken from Colijn et al. (1993). The incubator is constructed as a rectangular perspex tank (h*b*w=33*33*9 cm) with a turning wheel (max. 10 rpm, 18 cm in diameter) on which 12 experimental bottles (Greiner, tissue culture flasks, ca. 55 ml, 690160) are clamped. Water is recycled within the incubator by an aquarium pump causing the revolution of the turning wheel, with the bottles acting as paddles. On board ship the incubator should be closed accurately with a perspex cover to avoid overflowing and short-circuiting.
Illumination is provided by 10 Philips 8 W fluorescent tubes (TLD 8W J8, no. 33) which can be switched off/on separately.
Water temperature can be controlled using an external cooling device or with a running seawater system. Because we wanted to cool 4 incubators simultaneously a copper tube outside the light field along the narrow vertical walls and the bottom of each incubator was used; the copper tubes were parallel connected to the thermostat (Colora). In this way we reached similar levels of water temperature in the 4 incubators (see Table 1) without the risk of contaminating the cooling device or the 4 incubators at the same time.
Sensor construction and calibration
Knowledge on irradiance measurements is of great importance for P-I measurements. Therefore, a new small spherical irradiance sensor was constructed, consisting of a Si photodetector in front of which a green filter is mounted and surrounded by a spherical collecting element made of diffuse epoxy-resin. With a stopper, through which the wire passed, it can be fixed in the centre of an incubation bottle.
Detailed information of the measured typical spectral and spatial sensitivity of this type of sensor is given in ZEMOKO (1994).
For the absolute calibration of the sensor in W.m-2 or mmol.photons.m-2.s-1 a spectroradiometersystem was used, consisting of a spherical collecting element, an optical fiber, a Jarrell Ash gratingmonochromator and a Si photodetector. Furthermore a standard tungsten striplamp as a wellknown radiance source was used.
The obtained calibration factors (multipliers to get W.m-2 or mmol.photons-m-2.s-1) hold only for the combination of this sensor and TLD33.
With the sensor clamped to the turning wheel it was easy to make a complete rotation-angle of 360o and to calculate the average irradiance and standard deviation. The 4p sensor was calibrated using a tungsten strip lamp and a LICOR-1000 lightmeter. The obtained calibration factors (multipliers to get W.m-2 or mmol.photons.m-2.s-1) hold only for the combination of this sensor and TLD33.
Neutral density filtercoating
Different levels of irradiance were created by applying different layers of epoxy-resin (in which dark pigments are mixed in different ratios) as neutral density filters on the surfaces of the incubation bottles. The side walls and the necks of the bottles were covered with black epoxy-resin. The reason that we chose this material is our experience that nettings, grids, and even some neutral density filters seriously influence the relative transmission between 400-700 nm. Determination of transmission values in the 400-700 nm range was performed by means of a halogen lamp with daylight-filter and a monochromator. The tubes have the lowest absolute irradiance in the blue and green parts and the highest absolute irradiance in the yellow and orange parts of the 400-700 nm range (data not presented here).
Four series of bottles were available with the following transmission values (in %):
0 1.0 2.5 9.4 18.0 22.9 28.5 31.5 42.5 51.0 70.6 100
0 1.1 2.6 9.8 18.9 23.5 28.7 31.6 42.8 51.5 71.0 100
0 1.5 2.9 9.9 19.1 23.6 30.5 32.9 43.2 53.1 72.1 100
0 1.5 2.9 9.9 19.3 24.3 31.4 35.7 43.3 54.1 72.9 100
Figure 1 shows the relative transmission of 3 and 1.5 % filters of the used epoxy-resin. This material is most suitable in the very low transmission range (thick epoxy-resin layer). In the high transmission range (thin epoxy-resin layer) it must be even better.
The procedure to make the desired epoxy-resin/dark pigment composition and to fix the layers on the incubation bottles is not given here. The reason is that this work was done by a consulting firm that spended some research on this subject. On request the firm is willing to construct on a commercial basis (a restricted number of) series of incubation bottles with known irradiance levels (ZEMOKO, Maritiem technisch bureau, Dorpsplein 40, 4371 AC Koudekerke, The Netherlands, Tel/Fax 0031-0118-551182).
Figures 2-5 give examples of light measurements performed with the 4p sensor. In these figures rotation-angle 0 corresponds with the highest position on the turning wheel. The small and negligible nipple-shaped structures at the tops in Figures 2-5 are measured when the 4p sensor approaches the vertical parts of the copper tubing. Figure 2 illustrates the insignificant difference between the four TL-sets (with coated bottles and white polystyrene foam against one of the outer walls). Figure 3 gives the absolute irradiance distribution with clear bottles and with and without polystyrene foam. It can be seen that using the polystyrene foam substantially increases the amount of available irradiance in the incubator. Surprisingly, however, the difference between minimum and maximum values increased. Figure 4 illustrates the light-absorbing effect of all coated bottles in position on the turning wheel with 2, 4, 6, 8 and 10 TL tubes used. The most flat irradiance distribution was obtained using 6 TL tubes. Finally, Figure 5 gives the results with coated bottles and two sets of 10 TL tubes in parallel and crossed position. In parallel position the mean irradiance during one rotation is ca. 940 mmol.photons.m-2.s-1 and in crossed position ca. 960 mmol.photons.m-2.s-1, see Table 3 in ZEMOKO (1994). It should be preferable to have also one or two higher irradiance values in the more inhibiting part of the P-I curve. Higher (and more uniform distributed) irradiance values might be obtained by using circular fluorescent tubes at both sides of the incubator. Using a white epoxy-resin instead of black epoxy-resin to reach higher irradiance values might be possible. In that case attenuation is achieved by diffuse scattering/reflection instead of absorption. However, the spectral properties (relative transmission in the 400-700 nm range, see also Figure 1) of black epoxy-resin seem to be better than those of white epoxy-resin.
A series of 3 consecutive incubations were performed in all 4 incubators with changing users per incubator. A culture of Phaeodactylum tricornutum, grown in a 2000 l indoor pond with enriched seawater under continuous light (6 * Philips 60 W) at Chl-a concentrations of ca. 150 mg/l, was used. It was diluted tenfold with 0.2 mm filtered Oosterschelde water 24 hours before the experiment. Water temperature in the indoor pond was ca. 11oC, but is known to fluctuate during day and night. At the experimental day nutrient concentrations were P-o-PO4: < 0.03 mM; Si-SiO2: 18 mM; N-NH4: 1.5 mM and N-NO3+NO2: 48 mM. The low phosphate concentration and very high N/P and Si/P ratio's suggest phosphate-limited conditions.
For the experimental procedure we followed the standard protocol with a few modifications due to the lab facilities. Thus the incubation bottles were filled with 55 ml of the sample and to each 20 ml with 2 mCi was added. The bottles were always incubated for two hours. After incubation the samples were filtered over 47 mm GF/F at a reduced suction pressure of < 15 kPa. The filters then were put in scintillation vials. Up till here all manipulations were done by the different users; the rest (preparing the scintillation vials) by one user. To each scintillation vial 10 ml demineralized water was added. After addition of 0.5 ml 2 N HCl they were bubbled with air for 20 minutes. Previous experiments had shown that this period is long enough to remove all the inorganic 14C. After addition of 10 ml InstagelR the samples were counted for 10 minutes or to 1 % accuracy. Added activity was counted in the same mixture without addition of HCl.
In all samples a Chl-a value was determined using the HPLC method of the laboratory in Middelburg. Filtration was done over 47 mm GF/F at a suction pressure of < 12.5 kPa. SCO2 was measured by titration according to standard procedures; the measured SAlkalinity in some of the samples was 2.263. From each sample 20 ml was taken for cell counts (if needed) and preserved with 50 ml acid Lugol's solution.
The objective was 1) to examine the error in measured primary production parameters if a certain protocol was used by different users working in identical incubators and 2) to check the reproducibility of a measurement.
When determining the error one should take account of different sources of variability:
- variability as a consequence of subsampling,
- variability by the use of different, but in principle identical incubators,
- variability introduced by the inevitable differences in times of starting the incubations (Exp1-3, see below),
- variability by different users.
To attain the first objective a standard Latin Square Design as experimental set-up was chosen. This set-up can be illustrated with the following scheme:
A, B, C and D are the different users. Inc1, Inc2, Inc3 and Inc4 the different incubators and Exp1, Exp2 and Exp3 the 3 successive experiments. Allocation of the incubators (except Inc4) was ad random as was also the case with the distribution of the samples between the users. With this set-up it is possible to take full account of possible error effects within incubators and within experiments, in such a way that a possible user effect can be distinguished.
The first series of measurements (Exp1) started between 9 and 10 a.m., the second (Exp2) between 12 and 13 p.m. and the third (Exp3) between 15 and 16 p.m. In between samples were kept in the dark in cool boxes.
The photosynthetic parameters Pmax, Iopt, Ik and a were derived after fitting the data to the equations of Eilers & Peeters (1988), Jassby & Platt (1976) and Platt et al. (1980). Dark values were not subtracted in the productivity calculations; all dark values except one were ca. 1 % of the maximal photosynthetic rate.
To attain the second objective, reproducibility of a measurement, one user (D) always used the same incubator during Exp1-3 (see scheme above). Unfortunately these results deviated so much from the results of the other three users that a separate consideration was necessary.
Some general information on water temperatures and speed of the turning wheels during the experimental day is given in Table 1. It follows that these characteristics hardly changed during the experimental day.
The mean chlorophyll-a concentration of the nine used samples was 25.6 mg/l and the coefficient of variation 6 %. We thus can conclude that subsampling did not contributed much to variability.
From the analysis of the Latin Square Design it appeared that (except for the slope a determined with the Platt-Gallegos-Harrison model) the incubator (INC) effect was not significant (p>0.05) as was also the case for the time (EXP) effect. After correction of the 'disturbing' factors incubator and time there was no user effect (p>0.05). This means that for determination of the magnitude of the different parameters from the different P-I models the general mean can be used and that the magnitude of the error can be calculated from all measurements. The results (averaged values for all users) are depicted in Table 2.
Furthermore it appeared that differences could be found in a derived from the three P-I models both according to the number of the experiment and the number of the incubator; see Table 3. This table presents the averaged values for all users. The differences are small, but can be demonstrated with a design like this. For the other parameters the variation after correction for the 'disturbing' factors is to such an extent that differentiation is not possible.
From Table 2 it appears that Pmax has the smallest coefficient of variation and thus can be determined most accurately. Iopt is most variable, while Ik seems to be much more stable; especially for the Platt-Gallegos-Harrison model. The values for Pmax, Ik and a are reasonably comparable for the different P-I models.
Table 4 gives the results of the fourth user. Comparison with Table 2 shows clearly that this user's measurements differed from those of the other three. Only during the third measurement results were similar.
Table 5 gives the mean values with the standard errors and coefficients of variation for all P-I models used. These results were obtained from Table 2.
The general conclusion is: by handling of a fixed protocol a very precise production measurement can be performed.
Colijn, F., G.W. Kraay, R.N.M. Duin & M.J.W. Veldhuis, 1993. Design and tests of a novel Pmax incubator to be used for measuring the phytoplankton primary production in ICES monitoring studies. Annex 5 in Report of the Working Group on Phytoplankton and the Management of their Effects, Copenhagen, 28-30 April 1993, C.M.1993/ENV:7 Ref.:L.
Eilers, P.H.C. & J.C.H. Peeters, 1988. A model for the relationship between light intensity and the rate of photosynthesis in phytoplankton. Ecological Modelling 42: 199-215.
Jassby, A.D. & T. Platt, 1976. Mathematical formulation of the relationship between photosynthesis and light for phytoplankton. Limnol. Oceanogr. 21: 540-547.
Platt, T., C.L. Gallegos & W.G Harrison, 1980. Photoinhibition of photosynthe-sis in natural assemblages of marine phytoplankton. J. Mar. Res. 38: 687-701.
ZEMOKO, 1994. Irradiance Differentiation and Control in the ICES incubator. Needed in order to be able to accomplish reliable P-I measurements with respect to phytoplankton primary production studies. Unpublished report.
Table 1. General information on water temperatures and speed of the turning wheels during the experimental day.
|Water temperature (oC)||Speed (rpm)|
Table 2. Mean values, standard errors and coefficients of variation (defined as mean/standard deviation) of several measured parameters. pe=Ei- lers-Peeters model; jp=Jassby-Platt model; pgh=Platt-Gallegos-Harri- son model. Pobs is measured maximal production. Pmax and Pobs in mgC.mg-1Chla.h-1; Iopt and Ik in W.m-2; a in mgC.mg-1Chla.h-1.W-1.m2.
|Mean||Standard error||CV (%)|
Table 3. The slopes of the P-I curves calculated for the different experiments and incubators. EXP stands for the number of the experiment and INC for the used incubator. The measurements are arranged in order of magnitude (except for the incubators under ape, these gave a diffe- rent result when compared with the two other models). All values are mean values for the three users. Legend: see Table 2.
Table 4. The results of the fourth user. * points to a very high value resul- ting from not-saturated P-I curves. The figures are based on three measurements performed simultaneously with the three other users. Legend: see Table 2.
|Mean||Standard error||CV (%)|
Table 5. The mean values for the three different users and the different P-I models used. Legend: see Table 2.
|Mean||Standard error||CV (%)|
Figure 1. Relative transmission of 3 and 1.5 % epoxy-resin filters in the 400- 700 nm range.
Figure 2. Absolute irradiance distribution of four different TL-sets, 10 TL tubes, with polystyrene (PS) foam layer and with coated bottles.
Figure 3. Absolute irradiance distribution with and without polystyrene (PS) foam layer, clear bottles and 10 TL tubes.
Figure 4. Absolute irradiance distribution with polystyrene foam layer, with coated bottles and 2 (xxoxxxxoxx), 4 (xoxoxxoxox), 6 (xoxooooxox), 8 (xoooooooox) or 10 TL tubes.
Figure 5. Absolute irradiance distribution with coated bottles and two 10 TL- sets parallel (P) and crossed (C).
Experts to add figures
Needed in order to be able to accomplish reliable P-I measurements with respect to phytoplankton primary production studies.
Performed by ZEMOKO, specialized in Radiometry and Marine Optics. By order of RIKZ, March 1994.Contents
Neutral density filtercoating
Sensor construction and calibration
Tolerances and imperfections
Irradiance measurements in four incubators
Discussion and recommendation
A full discription of this incubator is given by Colijn et al. (1993). In essence the incubator consists of a TL-illumination set with ten TL33 fluorescent tubes inside, in front of which twelve rectangular shaped incubation flasks, with the flaskneck fixed on a turning wheel, are rotated. The whole flask assembly (see photo 1) is immersed in a ectangular clear perspex water tank. In principle a single-side illumination is foreseen, using a TL-illumination set at one side of the tank (see photo 2) and for higher efficiency a diffuse reflecting polystyrene foamlayer at the other side. [A higher irradiance can be achieved by a dual-side illumination, using one TL-illumination set at each side of the tank]. In both cases, even with all fluorescent tubes switched on, the irradiance distribution in the plane of the flasks is not uniform. That means that during revolution a time\position depending irradiance exists in the flasks and a time\position averaged value should be determined.
Photo 1. Flask assembly. Twelve rectangular shaped incubation flasks with the flaskneck fixed on a turning wheel to be immersed in the rectangular clear perspex water tank of the incubator. In the basic concept of the ICES incubator the possibility of irradiance differentiation is not provided. Only a maximum level can be chosen by the number of fluorescent tubes switched on. For measuring P-I relations an irradiance gradient or a proper sequence of different irradiance values is needed. Changing the basic concept of the incubator as less as possible, the easiest way to achieve this irradiance differentiation is making a sequence by coating the incubation flasks, each in a different transmission density (see section 2). The light attenuation using such a density filtercoating should be preferably neutral, otherwise one needs a proper quantum irradiance measurement including a photosynthetic action response.
When neutral attenuation is achieved, the irradiance measurements can be done with an irradiance sensor having in fact any spectral response within (or partly in) the TL33-spectrum. For measuring total (scalar) irra- diance, a sensor with spatially uniform sensitivity which is small enough to pass through the neck of an incubation flask, is needed (see photo 3). Absolute irradiance measurement is attained by a proper calibration of the immersed irradiance sensor (see section 3). Photo 2. The complete incubator. In principle a single-side illumination is foreseen using a TL-illumination set at one side of the tank and for higher efficiency a diffuse reflecting polystyrene foam-layer (not shown) at the other side.
In the above indicated way one has a proper irradiance control, although some possible sources of error must be kept in mind:
The position of the sensor within the flask is rather critical. To find a proper average irradiance value a well defined central position in the flask should be achieved. The ater in the incubator and flasks should be kept free of airbubbles which cause lightscattering in an unpredictable way.
Influence of ambient light should be controlled, especially in daylight situations. With a polystyrene foam-layer around the incubator an ambient light shield can be realized but each shield may influence the light situation as well, due to reflections. Long term stability of the TL33-light output is still unknown. Besides a reduction in the absolute
irradiance value also variation in the spatial irradiance distribution of the tubes has to be expected. The irradiance value and uniformity in the flasks' plane will change after some time. The fluorescent tubes should therefore be frequently replaced, preferably before the tube-ends are burned in, to maintain the hereafter described situation (section 5). Last but not least there are always some optical imperfections and tolerances in sensor calibration as well (see section 4).
Photo 3. The mini spherical irradiance sensor. A provision is made to position the sensor in a more or less fixed central position in the incubation flask.
NEUTRAL density water-resistant filtercoatings or filtermaterials are rather rare. After some research a good result is achieved with epoxy-resin in which dark pigments are mixed in different ratios. Compared with commercially available filtermaterials spectral neutrality is in most cases better with the advantage that every density value needed can be made.
A proper transmission density sequence for P-I measurements is made in epoxy filterlayers only 0.8 mm thick. The layers with the different transmission densities are integrated on the front- and backside of each flask. The other sides are made totally opaque. The spectral transmittance of the epoxy filterlayers is measured using a spectroradiometersystem and a tungsten halogen radiant source with daylight filter.
The specified transmittance is actually the quotient of two spectral measurements, one with the filterlayer and another without the layer.
Fig.1 gives the relative spectral transmittance measured for the highest density filterlayer with a transmissionfactor of about 2%. This spectral transmittance can be considered as neutral as necessary for irradiance measurements with an arbitrary spectral response.
Additionally the measured relative spectral transmittance of a filterlayer with a 30% transmissionfactor is given.
The lowest line in Fig.1 is the relative spectral transmittance measured on a complete flask with filterlayers on each side with a transmissionfactor of 50% per layer. These layers are slightly polluted by the primer used for the opaque sides of the flask, resulting in a small increasing red transmittance.
Experts to add figures
The thickness used for the epoxy-layers is rather small for a good control of the higher density values of these filterlayers; in fact the thickness forms a critical factor for these filters. Because the flasks have rims on both sides, it was most practical to use these rims to define the thickness of the epoxy-layers.
After only a small correction both rims where brought to an equal height of about 0.8 mm. A thickness of for instance 3 mm would be far less critical. These filters should then be made separately in the proper dimensions to fit on the flasks afterwards. Four identical filter/flask series are made to be used in four identical incubators. The 4 times 12 flasks are numbered 1-0 to 4-11, with the first number for the series and the second number indicating the transmission-step. Increasing numbers correspond with increasing transmittance- and irradiance values. So -0 corresponds with a totally opaque flask, -11 with a clear flask (front and backside only). The weight of these two extreme flasks, -0 and -11, is about 4 g lower as the weight of the other coated flasks because of the lack of filterlayers.
To balance the turning wheel these flasks shoud be mounted in opposite position. Fig.2 is a plot of the aimed transmission sequence together with transmissionfactors measured inside the flasks of the four separate filter/flask series in the incubator. The mutual differences are mainly caused by the above mentioned effect concerning the critical thickness of the filterlayers. The transmittance holds for the diffuse lightcondition in the incubator. Due to scattering in the somewhat diffuse filterlayers the transmittance is somewhat lower than expected. Nevertheless, the aimed
trend is fairly good achieved. Table 1 gives an overview of the attained transmissionfactors for the four filter/flask series.
With these filterlayers transmission density is achieved by absorption. Exposing them to high irradiance values may result in an increasing temperature and distortion of the material. However, in a (cooled) water tank this effect does not appear.
In general the mechanical resistance of these filterlayers will be no problem at all, even at low temperatures (storage at -6 degrees Celsius for weeks of a coated flask did no harm).
Fig.3 shows the relative spectral reflectance of the 18 mm thick polystyrene foam-layer, used for higher efficiency of the TL-illumination and measured to be sure that the spectral influence is negligible, which indeed is the fact.
Three mini spherical irradiance sensors are made, consisting of a silicon photodetector in front of which a green filter is mounted and a spherical collecting element made of diffuse epoxy. A provision is made to position the sensor in a more or less fixed central position in the flask (see photo 3). Fig.4 gives the measured typical spectral sensitivity for this type of sensor, restricted within the TL33-spectrum and Fig.5 the typical spatial sensitivity measured in two planes in front of a TL-set: one radial plane
around the sensor, the other axial through the top of the sensor. In the axial plane the influence of the obstruction, caused by the connecting cable, is obvious.
Fig.5. The typical spatial sensitivity of the sensor, measured by rotating the sensor in two planes in front of a TL-set: one radial plane around the sensor, the other axial through the top of the sensor.
For the absolute calibration of the sensors in Watt*m-2 and uEinst.*sec-1*m-2, the spectroradiometersystem is used again, consisting of a spherical collecting element as well, an optical fiber, a Jarrell Ash gratingmonochromator and a silicon photodetector. Furthermore a standard tungsten striplamp as a wellknown spectral radiance source. This source is used for calibration of the radiometersystem first. Immersed in water the mini sensors are intercalibrated for the TL33-spectrum using the calibrated spectroradiometersystem and a TL-illuminationset at 10 cm distance.
Fig.6 is the absolute spectral irradiance distribution of the TL33-set together with the integrated irradiance value measured at the calibration distance of 10 cm. From this integrated irradiance value the calibrationfactor or "multiplier" for the mini sensor results, as indicated on the labels of these sensors.
Table 2 specifies the same multipliers as indicated on the labels for the three different mini irradiance sensors delivered. These multipliers have to be set in the photocurrentmeter used. There is a difference in relative sensitivity especially for the first sensor compared with the other two. With the appropriate multiplier the absolute sensitivity of the three sensors agrees within 2%.
The measured spectral distribution of the TL33-set presented in fig.6 corresponds with manufacturers data. The spectral spikes on the curve result from the mercury discharge inside the tube.
Only an estimated maximum overall sensor calibration inaccuracy of +/- 15% can be mentioned, which is not an extreme tolerance in radiometric calibration (the overall inaccuracy can be roughly differentiated in: a 5% inaccuracy for the emission of the standard radiance source, 5% for the calibration of the radiometersystem and 5% for the sensor intercalibration). On the other hand, based on comparison with other calibrated irradiance sensors and a different radiometric system, the experience is that mostly the agreement turned out to be better than presumed by the above mentioned inaccuracies. So, the general experience is that the reliability of the present irradiance measurement is fairly good.
Because of aging effects or, in general, long term sensor instability, it is advisable to recalibrate sensors after at least a period of some years, depending on the aimed accuracy and general state of the sensor. Calibration is carried out at a temperature of 18 degrees Celcius. These sensors have a small negative temperature coefficient of about 0.17%/degr. for the selected wavelength, caused by the silicon detector incorporated. Because of the rather small temperature range in the incubator this effect can be ignored in most cases.
To be sure of a good linear response, these sensors should be used in combination with a low input impedance photocurrentmeter such as for instance a LI-COR LI-1000 or LI-189.
From the sensor configuration and the measured spatial sensitivity, it is clear that this sensortype is affected with an obstructed, incomplete field of view, caused in fact by the electric output connection. The sensor is positioned (in the flask) in such a way that this less sensitive direction is orientated in the dark neck of the flask to minimize the influence of this imperfection. By this the influence is negligible, although the restriction remains.
The sensors should be kept in good condition by storing them always mounted in an empty flask so that the sensors are kept free from dirt, scratches
and/or other damages.
With respect to the neutral density filterlayers the only requirement is in fact the spectral neutrality which is, as stated before, sufficient for reliable irradiance control with the spectral response of fig.4.
The only imperfection in the application in the ICES incubator is actually the absorption due to which part of the available light is lost (in general, local light attenuation may be achieved more efficiently by light scattering instead of absorption, see discussion).
Changing the filterflasks partly by clear ones will influence the average irradiance more or less, which holds for changing every absorbing and/or reflecting element in or even around the incubator.
For completeness, a general error resulting from spectral differences between the incubator lightsource and the actual daylight (underwater) spectrum, actually the radiation effectiveness for phytoplankton primary production, should be mentioned. The TL33-spectrum is expected to be less effective compared with most natural spectral situations. The choise of the TL33-type fluorescent tubes is therefore questionable. The available daylight types of fluorescent tubes will be generally more effective.
Four complete incubators were compared with respect to the irradiance in the filter/flask series during revolution in front of the TL- illuminationsets. For convenience the measurements were carried out in one incubator, using four separate TL-illuminationsets and four separate filter/flask series. First a registration was made of the (spatial)
irradiance distribution during one revolution for different illumination conditions. The irradiance is measured in absolute values (expressed in uEinst.*sec-1*m-2) with the mini sensor positioned in the centre of the filterflask 1-11 and all the other filterflasks in position, each one filled with 55 ml water. The same measurement for one illumination condition was carried out with all the twelve flasks clear (uncoated).
Top-position of the flask containing the sensor, during revolution, is indicated as zero-position and clockwise increasing to 360 degr. which is top-position again.
At zero-position the transmission sequence of the four filter/flask series was measured, and the result plotted in fig.2.
Fig.7 is the registration of the irradiance distribution with the coated flasks and the polystyrene foam-layer with respectivily 2,4,6,8 and 10 fluorescent tubes switched on in a pattern as indicated below:
Fluorescent tubes switched on (O):
2 ----------------------- : XXOXXXXOXX
4 ----------------------- : XOXOXXOXOX
6 ----------------------- : XOXOOOOXOX
8 ----------------------- : XOOOOOOOOX
10 ----------------------- : OOOOOOOOOO
Maximum irradiance is reached when the flask passes the middle of the fluorescent tube(s) where the illuminance of the tube is maximum. This characteristic holds for most of the the curves measured. The pattern for six fluorescent tubes switched on results in the most uniform irradiance distribution. It was found that the small spikes on the maxima of the curves are caused by reflection from the coolingpipe situated in the corners of the water tank on the opposite side of the illuminationset.
Figs.8 and 9 show the comparison of the four illuminationsets with (fig.8) and without (fig.9) the polystyrene foam-layer used and with coated flasks.
No big difference exists in the output and spatial distribution of the four different sets. The obvious effectiveness of the polystyrene foam-layer is demonstrated, although uniformity of the distribution certainly is not improved as one would expect at first.
Fig.10 shows the irradiance distribution when dual-side illumination is used.
With the tube direction of both sets in crossed position the irradiance becomes 50% higher, with the advantage of better uniformity, compared with the single-side illumination with the polystyrene foam-layer in position.
Fig.11 demonstrates a 50% irradiance gain when (12) clear flasks are used instead of (12) coated flasks (compare fig.11 with figs.8 and 9).
The effectiveness of the foam-layer is obvious again. The maximum irradiance is only 10% lower in comparison with dual-side illumination with parallel tube direction (compare fig.11 with fig.10).
In table 3, the statistics for the measured irradiance distributions have been listed; all statistical values refer to one complete revolution comprising 360 samples. Table 4, is a specification of the average absolute irradiance in the four different filter/flask series. These values result from the averages specified in table 3, multiplied by the easured relative transmission factors, determined in the flask zero-position for the different filter/flask series (see fig.2 and table 1).
Within certain limits the aimed light differentiation and control is achieved. A general restriction concerning the lightlevel remains. When only one illuminationset is used the maximum averaged absolute irradiance is rather low, even with the polystyrene foam-layer in position. The coated flasks absorb part of the available light and due to the 0paque sides there is also a spatial limitation of the light-input in the flasks (see table 3, influence of flask-coating). However, the opaque sides were necessary for a good control of the transmissionfactors of the transmission sequence in the filter/flask series. In this way the transmission is well defined by the absorbing front- and backside filterlayers only. For practical reasons the thickness of the filterlayers should be increased. These filters are not yet standard available. Probably a more efficient use of the available light is possible by coating the flasks with a neutral diffuse white epoxy-coating. Attenuation then is mainly achieved by diffuse scattering/reflection instead of absorption. As a consequence the transmissionfactor then is difficult to predict and to control, whereas a well defined transmission sequence then is hardly to achieve. Moreover, the now available neutral (diffuse) white epoxy is less neutral than the NEUTRAL black epoxy which is used now. Possibly also a well protected metallic coating could be used as a more efficient one.
Another point of view is to leave all the flasks clear with the advantage that one can more easily dispose them. It is recommended (in an eventually revised model of the incubator with a transmission/irradiance sequence) to mount the filters on two filterwheels on both sides of the clear flasks. Furthermore, more compactly mounted circular fluorescent tubes could be used dual-side for more effective and uniform illumination.
Even more than one concentric circular tube (or spiral tubes) could be effectively used.
In general, an optimal spectral characteristic of the fluorescent tube(s) with respect to the effectiveness for primary production is recommended.
The available daylight types of fluorescent tubes are expected to be more effective than the TL33-type. J. de Keijzer.
As a result of some special developments, irradiance differentiation and control in the ICES incubator, needed in order to be able to accomplish reliable P-I relation measurements with respect to phytoplankton primary production studies, is achieved. Due to these developments neutral density epoxy filterlayers could be made on four identical filter/flask series.
Besides that, it was possible to construct some mini spherical irradiance sensors, small enough to measure the irradiance inside the flasks. The filters and sensors proved to have good optical qualifications.
The maximum averaged absolute irradiance is rather low when single-side illumination is used. Therefore, when an incubator with an irradiance sequence is needed for P-I relation measurements, a more effective and uniform dual-side circular TL-illumination is proposed with the flasks between two filterwheels.
The author would like to thank Mr. L.P.M.J. Wetsteyn for useful comments on the manuscript.
Colijn, F., G.W. Kraay, R.N.M. Duin & M.J.W. Veldhuis, 1993. Design and tests of a novel Pmax incubator to be used for measuring the phytoplankton primary production in ICES monitoring studies. Annex 5 from Report of the Working Group on phytoplankton and the management of their effects. ICES C.M.1993/ENV:7 Ref.:L.
Tables 1, 2, 3 and 4 >>>>>.
Experts to add tables