High frequency sampling in the water column is essential to understand energy processes both in the pelagic and in the benthic communities. They have been overlooked because are expensive and time consuming, but we have to make the effort to do it if we want to have a clearer picture of the system funcioning. However, to asses the health of an ecosystem you have, in my opinion, to integrate, and short time cycles, high frequency sampling may no be the answer you need. In coastal ecosystems, sediment traps and some variables of the benthic fauna may be better to understand changes through the time and make robust spatial comparisons.

Primary production rates are challenging to measure directly, but were such data available they would be a useful indicator of ecosystem function...provided they are in the context of a lot of contextual data: i.e. time series, nutrient, light and grazing levels, compensation depth.

In absence of sich data, satellite imaging in combination with a primary production model (a pelagic mixing and growth model using nutrients data, light) might be a good alternative.

Phytoplankton production is best measured using the 14C incorporation method which is not hard but requires proper training and facilities (radio-isotope). Alternatively I have used light/dark bottle experiments with my classes to look at changes in O2 concentration as an indicator of primary production. The advantage is that with this approach you could also look at nutrient effects in tandem with new production (by amending the bottles with different nutrients. I agree with Les Kaufman though that production alone is not enough to assess the health of the ecosystem. Nutrient concentrations, dissolved oxygen and other factors should also be considered.

I think that the best approach to consider for using production as an indicator is high frequency (i.e., every 15 to 30 minutes) measurements of oxygen concentration. The "Odum" method can be used to estimate apparent daytime production via increases in O2, while the overnight decline in DO can be used to infer community respiration. There are some assumptions associated with the method that can make it better suited to some locations than others. The advent of reliable and biofouling resistent DO sensors make this practical for sustained monitoring and assessment. As others have suggested, measurements of nutrient concentration or chlorophyll-a concentration could be useful as well in an overall characterization of trophic state and ecological condition. In addition to providing estimates of production, diel O2 fluctuation may provide some insight into whether hypoxic conditions on some time scale could be affecting biotic condition.

I agree with the usefullness of high frequency primary production estimates to assess aquatic ecosystems health or functioning, when collected with data on potentially limiting factors like nutrients and light. To overcame difficulties with radio isotopes manipulation and oxygen low sensitivity I suggest the use of variable fluorescence measurements (PAM, FRRF...) and combination of PP modelling with oprtical remote sensing

I would favor this if we separate cyanobacteria from green algae, which can be done by monitoring the pigment phycocyanin. Chlorophyll is produced by many phytoplankton, but it is insoluble in water and is not as easy to detect as phycocyanin, which is soluble in water. This becomes particularly important when cells are lysed. Cyanobactera are favored by the greater acidity of surface waters due to more carbon dioxide being available in the atmosphere during global warming events.

High frequency data is critically important and actually essential for such fast growing organisms in order to minimize the biases and variations, and further, for being able to explain them, for the quality of the assessments. However, it is highly difficult to perform particularly in large environments. Today's technology can potentially overcome this difficulty but at what cost? This is the problem. Perhaps, they provide some products in near future. Beside all this, the obtain complementary data such as species composition, macro- and micronutrients, pigments, grazers, etc., is a "se qua non" condition and another challenging task to overcome. Consequently, both decision makers and scientists have to agree on developing a new vision for monitoring programs whcih are the main data collection efforts marine environmental researchs, particularly for those targeting to evaluate the state of ecosystem.

Primary production (pp) can be estimated from diverse methods as mentioned in answers above, and every approach has its pros and cons interms of cost, instrumentation and facilities required, ease, production of harmful waste products (e.g., 14C). The most convenient method would depend on the particular circumstances (i.e., facilities available) and the goal of your research Anyway I would stress that different methods may yield different pp estimators not directly comparable to each other, e.g., net vs. gross pp. Also very important is that an assessment of ecosystem health may hardly be derived from a single indicator like pp (whatever the method used); information on other ecosystem-level processes, chemical and physical properties would be required to have a reasonably clear picture.

High frequency sampling in the water column is essential to understand energy processes both in the pelagic and in the benthic communities. They have been overlooked because are expensive and time consuming, but we have to make the effort to do it if we want to have a clearer picture of the system funcioning. However, to asses the health of an ecosystem you have, in my opinion, to integrate, and short time cycles, high frequency sampling may no be the answer you need. In coastal ecosystems, sediment traps and some variables of the benthic fauna may be better to understand changes through the time and make robust spatial comparisons.

Satellite monitoring with LANDSAT TM data is able to sample the near surface waters with pixels that are approximately one-fifth of an acre, which means that between 4 and 5 pixels per acre are being measured by monitoring algorithms. There are two such satellites in orbit that are 8 days our of phase, each for an orbit that has a repeat cycle of 16 days, or an overall repeat cycle of 8 days, when using the two satellites LANDSAT 7 and LANDSAT 8. I have found quite a lot of variations from one part of a lake to another part, so the assumption that only a few in situ measurements are able to characterize the whole lake has been shown to be a poor one. The cost of entire lake coverage from satellite is about the same as the cost of a few (less than 6) in situ measurements. True monitoring algorithms do not require re-creation of the algorithms on every overpass, once they have been proved to have a reasonable accuracy. on a withheld data set (collected at another place on another date from which the original algorithm was built). I believe that only through satellite remote sensing can we afford to monitor our lakes and streams for water quality parameters.

I also do not agree with that the health of ecosystem could be indicated by the high frequency phytoplankton production. Some blooms are caused by toxic algae producing toxins in seawater. I think the algal toxins should be an important factor for the health of ecosystem. There are also dissolved toxins in the seawater although the concentration is usually very low but the total content is high. How to estimate the role of toxins for this issue?

Community index enumeration can give a better representation of aquatic health; particularly the stress index can be easily evaluated at an interval of 1 hour, provided an image analyzer is there.

As i am not expert in this area i can not comment but would be interested in reading the comments given by others

High frequency sampling of coastal water system can be integrated with remote sensors, satellite image of the chlorophyll and primary productive zone. Assessment of various physiological ecosystem process like of geochemical cycle, nutrient trap and sediment trap along with different mixing pattern in coastal zone would be integral part of the study. Study may be conducted in undisturbed area of the marine environment where there is less human and industrial influence. This study will help to identify primary productive area.

I think it's incomplete. First, in practice, in places where there is influence of sewage usually have high productivity of phytoplankton, however, some industrial run-off, the productivity could be low. Second, on the theory of ecosystem balance, based on thermodynamics, the ecosystem would be in equilibrium when production is almost equal to the breath (or consumption). If you have only on of these rates, you may not be able to access the ecosystem health.

Igor, this is why I use satellite mapping of phycocyanin (in cyanobacteria) and phosphorous content in water to show where the run-offs are occurring. Having spatial context helps greatly in determining the sources of phosphorous, which cause cyanobacterial blooms. However, there are only a few cyanobacteria that can live in salt water. I have been studing the fresh water of the Great Lakes by satellite. More algorithms are needed for salt water, but it is more difficult to get boats out there and to collect water samples at sea than it is in freshwater lakes. Phosphorous content mapping should be possible in both salt and fresh water, though I have not tried doing it in salt water yet.

A challenging question. In Australia the Federal Department of Environment has identified areas of high productivity as 'key ecological features' and would like to monitor their status. As a Node Leader for Australia's Integrated Marine Observing System we are keen to assist. We have looked at, and trialled, a range of options from various fluorometric methods to CO2 and O2 sensors on moorings and ARGO type floats right through to 14C. Given the spatial distribution of the regions of interest and their spatial extent it is very challenging to achieve in situ measurements in a cost effective manner. This seems to leave calibrating a biogeochemical model based on satellite estimates of ocean colour and we are working on this approach. I would invite any of my national or international colleagues to join us at the Australian Marine Science Association meeting in June this year for a special session on measuring 'ocean health' which will include some papers on primary production as a measure of ecosystem health.

Hi Robert, I agree that that we can use primary production find ecological problems. But, to me, the term health for ecosystems is problematic. I know that it is widely used, but is also widely criticized. Leaving aside the question that this term is applies only within the idea of a super-organism, a major problem is that a change problems in a community compartment may or may not reach the other compartments. Furthermore, different ecosystems have different resistances. This means that a given increase primary productivity for a given time may be a problem for a particular community (bringing drastic changes), while for others this productivity can be assimilated without problems. So, I believe that if you don't have enough knowledge about this specific ecosystem working, only primary productivity may not be enough.

The use of "high frequency phytoplankton production estimation " as an ecosystem health indicator"should be considered with other skills as reported above to specify health statement of an ecosystem and mainly should be considered the toxic forms and their effects on organisms. Else, surveys should consider lanfield specifities and geographical location, mainly land input wastes and impact on biomass production for aquatic ecosystems.

Hi Nathalie,

MSG/Seviri has 4 spectral bands in the Visible and NIR region. (Blue, Green, Red, NIR). This makes it possible for this sensor/platform to monitor ocean (coastal) sea colour at high mapping freqeuncies (15 minute shots). Combined with a dedicated radiative transfer model for the path - sun, atmosphere, ocean, atmosphere, sensor, one can correct for atmospheric impacts of aerosols, water vapour, and ozone on the rediation transfer form the top of theatmosphere to the sensor, after passing through the sea surface waters. The final goal is to estimate water surface leaving radiation especially in the blue, since at this wavelength the penetration depth in the sea is maximal and the interaction with suspended materials, but also chlorophylls and other pigments in algae as well.

This appraoch might give you a pretty good idea of algal and suspended material dynamics till a certain depth, actually the depth that algae can perform photosynthesis.

I would not use other platforms or sensors, because they don't catch the intra-daily phytyplankton movements like MSG Seviris does. The spatial resolution is about 4km², which is more than enough to map oceans as well coastal zone dynamics.

Keep me posted.

Cheers,

Frank

It boils down to the question: what is the purpose of the monitoring you are trying to set up. Usually, monitoring is keeping a finger at the pulse of the systrem in question. Trying to measure the actual productivity is no feasible goal for monitoring. Too time consuming and too expensive. What you need for monitoring is a reliable method that delivers consistent answers.

The issue 'ecosystem health' is a conceptual mine field. Don't walk that path. It is of no use; there simply is no objective 'health' level in operational terms, unless you define it yourself; which is a normative and anthropocentric choice anyway.

I would very much advocate the approach mentioned by Frank above. One issue here is that the watercolumn production needs to be calibrated against ocean colour. Modeling water column productivity as a dependant of nutrients, light and water column mixing, validated by ground truthing for your specific coastal waters is a good approach. Ground truthing can be done cheaply by flurometric methods (vertical CTD probe casts). Using phaeopigments versus chlorophyll a rations can be indicative for production cycle of algae. Highly productive algae contain more chl a than phaeopigments. The lower this ratio, the lower the production (senescent or dying algae). Of course it would be great to actually measure primary productivity (radioactively-labeled carbon incorporation), but that would be quite laborious!

If you actually want to improve your understanding of the causal mechanisms, a system approach is needed, incorporating actual measurements of bacterial activity (thymidine incorporation), sedimentation in traps (but be aware of sampling issues due to currents), benthic oxygen consumption (detritus input, bioturbation models, benthic chambers), etc. and construct a low-trophic 3D model inclusing the benthic environment.

Best

Arjen

The application of "high frequency phytoplankton production estimation " as an ecosystem health indicator is useful when other techniques are combined. Particularly in coastal ecosystem and anthropogenically influenced regions this should be considered. There are many new techniques available that could help to better assess the impact of biomass production for aquatic ecosystems.

I agree more withSergio the first contributor.It always a matter of scale and integration of many sites across the seascape. Any conceptual model must match the scale with its matching temporal scale. As Weis(1989) argues the model will suffer from pseudoprediction by not accounting for longer term emergent properties (nearly always!!). Consequently high frequency sampling, be it for dynamic river systems or marine ecology is primarily to obtain accurate integrated estimates at longer appropriate ecological scales. In otherwords accounting for possible aliasing and missing large short pulses of activity across the whole seascape (or statistical number of appropriate sites on the seascape). This makes high frequency sampling very expensive.

However, as in small lakes where a central single site sampling is representative of the whole water column, then as Carpenter suggested high frequemncy may be a useful indicator of the the edge of an change to an alternative stable state. Alternatively, the system maybe fractal (unproven) and the patterns at one scale may reflect the cause and form of patterns at larger scales (see Grahaem Harris and his book on managing ecoystem in an age of complexity _ not the exact title).

Like always, it will depend on the hypothesis your testing and for the moment I believe integrating time periods at very spatial resolution will provide more valuable information. For example the work on the Brisbane River in queensland used planted macrophytes 15N within the water column along the river and outside into Morten bay to soak up both pollution pulses and baseline flows of inorganic nitogen to understand the change and sources and extent of pollution.

Myself Im trying to put a program together using the integration of time from a large number of sediemnt cores across a seagrass landscape inorder to understamnd how the lanscape patch dynamic s affects the top dowm and bottom up control of seagrass meadows i.e matching temporal and spatial dynamics controlled from samller patch dynamics) over the meadows long term response

Regards

John Barry Gallagher

Data sets on lower Gangetic delta are available

Dear all

thank you very much for these highly interesting contributions !

The context of my question was the Marine Framework Directive of the European Union so the "good ecological status" that I associated to the notion of ecosystem health is more a juridic term than a scientific concept. However, because most propositions on coastal areas are turned toward high levels organisms (like the breeding succes of kittiwakes, the large fish index, i.e. the ratio of fishes > 40 cm or the Marine Trophic Index ie the trophic level of organisms at a level > 3.5), we are wondering how to complement this on the low trophic levels monitoring for the "Food Web" descriptor. If you have other ideas than PP, do not hesitate to propose.

Regards,

Nathalie

Measuring the macro & micro elements & compounds that propel phytoplankton production is another way of monitoring productivity of water sheets.Another method is to monitor foulers,settlers & Q & Q of marine larvae that are available in regular intervals of time,as they also feed on phytoplankton.

@Arjen,

Hi Arjen,

And what if you would use a productivity model and calibrate/validate it with the fluorescence method you suggest? The method looks pretty interesting and I figger accurate as well.

We are working as well along these lines but for terrestrial ecosystems, also with fluorescence measurements, but measured form an airborne platform to estimate Radiation Use Efficiency (RUE). I am not sure whether fluorescence is directly measurable overseas. Overland it works. Should be tested overseas.

Well actually if I read you well, you suggest to proceed along these lines no? I guess some good marine environment primary productivity (NPP) models must undoubtedly be available for marine environments as well?

But we would need an aircraft carrier for deep ocean flights ;-).

Doesn't the Dutch Marine not have a few spare ones in Den Helder? With a few crew members to cook for us.

Cheers,

Frank

Dear Nathalie,

very interesting question ...... but i guess such data is very much restricted. Since you are interested to know the lower level food-web dynamics and its relation to higher level. PP can give total production but will miss species specific roles particularly as grazers.... heterotrophic dinoflagellates those who contribute 50% in dinophyta can feed on eggs n larvae and cause considerable damage to the higher tropic level . To keep a track of such community structure .... the only way is regular monitoring from fixed time series station , using classical methods like microscopy and techniques like HPLC and FLOCAM can minimize laborious work involved.. There are few labs in the world those who have fixed time series station on phytoplankton studies ..For eg... gulf of naples (Italy) by SZN ... collecting data from last two to thee decades and dona paula (India) by NIO...collecting data on daily basis from last 5-6 yrs.....

such data set will be very useful to understand the shifts in community structure and its effect on higher tropic level changes...

@Frank

Within our institute, people are currently working on improving our PP model ( we have a North Sea 3D model hydrodynamic model including algal growth functions) adding sediment habitat functionality to it. In the same time, satellite observations are available on fluorescence, and we have CTD data from the NIOZ. Trying to combine these into a calibrated model is an objective, but it is limited to the coast and shallow southern North Sea. And I fear that the current round of economic 'reforms' will leave our Navy more or less crippled as is already the case with the air force and the army.

@Nathalie

We do and have done a lot of work (together with other institutes in the NL) on indicator and target development within the MSFD. One of the hardest nuts to crack are those for D1 and D4, Biodiv and Food web. Various countries are following OSPAR's advice to combine D1, D4 and D6 (Seabottom integrity) into one set of indicators and targets; the distinction between the (criteria and ) indicators mentioned in these descriptors is mostly conceptual and the background documents from the TSGs underline this. They should be better worked out, but the food web descriptor simply lacks a consistent approach; it's a bit of this and a bit of that, typical political solution. Most attention (because best data availability) is on high trophic level species such as birds and mammals, but these hardly constitute the marine food web. Neither are they proper indicators for the food web structure and functioning. Especially the lower trophic levels are underrepresented. My guess is that it is too complex for most people to formulate a consistent and easy to understand approach. This is why direct high-frequency PP measurements are not really an option for regular monitoring programs (unless you have a virtually unlimited budget). This is why many monitoring programs focus on sampling what is easy to sample/measure (an ultra positivist approach). But even for phytoplankton and zooplankton this is not done in an structured manner with a regular (spatio-temporal) pattern. Let alone the complications of offshore picoplankton, meroplankton (free-floating benthic larvae), and bacterioplankton.

Monitoring in the first place simply is not about understanding how the ecosystem functions, it is about trendwise determination of structural changes in the ecosystem that can be attributed to possible human activities and from that derive possible management options (to improve the environment usually). Of course you need a certain level of knowledge to make the causal connections, but at the moment the knowledge gaps are too large for some domains. The challenge here is to use the monitoring and research data to improve knowledge on ecosystem functioning and the human impact. From that an improved monitoring program can be set up, but still this will not be up to the level of the scientific research that focuses on specific details.

So my advice would be to at least try to understand how the lower trophic levels are organised (structurally) in your coastal waters, using standard sampling methods for phytoplankton and zooplankton, and then derive an algal production estimate, based on model development, pelagic nutrient and light penetration data and satellite ocean colour data. What is a large gap in marine ecosystem modeling is the mid-trophic level: the (production) connection from phytoplankton and small zooplankton to the larger zooplankton and smaller (planktivorous) fish. Some models may have a box-type solution (Ecopath), but from data we know that these models oversimplify the complexity of these trophic levels. Message here: don't put too much effort in trying to get this right for MSFD monitoring. Keep it simple, but don't give the message that this will solve the manager's problem. More work is needed to decrease the level of uncertainty for policy making and management of our resources. Uncertainty is something policy makers don't like, but if you want things for cheap, uncertainty and high risks is what you get...

good luck

Arjen

You might be interested in participating to this ICES group “Workshop on Food Web indicators” (WKFooWI) 21 March- 3 April 2014 at ICES headquarters, in Copenhagen.

http://prep.ices.dk/community/groups/Pages/WKFooWI.aspx

If you are dealing with relatively large scales, the monitoring of daily surface chlorophyll content will help you in understanding the direct influence of river plumes (potential eutrophication), of occasional blooms and of oceanic frontal systems. In my work on essential fish habitat, I use chlorophyll fronts as hot spots of marine productivity where top predators are effectively aggregating on. This requires the daily chlorophyll data since the integration in time will make these moving fronts disappearing.

Regards,

Jean-Noel

Speciation of phytoplankton (functional) groups may be relevant in ecological "health" state inference. They may reveal some sort of 'phenology' or relationship with changes in environmental conditions (light, temperature, nutrients, and so).

There is good chance to find optical measurements that can help to achieve some capabilities in phytoplankton clasification based on distinctive pigment ratios for several functional groups

Regards

After my experience with high frequency of sampling , If we sample in coastal area , the state of the sea influence the development of phytoplankton; only high frequency can detect these relationships. for exemple, In calm sea we have a normal development and a peak and when the sea state change , we have a drop of density of phytoplankton ; of course the morphology of the coastal area (shallow water , close bay or open, sea,...) can play also an important role;even if the nutrients are available; a some degree of stability are crucial to have a normal development.

i did a day- to- day sampling in two seasons spring and fall and results were similar with the difference on amplitude which was more important in spring.

I can send a copy of my papers for who are interesting in the results

the data was estimated by Haute performance liquid chromatogtraphy (HPLC) and by chemtax programm

I think its a great indicator of primary production but primary production is not necessarily an indicator of ecosystem health, stability or more importantly, resilience.

Without knowledge of primary production you basically do not know the potential maximum carrying capacity or your system. So I think we clearly need it. Unfortunately, the current techniques are not really suited for monitoring, but some very promising approaches are being developed right now. At the NIOZ we automated FRRF measurements and the setup can run automatically, and it is nearly ready to put on ships of opportunities, something where I think lies the future. However, FRRF measurements measure rate of photosynthetic electron transport, so some calibration is every now and then necessary, but this can be organized, and calibration can also be done using stable isotopes of C, i.e. 13C. This means no health and safety issues. We have done successful incubation with 250 ml samples on the Atlantic, so sensitivity is not a real issue here.

More important is to have a facility to treat the large datastreams, and we are working on that. The FRRF technique is especially suited for cruisetracks. However, for fixed stations (e.g. smart buoys) there are also some promising new oxygen methods!

So I think that new methodology is coming available to measure primary production at rather low costs at high temporal and spatial resolution and a follow up step should be to combine it with remote sensing to get a better synoptic overview. However, to the best of my knowledge, chlorophyll retrieval algorithms still fail in turbid coastal waters like those of the North Sea. However, if it works, it can be nice if phytoplankton functional types can be extracted from the RS radiometric data.

Of course primary production itself is not sufficient to evaluate ecosystem health of good ecological status. We need other data and proxies, and application of ecosystem modelling can be useful. However, I would not rely on a model only, and it is notoriously difficult to really predict phytoplankton primary production accurately.. Without knowledge of primary production some changes in the foodweb will be hard to understand. We now have the tools to start long term time series at rather low operational costs, we just have to do it!

Regards,

Jacco

very interesting subject. I am interested too. As my institute is looking after inland water resources, what should be our approach?

Hi Soma, in order to give some advice I need information about the infrastructure available: do you want to have a system which you can transport (i.e. something for on a ship) or do you have a fixed station (pontoon, jetty?) in mind. Maybe better to do this by mail: [email protected]

I have a regular monitoring in two stations one inshore station with only surface level and offshore station with surface, 40m and 60m depth since 1999 at monthly sampling with environmental parameters. only part of these results were published and for the moment I am collecting the whole data to write a recapitulative work

We tested the hypothesis that "PP measured in different marine systems constitutes an adequate proxy for predicting the amount of secondary and tertiary production (up to the fisheries level) in coastal marine system. In addition, irrespective of the PP, when the energy flows through the classical food web (i.e., large-sized phytoplankton and mesoplankton), differences in the ecosystem functioning and higher level in secondary and tertiary productivity/biomass and fishery activities can be observed than when the energy flows through the microbial food web (channeling through nano- and picoplankton)". Maybe our research could give you other aspect to be considered in this kind of studies: "Structure and functioning of two pelagic communities in the North Chilean Patagonian coastal system" http://link.springer.com/article/10.1007%2Fs10750-013-1576-8#page-1

Any high frequency sampling is bound to find interesting patterns, we tend to extrapolate a bit much from a few point samples. But just measuring production may not be enough.

Look at this approach:

http://link.springer.com/article/10.1007%2FBF02803563#page-1

The ecosystem health can't be determined by monitoring/estimating the phytoplankton/primary production alone. For an ecosystem health estimate, the levels of all trophic systems together with physical and chemical factors to be estimated and integrated. An integrated analysis of all the mentioned aspects (ecosystem modeling) will be helpful to trace out the ecosystem health.

High frequency monitoring of phytoplankton of the region definitely shows the indicator of the estimation of health ecosystem.Before you should consider the parameters from both physical and chemical states which the phytoplankton shows their variability in respect to the regional impact.And also the study of consumer to species specific of phytoplankton is important to check the health ecosystem.in my opinion there must be study required in community structure of phytoplankton in relation to certain influential parameters like temp.salinity.light and nutrients of high resolution data required and put in respective developed regional ecological modeling to study the ecosystem health.

It is also necessary to consider seasonal variation as most phytoplankton population are driven by seasonal changes, especially those at areas less influenced by anthropogenic activities. Initially you can conduct a high frequency sampling then if your analysis shows no significant variation in the parameter (e.g. composition, abundance, density, distribution, etc.) you can adjust your sampling frequency as necessary. Yes, also measure physico-chemical parameters as aside from seasonal variation, these will drive your phytoplankton population. You can stat tests to determine which parameters most likely affects your population. I think for productivity you can initially measure chl-a content. This is easily done in the laboratory.

A first and important question to be considered is how to define "environmental health"! Primary production is definitely an important indicator, especially when sudden (anthropogenically mediated) changes occur. However, since "health" implies a whole bunch of features and affected parameters, it only covers a particular aspect. Monitoring nutrients and potentially even other chemical parameters may be important. Further, the bacterial community and bacterial secondary production can be very sensitive to environmental changes, and may be considered as well. On the longer run invertebrates are used as bioindicators (especially in freshwater ecology) to characterize the quality of aquatic systems (streams, rivers, lakes). Again, to follow and document shifts and changes (long term monitoring) could reveal deeper insight to environmental alterations.

Peter, I agree wit you, "environmental heath" is a complicated issue, pp is only one of the measures that indicates biological production. Since doubling time of phytoplankton is very short, it is better to apply high frequency techniques, but depending on the availability of resources, Better to use some other indicators which is site-specific

Algae phytoplankton and zooplankton are very interesting to be studied because they have the principles and all the nutrientes ( or maigth be almoust all the nutrientes) of the ecosystem this is why the whales they feed by zooplankton and survive quite well.

The phytoplankton is only one of the indicators of a ecosystem health, such as harmful algae bloom, or eutrophication，but there are several other indicators, such as the functional groups change and sea water quality, oxygen etc.

Ecosystem health ? Yes you need to define it first see Mageaux ans Costanza papers for this.  Robustness to change The presence of HAB eutrophication need to be put in context to the systems natural evolutionary state e.g. A HAB system may be a perfectly healthy HAB system In otherwords separate human values from ecosystem functioning

hi

Article An Update on the Global Stressors and Constraints Affecting ...

Without setting time and space scales for the ecosystem you intended to monitor or analyze, it is not possible to decide on which method will meet your need! Otherwise your question will lead an academic debate :)

I recommend mapping phycocyanin, a blue pigment that is in blue-green algae (such as the microcrocystis strain) to map the toxin microcystin.  Bowling Green State University has a patent pending, of which I am the inventor, for mapping microcystin content from LANDSAT TM that has worked in Lake Erie for determining whether microcystin levels were at or above the 20 ppb level of the World Health Organization (WHO) advisory warning to get boaters off  those parts of a lake that equal or exceed that level. Tom Kornacki is the contact at the university in Bowling Green, Ohio, USA  for patent information. 

High frequency is necessary or not depending on the ecosystem that you want to study. Another thing is that you may prefer to do high resolution instead of high frequency if your system is constant in time but not in space. Example of high frequency can be HOT the Hawaiian times series.

 "monitoring of phytoplankton production in marine ecosystems" is a highly broad objective. You need to specify targeted time and space scales, especially when a regular monitoring is going to be set! This will help you to determine optimal data resolution in time and space, and hence you will be able to collect more precise opinions and comments!