Explanation by Triton of N-DOC testing

[Dan_P]

Hi All, it appears Triton have emailed customers with an explanation of N-DOC testing and it’s value to us, this may as well be Martian to me, my 1st question is does it make sense and is it worth doing? Also do you feel it is for systems using Algae beds? Wonder what it will imply for those who use carbon dosing and/or PO4 media for there main nutrient export. Thoughts chaps?
https://gallery.mailchimp.com/159b2...c45/Triton_Tech_Paper_TRITON_Ratios_final.pdf
@Lasse @MnFish1 @Randy Holmes-Farley

Read the paper. Here are my thoughts.

From the TRITON technical paper “The TRITON Reef Keeping Ratios”, the TRITON N-DOC service seems to be measuring dissolved organic nitrogen and dissolved organic carbon. Through their research, TRITON developed “desirable” levels or TRITON Ratios for carbon, nitrogen and phosphorous. TRITON sells additives to assist their customers achieving and maintaining the TRITON Ratios. The ratios are ratios of elements: (1) total or organic nitrogen to nitrate nitrogen, (2) total organic carbon to inorganic carbon, and (3) total carbon to inorganic carbon to nitrogen to phosphorous.


The technical paper is vague on the methodology used to develop the ratios. I came away with the notion that their one year of research involved splitting their ICP test samples into two and using one to develop their N-DOC procedure and N-DOC database. With this database of elemental analyses and customer observations about their system’s appearance, TRITON mined the database for correlations between element levels and system quality. The process might have gone something like “good systems have this level, problem systems have this level, therefore the ratio should be...” The technical paper does not say whether TRITON researchers put some customer data aside to verify the validity of the TRITON Ratios by predicting the customer’s system quality from its TRITON Ratios. I believe they did not because if they had and were successful, this information would have made for a powerful marketing and foundation for a scientific paper. If you are a cynic, you might say they did try to validate their ratios, failed, and went forwards.

The premise behind the “total or organic nitrogen to nitrate nitrogen” ratio is that the system contains sources of nitrogen (organic nitrogen) besides nitrate that can be digested by microorganisms to yield ammonia which can fuel the growth of cyanobacteria and algae. By keeping the ratio around 1:3, the system, on average, is at minimum risk according to TRITON of nuisance organism growth without starving the system.

The meaning of inorganic carbon to organic carbon ratio escapes me and seems like a situation where you produce a thing and then try hard to put it to use.

Organic carbon to inorganic carbon to nitrogen to phosphorous ratio is a repackaged Redfield Ratio. Here’s why. The TRITON Ratio Ci:Co:N: P is 11,150:1250:147:1. Let’s start by dropping inorganic carbon, Ci, leaving C:N: P 1250:147:1. Divide the amount of each element by 10 and the C:N: P ratio becomes 125:14.7:0.1. The Redfield ratio is around 106:16:1. Those ratios are close.

The actual numbers probably came from data mining their customer database. The carbon value reflects the carbon content of the average “best” system in the customer database, same for the 147:1 ratio for nitrogen and phosphorous, which corresponds to the not unreasonable nitrate and phosphate levels of 9 and 0.1 ppm, respectively. The actual nitrate and phosphare ppm levels would be somewhat lower depending on the amount of nitrogen and phosphorous tied up in organic molecules.

 

[Mortie31]

Read the paper.

I did that’s why I posted!!

 

[Mortie31]

The actual numbers probably came from data mining their customer database.

And what’s the point of this? Unless they had set parameters to mine against this is meaningless. How did they ***** the health of tanks and assign numbers? This whole exercise sounds to me like fitting numbers to some imaginary ideal..

 

[MnFish1]

It's not just an averaging of samples. That's an oversimplification The sample size required to get a 99% confidence level with a very tight confidence interval of 1 for a population of 100,000,000 tanks is about 16,500 participants. I don't know how many reef tanks are out there, but they said they had 23,000 participants. They have enough scale to get statistically significant results, if that's what the data shows, but that doesn't mean any results they got are actually statistically significant. To your point though, we don't know enough to determine the value of any of this yet. How they got their sample population matters. If only people with great healthy tanks participated, it would skew the data. It would be great to know what conditions existed in the tanks of the people to the extremes of the ratios; if a certain imbalance made nuisance algae proliferate, if a certain imbalance resulted in no algae but also no coral growth, if any of them aligned with acro RTN or STN, etc...

With 23,000 "users" it would seem there is a good chance someone on Reef2Reef participated. If anybody reading this was a participant, could you give us some idea of the information they requested and you reported?

How do you know? It has to be an averaging of 'some samples' otherwise they would not have a ratio. The question is 'which samples' are averaged to get the ratio they chose?

 

[MnFish1]

The technical paper is vague on the methodology used to develop the ratios.

Yes it is - thats the problem.

I came away with the notion that their one year of research involved splitting their ICP test samples into two and using one to develop their N-DOC procedure and N-DOC database.

But there is nothing in the information presented to suggest this.

 

[MnFish1]

The premise behind the “total or organic nitrogen to nitrate nitrogen” ratio is that the system contains sources of nitrogen (organic nitrogen) besides nitrate that can be digested by microorganisms to yield ammonia which can fuel the growth of cyanobacteria and algae. By keeping the ratio around 1:3, the system, on average, is at minimum risk according to TRITON of nuisance organism growth without starving the system.

OK - so ?

The actual numbers probably came from data mining their customer database.

Yes - probably - but who knows anything about their customer database. I mean - they analyzed 20,000 + tests - ok - but whats to say that that means anything. Some might be great at reef keeping - some maybe not so good.

I believe they did not because if they had and were successful, this information would have made for a powerful marketing and foundation for a scientific paper. If you are a cynic, you might say they did try to validate their ratios, failed, and went forwards.

Right - so it doesnt make for a powerful marketing and foundation for a scientific paper. It seems like the last sentence in the quote is in fact the truth

Organic carbon to inorganic carbon to nitrogen to phosphorous ratio is a repackaged Redfield Ratio. Here’s why. The TRITON Ratio Ci:Co:N: P is 11,150:1250:147:1. Let’s start by dropping inorganic carbon, Ci, leaving C:N: P 1250:147:1. Divide the amount of each element by 10 and the C:N: P ratio becomes 125:14.7:0.1. The Redfield ratio is around 106:16:1. Those ratios are close.

And they - in their article state that the Redfield ratio is meaningless (which it is to this discussion)

 

[Dan_P]

I did that’s why I posted!!

Sorry, I should have written that more clearly. It should have read “I read the paper” not “(you should) read the paper.”

 

[Dan_P]

And what’s the point of this? Unless they had set parameters to mine against this is meaningless. How did they ***** the health of tanks and assign numbers? This whole exercise sounds to me like fitting numbers to some imaginary ideal..

I agree. How they modeled the data is disappointingly vague. I am pretty sure they ran into the problem of noisy data and all chemical element levels were correlated. And yes, defining the health of a system is complex and subjective.

 

[Dan_P]

But there is nothing in the information presented to suggest this.

Here is the methodology section which describes how TRITON gathered their background for their ratios, their “research” using an element analyzer and their trade secret method to perform the analysis on a seawater matrix.

“A large number of analyses were made over the course of a year at TRITON Lab using a CHNS elemental analyser (N-DOC) and ICP-OES (ICP) with proprietary modifications by TRITON Lab to improve performance with seawater.”

Here is where TRITON got their water samples to create their database of element levels. I believe buried in here is where they got samples from “good aquaria”, public aquaria and their partners. Do regular customers report something about their system’s conditions either in ordering the service or on the website where their date is displayed?

“Dual sample N-DOC / ICP tests were conducted on seawater samples:
a. submitted from our global TRITON user base.
b. supplied by several Public Aquaria.
c. supplied by project partners Cairns Marine and
Horniman Museum.”

Is this muscle flexing or implying something about the database dample size?

“At the time of writing the database of analyses draws on 23000 users and >200 000 tests carried out by TRITON Lab representing the largest analytical data set on the behaviour of seawater in closed reef systems ever assembled.”

This standard analytical methodology stuff.

“Samples were handled under Laboratory conditions to minimise contamination. No dilution or acidification was used and all samples were tested within 3 days and again after 5 days to ensure no shift of organic carbon. Sample tubes were tested for suitability and the method validated to run undiluted samples.”

Buried in this awkwardly written paragraph is the “we put the numbers in a database and looked for correlations”. I don’t understand the last sentence. Does it cryptically tells how they linked element levels with how well a system was doing?

“The data were input to a database to calculate the TRTION Ratios (as described below) and visualise the results for comparison. The data were reviewed weekly and matched with biochemical observations from the supplier.”

 

[Dan_P]

OK - so ?

I thought the extended nitrogen cycle discussion did a good job explaining why nitrate level only gives us an incomplete picture about the nitrogen flux in our systems. TRITON did not say it, but they probably pinned their hopes on organic nitrogen level being a better diagnostic tool than nitrate in understanding cyanobacteria and algae breakouts as well as judging the “health” of a system. I feel the other ratios may have been invented after the date mining.

 

[MnFish1]

I thought the extended nitrogen cycle discussion did a good job explaining why nitrate level only gives us an incomplete picture about the nitrogen flux in our systems. TRITON did not say it, but they probably pinned their hopes on organic nitrogen level being a better diagnostic tool than nitrate in understanding cyanobacteria and algae breakouts as well as judging the “health” of a system. I feel the other ratios may have been invented after the date mining.

I agree = they say a lot. Whether what they say means anything or not remains to be seen. there is certainly no evidence that anything they presented will help/prevent cyanobacteria. Perhaps they are waiting for some other perfect time to do this - im just surprised they didn't choose to release it now lol.. We can all guess why.

 

[laverda]

I have not sent in a sample to Triton for testing. I did consider it when my tank was not doing well recently. I would guess a large number of people are like me and only send in samples when they are having an issue. While these samples my provide useful information and my help determine proper ratios. I would think they should be excluded from any sort of averaging to come up with a optimum ratio. The only samples I would think of much real use would be ones sent on a regular basis from the same customers combined with detailed brake down on how the aquarium is doing for each sample. Which out this it is just guessing. Even with that it is going to be very hard to really pin down optimum numbers.

 

[Dkeller_nc]

By the way - just a side comment about data analysis and the scientific method. It appears that what Triton has done is a classic case of retrospective data analysis to uncover correlations. That's a very, very useful tool in science to get ideas when forming a hypothesis, and it's done all the time in fields like human nutrition and disease processes. One significant example is the observation that patients with cancer that develop a severe opportunistic and untreated infection in some cases actually clear their cancer with no other treatment. The first observations and publications of this effect occurred in the early 1800's, and is one of the inspirations for today's immune checkpoint inhibitors, such as Opdivo.

However, mining data for promising correlations is only the first step in any scientific investigation. The next step is to form a hypothesis about the cause of the correlation, design experiments to prospectively test the hypothesis, statistically analyze the results of the prospective experiments to establish that the effect really is caused by the factors being hypothesized, get peers in the field to independently review the experiemental design/results and (hopefully, if you're a graduate student!) publish the results.

In this case, Triton has done the first step - data mining for interesting correlations that might lead to insight about water chemistry and tank management. They've then apparently skipped the next steps and gone straight to marketing. However, I'm not going to be too hard on them for that. This is a hobby field, not cancer therapy, and there's very, very little in the way of controlled prospective testing for most of the guidelines we use on a day-to-day basis for reefkeeping. One of the reasons is the expense; this is still a niche hobby, even though it has grown massively over the 30 years I've been doing it. Most of the companies involved aren't getting rich; they're tiny, and they just don't have the $$$ to spend on controlled trials to justify their marketing claims.

So I guess I'd chalk up Triton's new method as a "might be useful, might not be". It'll take a few years of people testing their systems along with reporting of successes and failures to actually determine whether this sort of testing is useful.

 

[MnFish1]

By the way - just a side comment about data analysis and the scientific method. It appears that what Triton has done is a classic case of retrospective data analysis to uncover correlations. That's a very, very useful tool in science to get ideas when forming a hypothesis, and it's done all the time in fields like human nutrition and disease processes. One significant example is the observation that patients with cancer that develop a severe opportunistic and untreated infection in some cases actually clear their cancer with no other treatment. The first observations and publications of this effect occurred in the early 1800's, and is one of the inspirations for today's immune checkpoint inhibitors, such as Opdivo.

However, mining data for promising correlations is only the first step in any scientific investigation. The next step is to form a hypothesis about the cause of the correlation, design experiments to prospectively test the hypothesis, statistically analyze the results of the prospective experiments to establish that the effect really is caused by the factors being hypothesized, get peers in the field to independently review the experiemental design/results and (hopefully, if you're a graduate student!) publish the results.

In this case, Triton has done the first step - data mining for interesting correlations that might lead to insight about water chemistry and tank management. They've then apparently skipped the next steps and gone straight to marketing. However, I'm not going to be too hard on them for that. This is a hobby field, not cancer therapy, and there's very, very little in the way of controlled prospective testing for most of the guidelines we use on a day-to-day basis for reefkeeping. One of the reasons is the expense; this is still a niche hobby, even though it has grown massively over the 30 years I've been doing it. Most of the companies involved aren't getting rich; they're tiny, and they just don't have the $$$ to spend on controlled trials to justify their marketing claims.

So I guess I'd chalk up Triton's new method as a "might be useful, might not be". It'll take a few years of people testing their systems along with reporting of successes and failures to actually determine whether this sort of testing is useful.

On the surface this makes sense. However, what if the supplement they recommend is harmful (for some reason) - what if it causes problems? You're right - its certainly not cancer chemotherapy - but if the logic we're going to use here is 'these poor companies can't afford to have proof that their products work' - I think that might be a mistake. Showing some evidence that this ratio 'helps' would not cost millions of dollars.

I read the article again - I still don't see any evidence that they have even done anything but collect data - to me it would be nice if they had described what they did (i.e. on what basis did they choose these numbers). At this point - I don't see where they described the correlation (they might have) between the numbers and a successful reef tank. Its an interesting discussion.

Aren't there really a couple issues? 1. Does keeping a tank at Ratio XYZ really mean it will do better than Ratio ABC? 2. Do the products that are being sold keep the tank at Ratio XYZ? ETC. What I see in this hobby, is that people want to believe in a 'magic system'. What I have realized that every tank is different than another - and these 'magic systems' that work for 1 tank may very well harm another.

 

[Dan_P]

I agree = they say a lot. Whether what they say means anything or not remains to be seen. there is certainly no evidence that anything they presented will help/prevent cyanobacteria. Perhaps they are waiting for some other perfect time to do this - im just surprised they didn't choose to release it now lol.. We can all guess why.

I reflexively want to call BS when someone claims “this one thing” will control cyanobacteria or algae growth. I have to say that I still like the notion that the extended nitrogen cycle is a better picture to gave in your head when thinking about nitrogen in your system.

 

[Randy Holmes-Farley Verified Community Expert]

I read the article again - I still don't see any evidence that they have even done anything but collect data - to me it would be nice if they had described what they did (i.e. on what basis did they choose these numbers). At this point - I don't see where they described the correlation (they might have) between the numbers and a successful reef tank. Its an interesting discussion..

Exactly.

By the same sort of "analysis", I could measure the cage size that tigers are kept in at zoos.

I find the average is 1021 sg ft with an standard deviation of 57 sq ft.

I conclude that tigers are optimally maintained in a 1021 sg foot enclosure. Higher or lower may be undesirable and I suggest zoos change the cages to meet this optimal ratio determined scientifically.

Hopefully, that is not all they did. lol

 

[User1]

Welcome to big data be it DNA, search, thermostat, power, or ICP tests. Data is collected as part of a quality of life change or fad and repurposing ensues.

 

[Scott Campbell]

Exactly.

By the same sort of "analysis", I could measure the cage size that tigers are kept in at zoos.

I find the average is 1021 sg ft with an standard deviation of 57 sq ft.

I conclude that tigers are optimally maintained in a 1021 sg foot enclosure. Higher or lower may be undesirable and I suggest zoos change the cages to meet this optimal ratio determined scientifically.

Hopefully, that is not all they did. lol

What would be more helpful is releasing the data for "bad" results. Like if tiger mortality increased dramatically with an average cage size below 800 sq ft. That would be good to know. But arguing for a specific "optimal" cage size above 800 sq ft seems difficult.

For me I'm not sure I really care about the ratios Triton puts forward. I am more interested in using the test to see if there are ratios specific to my tank that are useful. I have a green plating montipora that bleaches back every few months for no apparent reason. Then it recovers with a vengeance. Then dies back again. I am going to send tests in when the bleaching event occurs and also when the coral is growing rapidly and see if the tests reveal anything helpful.

So it seems a different type of test for me than the ICP test. The ICP test I send in every 3 months. This test I think I will send in when my corals are growing especially well and when my corals are experiencing problems and not on any regular schedule. Just to see if I can better understand what is happening. Perhaps averaged ratios from other tanks will be enlightening but I believe I first need to understand what the ratios and variance in my particular tank happens to be.

Bottom line - there may be a lot of questions and doubts about Triton's conclusions and recommendations but I believe it will still be very useful to have more information about my specific tank.

 

[Mortie31]

What would be more helpful is releasing the data for "bad" results. Like if tiger mortality increased dramatically with an average cage size below 800 sq ft. That would be good to know. But arguing for a specific "optimal" cage size above 800 sq ft seems difficult.

For me I'm not sure I really care about the ratios Triton puts forward. I am more interested in using the test to see if there are ratios specific to my tank that are useful. I have a green plating montipora that bleaches back every few months for no apparent reason. Then it recovers with a vengeance. Then dies back again. I am going to send tests in when the bleaching event occurs and also when the coral is growing rapidly and see if the tests reveal anything helpful.

So it seems a different type of test for me than the ICP test. The ICP test I send in every 3 months. This test I think I will send in when my corals are growing especially well and when my corals are experiencing problems and not on any regular schedule. Just to see if I can better understand what is happening. Perhaps averaged ratios from other tanks will be enlightening but I believe I first need to understand what the ratios and variance in my particular tank happens to be.

Bottom line - there may be a lot of questions and doubts about Triton's conclusions and recommendations but I believe it will still be very useful to have more information about my specific tank.

I’m not sure if the doubts are about what there potentially proposing, for me it’s the lack of clarification and explanation, I for one would relish more information about my system but if I don’t understand there results and what they mean and how they have arrived at there conclusions, then I don’t see the point in doing the tests..

 

[Scott Campbell]

I’m not sure if the doubts are about what there potentially proposing, for me it’s the lack of clarification and explanation, I for one would relish more information about my system but if I don’t understand there results and what they mean and how they have arrived at there conclusions, then I don’t see the point in doing the tests..

I completely agree. I guess what I am saying is that I am ignoring their results, averaged ratios and reasons for doing the test. But I am going to use the test any way and see if I discover anything interesting about my specific tank. If the test has value then I should be able to see similar results when my tank is doing well and perhaps similar results when my tank has issues. If I struggle to find any meaningful trends then I expect I will stop using the tests. But I am curious!