Salinity Change Results in Slow Response to Conductivity Change

[Tobin VP]

I conducted a 50% water change a week ago in my Coral & Invertebrate Quarantine Tank (10 gal Fiji Cube). I have a Neptune Apex in this tank.

My salinity prior to the change was 33.75 ppt. I intended to raise it to 34.5 ppt with the change. The water change took approximately 30 minutes and during the removal of the tank water I vacuumed my sand bed. Apex probes remained submerged the entire time and temperature remained stable at 78 +/- 0.5 deg F.

After the water change it took at least 48 hours for my Apex to respond to the change. It finally stabilized at around 34.75 ppt as can be seen in the plot below.

It's my understanding that seawater mixing occurs much faster than this (within minutes or an hour?). Any ideas what may cause conductivity to slowly respond to a rapid salinity change?

LATER EDIT AFTER ORIGINAL POST: The statement above that I intended to make the salinity increase during the water change is not actually what I ended up doing after reviewing my log book. I ended up attempting to match the salinity during the change and forgot that this is what I did because it was not my original intent. At the same time during the water change I raised my tank water level and forgot to move my sensor. The two mistakes together resulted in the appearance that the mixing took a long time. The 48 hour increase in salinity here ends up being the result of evaporation do to a tank overfill as explained in posts #18 and #19. So if interested on how this got figured out ... read on.

 

[blaxsun]

That's usually longer than I normally see, but depending on the volume, flow rate and other factors it isn't entirely unheard of. Am I correct that the Neptune salinity probe and your own salinity test are pretty much in-sync now?

 

[Randy Holmes-Farley Verified Community Expert]

Seawater mixing is nearly instantaneous. Longer times of mixing are needed to equilibrate salts water that are in sand and in rock pores.

 

[Tobin VP]

That's usually longer than I normally see, but depending on the volume, flow rate and other factors it isn't entirely unheard of. Am I correct that the Neptune salinity probe and your own salinity test are pretty much in-sync now?

Yes I have cross checked now with Marin Tropic high precision hydrometer. They are in agreement now.

 

[Tobin VP]

Seawater mixing is nearly instantaneous. Longer times of mixing are needed to equilibrate salts water that are in sand and in rock pores.

Yeah - my gut suggested that it would be very quick. When I pour RODI water into seawater I can see twinkling for a bit from the differing refractive indexes - but this disappears within seconds.

So why did my conductivity behave this way? The only other possible thing I can think of (and this seems pretty far fetched) was it possible that an algae was coating the probe and this slowly changed (dissipated) for some reason following the change ... and it took about 48 hours. Notice that the response is inversely exponentially decaying to the new value ... does not look random over that period ... it looks like a fairly well behaved response.

Or is there a possible electrical issue in my tank?

Two other observations that are noteworthy:

1. I have only snails in this quarantine tank right now (getting ready for some coral). Prior to the water change and this conductivity response my 14 Astrea snails were very active and would hang out at the water line often and move all over. Since this water change most of my Astrea snails have barely moved and have stayed on the sand bed in one general area. This is a vastly different behavior than before. Still scratching my head on this too ... seems like they couldn't be related ... but thought I would throw it out nonetheless.

2. Following the water change I wanted to bring the salinity back down a bit. I added a little RODI water. I saw an immediate conductivity change (next sample) and it was accurate. So that mixing showed correctly on the Apex. Also Neptune Systems indicated to me that there is no additional filtering or algorithms applied to the conductivity probe and it reads/reports conductivity measurement directly. They have offered no further explanation though on the observed response.

 

[Randy Holmes-Farley Verified Community Expert]

Aside from some sort of averaging by the software you are using, the answer must be what I mentioned above about equilibrating water that is not in immediate equilibrating contact with the bulk. like down in sand.

Nothing on a conductivity probe will cause a long response delay, unless there is a huge amount of crap keeping it out of contact with the bulk water.

 

[Tobin VP]

Aside from some sort of averaging by the software you are using, the answer must be what I mentioned above about equilibrating water that is not in immediate equilibrating contact with the bulk. like down in sand.

Nothing on a conductivity probe will cause a long response delay, unless there is a huge amount of crap keeping it out of contact with the bulk water.

OK. Thanks Randy for your input.

It indeed looked like a moving average applied to the probe in software ... so I quickly emailed Neptune. They assured me no such thing exists ... and I confirmed that too with my RODI addition a couple of days later.

I can see your point on seawater deep in sand/rock in low flow/mix areas. But in my case that would mean that I would have needed higher salinity in any such area to get the kind of response that I saw. So I think I can rule that out. Also, I have no rock and 1/2" only of sand.

I guess for the time being it shall remain a mystery. I'll add it to my running list of UNSOLVED TANK MYSTERIES ...

 

[taricha]

Unlikely possibility: your new saltwater wasn't 100% dissolved, and the measured conductivity reflects slow dissolution of the last bit?

 

[blaxsun]

The salinity in my 200-gallon system has been gradually diminishing over the past few weeks with the addition of various fish and a quite massive addition of cleanup crew, so when it was reading 35.1ppt earlier today I added several gallons of Tropic Marin saltwater that had been pre-mixed to slightly above 36.0ppt. This was added directly to the display at just after 3pm.

Within 35-45 minutes the Neptune salinity probe had peaked at 36.1ppt and over the next few hours settled in at around 35.7-35.8ppt, which was the last reading as of 5:30pm. Within the next 24 hours I will probably see it finalize at somewhere around 35.3-35.5ppt once the water volume in the system balances out (since I added more than a day's worth of RO, I have to wait for some of the excess water to evaporate). It's currently reading 35.5ppt.

 

[Randy Holmes-Farley Verified Community Expert]

These devices sometimes do a really poor job of temp correction. Might temp be a factor in the slow rise?

 

[blaxsun]

These devices sometimes do a really poor job of temp correction. Might temp be a factor in the slow rise?

Could be. There's a semi-hidden temperature conversion setting for the salinity probe that isn't enabled by default (and most don't realize is there).

 

[Tobin VP]

Could be. There's a semi-hidden temperature conversion setting for the salinity probe that isn't enabled by default (and most don't realize is there).

Thanks for pointing that out. I had indeed thought of that and checked it. Below is an overlay with temperature during that time.

I had fairly normal diurnal temperature fluctuations (approx. 1 deg F day to night) during that period with a very slight decrease of temp but nothing out of the ordinary. I have my temp correction settings at 2.1% and is a typical value for seawater.

 

[Tobin VP]

The salinity in my 200-gallon system has been gradually diminishing over the past few weeks with the addition of various fish and a quite massive addition of cleanup crew, so when it was reading 35.1ppt earlier today I added several gallons of Tropic Marin saltwater that had been pre-mixed to slightly above 36.0ppt. This was added directly to the display at just after 3pm.

Within 35-45 minutes the Neptune salinity probe had peaked at 36.1ppt and over the next few hours settled in at around 35.7-35.8ppt, which was the last reading as of 5:30pm. Within the next 24 hours I will probably see it finalize at somewhere around 35.3-35.5ppt once the water volume in the system balances out (since I added more than a day's worth of RO, I have to wait for some of the excess water to evaporate). It's currently reading 35.5ppt.

This is closer to the vicinity of what I would expect in terms of time responses ... something on the order of an hour or less ... when also considering sampling rates etc. In my case it was a very odd 48 hours of response.

But I'm surprised that you saw any change at all. A 1-2% volume difference (3 gal on 200 gal) of 1 ppt difference should have virtually been undetectable. You can see this by using:

Resulting Salinity Calculator

Aquarium calculator; Estimate the resulting change in salinity when water is added, replaced, or evaporated.

www.hamzasreef.com

Your situation should have resulted in almost no salinity increase - especially when considering the signal-to-noise ratio and precision of the probe. Assuming rapid mixing, I would have expected "maybe" a very slight one or two sample blip and that's it. Something that you probably couldn't even discern in the time series data. Yet you experienced a rapid 1 ppt rise and then somewhat of a decline over 2+ hours.

Puzzling!

Can you provide the readings for the next 24 hours?

One explanation (albeit unlikely) would be that the new 36 ppt seawater circulated for a time at or near your sensor (your reading showed this) then slowly began mixing back with the rest of the tank water which would represent the lowering value over the next couple of hours. I would then expect it to go back to nearly where you had started.

But even if it does go back to where you started as should be expected. This suggests the possibility for some interesting things. Are there some surprising transients and time responses at work with conductivity and/or mixing patterns that are very unintuitive?

Does this represent any kind of real problem for us in the hobby? Probably not. But I like to understand what is going on at a very deep technical level and know what we are measuring and not measuring. For me understanding/discovering the science behind everything in this hobby is at least half the fun!!!

I should note Neptune Systems is somewhat stumped as well. In my case, they told me that all they can say is something affected the conductivity in my seawater in the manner the data shows - but offering no explanation of what that could be.

 

[Tobin VP]

This all has me thinking of some simple transient/mixing experiments that could be done with my Apex salinity probe. If I get around to doing them I'll definitely report back.

 

[blaxsun]

Thanks for pointing that out. I had indeed thought of that and checked it. Below is an overlay with temperature during that time.

I had fairly normal diurnal temperature fluctuations (approx. 1 deg F day to night) during that period with a very slight decrease of temp but nothing out of the ordinary. I have my temp correction settings at 2.1% and is a typical value for seawater.

Note that if you do enable the TC feature you also have to recalibrate the salinity probe. I also recommend floating the (unopened) salinity calibration fluid in the sump for 10-15 minutes before recalculating.

This is closer to the vicinity of what I would expect in terms of time responses ... something on the order of an hour or less ... when also considering sampling rates etc. In my case it was a very odd 48 hours of response.

But I'm surprised that you saw any change at all. A 1-2% volume difference (3 gal on 200 gal) of 1 ppt difference should have virtually been undetectable. You can see this by using:

Your situation should have resulted in almost no salinity increase - especially when considering the signal-to-noise ratio and precision of the probe. Assuming rapid mixing, I would have expected "maybe" a very slight one or two sample blip and that's it. Something that you probably couldn't even discern in the time series data. Yet you experienced a rapid 1 ppt rise and then somewhat of a decline over 2+ hours. Puzzling!

Can you provide the readings for the next 24 hours?

Except I overfilled the tank by 3 gallons and simply let it evaporate to bring the salinity up (at which point the RO wouldn't be running, since the water level was temporarily higher). And yes, you are correct that the initial (higher) salinity of the new water most likely caused the salinity probe to spike, at which point it's settled in over the next 24 hours. Over the last hour it's reporting between 35.5-35.6ppt, with an occasional (minor) spike of 35.7ppt.

One explanation (albeit unlikely) would be that the new 36 ppt seawater circulated for a time at or near your sensor (your reading showed this) then slowly began mixing back with the rest of the tank water which would represent the lowering value over the next couple of hours. I would then expect it to go back to nearly where you had started.

But even if it does go back to where you started as should be expected. This suggests the possibility for some interesting things. Are there some surprising transients and time responses at work with conductivity and/or mixing patterns that are very unintuitive?

Does this represent any kind of real problem for us in the hobby? Probably not. But I like to understand what is going on at a very deep technical level and know what we are measuring and not measuring. For me understanding/discovering the science behind everything in this hobby is at least half the fun!!!

I should note Neptune Systems is somewhat stumped as well. In my case, they told me that all they can say is something affected the conductivity in my seawater in the manner the data shows - but offering no explanation of what that could be.

Yes, it did circulate near the sensor initially. Except I overfilled the tank, so I in actuality after the water evaporated I was probably adding much more salt than 36.0ppt. Hopefully that makes sense?

I'm not entirely sure what happened in your instance, as you'd think that at least within an hour in excess of 90% of the water would have thoroughly mixed already (perhaps even more). Do keep in-mind that with my tank it took around 12 hours for results to "settle somewhat" and closer to 24 hours for the results to "finalize".

 

[Tobin VP]

Note that if you do enable the TC feature you also have to recalibrate the salinity probe. I also recommend floating the (unopened) salinity calibration fluid in the sump for 10-15 minutes before recalculating.

Yes ... thanks ... good advice. I originally followed this post:

(Almost) Everything You Need to Know About Apex Salinity Calibration and TC Factor

I was trying to calibrate my new salinity probe a few weeks ago, and ran into all sorts of problems. Calibrations that were way off the mark, big fluctuations in readings that couldn't possibly be correct (and were verifiably not), etc. After searching around the internet, it was clear that a...

www.reef2reef.com

Also, it was interesting that he discovered that recalibration was definitely required when first turning on temp correction (since it had a zero valued stored) but when adjusting temp factor after if had been turned on recalibration was not needed each time since a multiplying factor is then reapplied to the stored value as it is edited.

Except I overfilled the tank by 3 gallons and simply let it evaporate to bring the salinity up (at which point the RO wouldn't be running, since the water level was temporarily higher). And yes, you are correct that the initial (higher) salinity of the new water most likely caused the salinity probe to spike, at which point it's settled in over the next 24 hours. Over the last hour it's reporting between 35.5-35.6ppt, with an occasional (minor) spike of 35.7ppt.

Ok sorry. I didn't think through well - of course! - your ATO would then not run and that water in the 3 gal add would evaporate for a couple of days. I just went back to Hamza's Reef Calculator - and changed to "add then evaporate" and now everything checks out in that after your 3 gal evaporation your tank would go to 35.6 ppt which you are now seeing.

So I think your transients make sense now. You got an initial spike for a couple of samples at 36 ppt (this would be a couple of samples where mainly the added seawater would be getting measured). Then a rapid drop back down (this would be the mixing of the two seawaters that initially took place). Then a slow increase up to around the 35.6 ppt level (this is the evaporation of the 3 gal add). This seems to make sense within the noise and precision of the device.

Thanks for pointing that out and sharing the data.

 

[blaxsun]

Yes ... thanks ... good advice. I originally followed this post:

Also, it was interesting that he discovered that recalibration was definitely required when first turning on temp correction (since it had a zero valued stored) but when adjusting temp factor after if had been turned on recalibration was not needed each time since a multiplying factor is then reapplied to the stored value as it is edited.

Ok sorry. I didn't think through well - of course! - your ATO would then not run and that water in the 3 gal add would evaporate for a couple of days. I just went back to Hamza's Reef Calculator - and changed to "add then evaporate" and now everything checks out in that after your 3 gal evaporation your tank would go to 35.6 ppt which you are now seeing.

So I think your transients make sense now. You got an initial spike for a couple of samples at 36 ppt (this would be a couple of samples where mainly the added seawater would be getting measured). Then a rapid drop back down (this would be the mixing of the two seawaters that initially took place). Then a slow increase up to around the 35.6 ppt level (this is the evaporation of the 3 gal add). This seems to make sense within the noise and precision of the device.

Thanks for pointing that out and sharing the data.

Fusion now prompts you to recalibrate when you enable that, so at least that's there now. No problem at all. I'm still at a loss to explain the length of time with your own salinity correction, other than it just took longer in your tank for whatever reason (type of salt, mixing, slower flow in the tank).

I usually allow at least 24 hours anytime I add saltwater. I don't do (any) water changes so it makes it reasonably easy to maintain salinity when I'm just occasionally topping up saltwater. I mixed up 5 gallons of Tropic Marin about a month ago and I've added maybe 3.5 gallons in total just to offset all the new fish and inverts I acclimated (about the only way I do end up indirectly performing a slight water change).

 

[Tobin VP]

@blaxsun Thanks! Your input here just solved my 48 hour transient mystery too!

Thinking through your data here triggered me to remember something (and I'm slightly embarrassed I did not catch this previously - ugh).

When I was doing my water change I decided to increase my water level slightly in my tank. But I do not recall ever going back to readjust my ATO sensor level. It's pretty clear now that I forgot to do so. This is VERY LIKELY the explanation for my 48 hour rise which shows an exponential approach to the the stable value. Very consistent with the dynamics of EVAPORATION!

It appears in the data that I got an approximately 0.2-0.3 ppt immediate drop (which would be the rapid time constant we would expect with mixing!). My attempt to match salinity during my water change was close but slightly low in my fresh mix - that's the little spike downward. Then I overfilled my tank slightly but didn't move my ATO sensor. Then that water evaporated taking me up to my resulting value!

It's all making sense now! My data shows an extreme case of what your data showed with the evaporation. NOTE: This is a QT tank 1/20 the size of your tank - so the surface area to volume ratio is such that very small overfills in my case show up much more pronounced.

In the future ... when I finally get back to instrumenting my display tank ... I plan to use one of Neptune's new water level sensors as well as monitor ATO pump cycling. By doing so it would have been immediately obvious what was going on.

Thanks guys!

 

[Tobin VP]

One last point here so anyone reading this thread is not confused.

My original post here stated that I intended to increase my tank salinity to 34.5 ppt with the change. That was my intent originally but I did not end up doing so. Because I was in the middle of testing several salinity measuring devices I just purchased against my new mix I spent several hours tweaking the mix salinity and testing the devices - this had me a bit distracted.

I just confirmed by reviewing my log book (log book finally has value!) that I did not refill with the higher salinity to raise to 34.5 ppt as was my original intent and what I wrote in the intial post. I ended up using a near salinity match in the change and the data indeed shows that with it being slightly low (the little downward spike).

Clearly I was doing too many things at once. I did not remember that I changed my mind along the way on matching salinity when I started noticing the slow rise in the data - I was too focused on my other device testing. And I also did not readjust my ATO sensor to the overfill state of the tank. Two mistakes on my part aligned to lead me down this rabbit hole ...

But very valuable for me in the end to gain trust in the Neptune Apex conductivity measurements in terms of being able to see all of this in the data ... and that mixing does indeed happens very fast. And anytime in the future that I see slowly increasing salinity (on the order of days) to immediately think EVAPORATION! ... and ATO dynamics!