Ambient pressure in widgets?

Not 100% sure I want to get into this discussion but what the heck...There are many many misunderstandings on the relationships between Absolute, Sea Level etc but it is actually simple in some ways and very complex in others. You may have some interest/need in knowing the absolute pressure but for most people (and most sport watches) the main requirement is altitude and altitude changes. Calculating altitude requires 3 things: Pressure as sea level, temperature and pressure at altitude. The watch can only determine the latter two values. This is where calibration comes in by telling the watch either the current elevation or, technically, the current sea level pressure (I say technically because this is what you are doing if calibrating the current barometric pressure). That's it - the formula is well established and no magic is needed. Here is a link to a calculator that also includes some details on how it is determined www.mide.com/air-pressure-at-altitude-calculator

So where does the "magic" come in? Here is a very old but still useful link to a discussion on ABC watches www.watchuseek.com/.../ . Bottom line is that the watch needs to figure out what is causing an absolute pressure change - due to changing elevation, weather changes, temperature, swinging your arm to the pressure increases?? A chart on the first link above shows how little a change in pressure relates to a substantial change in corresponding elevation. The various options in the watch (and I don't know the Instinct details - we have two in the family but I have a Fenix) are designed to "tell" the watch to prioritize one thing over the other - "assume most pressure change is atmospheric as I'm not moving much" as an example. I would say that it is these algorithms that have generally improved over the years - far from perfect but not sure how close you can get with a watch that needs to guess what is happening. Side note, a pilot will likely tell you that they recalibrate every 15 minutes or so - a fairly small change in ambient pressure (weather change) has a large impact on the corresponding elevation calculation - not good when flying! I remember the strange results I would get when entering one of those pressurized sports domes or using a high speed elevator in a tall office building.

So I'm not trying to question anyone's requirements or claim the watch will work perfectly, just hopefully provide some more details on how it all works. I suspect many people will tell you they rarely calibrate as they don't really care about the actual elevation, just the climb/descent during an activity (where it does not really matter is the start elevation is 100 or 300m).

Heath Hunter said:

I have a Casio watch with barometer reading which of course shows the real ("ambient") measurement. This is the pressure you actually experience. As a hiker why would you need the calculated SLP at 2500 meter high in the mountain? I guess meteorologists may need the SLP for getting the whole picture of relative pressure in a large area. But for me when hiking I need just the real measurement at the exact location point where I am.

I would suggest you would be much better off with SLP as opposed to Absolute (ambient) pressure.

The Casio watch is an excellent example of why Absolute pressure display is of little value when hiking.  Let's take an example of hiking in hilly terrain with consistent fair weather - i.e. weather conditions are NOT changing at all.

With the Casio watch, displayed Absolute pressure changes every time Elevation is gained or lost - no way to observe any weather trend.

With the Instinct, displayed SLP remains constant no matter the Elevation gain or loss - easy to recognize any weather trend.

HTH

I fully understand your position. I'm also not sure if there is any point in posting my five cents, but I do feel that you did not quite address the point in question. Instead, you addressed another one, which seems to be non-contentious. And, in my opinion, you also took for granted a couple of things that are not correct and further obscure the point.

A. Let me maybe start with your assumptions:

razmichael said:

Calculating altitude requires 3 things: Pressure as sea level, temperature and pressure at altitude. The watch can only determine the latter two values.

1. No, it cannot determine the latter two values. It can only determine the last one, not the value of temperature that the formula in your link requires, for a group of reasons. First, because if it is on your wrist the temperature readings of its own sensor will be severely biased. Secondly, because if it tries to avoid the previous problem by using the temperature information from the source from where it retrieves weather information, it maybe be looking at temperatures in locations that are miles and miles away (unless you are sitting upon the weather station being used, of course). Finally, even if you could properly address the previous problems by taking the watch off for a while, or by installing a proper weather station at your location, you would still be dead in the water and not have the temperature you need, because the temperature that you must use in that sea-level-reduction formula is the temperature at the reference level (in this case, at sea level), not at the altitude where the watch lies, as you can easily see on the formula you used. The watch clearly does not now that temperature. It may try to estimate it, but as we all understand, this estimate will necessarily contain errors, which will creep into the calculation and degrade the sea-level pressure estimation.

2. Calculating altitude does not require only those three things. It requires many more, to be done properly. Even if using the formula whose link you provided (and that formula is one of the simple ones; there are many others with increasing levels of complexity and precision), please notice that using it implies inputting values for say, Lp (temperature lapse rate), M (molar mass of the air), etc. However, the values of these variables depend, for example, on the humidity of the air. Therefore, you should know the humidity (in fact the humidity profile, but let us leave that alone) to properly use the formula. We can use their "dry air" tabulated values, but guess what: more errors will creep into the sea-level-reduction algorithm by doing that. The online calculator in your link (using only those three values) is therefore using a crude approximation to an already approximated formula.

Before proceeding, we must therefore conclude that the watch has no conditions to make a precise reduction to the sea value pressure. It may use mean values for the variables; it may assume the the air is dry; it may assume the standard temperature of 288.16 K for the sea level temperature; it may do the same calculations wether you are on a balloon at 2500 m or on top of a mountain at the same 2500 m (two several different conditions, because of the very different profiles between you and sea level); it may simply use a fixed mean value for the pressure change with height (e.g. 12 hPa/100 m);  but whatever it does will introduce additional errors that were not present in the original ambient pressure measurement. That is adding insult to injury, given that we already had sensor errors (and more) to cope with. That is why I claimed that I would trust a graph of ambient pressure more than a graph of the sea level pressure calculated by the watch. More errors (and dynamically varying errors, because these parameters are varying along the day/week) is a bad thing, never a good one.

Note: The previous points have focused on the formula presented on your link, but they are valid for whatever reduction algorithm you name. If anyone of us knows which algorithm Garmin is using, then it will be possible to point out the inaccuracies, approximations, and assumptions being made by that particular algorithm. Heck, we could then even make a proper sensitivity analysis of the algorithm. So far, however, I did not see any posts or any information explaining the exact way in which Garmin does this. We are all talking ifs and thens.

B. As for the attribution of pressure changes to either altitude variation or weather variation, that is indeed a difficult problem, as you very well said, but that is not under discussion here. One thing is the question: "Is this pressure variation a weather or altitude induced thing?" Another, posterior and totally different question, pertaining only to whatever has been attributed to weather and not to altitude, is: "Which atmospheric pressure am I going to show the user? The measured one (after taking out whatever has been accounted for altitude), or the one that I am going to compute based on it, even though at the expense of introducing further errors? These are two independent questions, and we were discussing the latter one, not the former.

C. Personally, I even prefer the SLP, because my brain is formatted to the average 1013.25 mbar, and I can related to that for my evaluations wherever I am. But that does not hamper the fact that whatever value for SLP pressure is being given to me by my watch is less reliable that the value for the local pressure (with the sensor errors and the weather/altitude attribution being a problem common to both, but the SLP incorporating also the further errors committed in the sea-level-pressure reduction algorithm)

That's it. I am now going into low-pressure mode (a bad pun, but I could not resist it).

Keep having fun with this great watch (and with these small debates :), too)

We all clearly agree with the fact that the pressure changes due to altitude variations must be discounted. But the question is: after having discounted those, what is more reliable? The value of pressure that is left after that, or a value calculated based on it in such a way that it introduces further errors?

So. Let's say that, after having played that game of deciding which changes are due to altitude, and what changes are due to the weather, the watch decided that your ambient pressure (A) is 950 mbar, and that your altitude is 500 m. The question is now: which is one is more reliable to evaluate what is going on with the weather? The value of A, or the value of RSL(A), where RSL() is the reduction-to-sea-level algorithm being used, which necessarily introduces further errors?

The value of A has two sources of errors: the sensor errors, and the eventual errors in the weather/altitude splitting algorithm.

The value of RSL(A) has those very same two source of errors, *plus* the errors committed by the sea reduction algorithm (using values that depend on humidity without knowing the humidity profile; using the altitude without knowing precisely the altitude; using temperatures without knowing neither the temperature at the reference level nor the temperature profile with height; using a formula that is most likely an approximation to start with; etc, etc.)

For me, it seems clear that A is more reliable than RSL(A). Having said that, I relate easier to RSL(A) just because it provides a more useful mental reference (1013,25 mbar) that does not change with my location. But that does not stop the fact of it being less reliable.

Best regards, Gaijin

PMO said:

1. No, it cannot determine the latter two values. It can only determine the last one, not the value of temperature that the formula in your link requires, for a group of reasons. First, because if it is on your wrist the temperature readings of its own sensor will be severely biased.

Well... no.

The main purpose of the temperature sensor in the Instinct is to measure the air temperature AT THE PRESSURE SENSOR in the Instinct.  This allows an accurate translation from Absolute pressure to SLP.

So, the watch knows accurate Temperature at the sensor and accurate Absolute pressure at the sensor which allows an accurate translation to SLP for display (assuming, of course, that the Altimeter was accurately calibrated in the first place).

Horses for courses...

HTH

gaijin said:

So, the watch knows accurate Temperature at the sensor and accurate Absolute pressure at the sensor which allows an accurate translation to SLP

I understand why you would think that. But that is not so. A quick search, both on this forum and on Garmin documentation, will show you that if you want to measure ambient temperature you must take the watch out of the wrist for a while and let it stabilize to the ambient temperature. Otherwise, your body temperature will be transmitted to the watch, therefore to the sensor, and the ambiente temperature readings will be biased. The only instance on which it shows a correct ambient temperature while on the wrist is while swimming. In that case, the water acts as a dissipator of the body temperature and the sensor becomes almost unbiased, and will correctly give you the water temperature.

Also, do not ignore the fact that, as I tried to make clear before, the temperature you need in formulas such as the one under discussion is the temperature at the reference height (sea level, in this case), not the temperature where you (and your watch) are. Therefore, even if you had a perfect thermometer with you, you would still not have the temperature to input into the formula. You would have to estimate it, and with estimation comes error. And we are not even mentioning at this moment the unknown humidity, the fact that all these formulas that do not consider full height profiles are approximations in themselves, etc.

We can milk the cow until it dies, but it will always be a cow.  (I could not come up with a proverb as nice as yours, so I had to make up something. ;) ). The estimation of sea level pressure from the local measured pressure will always create additional errors. It comes with the approximations that are always involved in such mechanisms. No way around it.

Best regards

PMO said:

I understand why you would think that. But that is not so. A quick search, both on this forum and on Garmin documentation, will show you that if you want to measure ambient temperature you must take the watch out of the wrist for a while and let it stabilize to the ambient temperature. Otherwise, your body temperature will be transmitted to the watch, therefore to the sensor, and the ambiente temperature readings will be biased.

You're missing the point.

The critical temperature is the temperature of the air column AT THE SENSOR (NOT ambient air temperature) at the time the sensor measures pressure.  That is why there is a temperature sensor paired with the pressure sensor inside the watch.  The pressure sensor provides the Absolute pressure, the temperature sensor provides the temperature of the air AT THE SENSOR, and that is all one needs to accurately calculate SLP along with a manually calibrated Altimeter.

The temperature AT THE PRESSURE SENSOR is NOT going to be the same as the ambient air pressure while the watch is on the wrist - obviously.  That is why there is a temperature sensor paired with the pressure sensor.

If one takes the watch off the wrist for 20-30 minutes or so, and allows the watch temperature to equilibrate to ambient temperature, then the measured temperature will change, the measured absolute pressure will change, but THE REPORTED SLP WILL REMAIN THE SAME - that's the beauty of using SLP.  Try that with your Casio and everything changes.

HTH

1. Wow. So many capitals in your text. You didn't have to. I am pretty sure I can follow your arguments and your line of reasoning (flawed as they may be) even if you write them in lowercase. Plus, there is this small thing called "net ethics", which equates capitalizing to the written equivalent of yelling. To be honest, I do not have the age nor the inclination to be yelled at.

2. I am missing the point? Yeah, right.

3. You are stating with certainty and capitals that the temperature measured at the pressure sensor location is being "paired" with  the readings of the pressure sensor. That is, that the pressure sensor on the Instinct is temperature compensated. Garmin says the opposite, in https://support.garmin.com/en-US/?faq=WlvNrOungC28xGtwB7hLY5: "Since barometric altimeters are not temperature-compensated, changes in ambient air temperature can affect altimeter accuracy." But maybe you know better.

4. Even if the pressure sensor were temperature compensated as you say (and in that case we should warn Garmin), that would only mean that a better pressure profile would be obtained while on the wrist, but would not mean in any way that the watch would be capable of knowing the air temperature in those conditions. These are two very different things. Your point is therefore not only wrong (Garmin says so) but also moot, because nobody in this thread argued about the accuracy of the pressure sensor. The important point is the watch's inability to determine even the local air temperature while on wrist, and that does not change a jota with the sensor being or not being temperature compensated.

5.You do seem confused concerning which temperatures are required in the sea-level-reduction algorithm. You are aware that the temperatures needed for sea-level reduction are atmospheric air temperatures, not "inside-the-watch", "near the sensor" temperatures, right? If you are indeed confused, you may want to look, for example, at the simple formula for sea-level-reduction whose link has been provided before in this thread, and notice that it does not mention the temperature of the sensor, only the temperature of air. That algorithm requires air temperatures. As it happens, air temperature is precisely what the Instinct cannot give you while on wrist.

6. You are trying to focus the discussion on the temperature sensor, trying to argue that its existence "near the pressure sensor" is enough for proper reduction of the pressure to sea level. If you happen to know for a fact that a formula such as the one discussed here is being used, then you are simply wrong (see the previous points). But if, as I strongly suspect, you have no idea of which algorithm is being used in the particular case of the Instinct to convert local pressure to sea-level pressure (as I do not), then you should take into account that it may not even use temperature at all (it may, for example, simply use the common mean value for the pressure change with height (12 hPa/100 m), use tabulated values for the region, use the standard sea-level pressure (288.16 K) in the said formula, use a fixed ratio as has been suggested in this thread by another poster, or whatever. That would be a disappointment and another serious setback, would it not? Talk about shooting from the hip, right?

7. Based on the fact that you keep trying to reduce the problem of the amount of imprecisions in the sea-level-reduction algorithm to the irrelevant (and wrong, besides) notion that the sensor pressure is temperature compensated, I cannot decide if you did not read my previous posts on this thread or if you simply chose to ignore them. I tried to be as thorough and clear as I could, but maybe I failed.

8. One last effort: you have the local pressure A (and it could be or not be temperature compensated, could be correct or wrong, could have large errors or be exact; it does not matter). The only question being addressed is: should we present this value of A to the user, or should you first pass it through the function B=SLR(A) (let us use the simple one being mentioned in this thread), which will necessarily add more errors (because of the unknowns, estimates, and approximations it requires), and show B instead? It is not as if A has all the errors in it, and that formula will correct them (look at the formula) and produce a value of B without the errors that A had to start with. Whatever errors are present in A will also be present in SLR(A) when such a formula is used. Using such a formula will only add additional errors, not decrease them. You have no difficulty with the math or the involved variables, have you?

9. In any case, and to be honest, the amount of capitals and confusions in this conversation has far exceed what I am willing to accommodate in a single conversation. I am sure that you meant no disrespect, but my threshold has nevertheless been exceeded. Therefore, as far as I am concerned, I will not touch this matter with you again. This conversation is over, as far as I'm concerned. I will be happy to meet you again in the future to discuss some other topic (hopefully with less capitals), but on this issue we are done. I tried to contribute what I could, and all I could do now was to repeat myself, anyway, and that I will not do.

My best regards. Enjoy your Instinct and your life.

PMO said:

Garmin says the opposite, in https://support.garmin.com/en-US/?faq=WlvNrOungC28xGtwB7hLY5

Link goes nowhere.