Rounding Errors Of Summit Coordinates

This question came about due to manually entering Summit coordinates in my GPS. There are several ways and places to acquire these coordinates. Where you acquire them can make more difference than I would ever have expected. My example is Summit W3/PD-029 Eagle Rock West. Going to the Summit Page on SOTAwatch indicates coordinates of 40 14 27 N, 76 22 22 W. These coordinates took me to the wrong place. Investigating the discrepancy lead to the conclusion that perhaps rounding and truncating errors are responsible.

A friend had acquired coordinates via the SOTA W3 Association GPX File ( 40.24113000, -76.37283000 ). The SOTA W3 Google Map matches those coordinates exactly. 40.24113000, -76.37283000 converts to 40 14 28.07 N, 76 22 22.19 W.

The SOTAwatch Google Map indicates coordinates of 40 14 28.0 N, 76 22 22.1 W ( Google also shows this as 40.241100, -76.372803 ). The SOTA Mapping Project shows these same coordinates rounded to 40.24110 N, -76.37280 W and 40 14 28 N, -76 22 22 W.

Using a piece of Geocaching freeware, FizzyCalc, the resulting conversion for the Decimal Coordinates is 40 14 27.96 N, 76 22 22.08 (or 22.09) W.

It seems that the Summit Page on SOTAwatch coordinates are then strictly truncated to the integer component. A 1 second error here is a difference of 101.195 feet, (thus out of our association AZ). Now that I have experienced this, I will not again rely on Summit Page on SOTAwatch coordinates. Is this a special outlier case not worthy of concern, or is it an issue of interest to SOTA? Which of these many mentioned coordinates is the true position of W3/PD-029 Eagle Rock West?

Glenn - AB3TQ

The Gospel is the database entry for the summit and this case it is 40.2411, -76.3728.


Now that you mention this, I also experienced a similar problem when I approached by car to hike and activate EA4/CU-001 last June. The SOTA Summit page GPS coordinates took me to a point in a dirt road at the bottom of a valley with two mountains one at each side of the road.
I initially started hiking the wrong one, but I quickly changed mountain, as the one at the other side of the road seemed to be a bit higher.
Once in the summit I saw that changing mountain was the right choice.
Not having payed attention to the height of both mountains when initiating the ascent may have taken me to the wrong one and a waste of effort.
Best 73 de Guru - EA2IF

My reading of what constitutes an activation is 100 feet VERTICALLY of the summit. The summit report page gives an elevation of the summit of 1174 feet. As long as you are at an elevation of greater than 1074 feet, you should be good.

The summit in question looks to have quite a bit of area that you could be in to be above that elevation. Here is a link to the Acme Mapper website with the location of 40 14 27 N, 76 22 22 W marked. ACME Mapper 2.2


*disclaimer: I have yet to activate a summit.

I have found penty of summits that I have activated to have some sort of location error. Darkie Point for example, a summit with its peak that is listed as over the edge of cliffs hundreds of meters high etc. The most challenging part of activating is the navigation I reckon. The radio bit is by contrast pretty easy.

Just use your common sense and as another poster has suggested being in the acticvation zone is the main thing to achieve.


I think the problem was confusion to the definition of the AZ. In the US where feet and inches are preferred, the AZ is defined as 100ft vertically from the summit. In Europe the vertical AZ limit is 25m. Had the question mentioned 25m not 100ft I’d have picked up on the confusion.



Thanks for the info. I am the activator who loaded the GPX file into my GPS receiver. Actually, N3TUQ and I had activated this summit for the first time last April; I later tagged along with Glenn during his first successful activation here.

This summit is actually just down the road on the same hill as our radio club site, which I have visited regularly for nearly 20 years. It is virtually the ideal summit for ease of activation, as there is a parking lot for a defunct radio site which extends into the activation zone (the summit itself is in the adjacent woods). The lot is large enough that one can park at the opposite end and walk a sufficient distance to satisfy the final approach requirement (in practice, we park at our site and walk significantly farther).

This is actually the top of a ridge, so there is indeed quite a bit of area which would fall within 100 vertical feet of the summit. However, we have been informed by our Association Manager that we must also be within a 100 foot radius of the summit. In the process of researching access logistics for another local summit, I asked this question as I could not find the answer spelled out anywhere. I believe I read that it is up to the discretion of each Association Manager. Is this different elsewhere?

At any rate, having already activated the summit, I knew where it was and went there. I was somewhat perplexed as to why Glenn was wandering down the hill while following his GPS receiver. :wink: We have noted inconsistencies and irregularities in the past while using the same devices to navigate to geocaches, and we have been victims of mistyped coordinates as well. At first, I assumed that he had loaded the same GPX file, and that his device was acting up. Once I learned that he had entered the coordinates manually, I assumed that it was the equivalent of a typo or something. When I entered the same coordinates into mine, and found the resulting waypoint about 100 feet south of the existing icon for the summit (from the GPX file), it got my attention. A good bit of subsequent research led us to the point which prompted him to start this thread.

This has been a definite learning experience. It is good to know which information is official.

I see several more replies have come in while I am typing this. Indeed, if the only constraint would be 25 meters/100 feet vertical, things would be a lot easier for us…

—73 Karl KA3RCS

Are you sure that you didn’t misunderstand what your Association Manager said? I checked the W3 ARM and it says “Operation must be within 100 ft. vertically of the summit”. I don’t see any requirements related to horizontal distance, nor have I heard of any like that in any other NA association.

As to coordinate accuracy, you may also want to verify that your GPS receiver is set to use the same map datum as was used to originally record the coordinates. For example, if your GPS receiver is set to use the WGS 84 datum and the coordinates were from a map using the NAD 27 datum (most older USGS maps use that datum) then the horizontal position error can be approximately 100 yards. See here for more info:


Eric KU6J

Free SOTA Spot Monitor Software + RBNGate FAQ:

Let me say that we are not at all confused by any coordinate Formats or Datums. While human error must always be considered a possibility, we are quite coordinate savvy. We are not that far from the AZ. We know the area. I just wanted to do a precise GPS confirmation. But the first coordinate I encountered on SOTA clearly had a 1 SECOND ERROR, due to truncation instead of rounding. There was no other issue. I just wanted to point that out with some Corroborating evidence. Nobody picked up on that. The focus shifted. Fine. I will never even look at the worthless truncated coordinates area again. I know where the “bible” is now.

I only added the magnitude of the 1 second horizontal error because I thought it supported my accuracy concern. That is what shifted the discussion to a more important issue. Activation Zone criteria. We received information from a trusted source via email. I will not be sharing that here. At this point it really doesn’t matter what it said. Thanks for all the helpful coordinate knowledge suggestions, but be assured that we could teach the subject.

And the Reflector has spoken. WE ARE NOT UNDER A HORIZONTAL Activation Zone RESTRICTION. Good to know. That subject came up in the context of private property and Summit access issues. It seems we have a lot more wiggle room than we might have been lead to believe. So some real good did come out of this discussion for us. I can find a precise AZ if I have accurate coordinates, but it seems the coordinates really don’t have to be as accurate as I thought. I am quite capable of following accurate coordinates to a nano size object. It is the nuances of the Activation Zone that are proving elusive. And the rules continue to be a moving target. This discussion has now driven us to possibly find new restrictive unknown to us info on our W3 Association web page. I thought everything we needed to know would be in one authoritative place. We do want to play by the rules, but we need to know ALL of them first. And there is no lack of effort to do so. We have also asked for a little private guidance as well.

Thanks for all those that made their best guesses about what my issue was, and then drove it in a more productive direction.

Glenn - AB3TQ

After reading this discussion, I’m left wondering how Joshua Slocum managed, nearly 120 years ago, single-handedly to circumnavigate the globe in a wooden sail boat with no GPS and only a “cheap tin clock” as timepiece…

GPS should be used as a guide not a means of accurate navigation - and batteries run down.
We will be running a navigation session at our walking festival this weekend and the first thing we will do is confiscate all gps receivers and insist on map, compass and common sense for navigation.
“When you reach the top everything around you is down”

Glenn, I did actually forget to say that the display on SOTAwatch does seem to be truncating and not rounding. It’s something for Jon @G4ZFZ to work on.


1 Like

This SOTA game is defined by the final authority, the General Rules, which can be downloaded as a pdf from the website. If there is a conflict between what the ARM and the GR says, the final authority is the GR. Note that an improved and clarified GR is in preparation, but for the moment the wording of rule is “The Operating Position must be within the permitted Vertical Distance of the Summit, as defined in rule 3.5. The terrain between the operating position and the actual Summit must not fall below the permitted distance.” The wording may be improved in the next edition of the GR but the requirement of the rule will be unchanged. As pointed out in an earlier post there is no rule defining a lateral distance from the summit, anywhere will do within the closed contour line of the permitted vertical distance. However, it should also be noted that there is no mention in the GR of any mandatory distance that should be hiked before the station can be set up. Any such distance mentioned in an ARM only has the status of a recommendation, it is not a rule. If there is a car park within the activation zone, some purists will hike to outside the activation zone and then back to the summit, but some activation zones have areas measured in square miles and may even be in urban areas! Basically, if the car is parked within the AZ then the GR only require you to set up an undefined distance away from the car, not using it in any way. Such summits are not common, but they have their uses for the less able amongst us. I only mention this because Karl KA3RCS said “The lot is large enough that one can park at the opposite end and walk a sufficient distance to satisfy the final approach requirement”, and no sufficient distance is defined in the GR.

SOTA is a game or sport, we can adopt any further limitations that we like for our own purposes, but those limitations defined by the GR are the requirements that we must heed.


Brian G8ADD

I was sure that I had read about the 50 foot final approach requirement for a summit accessed by vehicle somewhere, but now I cannot find any reference to it. I have found many variations on the theme of walking out of the AZ and back into it in such a case (all unofficial, of course). I suspect that the 50 foot thing was merely someone’s suggestion.

I had read the GR as well as the W3 ARM previously, and I have now revisited both of them. The only criteria for the AZ specified in the W3 ARM is that operation must be within 100 feet vertically of the summit. It says nothing about lateral distance, or about vehicle access. So far, so good.

As it turns out, a 100 foot vertical separation includes a large portion of the ridge, including our radio club site. Yes, this may be an unusual type of summit, but it was certainly the first one to get my attention when I discovered the programme. In light of this discovery, not only have all of my chaser contacts (as well as reflector posts) thus far been made from the AZ, but I first heard about SOTA itself while I was actually on a summit. :wink:

As stated previously, apparent additional requirements for our Association have been found. Not only were we informed of the 100 foot lateral distance (radius) requirement, but the W3 Association web page states the following:

All SOTA W3 Activators please note!
SOTA W3 ARM only states that all activations must be within 100 vertical feet of the summit.
SOTA W3 ARM does not state that you have to use non-motorized methods to the summit.
SOTA W3 Activators must walk a minimum of 100 vertical feet to their operating position.

I believe that Brian has provided sufficient clarification to invalidate these additional requirements, especially considering that neither the 100 foot lateral separation nor the 100 foot vertical walking requirement are even in the ARM.

No wonder we were confused…

Incidentally, I also found something interesting relating to the beginning of this thread. There is yet another source of coordinates, at least in the case of W3 summits: the W3 ARM. The entry for W3/PD-029 shows the coordinates as 40.24113, -76.37283. As summitslist.csv only uses four digits after the decimal point, I would suggest that these values were actually rounded and lost a bit of precision. The W3 GPX file and W3 Google map coordinates are identical to the W3 ARM coordinates. Of course, this is only a difference of about 14 feet, which we now know is of no consequence whatsoever… :wink:

—73 Karl KA3RCS

I think amateurs generally are very good at inventing and then believing rules and regulations that don’t actually exist. Sometimes there is an obvious explanation as a memory of a licence condition that was abolished sometime in the last century, but other cases seem totally inexplicable. Some just seem to enjoy being regulated (or imposing regulation on others, perhaps).

Although Andy said otherwise earlier in this thread, I was under the impression that the ARM was definitive and the database subservient to it. Of course what actually matters is the truth. I’ve spotted a fair number of errors and discrepancies over the years; sometimes one is wrong, sometimes the other, sometimes both.

Personally, I don’t see a lot of point in maintaining the summit lists in two different places. I wonder how many people bother to check the ARM rather than the online resources derived from the database.

Andy would be able to tell us for sure, but I would guess that they would be held in the database as floating point numbers and then formatted with 4 decimal places of precision when the CSV is generated.

The accuracy of summit coordinates varies enormously between associations. It has always been said that the coordinates are presented primarily for identification purposes are are not intended to give the precise location of the highest point. For example, the very first summit lists in the UK were defined as 6-figure national grid references, with an inherent precision of only 100m. When it was decided to standardise on WGS84 coordinates, I provided conversions which were loaded into the database; I think I offered 5 decimal places of precision but with the understanding that the accuracy was no better than about 3. I think that many of these coordinates have subsequently been adjusted to be far more accurate, but there are still plenty that are quite a bit off.

Given the de facto variation in accuracy, showing 4 decimal places of precision seems a reasonable compromise to me.

Ultimately there is no substitute for looking at the ground under your feet!

1 Like

Actually they are held as strings. As all computer programmers know, floating point numbers are inherintly unable to represent all real numbers. So the database saves the lat & long as a string such as “55.7687”. For some uses, the value is needed as a string and so it just gets displayed as is. In some cases maths is performed on the lat & long and the string is converted to a floating point number when needed. The differences are subtle between saving it as a string and converting to a float when needed or converting it to a float once and saving that. It’s just simpler to use strings, especially as we end up with the data in many files, spreadsheets and text editors during the association creation.

Case where maths is needed are when a lat & long is converted to a Maidenhead locator, e.g. IO83JA or when I do a bit of spherical trig to work out the great circle ditsance between 2 pairs of lat & long for the microwave awards or the S2S distance display.

The positions in the database should be used as gospel and we find that a 4 figure decimal gives more accuracy than needed for SOTA purposes. Most of the time your feet tell you if you are at the top. You don’t need to get to actual top for SOTA, much as that might be nice, because we have a vertical tolerance (25m/100ft). So you don’t need to know the position to more than 4 decimal places.

Association data consists of 2 files, the ARM and a summit list spreadsheet. The spreadsheet is used to load the database. We sometimes receive files with a figure like 44.323333333333 for a position, this is because a DMS figure has been converted to decimal and not rounded. We round the numbers in the spreadsheet. The same data is used to populate the ARM and so you may see 44.3233 in the database and 44.323333333333 in the ARM. Whilst the ARM may be factually more accurate due to possible greater precision, the database is the right value. We shouldn’t have the same data saved in 2 places because you have to keep all the values in sync. There is work under way to simplify the ARMs so that they consist of association background material, info on radio clubs, license authorities, access info, weather info etc. only with the summit list being a link to the spreadsheet that we use. That way when the summits are revised the ARM gets revised at the same time for free. We’'re not there yet.

So you may find multiple values for summit positions some from SOTA resourses (ARM, database, SOTAwatch) and other lists. Gross errors are probably typos and you should report these to the AM for the association. Otherwise the database is the right value to use.


So are strings!

Thanks for the clarification though. In principle then, you could give an approximate indication of the accuracy to which each coordinate is known by varying the number of decimal places stored. (This is not a serious proposal…)

I just checked the co-ordinates of one of my local summits in the database They are 55.5214, -1.8733 - I plotted this in mapping software and added waypoints at 55.5215, -1.8734 and 55.5213, -1.8732 - The mapping software tells me that each waypoint is 42ft from the database location. In highest possible resolution on a 25k map the separation is only just visible and this on screen resolution is better that you will see using any 25k paper map,

Clearly 4 decimal places of a degree is adequate.

Indeed. The resolution is a trifle more than 10m north-south; for east-west multiply that by the cosine of the latitude. Four decimal places are roughly equivalent to an 8 figure UK grid reference and so on, so it’s fairly easy to remember.

Doesn’t your database engine support the ‘decimal’ data type? Strings are entirely inappropriate.