Peace track with spikes and empty parts

It is probably not a problem of the GPS - I bet the track on the map is all right. The map track would show big deviations if the GPS did not work well. It is a problem of the algorithm Garmin uses for the pace estimation. For some reason, they rely more on the sensor data than on the GPS. So if you just glide, and do not paddle, the accelerometer in the watch (or in the HRM strap if you use one), does not detect any motion, and sends wrong data to the watch.

So now it depends whether you wear only the watch, or also a HRM strap. When using a watch, and keeping the hand with the watch busy, it should be sufficient to generate data allowing the calibration, and the estimation of the pace. If you use a HRM strap, you actually have to paddle so that the strap can sense the acceleration at each stroke. When you glide, without acceleration, the pace will drop to zero.

The worst scenario is when you use a HRM strap, and let the setting for the pace estimation on the option "always". It will then indeed always override the GPS data. When you use the option "indoors", the watch will use the accelerometer from the HRM only when the GPS signal drops.

Thanks for your help,

I m using the HRM belt. I have to check for "Always" or "Indoor" option. 

I  agree that GPS is fine. It seems that pace is based more on MEMS sensors instead of GPS signal as you said. I imagine that It is a matter always of two contributions: MEMS and GPS source. The weight of each contribution can change according to the activity and other paramenters like GPS signal strenght and accuracy. In seems that, because of low accuracy of pace measurement, post processing algorith in the app that extract the pace track have to add empty spaces in order to compesate the space measured with GPS and the speed over time (space) extracted by MEMS.

How to solve this two issues? I mean 1) empty spaces 2) smoothing of pace track. 

I have friends with other kind of Garmin devices.. their tracks of stand-up paddle boarding are much more smoothed.

Best regards 

Basically it works in this way:

1. When you select the HRM-Pro option for pace to Always, the pace will always come from the accelerometer and gyro sensors of the strap, regardless of the quality of the GPS signal, and will hence override the pace calculated from the GPS coordinates. This is highly problematic at SUP, since the HRM-Pro calibrates itself during runs and walks, and basically it just estimates the pace from the cadence. At SUP, it does not work well, since the accelerometer cannot measure the pace at glides. It can only detect acceleration or deceleration (strokes). The deceleration during glides is too low for getting the pace accurately. Hence it will strongly under-report the pace during the glides, resulting in the drops of pace on the graph.
   
   
2. When you select the option "Indoors" for pace at HRM-Pro, the watch will still use the HRM pace estimation whenever the GPS signal quality is below the top level (level 4). Depending on the conditions, it can still be rather often.
   
   
3. When you disable pace and distance from HRM-Pro, the watch will use the internal accelerometer in similar way.

Some users reported, the SUP activity on Instinct 2 returns a smooth pace curve, when they completely disable the pace from the HRM. Have a look for example here: Pace Graph is not a smooth graph - Instinct 2 series - Wearables - Garmin Forums

I try to explain what I got until now. It works as I didn't expect.

You mean: no strokes during gliding, low or no pace detected. 

So, pace is only extracted by GPS data if signal level is at top level redarless the use of HRM (or other accessories with MEMS) and the activity you set?

Running, Cycling, Stand-Up Paddleboarding dosen't use GPS if signal strenght is lower than 4? In this case, pace is extracted by cadence detected by MEMS of the watch or HRM or other accessories with MEMS?

And what if I'm descending a hill during a cycling activity and signal strenght is not 4? Does the measure should't be an integral of acceleration? I mean no change in acceleration, same speed..

Thanks for your support, it is so interesting.

Hope to go more in depth about this.