The SNRM drops because the graphed data is lower
. Not the other way round as your post implies.
The graphed data is lower because the noise is greater.
This effect has been known since ADSL was invented
. It’s why we monitor SNRM, as that’s what it tells us just by looking at the stats in the modern.
The purpose of the graphs you posted is to show up unusual patterns compared to the particular line’s usual one, or compared to other “normal” lines. Your two patterns are basically identical, suggesting there is nothing odd going on. There are known causes of many short stretches of such graphs having consistent unusual or sudden short gaps or very low values, so helping problem diagnosis.
The standard SNRM graph is more useful day to day, enabling you to spot time-patterns for significant drops other than the expected night-time ones. Then you might find the SNR per tone ones for the precise times affected useful.
My broadband basic info/help site - www.robertos.me.uk.
Domains, site and mail hosting - Tsohost
Connection - AAISP Home::1 80/20. 200GB. Sync 01/10/18 - 72382/13812Kbps @ 600m. (Disabled:BQMs - IPv4
If you never think of anything off the wall, you'll never think of anything original.