Why Na-Sc MH lamps have rapid lumen depreciation compared to MH lamps with other halides?
Is this because scandium causes faster electrode sputtering, causes rapid blackening of the arctube?

Because Scandium has an extremely high affinity for oxygen.  In the cheaper NaSc lamps the salts are dosed as NaI, HgI2 and the Sc as a solid metal chip.  During operation the HgI2 dissociates and releases free iodine which should then react with the scandium chip.  Hpwever, the scandium also melts and before it can be converted to iodide some of it will attack the quartz with which it is in contact.
4 Sc + 3 SiO2 ⇒ 2 Sc2O3 + 3 Si

Silicon is released.  Silicon is extremely soluble in tungsten and will dissolve in the hot electrode tip, reducing the electrode's melting temperature.  Additionally, some of the reacted ScI3 reacts with the ThO2 emitter which is used to dope the tungsten electrodes to give them good work function and high lamp efficacy.  This is actually very beneficial, the electrode temperature is reduced and tungsten vaporisation is also reduced.
3 ThO2 + 4 ScI3 ⇒ 3 ThI4 + 2 Sc2O3

However, some of the metallic scandium will exchange itself with the thorium atoms deep in the electrode structure.  The result is that some thorium metal is released.  That is excellent for boosting the red output and colour rendering of NaSc lamps, but as you may know from the old mercury lamps that used to have thorium emitter on their electrodes, that metal is quickly transported to the wall where it condenses as quite a heavy black layer.  However most of the blackening responsible for lumen depreciation is tungsten from the electrodes.

You now need to understand that NaSc lamps actually have their scandium dosed either as a metal chip, or as bi-component halide pellets of NaI+ScI3 as first invented by Thorn Lighting, and subsequently used by some other manufacturers especially for higher performing NaSc lamps.  One of the big benefits of dosing scandium in metallic form is that it functions as a getter to keep the arc tube purity high.  When changing to ScI3 dosing that beneficial effect is lost.  As a result many manufacturers did not make the change to ScI3 dosing.  It requires a substantial increase in chemical purity of the arc tube preparation, filling and pumping processes.  It is necessary to fill arc tubes in a glove box type environment which is slow and expensive.  Conversely, metal-dosed arc tubes can be processed on semi-conventional high speed automatic lampmaking machinery without need for a glovebox, so they are considerably cheaper.

However, for metal halide lamps containing Sc metal, due to its gettering action there is substantially no free oxygen in the arc tube.  That is a problem, a small amount of oxygen is beneficial because it reacts with free iodine and tungsten vaporised from the electrodes to form WO2I2.  That is a gas, and it prevents tungsten metal from being deposited on the wall and keeps the arc tube ends cleaner for longer.

Some of these challenges are discussed in the excellent paper that Wim van Erk of Philips presented at the Light Sources conference in Eindhoven in 2000.  See Transport processes in metal halide gas discharge lamps, https://www.pismin.com/10.1351/pac200072112159

Thorium iodide is very volatile

Thanks James!

Interesting read.

Sci-hub rules, of course

Never thought of thorium migrating from the electrodes as a reason for changing lamp's color balance. Will try to catch some characteristic Th lines in some seasoned lamps!

But electrodes of MH lamps don't have emitter at all, hence their brightly yellowish color at startup. Aren't they?

No, they are'nt.

Of course the tungsten is activated, by sintering a small amount of thorium oxide in this case. Thorium activation is less efficient than alcali earth oxide one, but is able to survive in MH fill, and work function is still lower than pure tungsten.

Interesting, if pure tungsten electrodes were ever practiced in MH lamps, for the reason of radiation scare, etc?

So why the electrodes in MH lamps glowing brightly during startup while in MV and HPS lamps they don't?

So why the electrodes in MH lamps glowing brightly during startup while in MV and HPS lamps they don't?

Thorium activation is less efficient that alcali earth one. So indeed tungsten has to be heated to somewhat higher temperature to sustain necessary discharge current. The same goes for older power electron tube with thoriated cathodes, glowing unusually bright.

So why mercury lamps with thorium emitter, don't have brightly lit electrodes as well?

Do you have any of such lamp in person? It seems to be a very old tech, hard to get any surviving lamps. And, cold MH lamp absolutely equals cold mercury lamp, just a discharge in pure Ar al low pressure.

The recent Osram/Ledvance MV lamps and most Chinese MV lamps, uses thorium emitter, and their electrodes don't glowing brightly during starting.