I hadn’t even thought about getting HH:MM from a sundial, that’s brilliant! Then getting seconds and the meter from a pendulum is just straight up elegant.
By careful geometry you could reduce tge size of your first container to make this easier, but keeping it big and then dividing the result minimises measurement errors.
I thought this was worth a callout for being a really important consideration in this thought experiment. Understanding that larger scale measurements generally reduce error, and perhaps also repetition with averaging of results, would be incredibly useful in fast tracking the redevelopment of precision.
If you have accesd to ice, letting it just melt gives you 0c.
This one I wondered about more because of the effect of atmospheric pressure(?) on melting point, such that I wondered if it would be worth using Fahrenheit’s Weird Brine ice slurry to get ~ -17.778 ° C instead. But that’s ofc also subject to air pressure influencing melting point so I’m unsure if it’d be worthwhile.
Relatively constant 9.81 m/s² gravity is also useful for deriving force as you mention, though it reminds me of learning, to my abject horror, in undergrad physics that gravity does vary quite a bit by geolocation :'D 9.81m/s² is a better starting point than nothing though