variance decomposition

[hyunjimoon]

edit: i'm not sure between $1/\sigma^2$ vs $\sigma^2$

@tomfid may I rephrase my question yesterday? Please share any thoughts via comment :)

From the above,

1. could we decompose the final variance of either time series vector or scalar? blue theta is structural parameter (ode), green theta is fixed parameter, red is estimated parameter. Challenge is var(theta) for blue and green is zero in current math but the frame I'm taking is an expression of how minuscule effect has.

2. variance can be decomposed with simplex vector but can it be done yo mean as well? ie the above math isn't sufficient for downgrade (not downplay) degeneracy problem

3. between the final outcome as scalar (vaccine value of each state - vector but with spatial decomposition (by_state) it becomes scalar) vs time series vector (future temperature), which is easier to downgrade degeneracy problem?