The Poisson distribution can be used to model the number of decays that happen in a fixed interval of time, given that these events happen at a known constant mean rate and independently of the time since the last event.
Given that 1,000,000 radioactive atoms decayed to 977,287 radioactive atoms over 365 days, we can first determine the number of decayed atoms:
\[ \text{Number of decayed atoms} = 1,000,000 - 977,287 = 22,713 \]
The mean number of decays, \( \lambda \), can be taken as the number of atoms that decayed, which is \( 22,713 \).
In a Poisson distribution, the variance is equal to the mean, so:
\[ \text{Variance} = \lambda = 22,713 \]
Now, rounding this to the nearest tenth, we find that the variance is:
\[ \text{Variance} \approx 22,713.0 \]
Thus, the variance of the decay of cesium-137 in this situation is \( \boxed{22713.0} \).