Following on from my suggestion that sensor.iss doesn’t work, I am now questioning sensor.season.
Mine didn’t change to spring until today 21st March 2019 but ‘The Internet’ suggest that it should have been 20th March.
Clearly I am questioning, perhaps two of the least used sensors in HA 
, but how would I go about raising this with someone who has the skills to investigate and perhaps fix?
(Presumably on GitHub? But where exactly?)
Changed here at 10pm (9:58PM) last night on 20/3 we are GMT +11h…
Where was the ‘suggestion’ from and what time zone/lat/long… varies with location I think so that’s a factor… (maybe)
Ah… ok.
The timing of your change might explain it.
The reason I know mine (seemed to be*) updated today on 21st is because I have an automation trigger on the state change but that is not actually acted upon until around 8am in the morning so if mine triggered after 8am yesterday it would mean that my automation would only have triggered this morning on 21st.
Which does then only beg the question, why did the sensor wait about 21 hours to trigger? Unless of course the change to Spring doesn’t actually happen at midnight, but at some ‘random’ time during that day based on astronomical measurements?
To be honest this is clearly not the most important thing to worry about so I am happy with this explanation. Thanks for responding and helping me clear it up (to my satisfaction even if it might not be strictly accurate).
*Unfortunately I can’t check the exact time mine changed now as I have now lost that information because I have been fiddling with the sensor manually to check my automation triggers - yeah, rookie error!
PS In case you’re interested, a season change triggers and update to an input_text which is then read out in the morning by my announcement. Like I said, not exactly going to stop my house functioning but I did get a comment today from my wife at breakfast that she thought Spring began yesterday!
Look in history… that’s how I knew my time…