revise SST climate forecasts: `mean(sst_sd)` annual, monthly (observed) vs annual (forecast)

[bbest]

• 1. fix standard deviation calculation:
  

  climate-dashboard/scripts/process_nc.R

  Lines 35 to 37 in 79ab543

  	# TODO: summarizing monthly to yearly; invalid to take mean of sd
  	sst_mean = mean(sst_mean),
  	sst_sd = mean(sst_sd))

• 2. get ranges to match with similar time steps for comparison; since not matching with switch from monthly (observed) and annual (forecast)
  
• 3. consider different temporal extents, eg since 1970 (not 1850), and either limit available data or zoom to this extent by default.

[bbest]

from Mitch Roffer:

• Error bars on graphs
• Error propagation in climate models | The Hump of the Camel

[bbest]

Frank: expect to see more uncertainty into future, ie expanding cone