I see your comment on 'noise' and raise you a 'complete balderdash'.
These charts are MONTHLY anomolies and therefore, in theory, should be much less prone to output variation and especially nearer the time.
If a whole month anomoly can change so much from day to day, confidence ranges from extremely low to non-existent.
Originally Posted by: some faraway beach
Are you sure? Why should monthly anomalies be any less prone to output variation than individual days on two different runs? If the synoptic situation generated for 0000hrs on 1 January is different on, say, the 06z run from that on the 00z run, then it's no surprise that the monthly anomaly produced as the next 31 days play out is also different.
This isn't a comment on the usefulness or accuracy of these long-range charts. I'm just not sure that manic monthly anomalies are necessarily a reason to ditch them.
Originally Posted by: nsrobins