StudentOfClimatology Posted October 12, 2014 Author Share Posted October 12, 2014 So a statistical parameter is a myth? These are numbers, any simple SPS or excel analysis can be completed on them with a defined uncertainty. You realized, based on your statement above, there is no point in defining a baseline because "averages" are a myth with temporally defined anomalies? It makes not an once of sense. But regardless, I'm not sure why you want to attempt to down play the global temperature record with a semantical argument. That's not what I said all. I believe the problem is you're trying to develop a singular parameter through equilateral interpolation. That's a technique used frequently in theoretical physics. It's not suitable for analyzing short term ups/downs in global temperature because there are no relative, underlying governing principles to relate the numbers back to. There's no way to verify the authenticity of the average. Link to comment Share on other sites More sharing options...
nflwxman Posted October 12, 2014 Share Posted October 12, 2014 I don't think this is correct at all unless you show the error is biased. +100. Statistics can be used how ever you damn well please. Assuming that the uncertainties and biases are well defined. Link to comment Share on other sites More sharing options...
StudentOfClimatology Posted October 12, 2014 Author Share Posted October 12, 2014 I don't think this is correct at all unless you show the error is biased. It has nothing to do with bias. First, the notion of a record being broken by 0.016 on a resolution of 0.05 is derived solely through a consolidating interpolation, and is not statistically legitimate until a record is reached outside the boundaries of uncertainty. Even then, the probability of a record being broken by 0.016 is not governed through the calculated "average"..you've got a huge range of potentialities here. The mean almost never occurs in actuality. Link to comment Share on other sites More sharing options...
Jonger Posted October 12, 2014 Share Posted October 12, 2014 September GISS comes in at +0.77: The hiatus marches onward. Link to comment Share on other sites More sharing options...
AvantHiatus Posted October 12, 2014 Share Posted October 12, 2014 The hiatus marches onward. Just by a thread tho. One thing is obvious that the way low years are almost gone now, a overall restructuring upward with less volatility month-to-month. Definitely a moderating ocean signal therein. Link to comment Share on other sites More sharing options...
nflwxman Posted October 12, 2014 Share Posted October 12, 2014 It has nothing to do with bias. First, the notion of a record being broken by 0.016 on a resolution of 0.05 is derived solely through a consolidating interpolation, and is not statistically legitimate until a record is reached outside the boundaries of uncertainty. Even then, the probability of a record being broken by 0.016 is not governed through the calculated "average"..you've got a huge range of potentialities here. The mean almost never occurs in actuality. Oh, by the way..I mispoke the record was broken by 0.04C (the old record is 73 in 2005). The Hiatus Continues!!!! PS- I see Jonger is in the trolling mood this afternoon. Link to comment Share on other sites More sharing options...
Msalgado Posted October 12, 2014 Share Posted October 12, 2014 It has nothing to do with bias. First, the notion of a record being broken by 0.016 on a resolution of 0.05 is derived solely through a consolidating interpolation, and is not statistically legitimate until a record is reached outside the boundaries of uncertainty. Even then, the probability of a record being broken by 0.016 is not governed through the calculated "average"..you've got a huge range of potentialities here. The mean almost never occurs in actuality. I get what you're saying but I am not sure its correct because it is not a single measurement. I'm not sure you're wrong, either, but its not just as simple as pointing to a margin of error. Link to comment Share on other sites More sharing options...
nflwxman Posted October 12, 2014 Share Posted October 12, 2014 I get what you're saying but I am not sure its correct because it is not a single measurement. I'm not sure you're wrong, either, but its not just as simple as pointing to a margin of error. Thats right. It's not about statistical significance. It's a set of global measurements each adding up to a collective margin of error. I made a purely qualitative statement, unrelated to statistics that for some odd reason StudentofClimatology had issues with. Breaking a record by a "good amount" is a relative, unscientific statement. I then basically defended my point by showing that the record that was broken was broken by a more than "average" amount. IMHO, that makes it a "good amount." Aside, this is a pretty irrelevent discussion, and it's just distracting from the fact that we have broken the September global temperature record. Perhaps that will change in the future with revisions, but that is where it stands now on GISS. Link to comment Share on other sites More sharing options...
StudentOfClimatology Posted October 12, 2014 Author Share Posted October 12, 2014 I get what you're saying but I am not sure its correct because it is not a single measurement. I'm not sure you're wrong, either, but its not just as simple as pointing to a margin of error. Not to be rude, but you are both sorely mistaken. Within that spectrum, that data-point could be anywhere. You're looking at a probabilistic bandwidth that's 400% larger than the differential. It just can't work. This is just the beginning, too. The average doesn't determine the probability of the deviating anomaly. You'd find a better correlation, in qualitative terms, with the mode. The "average" is derived through a highly variable series of data. Using the "average" here is utterly stupid and irrelavent to reality. Link to comment Share on other sites More sharing options...
Msalgado Posted October 12, 2014 Share Posted October 12, 2014 Not to be rude, but you are both sorely mistaken. Within that spectrum, that data-point could be anywhere. You're looking at a probabilistic bandwidth that's 400% larger than the differential. It just can't work. This is just the beginning, too. The average doesn't determine the probability of the deviating anomaly. You'd find a better correlation, in qualitative terms, with the mode. The "average" is derived through a highly variable series of data. Using the "average" here is utterly stupid and irrelavent to reality. EDIT: Actually can you elaborate on this? Where does the 400% come into play? Link to comment Share on other sites More sharing options...
StudentOfClimatology Posted October 13, 2014 Author Share Posted October 13, 2014 EDIT: Actually can you elaborate on this? Where does the 400% come into play? The average he gave me was 0.016, while the published error bars on GISS are +/- 0.05. So 400% was actually a generous interpolation on my part, if anything. I'd agree with nflwxman (somewhat) if the September average had been broken by 0.06C or more. Link to comment Share on other sites More sharing options...
Msalgado Posted October 13, 2014 Share Posted October 13, 2014 So whats the distribution of anomalies look like? The error bars of GISS are irrelevant to the expected distribution for the value of the monthly anomaly unless GISS is biased. More specifically, how often are records broken by this type of a margin? If it is a rare event, regardless of MOE, that says something. Link to comment Share on other sites More sharing options...
The_Global_Warmer Posted October 13, 2014 Share Posted October 13, 2014 GISS has been revised. What no one said a word or noticed? Where is the outrage? Where is the Stevengoodard blog posts? At the very least every month of 2014 was cooled. 2010 was cooled down to .66C+ for the year. 2014 is currently .651C+ on the year. CFS is currently 0.153C+ or a 0.70C+ monthly GISS equivalent. The forecast has backed off the cooler spell considerably. We will be sitting at 0.150C+ at the minimum thru October 20th. Enjoy it. It's the only climate you've got. Link to comment Share on other sites More sharing options...
StudentOfClimatology Posted October 13, 2014 Author Share Posted October 13, 2014 I didn't notice, actually, but I do now. Thanks for the heads up. One of the reasons I dislike GISS (along with their switch from IOSST/HADSST). Don't see the satellite data or HADCRUT4 changed as often as GISS/NCDC. Link to comment Share on other sites More sharing options...
nflwxman Posted October 13, 2014 Share Posted October 13, 2014 GISS has been revised. What no one said a word or noticed? Where is the outrage? Where is the Stevengoodard blog posts? At the very least every month of 2014 was cooled. 2010 was cooled down to .66C+ for the year. 2014 is currently .651C+ on the year. CFS is currently 0.153C+ or a 0.70C+ monthly GISS equivalent. The forecast has backed off the cooler spell considerably. We will be sitting at 0.150C+ at the minimum thru October 20th. Enjoy it. It's the only climate you've got. Yeah. We would need to average approx a 69 on GISS the rest of the way to break the global temperature record. This seems certainly doable at this point, but the revisions made it less likely than last month. Link to comment Share on other sites More sharing options...
The_Global_Warmer Posted October 13, 2014 Share Posted October 13, 2014 Conditions have definitely turned more favorable for ENSO warming to some degree as we head thru October. There will also be a burst of warming over the Indian Ocean this week. Bob Tisdale downloaded the global ssta from KNMI. Which doesn't have weekly data or the 1971-2010 climo. So he is using 1981-2010 from the raw data. Which is why the anomalies are different. But the data is still the same. September came in ever so slightly cooler than August. September of 2014 is the 2nd warmest monthly anomaly on record only behind August of 2014. Link to comment Share on other sites More sharing options...
The_Global_Warmer Posted October 13, 2014 Share Posted October 13, 2014 Yeah. We would need to average approx a 69 on GISS the rest of the way to break the global temperature record. This seems certainly doable at this point, but the revisions made it less likely than last month. Only 67 to tie it. Link to comment Share on other sites More sharing options...
AvantHiatus Posted October 13, 2014 Share Posted October 13, 2014 Only 67 to tie it. Additonally, one month can average 65-67 while another month averages 75'ish and we would still break the record. Link to comment Share on other sites More sharing options...
The_Global_Warmer Posted October 13, 2014 Share Posted October 13, 2014 Additonally, one month can average 65-67 while another month averages 75'ish and we would still break the record. I think the biggest obstacle will be December. I think October and November will be around .70C+ on giss. November might get close to .80C+ if things fall into place. But December can vary dramatically even with record SSTs Link to comment Share on other sites More sharing options...
nflwxman Posted October 13, 2014 Share Posted October 13, 2014 I think the biggest obstacle will be December. I think October and November will be around .70C+ on giss. November might get close to .80C+ if things fall into place. But December can vary dramatically even with record SSTs Given the statistics, we should average a little over 70 for OND if ENSO finally starts trending towards Nino status. Last year was a 65 with a slightly negative ONI. Link to comment Share on other sites More sharing options...
ORH_wxman Posted October 13, 2014 Share Posted October 13, 2014 Trying to differentiate between averaging 0.64 or 0.69 or 0.72 for an upcoming 3 month period is basically an impossible task. Link to comment Share on other sites More sharing options...
nflwxman Posted October 13, 2014 Share Posted October 13, 2014 Trying to differentiate between averaging 0.64 or 0.69 or 0.72 for an upcoming 3 month period is basically an impossible task. Yeah, the variance is incredibly high over a short 3 month period. The prediction is all in good fun really, using highly uncertain statistics to attempt to map it out. Link to comment Share on other sites More sharing options...
The_Global_Warmer Posted October 14, 2014 Share Posted October 14, 2014 Given the statistics, we should average a little over 70 for OND if ENSO finally starts trending towards Nino status. Last year was a 65 with a slightly negative ONI. Yeah for sure. CFS didn't even go below 0.15C+ on the dailies today. The GFS shows today as the coolest day of the week with slow warming the rest of the week. If that bears out we should see CFS warm up even more from where it is now. Sub-surface warming has started to pick up again around 180-160W. The NPAC had essentially no cooling overall from August to September. It has definitely cooled quite a bit more recently. But the NATL, ENSO, Indian, and W/SW/SPAC has warmed up to help off set that cooling. The one day progs show the Western Indian ocean has started to warm up considerably. Winds are favorable there for warming over the next week as well. Link to comment Share on other sites More sharing options...
StudentOfClimatology Posted October 14, 2014 Author Share Posted October 14, 2014 The NPAC had essentially no cooling overall from August to September. The NPAC has cooled by 0.8C on NESDIS after September 22nd. The WPAC has also cooled...so has the SATL..the entire NH has cooled, as a matter of fact. Only the IO & ENSO domains have warmed. OISST: Versus: Link to comment Share on other sites More sharing options...
Msalgado Posted October 14, 2014 Share Posted October 14, 2014 August to September does not equal September to October. I don't know if there is a cooling trend or not, but what you posted doesn't refute what was said. Link to comment Share on other sites More sharing options...
StudentOfClimatology Posted October 14, 2014 Author Share Posted October 14, 2014 August to September does not equal September to October. I don't know if there is a cooling trend or not, but what you posted doesn't refute what was said. I edited my post after not reading carefully enough..that's on me. I still have no clue why August-September is being brought up, though...those numbers are in the bag and have been for weeks. Link to comment Share on other sites More sharing options...
The_Global_Warmer Posted October 14, 2014 Share Posted October 14, 2014 It also doesn't change that it hasn't cooled off very much. But then again it's not going to. The tropics have warmed up substantially globally. Link to comment Share on other sites More sharing options...
chubbs Posted October 14, 2014 Share Posted October 14, 2014 Global winds were quite weak outside the tropics in September continuing the current weak wind regime. ftp://ftp.remss.com/wind/monthly_1deg/ Year mon -60.0/ -20.0/ 60.0 20.0 2014 1 -0.031 -0.0472014 2 -0.085 -0.0682014 3 -0.070 -0.1122014 4 -0.138 -0.0852014 5 -0.093 -0.1182014 6 0.037 0.1332014 7 -0.012 0.2482014 8 -0.338 -0.3942014 9 -0.486 0.050 Link to comment Share on other sites More sharing options...
donsutherland1 Posted October 14, 2014 Share Posted October 14, 2014 With respect to GISS: 2014 (year-to-date) has had the 4th warmest January-September period: 1. +0.674°C, 2010 2. +0.657°C, 1998 3. +0.653°C, 2007 4. +0.651°C, 2014 The three warmest years on record are: 1. +0.658°C, 2010 2. +0.652°C, 2005 3. +0.619°C, 2007 2014 will very likely wind up warmer than 2007. The kind of cool down that would be required to see the temperature fall to or below the 2007 figure has not been very common in recent years. In the face of a possible emergent El Niño event, such a cool down is very unlikely. Whether 2014 will wind up as warm as 2005 or 2010 remains to be seen. Below are the average October-December anomalies needed to reach each of those years and the percentage of months and percentage of months in the October-December timeframe that saw such anomalies since 2009. To reach +0.658°C: October-December anomaly: +0.680°C 28% months since 2009 had an anomaly of +0.680°C or above 33% of months in the October-December timeframe had an anomaly of +0.680°C or above To reach +0.652°C: October-December anomaly: +0.653°C 29% months since 2009 had an anomaly of +0.680°C or above 42% of months in the October-December timeframe had an anomaly of +0.680°C or above To fall to +0.619°C: October-December anomaly: +0.523°C 25% months since 2009 had an anomaly of +0.523°C or below 33% of months in the October-December timeframe had an anomaly of +0.523°C or below Link to comment Share on other sites More sharing options...
nflwxman Posted October 14, 2014 Share Posted October 14, 2014 Great analysis, thanks Don. Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.