|
Post by trbixler on Aug 28, 2010 1:55:06 GMT
|
|
|
Post by glc on Aug 28, 2010 8:36:15 GMT
If you can't cite your source, don't post it! The source is here: www.worldclimatereport.com/archive/previous_issues/vol4/v4n8/feature.htmScroll down to the last section (it begins 'Finally!') If anyone had wagered using IPCC climate models, would they have won? How about finding quotes from the other side what they predicted would happen after 1998. With you being a skeptic and all it is probably from the same source you cut and pasted. 1998 was a +2 sigma event. No-one expected the record to be broken in the short term. I really tried to be nice glc, but frankly for you do even conjure up something as daffy as thinking "initial conditions" meant a baseline period was being established by using 10 days in August 2010 to compare with January 2011 was....to be nice again....enlightening. Oh Dear - Magellan's got the wrong end of the stick again. Firstly, it's quite reasonable, for example, to compare a global anomaly in August with one in January. Global anomalies are effectively seasonally adjusted (or neutral). However I agree it is not appropriate in this case. Secondly, I am familiar with initial value (and boundary value) problems. I have solved hundreds of differential equations using both analytical and numerical techniques. I have also done some mathematical modelling. It is not unusual to use the initial values (or conditions) as a reference point or point of origin or baseline or whatever. For example, if I wanted to model the path of a projectile (e.g. missile) which has been fired from the surface of the earth it's not much use saying that the projectile has landed 2 miles west of Smallsville. We need to know it's path and landing point relative to the point at which it was fired - i.e. relative to the initial values. I questioned the charts because I couldn't really see the point of specifying the IC period if it's not relevant to the calculated anomalies. I still don't see the point. They could have just specified the date on which the model was run. The IC period doesn't help in understanding what the charts represent. The anomaly base period, on the other hand, would have been informative.
|
|
|
Post by icefisher on Aug 28, 2010 14:39:10 GMT
I questioned the charts because I couldn't really see the point of specifying the IC period if it's not relevant to the calculated anomalies. I still don't see the point. They could have just specified the date on which the model was run. The IC period doesn't help in understanding what the charts represent. The anomaly base period, on the other hand, would have been informative. I posted the relevant statement that one could infer the reason GLC. They run the model 120 times with 3 different consecutive 10 days periods of initial conditions making up the the previous 30 days. The IC is on the chart to identify which runs one is looking at. The latest 10 day runs are used for the forecast so those will always be the latest 10 days to the run date on the final published forecast but they are not the only runs utilized by the modeling team to put together the final forecase. Anybody who has done professional modeling understands the forecast run is not the only run you make. Sensitivity analysis runs are made to understand other aspects of your model outputs. Here many runs are made and analyzed for putting together a final forecast. The IC specifications are obviously there to identify which set of runs the analysts are looking at. The date of the run is not sufficient to convey that. Here is an example of the E1 E2 E3 and Eall from Aug 26 (notice how the IC's are necessary to identify which is which.) E1 E2 E3 EALL
|
|
|
Post by glc on Aug 28, 2010 15:53:00 GMT
I posted the relevant statement that one could infer the reason GLC. They run the model 120 times with 3 different consecutive 10 days periods of initial conditions making up the the previous 30 days. The IC is on the chart to identify which runs one is looking at. The latest 10 day runs are used for the forecast so those will always be the latest 10 days to the run date on the final published forecast but they are not the only runs utilized by the modeling team to put together the final forecase. Anybody who has done professional modeling understands the forecast run is not the only run you make. Sensitivity analysis runs are made to understand other aspects of your model outputs. Here many runs are made and analyzed for putting together a final forecast. The IC specifications are obviously there to identify which set of runs the analysts are looking at. The date of the run is not sufficient to convey that.
It's OK I 've found the source for the model runs and anomaly maps and it all makes sense now. When the original anomaly map was posted (by Walterdnes) I assumed it was just a sort of snapshot which is issued 'publicly' every so often. I've never really looked at what is available with the forecasted stuff before. You are right the date alone would not carry sufficient information.
|
|
|
Post by icefisher on Aug 28, 2010 18:04:36 GMT
Interesting that while posters on this blog are generally dismissive of models/forecasts we've just had a whole page of model forecasts posted. Are some models/forecasts more reliable than others? Or is it just that there are some we happen to like more? Re: J. D. McLean, C. R. de Freitas, R. M. Carter There is a correlation between SOI and average global temp but you wouldn't know it from the graph. You must not be able to read graphs then. There clearly is a relationship and the relationship even makes physical sense. When the Pacific goes into its ENSO program from wind shifts brought about by shifting highs and lows the SOI records those shifting pressure zones. The lag comes about from the time required for the winds to dig into the ocean and spread the water change across the entire surface at which time typically an ENSO event is noted. Then there is a secondary lag for the effects of changing surface temperature on global temperatures. You have acknowledged previously the secondary delay between ENSO and global temperature effects. Now all you have to do is acknowledge the delay between the shifting pressure fields and the resulting ENSO effect. The SOI is unique primarily because of long term monitoring stations but SOI like patterns that associate themselves with temperature could likely be realized at many locations in the ocean. It seems to me that the forcing that brings about shifts in pressure zones could be fairly weak and persistence is the major modifier in terms of long term results. A couple of degrees over long time scales seems to be a no-brainer possibility. Since the difference between surface temps and average ocean temperature is more than the difference of ice ages with proper feedback mechanisms like cloud and ice changes these changes could be prolonged and run very deep but our monitoring has only covered potentially one leg of such an oscillation. (the alleged LIA recovery) To think such a possibility doesn't exist is a great example of self centric type thinking that brought us 1500 years of the ptolemic system that everything is centered between our eyes. The ocean is a huge and vast source of change. There is probably enough cold water in the ocean to create an iceage, certainly a little one anyway. Something north of 95% of the water in the ocean is near freezing and relatively impervious to the normal seasonal changes we have seen in our lifetimes. Underwater ocean conveyor belt like currents run for thousands of miles, perhaps into the 10's of thousands. We see persistent high pressure zones on the map. Stuff so persistent only shifting seasonally we have no idea what the effects would be from a long distance shift in those zones. Its really interesting that just now we are seeing some very strange shifts of this nature and it seems to be associated to changes we haven't seen in the sun for a hundred years. I'm stocking up on popcorn and getting a new easy chair to watch the show with. You betcha GLC the SOI is real and it has a climate like impact, though the oscillations we have seen have been short lived perhaps bringing us the warming of the 1940's and some of the coolness of the early 20th century. The cold event we saw post 1940's was weak, maybe the positive SOI was weak also. I came into this topic a couple of years ago firmly believing in these climate events. It has been verified in both fishery histories and fossil records. It was also likely associated with the downfall of several pre-Columbian western civilizations. Its an area of study we know very little about. I see a number of scientists allude to this but without much knowledge of the ocean climate scientists are pretty much taking shots in the dark.
|
|
|
Post by glc on Aug 28, 2010 18:37:43 GMT
You must not be able to read graphs then.
A few points here
1. Are you sure you haven't gone to Page 1 of this thread by mistake. You are responding to a comment I made more than 2 weeks ago. 2. I agreed there was a correlation between SOI and global temperature (ENSO) - so I'm not sure what you're arguing about anyway. 3. I was merely remarking that the graph was not a particularly good illustration of the SOI/temperature correlation. I still don't think it is.
|
|
|
Post by trbixler on Aug 31, 2010 12:50:06 GMT
|
|
|
Post by nautonnier on Aug 31, 2010 14:45:07 GMT
Interesting comparison can be made... the most pessimistic of the models (above in the thread) said: and this actual observation said: Looks like a steeper drop than even the pessimistic model expected. Its a good validation. Does NCDC keep an archive of the models that you can run to see how they under/overestimate changes?
|
|
|
Post by yohoho on Sept 1, 2010 0:19:31 GMT
The NCEP El Nino weekly update (8/30), had Nino 3.4 at -1.5.
|
|
|
Post by yohoho on Sept 3, 2010 8:22:58 GMT
Here is the progression of monthly model forecasts for Nino 3.4 ssts: E1, E2, and the most recent (from Thurday) E3: E1: E2: E3: Notice that the most recent model (E3) has La Nina bottoming out deepest, at -2.5 from about Nov-Mar. I'd be interested to see how this will stack up against other La Ninas:
|
|
|
Post by glc on Sept 3, 2010 9:57:32 GMT
Not so long ago on this blog, regular monthly threads would be started which would discuss the recent month's global anomalies. About a year ago this stopped (I can't think why). However, as someone who accepts the existence of climate cycles, I believe they may make a return shortly.
My prediction is that the next monthly temperature thread will make an appearance in Jan/Feb 2011.
|
|
|
Post by nautonnier on Sept 3, 2010 10:00:34 GMT
Not so long ago on this blog, regular monthly threads would be started which would discuss the recent month's global anomalies. About a year ago this stopped (I can't think why). However, as someone who accepts the existence of climate cycles, I believe they may make a return shortly. My prediction is that the next monthly temperature thread will make an appearance in Jan/Feb 2011. I think you will find the reason is that the poster who was most often posting the information is no longer an active here. However, you will probably draw the same level of conclusion from that as well.
|
|
|
Post by magellan on Sept 3, 2010 14:25:53 GMT
Not so long ago on this blog, regular monthly threads would be started which would discuss the recent month's global anomalies. About a year ago this stopped (I can't think why). However, as someone who accepts the existence of climate cycles, I believe they may make a return shortly. My prediction is that the next monthly temperature thread will make an appearance in Jan/Feb 2011. Unlike others who would never use El Nino to establish a positive trend when previously did not exist. Nah, you'd never do that.....
|
|
|
Post by atra on Sept 3, 2010 19:09:52 GMT
|
|
|
Post by woodstove on Sept 3, 2010 23:56:15 GMT
high but not unheard of ... some examples of higher daily readings: 1998 342 (day of year) 1014.45 1000.90 51.02 2000 334 (day of year) 1014.33 1003.65 49.61 2000 335 (day of year) 1014.14 1003.85 47.14
|
|