Emanuel 2005 stated:
The accumulated annual duration of storms in the North Atlantic and western North Pacific has indeed increased by roughly 60% since 1949, though this may partially reflect changes in reporting practices, as discussed in Methods.
Speaking of “rudimentary statistics”, this seems like a rudimentary statistical statement. But there are wheels within wheels.
From the collation of unadjusted Best Tracks data, I calculated the number of quarter-days that each storm had reported wind speeds greater than 18 m/sec (a figure mentioned in Landsea’s Comment) , divided by four and calculated the sum – which I interpreted to be the “accumulated annual duration”, doing this for both North Atlantic and West Pacific. (This calculation is using pre-adjusted data.)
Figure 1. Accumulated annual duration. Left – Atlantic; center – W Pacific; right- total. Red – 2005-2006.
I’ve tried pretty hard to figure out how one can get an increase of “roughly 60% since 1949”. First, what is an operational definition of what periods are involved? As an exercise, I fitted a simple linear trend to the 1949-2004 periods for Atlantic, W Pacific and Total as shown below. Obviously there is not a “roughly 60%” increase using fitted values from a linear trend – which would seem to be the most logical means of estimating an increase. The closest that I could get to a “roughly 60% value was using the ratio of the segments shown in blue in the right graph – the 2000-2004 mean over the 1946-1950 mean. But even that is only 47%. Maybe the adjusted data is a little different.
Whatever the answer to this little conundrum, it illustrates the differences between audits and peer review. Obviously a Nature peer reviewer would not dirty his fingernails checking whether Emanuel’s “roughly 60%” figure was right. I don’t even think that they noticed how indeterminate this statement was – even allowing for “roughly”.
Script is here. The script uses R-tables; you’ll need to modify to use the *.txt files that I archived if I don’t get to that.
Climate Audit 
23 Comments
It looks like cherry picking your start and end dates to get the best (read: most dramatic) difference.
Re dates, 1949 is the date that Emanuel started with. Note this is essentially the peak of the AMO, so for the North Atlantic anyways, this is the least favorable one to start with if you are looking for an upward trend.
Re: #2
Data snooping would suggest starting in the 1970. Haven’t I seen references to changes/trends from the 1970s to present and justified by statements that the data were more reliable during that period?
But before one worries about things like AMO, where does the 60% come from? Does anyone at Georgia Tech know?
We don’t have any particular insights into what Emanuel did, and we have not tried to replicate his results We have stayed away from looking at any data that includes hurricane intensity prior to 1970 owing to problems with the data. The data since 1970 are more reliable than the earlier data, because of the presence of satellite data.
With the vast number of ships at sea during wwII and the recovery from same, and especially in the WP and NA, I would have thought observations would have been quite good and well dispersed. Was there some great leap forward in accuracy around 1949 that justifies not also using prior years?
I asked Kerry about this maybe a year ago, and i recall that he picked 1949 since that was the year that the NWP reanalysis products started (but it does not seem that he uses the reanalysis products in his paper). Routine aircraft reconnaissance flights of the TCs started in 1944, and prior to 1944 maybe the number of storms is ok but there are much greater uncertainties in intensity and the number of TC days. I suspect that the data from the first year or two of the aircraft reconnaissance may not be as good as the later data. So basically the issue is whether he should have started in 1944 or 1949. I would have started in 1944.
Judy, whatever you choose to publish, you ought to do a sensativityy analysis based on changing the data source. This whole feild is looking crappier and crappier and the practioners dumber and dumber. No offense…
I’m trying to visualize the density of shipping in the 40’s. Considering the slow speeds of most shipping back then, I would think even ships trying to avoid TCs would have a difficult time staying out of the way, and many more observations would be made by them than by the infant stages of the recon aircraft observations. Then again, I guess it could just be one of those choose your poison decisions. I really don’t know much about it.
I’ll bow to your expertise on the subject. Thanks for your insight.
Bob, I don’t have any particular great expertise on the subject, I am mainly citing Chris Landsea’s opinion on the quality of the NATL data; he says it is reliable since 1944 (with the caveat regarding the uncertainty in intensities prior to 1970).
steve, i am losing track with all the different emanuel posts. did you use the landsea correction or not in calculating the PDI? without the landsea correction, the trend is pretty small.
The whole selective use of data is very suspicious, Judy. When it helps you, you don’t use data. THen you don’t have enough to have a meaningful trend. But you publish anyway. In Nature. Sheesh.
Even today there is not a lot of ship traffic between Africa and the Caribbean. Father north, between Europe and North America, the traffic density is much larger.
My suspicion is that, pre WWII, some weaker storms in low latitudes between Africa and the Caribbean were missed. No satellites, no planes, few ships.
A ship that found itself in rising seas and winds had “sailors rules” for moving away from the storm. Few ships actually sampled the important center of a hurricane.
Old data is subjective and best-guesses.
One final point on the use of old storm counts: Emanuel and Mann used Atlantic storm count data from the 1880s and 1890s in their joint 2006 paper. The nineteenth century was full of detection and reporting problems, as mentioned, especially the storms that never touched land.
That dubious data was used in a graph that also included storm count from the late-20’th and 21’st century satellite era.
Apples and oranges in the same graph.
TCO is really on target here! All these pronouncements to the media and all the hype, and it turns out that the data sources and studies are being questioned and debated by the very scientists that are using those data bases and studies and yet they are still hyping their pet theory. Good grief! No wonder there are so many skeptics!
I was curious about aircraft range in the early days of hurricane hunting and I found the cruising range to be in the neighborhood of 2200+ miles. So coverage wasn’t total.
I also came across this interesting nugget. I have no idea how old the item is.
Here’s the link.
Check out the photo 1/2 way down the page. How accurate can recon have actually been?
Here’s a link to a description of the Beaufort Wind Scale.
If this is truly the way they were measuring wind speed, I have serious doubts as to the air recon wind records being useful for any accurate statistical evaluation of that period. I think ship observations where an actual instrument was doing the measurement would be more accurate.(whenever available) I think we’re going to have to wait until sufficient satellite records are accumulated.
I am assuming that sea-level wind speed is still used to determine hurricane strength.
Argghhh. Screwed up the links. Try again.
Wiki on Beaufort. http://en.wikipedia.org/wiki/Beaufort_scale
Photo and quote. http://www.hurricanehunters.com/navy.htm
Re #16 Bob, I think what they usually did was to fly into the center of a hurricane and then use a dropsonde to measure the center pressure. Then they use the pressure to estimate the winds. The sea surface condition was used as a secondary check.
Like the name implies, a dropsonde is dropped out of the plane and it falls to the ocean. It has a barometer and a radio broadcaster and tells the airplace what pressure it senses at the sea surface of the eye.
The plane drops another one at a distance from the eye, to get a storm-edge pressure. Then, the meteorologist plugs the two pressures, and the distance between them, into an equation which gives the windspeed estimate.
That approach works reasonably well, but the accuracy is not great. In fact, today it is not uncommon for the dropsonde pressures to suggest one windspeed while the aircraft flight level observations suggest a different windspeed and the satellite appearance suggests a third. The hurricane forecasters are quite up-front about this problem and often discuss it in the “discussion” section of their forecasts. The forecasters use judgment as to which windspeed they choose to believe and make official.
I believe that modern aircraft are able to measure windspeed at flight level (with help from GPS, I believe) and then use an equation to extrapolate that to a sea-surface speed. Different flight levels use different coefficients in the equations, and the coefficients were developed and improved based on observations (in other words, trial and error over the years).
My suspicion is that some of the current struggle with historical data has to do with the equations and coefficients.
There is room for error in many parts of this whole process.
Bob K, thanks for your post. You say:
As a long-time sailor and seaman, I think you are overestimating the instruments, and underestimating Beaufort. Why?
1) A ship’s very life depends on the crew’s ability to accurately judge wind speed. The Beaufort Scale is in fact very accurate if used by an experienced seaman … which we used to have … especially in heavy weather situations.
2) Most ships did not have anemometers until recently, and if the ship is rolling, or is heeling over, anemometers are inaccurate, probably no better than the Beaufort scale.
3) In any kind of seaway, there is much less wind in the trough than at the crests. While this affects anemometers, it does not affect the Beaufort Scale.
w.
There’s a beautiful book about the Beaufort scale:
Scott Huler, Defining the Wind : The Beaufort Scale, and How a 19th-Century Admiral Turned Science into Poetry, Crown, 2004, ISBN 1-4000-4884-2
The scale sought precision using definable objective criteria, and was a great success. The poetic quality of the decriptions can be seen in the wiki entry.
Large branches in motion. Whistling heard in overhead wires. Umbrella use becomes difficult.
Thanks for the clarification fellas.
I figured there would be some difference in Beaufort Scale assessments due to the perspective of the observer. i.e. at sea-level vs. 500-1000ft above. Still seems awfully subjective.
The exchange of ideas that happens here is great to see. I mostly lurk in the background, just to give this old brain some interesting exercise.:)
Bob K,
Your point is a good one. It would be difficult to assign Beaufort numbers from an aeroplane by visual observation. An estimate of windspeed would be trivial with GPS technology, but would have required a lot of arithmetic back at base decades ago.
RE: #19 – To wit: “Air filled with foam and spray. Sea completely white with driving spray. Visibility very greatly reduced.”
I would add to this, “Untethered on deck? You za gonner!”