Weatherinfo

I have a mystery.

The time series above compares the monthly RSS lower troposphere (”satellite”) temperature anomaly with the NCDC surface (”surface”) temperature anomaly since January 2000. What’s interesting to me, and mysterious, is how little difference there typically is between successive values. I expected to see large month-to-month swings but the chart shows month-to-month values which are typically close to one another. 

Perhaps this is simply noise or perhaps there is a physical basis for the satellite and surface anomalies to sometimes drift toghether and then drift apart. I don’t know, I have no conjecture. This is simply a mystery to me.

Anyone have thoughts or suggestions on this, please post. Perhaps there’s some way to statistically characterize the data to see if it’s noise or pattern. Perhaps there is a plausible physical basis.

There is a question as to why the reported tropical cyclone count in the Atlantic has increased over the last 140 years. Some say it is due to the reported higher sea temperatures. I suggest a simpler explanation. My belief is that the increase is largely due to improvements in detection and record-keeping.

Here’s a chart that shows cyclones categorized by probable detection method:

 

The time series breaks the storm record into three groups:

Group one (mauve line) are those storms which affected (hit or passed within 100 miles of) the US or Canada. These storms were close enough to be detected and were in regions of the world where record-keeping was decent even 140 years ago.

Group two (yellow) are those storms which affected (hit or passed within 100 miles of) Spanish-speaking regions of North America, including the Antilles and The Bahamas. Here, too, storms within 100 miles would likely be detected but record-keeping (with the possible exception of Cuba) was spotty before the 1880s.

Group three (blue) are the remaining storms, which by definition never affected (came within 100 miles of ) land. These would, with few exceptions, have been detected by ships.

The chart records the number of storms in each group, by calendar decade (1865-1974 to 1995-2004).

What it shows is

1.) the frequency of storms detectable from the US or Canada, places with pretty good detection and record-keeping, shows no upward trend over the last 140 years. (One can, though, see an apparent Atlantic Multidecadal Oscillation pattern.)

2. the frequency of storms detectable from Latin America or tropical islands shows oscillation and a slight upward trend, but I believe that is mainly due to under-reporting in this region (excluding Cuba) prior to about 1885. (There is evidence to support this conjecture.)

3. The frequency of entirely-at-sea storms shows little change until the 1950s, at which time the frequency increases even though tropical sea surface temperatures were little changed (versus earlier periods) until the 1990s. I think this is due to the use of aircraft and especially satellites by the eraly 1970s, and to increasing emphasis on hurricanes evidenced the formation of a full time US hurricane bureau after 1945.

In summary, the increase over the last 140 years has been in the entirely-at-sea storms, where detection methods improved after WW2. On the other hand, storms passing over or near land, where detection and record-keeping have been good for many decades, show little change.

As a second exercise, I looked at a particular type of storm. This storm is very weak and existed as a tropical storm for less than 24 hours. Such a short life is generally insufficient for the storm wind field to spread, meaning that the storm-force winds were confined to a small area for a short time. My conjecture is that these almost always require special modern means (modern satellites, Doppler radar, buoy network) to detect their time as tropical storms.  Prior to these methods these storms were likely just missed.

Whaty does the record show? When I look at the Unisys data, I see the following:

Weak, short-lived storms:

Prior to 1945-54: 0

1955-64: 1

1965-74: 2

1975-84: 4

1985-94: 2

1995-2004: 7

As thought, these very weak storms are clustered in modern decades, where detection is much-improved. Also, they tend to be very close to land and are almost all in the western half of the Atlantic basin.

A plot is given below, with the weak, special-detection storms removed:

As is widely known, the thermohaline circulation (THC) plays a significant role in the global distribution of heat. Warm tropical Atlantic water flows northward to the Arctic subpolar regions, where the water cools and sinks. It then travels southward along the ocean floors, ultimately upwelling in various parts of the Southern Hemisphere.

A good summary paper (Latif et al 2005), combining observational studies as well as model simulations, is located here:

http://luv.dkrz.de/publications_2005/pub_291_329.pdf

 Several points from the paper:

* an indicator of the THC strength (flow rate) is the difference in temperature between the subpolar North Atlantic and the subpolar South Atlantic. A large temperature difference between the two indicates a strong THC while a small temperature difference indicates a relatively weak THC.

* using this indicator, the strength of the THC in the 20′th Century can be estimated.  The black line in the chart below indicates the relative THC strength:

 (Click to expand)

*  there are two primary downwelling (sinking) regions in the subpolar Arctic: the Nordic Seas and the Labrador Sea.

* changes in the Nordic Sea downwelling appear to be sensitive to changes in density (= freshwater intrusion)

* changes in the Labrador Sea appear to be sensitive to changes in the North Atlantic Oscillation (NAO), an atmospheric phenomena. The NAO affects the Labrador cooling rate (and thus the sinking rate) via the amount of cold, windy air flowing across that Sea.

* The impact of the NAO on Labrador downwelling, and thus that part of the THC, involves a lag of 5 to 20 years, based on a comparison of 20′th century NAO and THC. The relationship with a 5 to 20 year lag is stated to be statistically significant.

* Estimates are that the Nordic freshening has decreased the THC by perhaps 1 Sv since 1970 while the NAO impact on the Labrador Sea has been to increase the THC by perhaps 2 or 3 Sv in recent decades. (Total THC is about 15 Sv. A Sv is a million cubic meters of water per second.)

* The historical NAO is shown in the figure above, which indicates a swing to positive especially in the 1980s and early 1990s. The THC responded to that by strengthening beginning in the early to mid 1990s. My SST numbers (not shown) show that the THC strengthening has continued into 2007.

* The more-recent NAO is shown below:

As indicated, the NAO has weakened to a near-neutral state, so the expectation is that (should this NAO trend continue) the THC should peak in the next 5 years or so and then begin to weaken. This has temperature (cooling) implications for portions of the Northern Hemisphere and perhaps sea ice extent.

The next 10 to 15 years could be interesting if the globe sees a weakening THC, cool-phase PDO and the speculated solar activity decline all about the same time.

An interesting aside is that the dark line (SST difference between the North and South Atlantic) in the first thumbnail also roughly approximates Atlantic hurricane activity (the AMO) as well as approximating the THC strength. Bill Gray hypothesized a correlation between the two.

The purpose of this blog is to store various useful meteorological/climatological links and to offer commentary in selected topics.

Welcome to Auditblogs.com. This is your first post. Edit or delete it, then start blogging!