Tropical Tidbits from the Tundra

Yasi Threatens Queensland; Northern Hemisphere pattern changing?

By: Levi32, 6:04 PM GMT on January 31, 2011

If you can, playing the video in HD makes it much easier to see things. The video will play in low quality by default. If HD quality isn't available, then it will be in a few minutes. Let me know if you have problems or questions about the video.

200mb Vertical Velocity (green areas represent upward motion associated with the MJO):

A glancing comparison of this winter's Atlantic SST with previous first-year La Ninas

By: Levi32, 5:27 PM GMT on January 27, 2011

If you can, playing the video in HD makes it much easier to see things. The video will play in low quality by default. If HD quality isn't available, then it will be in a few minutes. Let me know if you have problems or questions about the video.

200mb Vertical Velocity (green areas represent upward motion associated with the MJO):

The Problem with Measuring and Understanding Global Temperature

By: Levi32, 6:55 AM GMT on January 09, 2011

Temperature is one of the most important and most closely monitored aspects of meteorology. In recent decades it has come flying into the world-wide spotlight due to the theories about anthropogenic global warming. There are several reconstructions of global average temperature since the late 1800s, with the three main ones produced by the Goddard Institute for Space Studies (GISS), the Met Office Hadley Center (HadCRUT), and the National Oceanic and Atmospheric Administration (NOAA). In this post I will use data from the GISS data set, as it is the most frequently used and cited in my experience.

Global temperature is almost always expressed as an anomaly from the mean of a base period (GISS currently uses 1951-1980) in degrees Celsius. This method seems straight-forward enough, and is easy to understand, but it turns out that measuring global temperature in this way causes a problem.

Something that most lay people do not understand is that temperature is a measure of energy. Specifically, the average temperature in degrees Kelvin of a parcel of air is directly proportional to the average kinetic energy of the air molecules inside the parcel. When measuring the average global temperature, we are measuring the average total amount of energy in the troposphere, using the temperature near the surface as an approximation.

The problem with measuring the average global temperature with just straight anomalies is that the temperature varies in different amounts around the globe. Have you ever noticed how those blocking highs over the arctic can put up weekly anomalies of over 20C? And yet anomalies of greater than 5C are very rare in the equatorial regions, even over land areas. A good example of this was the recent winter of 2009-2010. One can see the 3-month anomalies in the arctic ranging from -9C in the heart of Russia to +11C over northeast Canada. However, for the same time period, the equatorial regions failed to register any anomalies greater than 3C.

Figure 1. Average surface temperature anomalies during the period from December, 2010 to February, 2011 for the arctic (top) and the equatorial region (bottom). Image courtesy: NOAA Earth System Research Laboratory

This high difference in variation between the equatorial region and the poles can be verified by looking at any temperature data set. Using GISS data, the annual standard deviation for the equatorial regions (<|24| latitude) is 0.2C, but the annual standard deviation for the polar regions (>|64| latitude) is more than four times that at 0.9C.

Now here is where the key difference is, and where the understanding of the Earth's atmospheric system comes into play. Where you have more variance in energy (temperature measurements) relative to somewhere else, you must also have more variance in energy input and/or output relative to that somewhere else. In the case of the Earth system, the equatorial regions receive relatively constant energy input (solar radiation), and since net input is at a maximum on Earth at the equator, the planet's natural atmospheric circulation prevents many invasions of foreign air masses, and thus the outgoing (output) energy near the equator is also relatively constant. The result is relatively low variance in temperature over short time periods. In contrast, the polar regions receive very little sunlight, yet they are able to experience drastic increases and decreases in temperature over short time periods because energy can be easily added or subtracted from these areas by the Earth's natural atmospheric circulation system. The mid-high latitudes are constantly exchanging different air masses, and since the polar regions contain very little energy on average compared to the tropics, it requires less effort to warm them up.

This concept of energy exchange is the same reason why temperature varies more over land than over water. The land, having a lower specific heat (ability to retain heat), radiates a lot more energy (output) than the water, resulting in greater temperature variations.

The important ramification of differences in temperature variance between latitude bands when it comes to measuring average global temperature is that it "means more" to change the temperature in the tropics than in the arctic or antarctic. In other words, relative to the average variation in temperature, warming or cooling the equatorial region by 1 degree Celsius is much more significant than warming or cooling the polar regions by 1 degree Celsius. Some might argue that the average temperature change globally is what it is and it doesn't matter where the anomalies contributed geographically.

Ah, but it does. As mentioned earlier, and as we all know, the equatorial regions receive and absorb the greatest amount of sunlight on average per year. This is shown in figure 2.

Figure 2. Average annual solar flux: (a) Downwelling shortwave radiation (solar radiation striking the surface) and (b) Net shortwave radiation (solar radiation absorbed by the surface).

An important point has to be made here, and that is that the vast majority of the Earth's tropospheric energy is contained within 30 degrees latitude of the equator. The polar regions receive very little energy input from the sun, and yet experience drastic variations in temperature. This energy has to come from somewhere. With the maximum energy input in the Earth system occurring in the equatorial regions, the majority of the energy put into play within the planet's weather patterns comes from the tropics. Due to the natural atmospheric circulation system, this energy is constantly transfered away from the equator towards the poles. This means that the primary energy source region for the polar areas is the tropics. Most of the incoming energy into the arctic and antarctic ultimately comes from the equator. It is this area of maximum energy input, the equatorial region, that drives the whole system. Thus, changes in total energy in this region will have more significant consequences down the line than changes in areas that are already fed by the source region to begin with.

The significance of temperature changes in different areas of the globe must be considered when measuring average global temperature and analyzing the long-term trends. Some of this difference is already taken care of automatically by the fact that the polar latitude belts are much smaller in surface area than the equatorial belt, and thus are weighted less in the global average, but that doesn't guarantee an accurate representation. One way we can accomplish this is by standardizing the temperature anomalies by different latitude belts. This means expressing the anomalies as standard deviations from the long-term mean. Here I have plotted the GISS raw global temperature anomalies alongside the same temperature data set, but divided into different zonal latitude belts, using each belt's standard deviation to contribute to a global mean. I of course weighted each belt by its surface area. The resulting graph is shown below.

Figure 3. Raw GISS global data vs. mean zonal standard deviations. GISS zonal means are available publicly here.

Notice the differences between the two data sets, with the adjusted graph tending to be warmer than the raw data in the late 19th and early 20th centuries, and then cooler than the raw data in recent decades. This can be put more into perspective by graphing the difference between the two data sets.

Figure 4. Difference between GISS raw data and zonal mean standard deviations.

One can plainly see that the trend in the raw data has been to be warmer than the adjusted mean deviations as time has gone on, with GISS values for recent years peaking as high as 0.5 standard deviations above the zonally-weighted values. This is a significant difference. This trend makes sense because the greatest warming over the last 130 years in the GISS data set has been in the polar regions. This can be shown by graphing the difference between the polar anomalies and the anomalies everywhere else.

Figure 5. Difference between raw GISS polar anomalies and the rest of the globe.

The trend here has also clearly been up throughout the span of the data set. When more of the tropospheric energy globally gathers over the poles rather than at the equator, it loses significance in the global mean when the zonal mean deviation adjustments are applied. Again, some might say it doesn't matter where the anomalies are, but remember that just as it is very easy to transport excess energy to the polar regions, it is also very easy to get rid of it. That energy came from the source region to begin with, and once it's up at the pole, it's going to be lost from the system at a much faster rate than if it was still at the equator. Changes at the location of the greatest energy concentration are the most significant.

Now the adjusted data set I have plotted here certainly doesn't reverse the apparent global warming trend in the GISS record, but it does show that the GISS estimates of the global energy balance may be too high with the methods they currently use, sometimes by a significant amount, and they are increasing as time goes on. There are of course other issues with global temperature data sets which can compromise the integrity and trustworthiness of the data, but that is not the point here. The point is to show that the current methods of recording global average temperature don't necessarily reflect correctly the energy balance of the Earth, and that measurements of the average global temperature must be done in standard deviation units for different latitudinal regions in order to preserve the significance of energy changes in different areas of the globe. Hopefully more accurate methods will be used in the future.

We shall see what happens!

The views of the author are his/her own and do not necessarily represent the position of The Weather Company or its parent, IBM.

Tropical Tidbits from the Tundra

About Levi32

Masters student in tropical meteorology at FSU. Raised in Alaskan blizzards, but drawn toward tropical cyclones by their superior PGF.

Recommended Links