Poleward motion of storm tracks

A new study analysing cloud cover measurements has noted storm tracks and the cloudiness related to them migrating polewards during the last few decades. Additionally storm tracks seem to have become narrower. Observed changes in cloudiness have an amplifying effect on global warming.

In mid-latidudes of both hemispheres there is a region where large scale atmospheric wave activity transfers heat and moisture polewards. This causes a band of low-pressure systems, where cloudiness and rainfall is common. These bands are called storm tracks. Storm tracks are an important part of weather system and they also have an effect on climate by their radiative cooling. They also have an effect on large scale weather patterns (for example on North Atlantic Oscillation and on El Niño).

Based on the simulations of climate models, storm tracks should move towards Earth poles when climate warms. However, the expected change is so small that it is still difficult to observe it at this time. Despite of numerous studies it is not clear exactly why the storm tracks move polewards in models. Greenhouse gas warming is probably one factor but also aerosols might be involved.

A new study seeks to find if the shifting of storm tracks is already evident in the cloud cover measurements of last few decades. Study used ISCCP (International Satellite Cloud Climatology Project) measurements between 1983 and 2008. During that time global surface temperature has risen about 0.5 degrees C, mich might be big enough for the detection of storm track shifting. Study also used data from some other sources (ERBE, CERES and MODIS). Additionally situation was simulated with climate models.

ISCCP problems

There are known errors in ISCCP data, of which particularly the changes in satellite viewing angle makes the interpretation of the data problematic. It seems that these days every new study using ISCCP data is also a study on satellite viewing angles. Satellite viewing angle changes, when a satellite is added to the satellite network, satellite location is changed, or a satellite is removed from the network. A change in the viewing angle affects the amount of measured cloudiness. A proper solution to the problem does not exist yet, but this study makes an effort to go around the problem with couple of different methods.

Easiest method for going around the problem is to leave out the data from those regions which are affected the most. The weakness of this method is that the problem still exists in other regions at least to some extent. In this study the region of Indian Ocean was left out because it was between the borders of the measuring areas of two satellites and there were no supporting data available from other satellites before 1997.

The effects of satellite network changes can be tried to remove by regression analysis. This method was used on the global cloud data over the oceans. After the correction, the decreasing cloud cover trend in ISCCP data got smaller from 1.5 % to 0.4 % per decade. In the storm track areas the effect of the correction is smaller.

Satellite viewing angle problems are especially bad in thin clouds, because they can easily fall out of detection limit after viewing angle change. It might be possible to reduce the problem by concentrating on thick clouds only, but this method has its own problems, and the method doesn’t give truthful picture of total cloudiness.

Shifting of storm tracks

Studied storm track regions show clear cloud maxima which can have over 90 % average annual cloud cover. Observations also show a poleward shifting of storm tracks during the last 25 years. Shifting has been stronger in the equator side of storm tracks, which has caused a narrowing of the storm tracks. All studied regions show similar changes in storm tracks. Narrowing of storm tracks is especially strong in the North Atlantic, which may suggest that the region is strongle infected by the satellite viewing angle problem, which also was to be expected beforehand.

In ISCCP data, storm tracks have shifted 0.15 – 0.17 degrees of latitude per decade in other regions except in North Atlantic, where again is anomalously large change (0.6 degrees of latitude per decade).

With storm track shifting there is an associated decrease in cloudiness in storm track areas. Observations show couple of percent decrease in cloudiness in storm track areas (and again North Atlantic has clearly bigger change). Most of the cloud cover decrease occurs in low clouds. Middle clouds and chigh clouds have even increased a little.

The climatic effect of storm track shifting

The changes in longwave and shortwave radiation (measured in ERBE and CERES projects) correlate with cloudiness sufficiently well, that the effect of clouds to radiative fluxes can be estimated from ISCCP data. In the case of shortwave radiation, clouds change the radiative flux by reflecting the radiation from the Sun back to space. The effect to shortwave radiation can be seen in the data especially in the storm track areas. Clouds affect the radiative flux of the longwave radiation through their greenhouse effect, which also is seen in the data.

Total cloudiness has decreased 2-3% in storm track areas. This has caused the reflected solar radiation to decrease, which means that in storm track areas more solar radiation reaches the Earth’s surface. This change causes a warming effect (of about 3 watts per square meter). Middle and high clouds have increased by 2-5%. This reduces the outgoing longwave radiation, which means that bigger part of longwave radiation stays on Earth and causes warming (of about 1 – 3 watts per square meter).

The changes in cloudiness cause a warming effect both by longwave and shortwave fluxes (total effect is about 3 – 5 watts per square meter). Therefore it seems that cloud changes in storm track areas have amplified global warming (this is so called positive feedback). However, this result doesn’t necessarily reflect the global situation, because possible compensating effects in other areas were not studied. Additionally, ISCCP data problems may contribute to the situation considerably. For example, without thin clouds even a small negative feedback is possible (total effect in this case is -1 – 2 watts per square meter, so even in this case the most probable value of feedback is positive, even if uncertainty bars allow the possibility of a small negative feedback).

20 climate models were investigated and seven of them showed poleward storm track shifting with global warming. However, the amount of shifting was considerably smaller than what ISCCP data shows. The models, where storm tracks shifted the most, meaning that they were closest to observations, climate sensitivity was greatest. Models with climate sensitivity less than 3 degrees Celsius don’t even show statistically significant storm track shifting in all storm track areas. This suggests that also model results support positive cloud feedback associated with storm tracks.

Reference: Frida A-M. Bender, V. Ramanathan and George Tselioudis, Changes in extratropical storm track cloudiness 1983–2008: observational support for a poleward shift, Climate Dynamics, DOI: 10.1007/s00382-011-1065-6. [abstract]

Additional information:

Viewing Angle on ISCCP Problems
Papers on Hadley Cell expansion