As noted, column ozone measurements are most frequently reported in the media. Interest in these measurements is generated by their relevance to biological life forms at Earth's surface, since as we've mentioned before, ozone absorbs harmful UV radiation. The impact of UV radiation on life is discussed further in Chapters 4 and 8.

Figure 3.09 shows a global map of total column ozone for the same day as the SBUV maps in Figures 3.06, 3.07, and 3.08.

This figure shows total column ozone at the end of the northern summer/southern winter. We see low values of total ozone in the tropics, a small region of low total ozone in the North Polar region, and a large region of very low ozone over Antarctica. This latter low ozone region is known as the Antarctic ozone hole (see Chapter 10). Higher amounts of ozone fill the midlatitudes in both hemispheres, but the collar region in the southern hemisphere is especially prominent.

The collar region exists owing to complicated motions (dynamics) in the atmosphere. Why these motions exist is the subject of Chapters 2 and 6. Suffice it to say that these motions exist. The first is a wintertime equator-to-pole circulation of ozone known as the Brewer-Dobson circulation. It is described in detail in Chapter 5. The second is the wintertime west-to-east circulation around the poles, called the polar night jet. It is described in more detail in Chapter 2 Section 4.1.3. The polar night jet sets up along the day-night terminator line, inside of which occurs 6 months of polar darkness. The region inside is called the polar vortex, and it gets very cold there. The more isolated the vortex is, meaning the less outside influence from warmer air in the middle latitudes, the colder it gets. This also means that less ozone can get inside the vortex.

The Brewer-Dobson circulation transports ozone out of its tropical source region in the mid- to upper stratosphere and into the polar regions in the lower stratosphere. In the southern hemisphere, the polar night jet is stronger. Winds blow more west to east (zonal) and there is less north-south mixing of air than in the northern hemisphere. This is why it gets so very cold in Antarctica. The Brewer-Dobson circulation, rather than extending all the way to the south pole, stops at the edge of the polar vortex. Ozone accumulates along the edge of it, but not inside; hence, the collar region of higher ozone surrounding lower ozone air. In the northern hemisphere the polar night jet is weaker, and there are incursions of warmer air from lower latitudes. The Brewer-Dobson circulation cell extends all the way to the North Pole; hence, no collar region exists.

When temperatures get sufficiently cold, the sorts of chemical reactions described briefly in Chapter 1 and in more detail in Chapters 5 and 11 occur. The chlorine that exists in the stratosphere because of manmade chlorofluorocarbon compounds undergoes the so-called heterogeneous chemical reactions that bring about ozone loss at certain times of the year.

Comparing Figures 3.06 through 3.09, how do they differ? Figure 3.09 is a map of the total column amount in Dobson Units (see Section 3) while Figures 3.06, 3.07, and 3.08 show the amount of ozone in ppmv at different altitudes. Many differences appear between these figures. The filament of low ozone visible in Figure 3.07 is not apparent in Figure 3.08. Furthermore, the tropics are low in ozone in Figure 3.06 (at 50 mb) , but are high in ozone in Figure 3.07 (at 15 mb), and low in total column ozone shown in Figure 3.09. Meanwhile, Figure 3.08 (at 0.5 mb) contains almost no features in the ozone distribution whatsoever. We conclude from this that looking at ozone distribution at a single altitude does not tell us what the total ozone distribution really looks like.

7.1 Average Total Column Ozone Field

Figure 3.10 shows a global map of total column ozone as compiled from an average of data from the Total Ozone Mapping Spectrometer (TOMS) instrument aboard the Nimbus-7 satellite gathered between November 1978 and May 1993. (The TOMS instrument is explained in Chapter 7.)

We see that the tropics (roughly the region between 30° N and 30° S) are relatively low in ozone, with total column amounts between 250 and 270 Dobson Units (DU). If we recall the shape of the number density profile curve in Figure 3.04, it makes sense that total column amounts in the tropics are relatively low. As we move northward away from the equator, typical midlatitude values fall between 300 and 350 DU. At the highest northern latitudes we see the highest values, almost 400 DU.

The general pattern revealed by Figure 3.10 of lower-to-higher column ozone from equator to poles is attributable to a broad atmospheric circulation known as the Brewer-Dobson circulation. It is explained in detail in Chapter 6. In brief, rising motion in the tropics elevates the height of the peak ozone layer while poleward transport carries the newly created molecules away from the tropics. These processes combine to elevate the peak and to reduce the overall magnitude of tropical ozone profiles. At high latitudes, sinking motion lowers the height of the ozone peak, while transport brings in air rich in ozone leading to an increase in the magnitude of ozone.

The southern hemisphere, however, appears quite different. While the southern midlatitudes again show ozone column amounts in the range of 300-350 DU, the southern polar region shows values less than 300 DU- a marked difference from its northern polar region counterpart. This difference is what is known as the ozone hole over Antarctica during the southern or Austral spring. No corresponding ozone hole exists over the high Arctic (North Pole) during the northern spring.

7.2 Monthly Means

We've seen that ozone varies as a function of altitude and latitude. It is no surprise then that ozone also varies as a function of season. Again, the creation and destruction of ozone depends on the amount of sunlight (i.e. UV radiation) available. Sunlight varies throughout the year, most noticeably at the poles (which exist for about 6 months in darkness and 6 months in light).

Figure 3.11 shows the monthly mean column ozone for each month of the year as derived from the TOMS Nimbus-7 record for 1978-1993. The gray regions represent areas in which TOMS was unable to make observations due to the lack of sunlight (i.e. regions in polar night).

While the gross features are quite similar to those seen in Figure 3.10, we notice quite a bit of variability from that data. If we were to average all 12 plots in Figure 3.11, however, the result would appear very much like Figure 3.10.

Examining the images, what do we find?

To being with, in the northern winter we see very large ozone column amounts (in excess of 400 DU) at high northern latitudes. The tropics are simultaneously at their lowest at the same time with less than 250 DU. The midlatitudes in both hemispheres range from 300 to 350 DU on average. Notice that in the northern hemisphere midlatitudes, no region of unusually low ozone appears for any monthly average from 1978 to 1993 average.

Moving into the high northern summer latitudes, we find ozone values falling off from winter time values of >400 DU to the 300-350 DU range by August. Meanwhile, tropical ozone values have increased to 250-270 DU.

Turning our attention to the southern hemisphere in September and October (the southern spring), we see regions of ozone around 400 DU at midlatitudes that surround a region of very low ozone, under 200 DU, over Antarctica. This average value is derived for the 14 available Septembers and Octobers. By far, the ozone values observed in the Antarctic are the lowest values observed on the planet in any given year. This is a direct result of the Antarctic ozone hole.

By December, the ozone hole has disappeared and ozone values at high southern latitudes have recovered to 300-350 DU. Ozone values begin falling again over the tropics while increasing at high northern latitudes.

7.3 Mean Annual Cycle

An easier way to see the seasonal changes in the total ozone field is to take a zonally averaged plot of ozone versus time. That is, we average in both space and in time the 12 plots in Figure 3.11. We begin by drawing lines on a globe parallel to the lines of latitude on Earth. Our lines are spaced an equal number of degrees of latitude apart. Such lines are typically spaced anywhere from one to five degrees of latitude apart. Within any two lines we have our "latitude band." We do the spatial averaging by averaging together all the TOMS measurements falling between each pair of lines (i.e. between those latitude bands). We do this for each month. The result is a unique column ozone value for each latitude band (from 90° N to 90° S) for each available month in the period. Next, we average in time by taking a monthly average. That is, we average all the available Januarys together, all the available Februarys, etc. We now have a single "zonal mean" value of ozone corresponding to each latitude for each month of the year in the period, which in this case is 1978-1993 (refer to Figure 3.10). Such a plot is shown in Figure 3.12.

This figure shows the average amount of column ozone at a given latitude band (the y-axis) as a function of time (the x-axis). It is an average of all the monthly mean TOMS data taken between November 1978 and May 1993 (the period in which the TOMS Nimbus-7 instrument was operational). The gray patches in the corners again represent regions in which no data were available due to the polar night (see Chapter 7).

Examining Figure 3.12, we can identify the same latitudinal differences and the seasonal variations much more readily than in Figure 3.11. Among the findings:

The TOMS data clearly indicate an annual natural cycle to ozone. Ozone variations of up to 40 percent occur naturally over the course of a year due solely to differing amounts of sunlight. The annual cycle, however, cannot explain the extremely low values of ozone found over Antarctica in recent years. To generate such low values of ozone, chemical destruction of ozone is required.

Other cycles can be seen in the TOMS data as well. In addition to the annual cycle, ozone displays a semiannual oscillation and a quasi-biennial oscillation (QBO), named for the periods associated with the variability. A more detailed discussion of these variations can be found in Chapters 2, 6, and 8.

7.4 Total Ozone Trends

Hidden in the averaging to produce Figures 3.11 and 3.12 are the long-term changes in ozone that have been occurring with time, particularly in the southern hemisphere winter. We noted that the lowest values in the graph of the mean annual cycle from the TOMS data were just under 200 DU in September and October at high southern latitudes. Yet we also know that recent observations of the ozone hole have indicated values half that, around 100 DU.

Why the discrepancy?

Well, let's look more carefully at the various Octobers (sixteen in all) that we included in our averaging to produce Figure 3.12. Figure 3.13 shows monthly mean TOMS data for all the Octobers between 1979 and 1994. A TOMS instrument measured all sixteen Octobers, with the first fourteen (1979-1992) produced by the Nimbus-7 TOMS and the last two, 1993 and 1994, produced by the METEOR-3 TOMS instrument.

The first plot of October 1979 data reveals no ozone hole. Values in the polar regions in 1979 remained around 300 DU during the Austral spring season. However, as we move through the '80s, ozone in the south polar region steadily decreases, i.e. the hole opens up. By 1993, values as low as 110 DU are recorded. Figure 3.13 indicates that the amount of ozone in the southern polar region has been changing quite dramatically since TOMS starting making observations. When we average all of these plots together, the result is the October portion of Figure 3.12. The steady decrease in October monthly ozone over Antarctica is lost in the time-averaging of Figure 3.12.

7.5 Historical Measurements -- So How Did Ozone Look Before TOMS?

The TOMS instrument has only been measuring ozone since 1979. It could be the case that the sharp drop in Antarctic ozone in the Austral spring that shows up in its record is due to some cyclical variation with a frequency of decades or centuries. The TOMS data set does not exist far back enough to capture this. In addition to the manmade chemistry that has been shown to destroy stratospheric ozone over Antarctica (see Chapters 5 and 10), there could be a longer-term cycle at work. So the question becomes: how did ozone look before TOMS?

7.5.1 The Swiss Ozone Record Since 1926 -- To answer that question, we need to examine ozone data from one of the pioneers in ozone studies. G.M.B. Dobson (for whom the unit of column ozone is named) had been observing ozone since the 1920s with instruments of his own design. He compiled a long record of data from the northern hemisphere and knew well how ozone behaved over Europe.

Figure 3.14 shows the data from one of Dobson's instruments deployed in Arosa, Switzerland. This data set represents the longest existing continuous record of ozone measurements in the world.

Each data point in Figure 3.14 is an average of all the measurements made at Arosa that year. The record shows significant variation from year to year, but clearly indicates a downward trend over the last 20 years or so. The downward trend here has nothing to do with the ozone hole and will be touched upon in later Chapters.

7.5.2 Antarctic Ozone Measurements for 1957-65 -- In 1957 one of Dobson's instruments was deployed in Antarctica as part of the International Geophysical Year (IGY). IGY was a coordinated effort among a variety of Earth scientists to make measurements relevant to the Earth system in as many places as possible at the same time during 1957 and 1958. By the IGY, Dobson's instruments had been set up at a variety of locations around the world. Dobson set up ozone measuring instruments at Halley Bay in Antarctica on the coast of the Weddell Sea and also at Resolute Bay, Canada on the Arctic Ocean. Dobson continued taking measurements until 1965.

In 1966, Dobson published a report on the data he had gathered. To his surprise, he observed that ozone behaved quite differently over Antarctica than over high northern (Arctic) latitudes. Figure 3.15 shows Dobson's data measured at his two ground stations at Earth's opposite poles. One is shifted relative to the other by 6 months to account for the hemispheric seasonal phase differences in ozone.

This figure demonstrates that during the southern winter, high latitude ozone values over Antarctica are much lower than in the corresponding period in the northern hemisphere over the Arctic. A clear difference in the behavior of ozone between the two hemispheres is thus evident from his data. Beyond this hemispheric difference (the meteorological causes for which are explored in Chapter 10), the figure indicates that the low levels of ozone Dobson measured over the Antarctic in 1957 and 1958 were still significantly higher than those found today over the same regions. The amount of ozone over the Antarctic in late southern hemisphere winter and early southern hemisphere spring has decreased dramatically over the latter half of the twentieth century. We explore the causes of these changes in Chapter 10.

7.6 Global Mean Changes

Long-term changes in ozone can be understood by examining the global mean ozone as a function of time. Figure 3.16 shows an average of all the TOMS data taken between 60° N and 60° S as a function of time.

Clearly visible in the record is an annual cycle, with the highest values in the northern spring and the lowest values in the northern winter. There also appears to be an overall downward trend to the oscillations, indicating a decreasing amount of total ozone. The data record, however, is far from periodic. We can see that there are year-to-year variations in the magnitude of ozone, and some higher order (shorter term) variability in the data as well. Analyzing such "signal" in a data record and determining trends can be complicated (see Chapter 8).

If we now look at similar average records, but over more limited latitude ranges, the picture becomes somewhat different. Figure 3.17 shows the global mean (60° N to 60° S) in black, an average between 30° N and 50° N in blue, and an average between 20° S and 20° N in red.

Note that the amplitudes of both these curves are greater than that of the global signal. The downward trend evident in the global average is less obvious in the 20N to 20S (tropical) range. Also notice that the variability of the red (tropical) data is much less than that of the blue (northern midlatitude) data. Can you explain why this might be the case? We'll tackle this issue in Chapter 8.

7.7 Day to Day Variability

We now have a good sense of how ozone varies vertically, seasonally, and with latitude. We might also be interested in how ozone varies from day to day over some particularly place on the globe, such as over the city we live in.

Figure 3.18 shows such a record of ozone measurements over Fresno, California, for 1992.

In green are data from the TOMS instrument aboard the Nimbus-7 satellite. In red are data measured by a ground-based instrument known as a Dobson Photospectrometer. The scales for the two measurements have been offset from one another to facilitate comparison. First of all, we note that the two measurements agree quite well with each other: TOMS and the ground-based observations seem to observe high ozone events simultaneously and low ozone events simultaneously. The two instruments also report similar magnitudes for column ozone of about 300 DU. Finally, they both report day-to-day variations of a similar magnitude. The largest excursions seem to be about 25 percent from one day to the next, but more typically both instruments report a day-to-day variation of under 5 percent. The importance of the comparisons between the ground based Dobson Photospectrometer and the TOMS measurements is that we have a ground based instrument capable of verifying the accuracy of the satellite instrument.

Figure 3.19 shows the frequency distribution of ozone variability from day-to-day in 1992 (over the same period as the data in Figure 3.18).

Notice that the distribution is sharply peaked around 0 percent change (meaning that on most days, column ozone variability was very small). The ends or "tails" of the frequency distribution curve extend to about plus or minus 25 percent change. Such events were rare in Fresno in 1992, located at 36.6 °N, well into the northern hemisphere midlatitudes. This is in accordance with what we learned earlier about northern midlatitude ozone variability during the course of a year. While ozone can vary by as much as 25 percent from one day to the next at these latitudes, the occurrence of such dramatic changes is rare. Figure 3.18 also suggests that variability is greater during the winter months than during the summer months. The difference in day-to-day changes in ozone can be attributed to a more active weather pattern in winter than in summer. This will be discussed further in Chapter 6.