- The TOMS Instrument
- Determining Ozone
- Errors and Uncertainties
The best long term global dataset of atmospheric ozone was collected by the Nimbus-7 TOMS. Nimbus-7 was launched on
October 24, 1978, and carried a payload of experimental instruments to observe Earth's environment.
Among these were SBUV and TOMS, a pair of sensors that shared several components and were
designed to study ozone concentrations in Earth's atmosphere. TOMS became operational on October 31,
1978, and provided global ozone data almost every day until May 6, 1993.
TOMS operated by comparing incident solar radiation with radiation reflected from Earth's
atmosphere. Six UV wavelengths were observed, each with a bandwidth of 1nm:
The sensor scanned with a mirror 51 degrees on either side of the satellite, perpendicular to its direction of
travel, by stepping in 3 degree increments. 35 measurements were taken during each scan with a field of view
of 3 by 3 degrees, corresponding to 50 x 50 km directly underneath the satellite and 125 x 280 km at the 51 degree
extreme scan angle. Light, reflected by the mirror into TOMS, was split by a monochromator into
individual wavelengths, and the six to be measured were separated out by a rotating chopper disk. As
openings in the chopper disk aligned with exit slits, individual frequencies of radiation were measured
by a photometer. Once a week solar irradiance was measured by reflecting sunlight into the instrument from an aluminum diffuser plate.
Because TOMS relied on sunlight no measurements were
made at night. This prevented wintertime observation of the poles.
TOMS was calibrated before launch and in flight, and its measurements were also compared with
those of ground based Dobson photometers. Ground calibration was conducted using light sources of
known wavelength and intensity. A mercury-argon lamp carried aboard Nimbus-7 was used for
in-flight spectral calibration, and no degradation was observed over the 14 year life of the instrument. A
wavelength where solar radiation varies by only .3% was used to calibrate the instrument to correct for
changes in angle during orbits that affected solar irradiance measurements. The diffuser plate that
reflected solar irradiance into the instrument suffered more degradation than was initially expected, and
an algorithm was developed to correct for it. Near simultaneous measurements with the world standard
Dobson spectrometer at Mauna Loa, Hawaii confirmed the stability of TOMS.
Levels of total ozone were determined by comparing direct solar radiation with backscattered solar
radiation at the six wavelengths observed by TOMS. Tables were created which relate theoretically derived backscattered radiances to total
ozone for several independent variables, including:
atmospheric scattering and surface reflection properties
ozone climatological profiles
geometry of the sun, satellite, and backscattered radiation
ozone absorption properties
For a given set of observed conditions, total ozone is inferred by comparing the measured TOMS radiances with
those in the table and extracting the appropriate value of ozone associated with a match.
Atmospheric Scattering and Surface Reflection
Light received by TOMS is both scattered from Earth's atmosphere and reflected from its surface.
Rayleigh scattering governs the amount of backscattered radiation due to molecules in the atmosphere.
Radiation that penetrates the atmosphere can also be reflected by clouds and aerosols, or surfaces such
as land, sea, or ice. The amount of UV light absorbed by ozone or backscattered by atmospheric
constituents varies with the depth at which it was reflected, so an estimate is made for the average
tropospheric altitude at which the majority of reflection occurred. Sea-ice maps, composed from
separate satellite measurements, are used to account for the high reflectivity of snow and ice.
Climatological profiles are standard datasets used to estimate changes in atmospheric ozone content.
The profiles are used to calculate total absorption of radiation in the atmosphere and the amount of
backscattered radiance measured by a satellite. The backscattered radiance would be the amount
actually measured if the true ozone profile corresponded to one of the standard climatological profiles.
Climatological ozone profiles were developed from SBUV data and balloon
ozonesonde data. When ozone levels in the lower atmosphere are obscured by cloud cover, the profiles
are used to account for the missing information, and provide an estimate of total ozone. Climatological
temperature profiles are standardized tables of temperature versus altitude, and are used because ozone
absorption is dependent on temperature.
The relationship between the positions of Nimbus-7, the sun, and the area being scanned by TOMS
also affects measured radiances. Important aspects of the geometry are the solar zenith angle, satellite
zenith angle at the Instantaneous Field of View (IFOV), and the angle between the solar vector and the TOMS scan plane at the
IFOV. These angles affect the amount of radiation scattered by atmospheric molecules or absorbed by
ozone - larger angles imply a longer path through the atmosphere which the radiation must penetrate.
Ozone absorption of UV light varies with wavelength. In the TOMS instrument the four short
wavelengths (312 nm - 339 nm) are influenced by ozone absorption, while the longer wavelengths (360 nm
and 380 nm) are insensitive to ozone absorption. Instead of using individual frequencies for calculation
of total ozone levels, the ratio of radiances at two different wavelengths (one more sensitive to ozone
absorption than the other), called a pair value, is used, since this minimizes instrument calibration
errors and wavelength independent effects that are not related to ozone. A series of calculations, detailed
in the NIMBUS-7 TOMS Data Product User's Guide, relates the measured pair values to a level of
total column ozone expressed in Dobson units. It is based on comparing the measured radiances of pair
values with those derived from the climatological ozone profiles.
As with any scientific instrument, TOMS is subject to errors and random uncertainties. Some of these
errors come from the instrument, and others from environmental phenomenon. Aside from some
known problems at specific times and locations, accuracy is believed to be within 3-4% of actual ozone
A degree of uncertainty in measurement is present in TOMS itself. Errors in either the initial calibration
of the instrument or accuracy of the lamp used for in-flight calibration would affect TOMS accuracy.
Degradation of instrument components, such as the diffuser plate, and the spectrometer itself, affected
TOMS. Over time, changes were made to the ozone retrieval algorithm to minimize this error. A small
uncertainty is introduced by the digitization of values from measured (analog) intensities. Perturbations
in the orbit of Nimbus-7, and small changes in orientation, add some error. Finally, random noise in
the electronics of TOMS contribute some uncertainty.
Physical phenomena also affect the ozone values determined by TOMS. Some uncertainty is present
in the assumed value of ozone absorption coefficients. Likewise, there is error in the calculated amount
of atmospheric scattering. Temperature change in the atmosphere affects ozone absorption, and
therefore measured ozone levels. Differences in the amount of ozone below clouds and other reflecting
surfacesfrom the assumed amounts also contribute uncertainty. Sulfur Dioxide (SO2), injected into
the stratosphere by volcanic eruptions, absorbs at some of the same frequencies of ozone, and affects
backscattered radiances from which total ozone is derived. This SO2 gas is subsequently converted to
sulfuric acid aerosols which enhance the backscattered radiation. Fortunately, data affected by SO2 can
be detected and flagged. Measurements obtained during solar eclipses can not be used due to the
change in incoming irradiance, so these data have been removed from the dataset. Polar stratospheric
clouds create a reflecting surface far above that assumed by the TOMS algorithm, and result in
an erroneously low value of total ozone in these regions. There has been no correction for this error,
but the effects occur over only a small area, and for brief periods.
Data that falls outside a certain range are flagged by the ozone algorithm and removed from the dataset.
These data flags were instrumental in the discovery of the ozone hole. Large numbers of flags for low
ozone levels were produced during the processing of data, and these were checked with ground-based
measurements and found to be correct. The algorithm was then adjusted to allow for a wider range of
ozone values. Checks are made for ozone levels that are too high or too low, extreme reflectivity
values, and similarity between different wavelength pairs.