The MAXDOAS instrument and measurements
All stations of the BREDOM network are equipped with Multi Axis Differential Optical Absorption Spectroscopy (MAXDOAS) instruments. These instruments are basically UV/visible spectrometers observing scattered light in different viewing directions.
The requirements for a MAXDOAS instrument are
The Bremen MAXDOAS instrument consists of a grating spectrometer equipped with a cooled CCD detector and a separate telescope unit connected to the main instrument via a quartz fibre bundle. The spectrometer is temperature stabilized to avoid wavelength drifts. Although the CCD used is a 2-dimensional detector, it is operated in full vertical binning for optimal signal to noise. The quartz fibre bundle efficiently depolarizes the incoming light and also provides flexibility for instrument set-up.
To optimize spectral coverage and resolution, most BREDOM stations are equipped with two spectrometers, one for the UV and one for the visible part of the spectrum. The instruments share the telescope unit through a quartz fibre bundle that splits into two ends on the spectrometer side.
Two different telescope units are being used for Bremen MAX-DOAS instruments, an older one with a fixed azimuth scanning only in the vertical direction and a newer one on a pan tilt head which can be pointed into essentially any direction.
Single Direction Telescope
The single direction telescope unit (see figure below) has two viewing ports, one to the zenith and one to the horizon. The viewing direction is selected via a motorized mirror. In the lower part of the housing, two calibration lamps (a HgCd line lamp and a white light source) are integrated for calibration measurements. To limit the field of view of the instrument, a small lens is installed in front of the quartz fibre bundle. To prevent direct sunlight from entering the telescope, additional shades are mounted on both the zenith and the horizon ports (not shown in the figure).
During operation the instrument scans sequentially several viewing directions from the horizon to the zenith, taking about 1 minute of measurements in each direction. The angle selection depends on location and the focus of the measurements and integration times increase towards twilight. At night, calibration measurements (line lamp for wavelength calibration, white light source for throughput monitoring, dark measurements for dark current characterization) are taken. As the calibration unit can be closed by a shutter, the calibration is also possible during polar day.
Pan Tilt Head Telescope
The current version of the IUP-Bremen MAX-DOAS instrument is equipped with a smaller telescope housing which is mounted on a pan-and-tilt head that allows a direct pointing in any viewing direction. Therefore, only one viewing port is necessary and there is no need for a mirror which might adversely affect the measurements.
Light is entering the telescope housing through a quartz glass window (A) and is focused by a lens (B) limiting the field of view onto an optical fibre bundle (C). A shading tube can be attached to the entrance window to prevent direct sunlight entering the telescope (not shown in the image). In the telescope housing, a gravity driven shutter (D) is installed that claps into the optical path between entrance window and focusing lens blocking the light if the instrument is pointing straight downwards (in this direction, only calibration measurements are performed). This allows measurements for dark current characterization of the system. The gravity driven shutter's side facing the lens and the optical fibre bundle is covered by a white PTFE (Teflon) plane.
This plane can be illuminating by a HgCD-line lamb (E) in order to perform wavelength calibration measurements. In the operating mode both, dark measurements as well as wavelength calibration measurements are performed automatically during night time. For viewing condition surveillance (clouds, fog, etc.) and investigation of events during the data analysis, a video camera (F) is installed in the instrument pointing always in the direction of the measurement's line of sight. Snapshots of the observed scene are taken and saved automatically for each measured spectrum. The accurate pointing of the instrument (elevation angle) can be quality-controlled by an inclinometer (H). This is especially important for operation on unstable platforms, e.g. during ship-borne measurements. A silica gel containment (G) reduces the humidity inside the telescope to prevent possible condensation on optical components or electrical problems/shortcuts.
For stratospheric observations, usually the zenith viewing direction is used. If the species of interest is known to be negligible in the troposphere, any viewing direction can be used and in fact measurements towards the bright part of the sky have been used in some studies of e.g. OClO in Antarctica to improve signal. The sensitivity of the instrument is largest at twilight as a result of the long light path in the stratosphere. This is illustrated in the figure below, where a simplified light path is shown at high and low sun.
In the real atmosphere, the situation is more complex as a result of the curvature of the atmosphere, refraction and the effects of multiple scattering. However, the basic idea remains valid, and measurements of stratospheric constituents are therefore based on comparison of data taken at dawn and dusk with a reference taken around noon. If the instrument is stable enough, a background spectrum from another day can be used to improve consistency or signal (important in high altitudes where the range of solar zenith angles (SZA) available is small during some times of the year).
An interesting aspect of the above sketch is that the light path in the lower troposphere (red) is the same for measurements at high and low sun. Thus, a constant tropospheric contribution will cancel when comparing noon and twilight measurements of the same day, making the measurements even more adequate for stratospheric research. However, if the tropospheric concentrations vary over the day or if the tropospheric light path changes e.g. due to clouds, the stratospheric measurements can be affected.
While sensitivity is highest at twilight, this time of measurement is not always desirable for other reasons. Many atmospheric species of interest undergo rapid photochemistry, and concentrations can change strongly during twilight making interpretation difficult. On the other hand, in combination with model calculations, such changes can provide additional information on the chemistry of the absorber. For satellite validation, the quantity needed is the column at time of overpass, and this is often not at twilight. For such cases, the measurements must either be interpolated (possibly using a chemical model) or a reduced sensitivity of the measurement at higher sun has to be accepted.
For tropospheric observations, the MAXDOAS instruments use the horizon viewing measurement directions. As illustrated in the figure below, the light path through the upper atmosphere (dark blue, above the scattering point) does not depend on viewing direction, while in the lowest atmospheric layers, the light path (dark red) increases as the viewing direction approaches the horizon.
The length of the light path in the lowest layers depends on geometry (the elevation angle used) but also on the mean free path of the photon. In the sketch, the scattering point is on the circle indicated by the dotted line. If the scattering point is above above the surface layer then the light path is determined by geometry only. If scattering probability increases e.g. at higher aerosol loading or for measurements in the UV (more Rayleigh scattering), the last scattering point is closer to the instrument and the light path for the lowest viewing directions reduces (lowest line in sketch). In the extreme case (fog, snow fall), there is no difference in where the instrument is pointing. An important boundary condition for the interpretation of the measurements is the assumption, that horizontal gradients can be neglected. In real measurements, this is not always the case and results can become ambiguous with respect to whether a signal variation results from vertical or from horizontal changes.
As the vertical sensitivity is a function of elevation angle, the combination of all measurements can be used to retrieve vertical profiles of absorber concentrations. The vertical resolution of such profiles depends on SZA, aerosol loading and surface albedo but is in the order of 3 - 5 layers for the lower troposphere. For such an inversion, a good estimate of the aerosol optical depth and vertical distribution is needed. This can be retrieved from measurements of species with well known vertical distribution such as O2 or O4, and in fact aerosol optical depth is another output of the inversion algorithm. If data are taken not only for different elevation angles, but also at various azimuth angles, information on the aerosol phase function and thus aerosol composition can also be obtained.
For satellite validation, the tropospheric column is often the quantity of interest. This can be determined either by integrating the retrieved profile or - more quickly - by using a viewing direction of 30° or 45° elevation where the last scattering point is almost always above the layer with high concentrations.