Reading notes-17

AIRS measures approximately 200 channels in the 7.66 um absorption band of CH4, of which 71 channels are used to retrieve CH4. The retrieval algorithm of CH4 is similar to the AIRS ozone retrieval and CO retrieval.

In these retrievals, the atmospheric temperature profile, water profile, surface temperature and surface emissivity required as inputs are derived from other AIRS channels. FOr the CH4 retrieval, these data plus a first-guess profile of CH4 are used as inputs to the forward model to compute the upwelling radiance. The difference between the computed radiance and the cloud cleared radiance.

Because the information content of the 200 AIRS channels near 7.6um is highly redundant, the 71 channels primarily sensitive to CH4 but less sensitive to water vapor and HNO3 were selected.

This channle selection was optimized to be applicable for all conditions from the tropics to the polar region. With the limitation of the information content, it is impossible for us to derieve fine structure in the vertical variation of CH4, thus the current CH4 retrieval divides the tropospheric CH4 profile into a series of seven vertically overlapping trapezoidal functions.

More information regarding the “trapezoidal functions” used in the retrieval can be referred to Susskind et al. A thicker layer near the surface was selected due to the lpwer sensitivity of AIRS in this region.

As pointed out by Susskind et al, it is desirable to use channels with sharp localized weighting functions, or kernel functions, which define the sensitivity of the brightness temperature of AIRS channels to the change of CH4 in different layers and is calculated by a finite difference using the operational forward model.

As examples, the change of brightness temperature for some CH4 channels near 7.6um for a 2% change of CH4 in different layers for the polar, tropical and midlatitude profiles are shown in Figure 1.

It is evident from Figure 1 that the most sensitive layers are in the mid-upper troposphere, and the information in these channels is redundant as the weighting functions are very similar for many channles.

Given that the methane band resides inside the board 6.7um H2O band, the moisture optical depth pushes the CH4 peak sensitivity upwards, so the most sensitive layer in topics is at a higher altitude than that in the midlatitude, and the midlatitude is higher than the polar region.

In the tropics the most sensitive layer of AIRS channels to CH4 is at about 200-300 hPa, and it decreases in altitude to about 400-500 hPa in the polar region. Moreover, the sensitivity in the polar region is usually smaller than in the midlatitude and tropics.

Since the temperature and moisture profiles significantly impact the weighting function, the vertical sensitivity of our retrieval has significant geographic and season variability. Larger sensitivities in the mid-upper troposphere suggest the variability of CH4 in these regions can be better retrieved from AIRS, while smaller sensitivities in the lower troposphere and stratosphere indicate that the retrieved CH4 in these layers will be mainly determined by the first-guess.

Since the selected AIRS channles for CH4 retrieval are not independent of each other, it is important to determine the number of independent pieces of information in the retrieval. This can be obtained by finding the singular vectors and singular values of the covariance marix of the sensitivity weighted by an inverse of the estimates of the precision and accuracy of our radiative transfer model and the errors and noise in the measurements.

The corresponding eigenvalues give an indication of the usefulness of each component and are used to determine the damping variable in the retrieal. We noted that, in general, in the tropics one to three eigenvectors are used but the second and third ones are heavily damped, and in the polar region one or two eigenvectors are used and both may be damped.

The degrees of freedom(DOF), defined as the fractional number of significant eigenfunctions used in the retrieval process, on average, are about 0.6-0.8 in the high northern hemisphere, 0.8-1.0 in the midlatitude, and 1.0-1.4 in the tropics. All these demostrate that there is ~1 piece of information in the tropics and less than one in the polar region.

We noted that the retrieved profiles with the DOFs less than 0.4 were no trustable, and suggested to rule them out. These profiles are usually over frozen surfaces and the thermals signals are very low.

引自《Characterization and validation of methane products from the Atmospheric Infrared Sounder (AIRS)》