Data processing

SMILES data processing

SMILES data processing can be largely divided into three steps. Level 1b: observed emission spectrum of the submillimeter-wave region. Level 2: the abundance altitude distributions of each molecule. Level 3: visualized information desired by researchers such as climatology data, diurnal and seasonal variation, which are provided on the Quick look page. The Level 2 and 3 processings are being performed in the NICT Science Cloud computer system.

Level 1b: Submillimeter-wave limb-emission spectrum

Submillimeter-wave radiation from the atmosphere is acquired as an electrical signal by SMILES. This electrical signal is then converted into physically meaningful spectrum data. Specifically, a graph whose vertical axis is for frequency and horizontal axis for radiance is created as a product of this process. Each of atmospheric molecules is known to emit a unique submillimeter wave at a specific frequency (for example, ozone emits a submillimeter wave at 625.371 GHz), so the concentration of the molecules can be estimated by measuring radiance of each frequencies.

Level 2: Vertical profile of atmospheric molecular volume mixing ratio

The L2 processing retrieves estimated altitude distributions of abundances for each atmospheric molecule from Level 1b submillimeter-wave limb-emission spectrum. SMILES observes the atmosphere at the atmospheric limb, so the spectrum contains mixed signals from various attitudes. Because the spectrum can not be decomposed fully into attitudes, a strict solution can not be found by analytic method. Accordingly, we estimate the most probable altitude distributions by the use of probabilistic technique called by maximum a posteriori estimation method.

Level 3: Visualized data (Ex. Climatology, seasonal variation and global distribution)

Level 2 data has information of the altitude distributions of atmospheric molecules at a certain time in a certain area. By combining the data, we can obtain time variation of molecule abundances in a certain region, and a global map of molecule abundances in a certain time zone, for example.


Evaluation of data reliability

It is important for both data provision side and user side always to be aware of data reliability. Generally in earth observation, unlike an experiment done in a laboratory, we can not confirm the reproducibility of data. Therefore, we evaluate reliability of data by comparing SMILES data with data measured by other instruments at the same place and time, or by investigating factors giving an error to analysis results through simulation.
SMILES mission has no official validation program, SMILES team has conducted data verification on the research basis.

Schematic flow of SMILES data validation

Back to top of this page