|
|
Aura-TES L1B Products: Version 2 Quality Description |
TES L1B V002 data products are a significant improvement over V001 (Beta release). Several systematic errors have been resolved giving much better agreement with Aqua-AIRS radiance measurements of the same homogenous target scenes (see below). The error estimates given below are representative of TES nadir data. Errors specific to each target spectrum are available within the data products.
Precision estimates are given in the NESR (Noise Equivalent Spectral Radiance) part of the L1B product, available with each target spectral radiance. The NESR is estimated for each measured spectrum using the noise extracted from the spectral range outside the signal region allowed by the TES optical filter used for the measurement. The NESRs have not changed significantly due to algorithm improvements, however, data acquired after December 7, 2005 (run 3202 and higher) have better NESRs (around 50 nW/cm2/sr/cm-1) due to improved optical alignment following the warmup of the TES optical bench to a higher operating temperature.
| Filter | Freq Range (cm-1) |
Nadir NESR (nW/cm2/sr/cm-1) |
|---|---|---|
| 2B1 | 650 - 930 | 700 |
| 1B2 | 920 - 1160 | 200 |
| 2A1 | 1090 - 1350 | 150 |
| 1A1 | 1890 - 2260 | 100 |
Based on our validation with AIRS and our L2 retrievals, we estimate our systematic errors to be less that 0.5 K in brightness temperature. A known remaining error source is due to velocity jitter that affects our interferogram sampling. This sampling error produces the largest uncertainties (<1%) near the edges of our frequency filter bands. To mitigate this error, we suggest only using L1B data that is about 30 cm-1 away from the spectral range boundaries. For example, 950-1130 cm-1 for filter 1B2.
We have compared nadir TES L1B calibrated radiance spectra to Aqua-AIRS radiances by first convolving TES spectra with the AIRS spectral response function (SRF). Mean and RMS AIRS-TES differences in observed brightness temperature for homogenous targets (as determined by TES) are <0.5K. We see similar (small) differences in our comparisons to S-HIS (Scanning - High Resolution Spectrometer) measurements taken from the WB-57 during the first AVE (Aura Validation Experiment) Oct-Nov. 2004, and during CR-AVE, Jan-Feb 2006.
