J18.2
New Methodology for ATMS Noise Characterization

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner
Wednesday, 7 January 2015: 4:15 PM
230 (Phoenix Convention Center - West and North Buildings)
Miao Tian, Earth System Science Interdisciplinary Center, Univ. of Maryland, College Park, College Park, MD; and F. Weng

Characterization of a microwave radiometer noise during the pre-launch and post-launch phases is very important not only for the instrument calibration and validation, but also for data applications. In the past, the so-called noise equivalent delta temperature (NEDT) is determined by computing the standard deviation of the brightness temperature of the warm calibration target at a specific channel. It has been widely applied with great success in the noise sensitivity characterization for many microwave instruments, such as the Advanced Microwave Sounding Unit (AMSU-A), the Microwave Humidity Sounder (MHS), and the Advanced Technology Microwave Sounder (ATMS). However, recently, researchers have shown that NEDT is not always a proper way for characterizing the noise sensitivity for a microwave instrument; especially when there are slowly varied, multi-scale, long-term drifts (or trends) existed in the measurements. Hence, to better describe the noise sensitivity the so-called Allan Deviation is brought into consideration.

This research first conducts a series of comparisons of the performances of NEDT and Allan Deviation applied on a few different datasets, which include several artificially constructed noisy data that can represent stable, quickly and slowly fluctuated measurements, respectively. The Allan Deviation given in this comparison includes two forms: the so-called overlapping Allan Deviation and two-sample Allan Deviation, and each form coming with two variances. The results are discussed and recommendations are provided based on the analysis. Overall, the Allan Deviation serves better for sensitivity characterization for all the tested cases. Further, stability of the overlapping Allan Deviation is studied by altering the size of data and the Allan interval (e.g., the averaging factor in some literature). In each case, the overlapping Allan Deviation quickly converges as either the size or the Allan interval increases. In the end, both NEDT and Allan Deviation are compared by using the real ATMS measurements, ground and on-orbit.