Statistics – Computation
Scientific paper
May 2003
adsabs.harvard.edu/cgi-bin/nph-data_query?bibcode=2003dps....35.1411h&link_type=abstract
American Astronomical Society, DPS meeting #35, #14.11; Bulletin of the American Astronomical Society, Vol. 35, p.936
Statistics
Computation
Scientific paper
Mars Global Surveyor has produced a wealth of atmospheric data, including hundreds of millions of Thermal Emission Spectrometer infrared spectra. The scientific value of these individual observations is enhanced by data assimilation, wherein retrievals of the spectra are constrained by the laws of physics as embodied in an atmospheric general circulation model (GCM). The result is a four-dimensional, global dataset including temperatures, winds, and surface pressure. We have previously used a four-dimensional, variational assimilation technique. Data is accumulated in 12-orbit (approximately one sol) periods, and the initial state of the model that can best reproduce the observations over that time is determined. Observed spectra are accurately modeled; other fields agree with independent atmospheric measurements and are consistent with the meteorology generated by free-running GCMs. However, there are some drawbacks to this methodology. The division of data into 12-orbit chunks is artificial. Each sol's analysis is independent of the previous sol, causing discontinuities in the assimilated fields. The computed diabatic forcing overcompensates for errors in the model state. We are therefore turning to a sequential approach to data assimilation. All data are used to update the model at the time of observation. The GCM is run only in the forward direction. The process cycles between observational updates and model predictions. We expect improved forecast skill from this approach and that all quantities will converge to their true physical values. By calculating model forecast errors in the observation space, we avoid the lengthy computations normally associated with sequential assimilation using the Kalman filter. This procedure can be applied to the most complex GCMs. Using it, we expect to produce the most accurate possible picture of Martian meteorology, suitable for detailed intercomparisons between different observation systems and models.
Supported by the Mars Data Analysis Program under RTOPs 344-34-21-04 and 390-90-21-08.
Houben Howard
Weeks Colin L.
No associations
LandOfFree
Mars Global Surveyor Meteorology by Sequential Data Assimilation does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.
If you have personal experience with Mars Global Surveyor Meteorology by Sequential Data Assimilation, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Mars Global Surveyor Meteorology by Sequential Data Assimilation will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-1374448