Statistics – Computation
Scientific paper
May 2004
adsabs.harvard.edu/cgi-bin/nph-data_query?bibcode=2004agusm.u44a..06h&link_type=abstract
American Geophysical Union, Spring Meeting 2004, abstract #U44A-06
Statistics
Computation
3337 Numerical Modeling And Data Assimilation, 3346 Planetary Meteorology (5445, 5739), 5409 Atmospheres: Structure And Dynamics, 5445 Meteorology (3346), 6225 Mars
Scientific paper
There is now a small fleet of spacecraft orbiting Mars with instruments that make observations relevant to the atmosphere. Chief among these is the Mars Global Surveyor (MGS) Thermal Emission Spectrometer that takes up to 6 nadir infrared spectra every three seconds from a low, circular, polar orbit, with occasional limb scans. The MGS also includes a Horizon Sensor that makes side-looking broadband 15 micrometer measurements, and a Radio Science experiment that determines the atmospheric structure from occasional radio occultations. From a different orbit, at a different time of day, the Mars Odyssey THEMIS instrument makes downward looking broadband 15 micrometer measurements. The Mars Express will add infrared and ultraviolet observations of the atmosphere from a highly elliptical orbit. All of these spacecraft carry cameras that observe ice and dust clouds. While none of the instruments or observing patterns is optimized for atmospheric science, the sum total of the data is more than enough to specify all of the parameters in a low resolution Martian general circulation model. We can therefore make use of data assimilation techniques (like those used in operational weather prediction on Earth) to deduce the full atmospheric state (4-dimensional temperatures, geopotential heights, winds, water vapor, dust, clouds, and surface pressure). The payoff is enormous: retrievals of atmospheric parameters are no longer independent of each other (and underdetermined), but are constrained by physical laws; the data assimilation product is a compact physical state that can reproduce the much more extensive spectral data (to within the observational errors); calibration can be addressed from the internal consistency of the observations of a given instrument; validation (in the absence of ground truth) is performed by detailed comparison of the data from different instruments and different platforms (even when there are no co-incident observations); data quality control emerges naturally from the observation weighting scheme which rejects data that disagrees violently with both other measurements and the forecast model; and real-time weather forecasts can be made available for a host of operational purposes (aerobraking, aerocapture, gliding, ballooning, dust storm warnings). All of these are made practical---given the limited computational resources that can be devoted to Martian weather forecasting---by an observation space sequential assimilation technique, using a transformed extended Kalman filter that weights both model forecasts and observational errors by agreement with the data. Forecast errors are less than 4 K and should improve with more sophisticated predictive models.
No associations
LandOfFree
Martian Weather Analysis and Forecasts from Multiple Spacecraft Observations 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 Martian Weather Analysis and Forecasts from Multiple Spacecraft Observations, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Martian Weather Analysis and Forecasts from Multiple Spacecraft Observations will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-1178125