Modeling and Propagation of Physical Parameter Uncertainty in a Mars Atmosphere Model


This paper demonstrates the use of Non-Intrusive Polynomial Chaos to propagate uncertainty in a single physical parameter throughout a Mars Atmosphere model. The deterministic samples used to conduct this stochastic analysis were obtained using Mars WRF, a Mars-adapted version of the Weather Research and Forecasting code. A parameter governing the diffusivity profile shape in the boundary layer due to Troen and Marht was taken to be the uncertain variable. A 30-day run in Northern Hemisphere Summer was conducted to explore the propagation of boundary layer uncertainty throughout the Mars atmosphere model. All other sources of uncertainty were neglected. It was found that uncertainty generated in the Mars atmosphere was significant (wind speed coefficients of variation greater than one) over large regions of the model domain and was not limited to the lower regions of the atmosphere near the boundary layer. The greatest uncertainties were found to occur in the region of a winter-time Southern Hemisphere polar jet. Two rigorous measures of stochastic analysis convergence, the probability density function and the cumulative distribution function of wind speed, were used at three sample points within the atmosphere. It was found that a fourth order chaos expansion was sufficient for convergence of the wind speed probability distribution at these three points. Unfortunately, the discovery of a numerical instability in three of the sixteen sample deterministic runs may cast some doubt on the conclusions of this paper. However, once this source of this instability is uncovered, it is entirely likely that the findings of this paper regarding the propagation of boundary layer uncertainty will be confirmed. Regardless, this paper successfully demonstrates the use of Non-Intrusive Polynomial Chaos for a problem of a scale previously unattempted and is a first step towards the regular use of Non-Intrusive Polynomial Chaos to propagate uncertainty in a Mars Atmosphere Model.

Meeting Name

46th AIAA Aerospace Sciences Meeting and Exhibit (2008: Jan. 7-10, Reno, NV)


Mechanical and Aerospace Engineering

Keywords and Phrases

Mars Atmosphere Model; Modeling; Physical Parameter Uncertainty; Propagation

Document Type

Article - Conference proceedings

Document Version


File Type





© 2008 American Institute of Aeronautics and Astronautics (AIAA), All rights reserved.