Team Number: 005
School Name: Alamogordo High School
Area of Science: Theoretical Astrophysics
Project Title: S.U.N.
We have completed the research phase. Our project will involve a computer program written by Dr. Hans Uitenbroek, a scientist at Sunspot National Observatory. The program mathematically models the SunÕs atmosphere, including elemental composition, density, particle flow, electromagnetic currents, among other variables. The mathematical model was generated based on our most accurate prediction of the conditions of the solar atmosphere. This model is capable of generating various forms of output. We are interested in the solar spectrum output reported by the model.
Every element has a specific spectrum, or set of wavelengths of photons it can absorb and emit. The sun is made up of many different elements, all contributing to the overall solar spectrum. The overall spectrum is not only affected by the atomic makeup of the elements, but also the density and electromagnetic conditions of the atmosphere. Additionally, the Doppler Effect can alter the overall spectrum, shifting certain spectral lines. UitenbroekÕs mathematical model attempts to accurately model the atmosphere of the sun by generating a virtual solar spectrum. The program performs this by calculating the most probable state of the solar atmosphere at 80 points in a single column (from the inner part of the sunÕs atmosphere to the outer part). The virtual spectrum is then compared with actual spectral data recorded by Earth-orbiting satellites.
Currently, the program has several flaws. The most notable of which is that the model doesnÕt work. Although the atmospheric modelÕs variables can be manipulated to produce accurate spectral lines in one interval of the spectrum, such modifications cause spectral lines in a different wavelength interval to be entirely different than the anticipated, or actual, data.
Ideally, this problem could be solved by inputting a large domain of variables into the program to produce a large number of virtual spectrums, which could be compared to the actual solar spectrum for analysis. However, the program is extremely complex. On UitenbroekÕs dual-processor Sun workstation, the program takes approximately 10 minutes to calculate the spectral lines for one solar atmosphere. During phase 2 of the project, we wish to take UitenbroekÕs source code, and modify the code to make it more efficient. We then plan to run the modified code on a multi-processor supercomputer, to try to reduce the amount of time the program takes to calculate the spectral lines for one atmosphere. During phase 3, we will use the utilities on the supercomputer to determine the amount of time the program takes to execute, and attempt several different strategies for making the program more efficient, until the best solution is discovered.
If the program is made more efficient for Dr. Uitenbroek, he has the capabilities to try more solar atmosphere simulations that with his current code. An accurate model of the sun may then be produced.