AAREADME.txt Paul Withers, 2006.01.10 Center for Space Physics, Boston University Version Number ============== 1.0 Aim === The aim of this delivery is to provide ODY ACC data products to the MRO aerobraking team. These data products have been generated very rapidly in order for them to be useful for MRO aerobraking, so many crude approximations have been made. List of files ============= AAREADME.txt (this file) equipot_radius.dat (an possible areoid) gen_areoid.pro (IDL file to generate equipot_radius.dat) makeodyaccresultsformro.pro (IDL file to generate results) Nav_reconstr_thru_P338.xls (Excel spreadsheet with orbital information) navstuff_v01.csv (Stripped-down CSV version of the Excel spreadsheet) odymass.txt (List of ODY masses for each orbit) odyorb.txt (List of ODY orbit numbers for each orbit) odyperiod.txt (List of ODY periods for each orbit) outdir/ (Directory with results in it) readodyaccresultsformro.pro (IDL file to read the results) validation.txt (Discussion of validation efforts) outdir/ allalt.txt (Results at all constant altitude levels) ancinfo.txt (Archive of ancilliary information) pXXXprof.txt (Results for specific profile) perires.txt (Results at periapsis) These results files are described in readodyaccresultsformro.pro The generation of the results is described below and in makeodyaccresultsformro.pro The raw ODY ACC data files are not included here Sources of data =============== (1) file on JPL server MMONT1 called M01-AB/NAV/Reconstruction_Data/Nav_reconstr_thru_P0338.xls and (2) pXXXacc.txt from Jim Murphy at PDS Determine trajectory ==================== Inputs are all from JPL NAV file a, e, i of orbit periapsis date/time, lat, lon Use geometry of elliptical orbit to find r, lat, lon as function of time after periapsis Neglect rotation of Mars, so longitude might be slightly inaccurate. Find v of ODY in inertial frame from orbit. I don't know what areoid was used for altitude in PDS dataset ODYA_0001, so comparison of my results with that dataset is impossible. I've used a very crude areoid to obtain altitude from radius. Obtain ACC data =============== Use ay column from pXXXacc.txt as acceleration. This has sampling rate of 1 Hz. This is NOT an average of the 200 Hz high-rate data. Data quality is significantly lower than that. It isn't clear what it is, but it could be based on a 1 Hz sampling of the ACC. I haven't done any processing of the densities. I don't know what processing has been done to them. They don't appear to have any significant zero offset. I haven't removed any linear biases, the effects of thrusters or angular motions. They may already have been removed, or they may not. If not, they are a source of error. Estimate uncertainty in these measurements from the standard deviation of pre-pass ACC data. Generate 7-point and 39-point running means of the raw 1 Hz acceleration data. Use square root of N to get their uncertainties. Determine density profiles ========================== rho = 2 m acc / (C A v2) mass m given in JPL NAV file area A is published, 11 m2 drag coefficient C is estimated as 2. I don't have the aerodynamic database yet. atmosphere-relative speed v is approximated using the orbital velocity instead. This neglects the rotation of Mars and any winds. I calculated rho using raw 1 Hz accelerations, the 7-point means, and the 39-point means to obtain three distinct density profiles. Determine constant altitude properties ====================================== Since my altitude scale is unreliable, these results will also be suspect. Levels are 100, 110, ..., 140 km. Data between 95 and 105 km are used for the 100 km fit, and so on. I used an exponential fit to obtain density and density scale height at the relevant altitude, plus uncertainties. Validation ========== This is addressed in more rough detail in validation.txt Densities as a function of time compare well to other published data products. Densities as a function of altitude do not. A few km difference in the areoid can really make a difference here. Periapsis densities have been compared to the ODY AAG QLR values. I have three density products (1 Hz, 7-pt, 39-pt). I don't know which of these was used in the ODY AAG QLRs. The mean relative difference is 1-2% for all three possibilities. The standard deviation of the relative difference is 5% for the 39-pt mean, 9% for the 7-pt mean, and 13% for the 1 Hz data. Caveats ======= Results would be more accurate if I had used high rate data instead of the less accurate ACC data. If comparing these results to anything else, use your own preferred areoid to convert these rho(r) profiles into rho(z) profiles. Comparing two datasets that use two different areoids is a very bad idea. Densities as a function of time or radius compare better to other results than densities as a function of altitude do. I haven't generated LST for anything beyond periapsis, but it is easy to do since (LST - pLST)/24 = (lon - plon)/360. Please calculate it yourself from these known quantities. Errors are based solely on the scatter in the pre-pass datapoints for each orbit. Neglected sources of error include: Thruster contributions to ACC data Angular motion contributions to ACC data Atmospheric motion contributions to relative velocity Rounding errors on A=11 m2 Errors in ODY mass Errors in drag coefficient, probably about 5% It is difficult to estimate the accuracy of the results without thinking about these factors in detail. If the thruster and angular motion effects have not been removed from the raw datasets I've been using, then I estimate the uncertainty in the density to be 1 kg km-3. If they have been removed, then the uncertainty is likely to be about 10% (based on remaining terms in list above) or the noise level, whichever is greater.