Bruce L. Gary

Navigation System Corrections

In another web page (ER2 Nav Calibrations) I determined that during the CAMEX4 mission the navigation system had calibration errors in both the static air temperature, OATnav, and pressure altitude, ZPnav.  Using RAOBs to establish a CAMEX4-specific corrections, I concluded that the following corrections are required:

    ZPnavCOR [meters] = 0 + 85 * (utsec - utsec_takeoff)/14400 + 162 * (ZPnav/20 km)2 + 105 * (ZPnav/20 km)10

and that outsdie air temperatures be adjusted using the following offset:

    OATnavCOR [K] = -1.5 +/- 0.4 K

Window Correction Table

Using the above OATnavCOR, I have processed raw MTP data to produce TB values based on gains that produce a match of OATnav (corrected) and TB(horizon).  I also have adopted the ZPnavCOR equation, above.  Four flights, affording 30 RAOB/MTP comparison occasions, were used in determining the CAMEX4 WCT:  ER010903, ER010909, ER010919 and ER010926.

Figure 1.  Window corrections for MTP cahnnel #1, indicated by 30 comparisons of RAOB-based TB calculations and measured TB (using gains based on OATnavCOR = -1.5 K).

Figure 2.  Same as above figure, except for MTP cahnnel #2.

In both figures there's one "outlier" having extreme positive values.  It's produced by the third 2001.09.09 comparison, which is also the RAOB flyby that produced an extremely negative value for OATnavCOR (-6.4 K).  I have invoked "Pierce's Criterion" (1869) and have rejected this RAOB/MTP comparison in the calculation of an average WCT profile, shown below:

          60   0.26  -0.51
          44   0.80   0.29
          30   0.26   0.13
          18  -0.41  -0.84
           9  -0.19  -0.34
           0   0.15   0.11
          -9  -0.71   0.01
         -21  -0.80  -0.90
         -37  -0.07   0.09
         -58   0.56   0.05

The SE of these entries is approximately 0.2 K, which is based on a population scatter of about 1 K and a sample size of 30.

MTP/ER2 gains for these 4 flights, based on OATnavCOR = -1.5 K, formed scatter diagrams with straight-line fits summarized by the following equations:

    G1 [cts/K] = 18.45 (1 - 0.04 (tMXR - 43.4 C))
    G2 [cts/K] = 16.10 (1 - 0.06 (tMXR - 43.4 C))

During times when tTGT is too close to OATnavCOR, when gains cannot be derived on the basis that they produce agreement of TB(horizon) with OATnavCOR, these equations are available for use.  In this analysis I have not used tTGT leads or tMXR lags (which affect mainly ascent and descent data).

Comparison of WCT With SOLVE

It is important to determine if the WCT changed between the SOLVE and CAMEX4 missions, spaced 1.5 years apart.  The following figures are the final SOLVE WCT for Channels 1 and 2.

Figure 3.  SOLVE WCT for Ch#1.

Figure 4.  SOLVE WCT for Ch#2.

In reviewing the two pairs of WCT graphs, there is a noticeable change between SOLVE and CAMEX4.  For both missions, the two channels resemble each other more than a given channel resembles itself between missions.  This implies that the radome window actually changed between the missions.  For CAMEX4, there is a possibility that hoar frost formed at altitudes above 15 km, since that's where the "Loc 8 TB feature" appeared.  The CAMEX4 WCT has a large negative correction at Loc 8 (-21 degrees), and this WCT feature is undoubtedly produced by the "Loc 8 TB feature."  Whether the "Loc 8 TB feature" is produced by hoar frost, or something else, I can't say.  But it certainly dominates the shape of the CAMEX4 WCT.

Comparison of OATnavCOR With SOLVE

On another web page (MMS and NAV "Outside Air Temperature" Corrections for POLARIS and SOLVE ) I describe an analysis of OATnavCOR for a combination of the SOLVE and earlier POLARIS missions.  I concluded that OATnavCOR was dependent upon both altitude and OAT:

    OATnavCOR [K] = -0.69 + 0.259 * (Pressure Altitude [km]- 18 [km]) + 0.126 * (OAT [K] - 220 [K])

If this equation is used for each of the 30 CAMEX4 RAOB/MTP comparison events, ...


This site opened:  October 20, 2001 Last Update:  March 12, 2003