Steps In MTP Post-Campaign Data Analysis

MJ Mahoney

8. Determine the Window Correction Table

The term Window Correction Table (WCT) came about years ago when it was found that measured brightness temperatures could be made to better agree with those predicted based on radiosondes launched near an aircrafts flight track if elevation angle corrections were applied to the measured TBs. Initially it was thought that the HDPE window that the scan mirror views through was having a small effect, hence the term. It was also thought that back lobes might be seeing elevation angle dependent temperature structure. During the 2003 SOLVE-2 campaign it appeared that the magnitude of the WCT was pressure altitude dependent, especially for the most optically thin frequency (CH1). In fact, for the recent PAVE campaign this also seemed to be the case. Careful examination of the data however showed that a lot of structure in the WCT was due to temperature inversions near the flight level. When radiosondes with significant temperature structure near flight level were excluded, the pressure altitude dependence in CH1 went away.

To calculate a WCT, you will need the MISSION_RAOBSs.RAOB2 file, and of course all the flight data must have been processed with a null WCT. You will also need the MISSION_RAOBrangeAll.txt file, which contains the list of all the radiosondes that the aircraft flew close to. This file should be copied to the MISSION_RAOBrangeWCT.txt file. Initially, a pass will be made using all radiosondes, and then additional passes will be made with editted versions of this file which has removed radiosondes showing temperature structure near flight level.

MTPsimTAs.png
Figure 1. The MTPsim TAs tab.

The WCT is calculated using the MTP simulation program called MTPsim. Launch this program and select the TAs tab as shown in Figure 1 You can use the Open RAOB File button to the right, to open the MISSION_RAOBSs.ROAB2 file, or simply click the Calculate WCT button and an Open Dialog box will appear as shown in Figure 2.

MTPsimOpenRAOBfile.png
Figure 2. The MTPsim Open Dialog box.

Select the MISSION_RAOBSs.RAOB2 file (in this case PAVE_ROABSs.ROAB2) and click the Open button. The program will then use the MISSION_RAOBrangeWCT.txt file to determine which comparisons should be made, calculating the predicted brightness temperatures (observables) for each radiosonde, and then search the ACYYYYMMDD.REF file for the measured brightness temperatures. These observables are entered into the two leftmost tables of the TAs tab, and the difference, Predicted - Measured, is calculated and put in the rightmost table and written to the WCT.txt file in the mission folder.

OpenWCT1.png
Figure 3. Step 1 of 3 of the Excel Open Dialog

After all the radiosondes have been processed, open an Excel workbook, and then use it to open the WCT.txt file just written. The first step is shown in Figure 3. Make sure that the Delimited option in the Original data type frame is selected, and then click the Next button.

OpenWCT2.png
Figure 4. Step 2 of 3 of the Excel Open Dialog

Step 2 of 3 (shown in Figure 4) will appear. Make sure the Tab and Space check boxes in the Delimiters frame are selected, and click the Finish button to import the text file into Excel.

WCTtxtFile.png
Figure 5. The WCT.txt file after being imported into Excel.

After importing the WCT.txt file into Excel, open the MISSION_WCT.xls file from the previous campaign on this aircraft and save it to the MISSION_WCT.xls file in the mission folder for the current campaign. Our example, shown in Figure 6, is for the PAVE_WCT.xls.

WCTavg.png
Figure 6. The WCTavg tab in PAVE_WCT.xls workbook.

Next select all the data in columns A:F in the WCT.txt window, and cut and paste it at cell B3 (shown in red in Figure 6) in the PAVE_WCT.xls window. Column A contains sequential numbers for each radiosonde comparison, and each comparsion contains 13 lines of information in 6 columns. Referring to cell C3, we have the radiosonde name and the flight data, NKX20050114. Cells B4:F4, as  suggested above in row 2, contains the UT(ks) of the comparison, the average aircraft pressure altitude (aZp), the average tropopause pressure altitude (aZt), the average flight level temperature (aTz), and the average tropopause temperature (aTt). These data can be used to flag comparisons close to the tropopause and possible large inversions which can corrupt the comparisons.


The WCTavg tab of this workbook uses the Excel INDEX function to move information in columns D:F into a better organized and more versatie format. Cell R6 for example contains =INDEX($C$3:$D$678,$C6+R$4+2,2). It says to index into the worksheet area C3:D678 and copy the information in row C6+R4+2 and column 2 of this area into the cell R6. Since C6 =1 and R4 = 1, C6+R4+2 = 4, so that relative to the starting row of the area, the information in row 6 colum D (-7.05) appears in cell R6. You need to make certain that the bottom row (678 in this case) corresponds to the last row of the imported WCT.txt information. Using the bottom right hand corner of cell R6, copy this information into cells R7:R15, and then select cells R6:R15 and copy this information into the all the cells to the right of column R, creating one column for each sounding. This sets up the data needed to calculate the CH1 WCT. This data should not be copied and pasted into cell R19. We do this because plots are done for all the soundings as well as an editted subset. Finally repeat this process for CH2 and CH3.

Columns I through M contains the needed statistics for uneditted and editted WCT data for the three frequency channels: the average WCT (Avg), the standard deviation about the average (RMS), the minimum value (Min), the maximum value (Max), and the standard error (SE) on the determination of the average. For the latter to be correct, cell M4 must manually be filled with the total number of soundings, 52 in this example. This number must be reduced when sounding comparisons are editted out. That's why cell M17 contains 49 -- three soundings were editted out.
WCT1e.png
Figure 7. Uneditted WCT for CH1 PAVE
Figure 7 shows the results for all 52 of the soundings used to determine the PAVE Channel 1 window correction table.
WCT1.png
Figure 8. Editted WCT for CH1 PAVE
Figure 8 shows the results for the 49 editted soundings used to determine the PAVE Channel 1 window correction table.


PAVE_WCT.png
Figure 9. Average WCT for PAVE.
The average window corrections for PAVE are shown in Figure 9. The standard error is also shown for each measurement.

ObservableNoise.png
Figure 10. The calculation of Observable Noise on the OB Noise tab

When retrieval coefficients are calculated, it is important to have an accurate estimate of the observable noise. For example, it the noise estimates are too large, the information content of the measurements will be unfairly downweighted. The OB Noise tab contains the WCT from the WCTavg tab, as well as the standared error (SE) on each of the WCT measurements from column M of the WCTavg tab. As shown in row 27 of Figure 10, we estimate the radiometric noise to be 0.3 K in each channel. The Observable Noise is the quadrature sum of the Radiometric Noise, and the SE on the WCT. This information should be copied into the MISSION_RMS.txt file in the mission Setup folder, 1 line for each RMS starting with Channel 1 and the highest through lowest elevation angle, then Channel 2 and Channel 3. In addition the WCT information should be saved in the MISSION_WCT.txt file in the mission Setup folder.

Previous  |  Next  |  Index  |