APPENDIX E

---

DETERMINATION OF CORRECTIONS TO N TABLES

FROM STANDARD LAMP TEST DATA

          Each calibrated spectrophotometer should have associated with it information to be used in maintaining a continuing check on the calibration of the spectrophotometer. This information should be contained in dated reference tables entitled Table of Settings of Q, Standard Lamp Tests, Wedge Calibration Tests, and Sensitivity Tests.

          In practice it may be found that new standard lamp test data are not in agreement with the data presented in the table entitled Standard Lamp Tests. Discrepant data arise, for example, when the spectrophotometer optics require cleaning or when a new photomultiplier is installed into the instrument. It becomes necessary, then, to determine appropriate corrections to the N tables to effectively recalibrate the spectrophotometer on an absolute scale.

          The method for determining corrections to the N tables is as follows. Assume, for example, that reference standard lamp readings for a particular spectrophotometer and lamp are the following:

Wavelengths      Mean Rref.
-----------      ----------
     A             27.2    
     C             27.2    
     C'            73.8    
     D             26.9    

Suppose that the observer obtains the following results when performing a standard lamp test using the same lamp:

Wavelengths      Mean Rexp.
-----------      ----------
     A             24.8    
     C             24.9    
     C'            76.4    
     D             24.4    

Obviously the two sets of data given above are significantly different. After repeating the tests carefully and also using another standard lamp to ensure that it is indeed the spectral response of the spectrophotometer that has altered and not the spectral characteristics of the lamp, the required corrections to N tables are obtained simply by subtracting mean R_exp. values from mean R_ref. values. For the example shown above the corrections are as follows:

   N Tables             DeltaR = Mean Rref. - Mean Rexp.
   --------             ----------------------------
To values in NA table      Add 27.2 - 24.8 = ±2.4   
To values in NC table      Add 27.2 - 24.9 = ±2.3   
To values in NC' table     Add 73.8 - 76.4 = ±2.6   
To values in ND table      Add 26.9 - 24.4 = ±2.5   

Notice that the experimental A-wavelengths standard lamp reading, for example, is too low by 2.4 degrees. This means that when routine ozone measurements are made the R_A and, hence, N_A values will be too low by about 2.4 units. It is necessary to compensate for the change in the spectral response of the instrument by increasing N_A values.

          The optical wedges of most spectrophotometers are designed in such a manner that a change of one R unit along the wedge corresponds very nearly to a change of one N unit along the wedge. Therefore, corrections to the N tables can be determined simply by comparing standard lamp test R-dial readings as has been done above; otherwise, corrections should be computed in terms of DeltaN rather than DeltaR.

          Ozone data that have been computed may occasionally have to be recomputed or corrected when an observer learns, after performing a standard lamp test, that the spectral response of his instrument has changed. The observer will have to use his discretion in deciding how much of the back data needs revision. For this reason it is important that standard lamp tests be performed at least once per month and that notes be kept regarding the occurrence of unusual events that may have a bearing on the spectral response of the spectrophotometer.

---

Return to Table of Contents

Forward to Appendix F. Computation of Cos Z and mu for Sun and Moon