6.8.1 Substitution Procedures

This section provides general guidance on substitution procedures for use in completing meteorological data bases prior to their use in modeling. It is intended for use by applicants and reviewing agencies in the development of substitution protocols for application to regulatory air quality dispersion modeling. Substitution protocols should be included in a modeling protocol and submitted for approval to the reviewing authority prior to the modeling analysis.

Substitution procedures will vary depending on the nature of the application, the availability of alternative sources of meteorological data, and the extent of the missing or invalid data. If the data base is such that there are relatively few isolated one-hour gaps, then an interpolation procedure, which is easily automated, may provide the most practical method of substitution. However, it there are lengthy periods with missing or invalid data, then application specific procedures will generally be necessary.

The goal of substitution should be to replace missing data with a “best estimate” so as tominimize the probable error of the estimate. The following suggestions have been prioritized in order of increasing probable error.

Substitution procedures which are considered to be “best estimators” include the following:

• Persistence - Persistence is the use of data from the previous time period (hour). This procedure is applicable for most meteorological variables for isolated one-hour gaps; caution should be used when the gaps occur during day/night transition periods.



• Interpolation - This procedure is applicable for most meteorological variables for isolated one-hour gaps and, depending on circumstances, may be used for more extended periods (several hours) for selected variables; e.g., temperature. As in the case of persistence, caution should be used when the gaps occur during day/night transition periods.



• Profiling - Profiling (profile extrapolation) refers to the procedure in which missing data for one level in a multi-level data base (e.g., data from a meteorological tower) is replaced by an estimate based on data from an alternative level or levels in the same data base. The probable error of the profiling estimate does not increase with the duration of the missing data, as is the case for persistence and interpolation. Consequently, profiling becomes a better estimator compared to persistence and interpolation as the length of the missing data period increases. Profiling based on a power-law should be used for extrapolating wind speed with height; the stability dependent procedure discussed in Section 6.2.5 is recommended. Profiling based on lapse rate should be used for extrapolating temperature with height. Alternatively, with the approval of the reviewing authority, applicants may use site-specific profiling procedures for wind speed and temperature.

Substitution procedures which provide estimators with moderate probable error include the following:

• Substitution from sensors located at comparable levels at nearby locations with similar site-specific (surface-specific) characteristics.



• Persistence when used for more than several hours.



• Interpolation when used for more than several hours.

Substitution procedures which provide estimators with high probable error include the following:

• Substitution from measurements at nearby locations with dissimilar site-specific (surface-specific) characteristics.



• Substitution of a climatological value for a particular time period; e.g., a seasonal ormonthly average.



• Substitution of simulated meteorology based, for example, on a boundary layer model.



• Substitution of “dummy data” such as a constant value for a variable.

6. METEOROLOGICAL DATA PROCESSING
  6.1 Averaging and Sampling Strategies 
  6.2 Wind Direction and Wind Speed 
      6.2.1 Scalar Computations 
      6.2.2 Vector Computations 
      6.2.3 Treatment of Calms  
      6.2.4 Turbulence 
      6.2.5 Wind Speed Profiles  
  6.3 Temperature 
      6.3.1 Use in Plume-Rise Estimates  
      6.3.2 Vertical Temperature Gradient 
  6.4 Stability 
      6.4.1 Turner's method  
      6.4.2 Solar radiation/delta-T (SRDT) method 
      6.4.3  _E method 
      6.4.4 _Amethod 
      6.4.5 Accuracy of stability category estimates 
  6.5 Mixing Height 
      6.5.1 The Holzworth Method  
  6.6 Boundary Layer Parameters  
      6.6.1 The Profile Method 
      6.6.2 The Energy Budget Method  
      6.6.3 Surface Roughness Length 
      6.6.4 Guidance for Measurements in the Surface Layer 
  6.7 Use of Airport Data 
  6.8 Treatment of Missing Data  
      6.8.1 Substitution Procedures 
  6.9 Recommendations