AERMOD View – Long Term Averages
February 2015
The AERMOD air dispersion model can calculate three different long-term averages: Month, Period, and Annual. The equation below demonstrates how these averages are calculated for a model run spanning a single year.
where
i = Index ranging from 1 to Nhrs
Ci = Concentration at hour i
Nhrs = Number of hours in the averaging period
Nclm = Number of calm hours in the averaging period
Nmis = Number of missing hours in the averaging period
The Period and Annual averages will produce identical results for the single-year case. However, the results differ if multiple years are modeled together.
One major difference is that the Period average can be run for any start and end date. Whether the period is 1 hour, 1 day, 1 week, 1 month, 1 year, or 10 years, the Period average will report the concentration for the number of hours processed by the model. Annual averages, on the other hand, can only be run for periods containing complete years. The model issues a fatal error and stops processing if the period does not span the full 8,760 (or 8,784) hours.
This difference manifests itself in averages across multiple years. The Period average is effectively weighted according to how many non-calm, non-missing hours are present in each year. The Annual average across multiple years is performed by summing the individual annual averages and dividing by the number of years as shown below.
i = Index ranging from 1 to Nhrs
Ci = Concentration at hour i
Nhrs = Number of hours in the averaging period
Nclm = Number of calm hours in the averaging period
Nmis = Number of missing hours in the averaging period
The Period and Annual averages will produce identical results for the single-year case. However, the results differ if multiple years are modeled together.
One major difference is that the Period average can be run for any start and end date. Whether the period is 1 hour, 1 day, 1 week, 1 month, 1 year, or 10 years, the Period average will report the concentration for the number of hours processed by the model. Annual averages, on the other hand, can only be run for periods containing complete years. The model issues a fatal error and stops processing if the period does not span the full 8,760 (or 8,784) hours.
This difference manifests itself in averages across multiple years. The Period average is effectively weighted according to how many non-calm, non-missing hours are present in each year. The Annual average across multiple years is performed by summing the individual annual averages and dividing by the number of years as shown below.
where
a = Index ranging from 1 to Nyrs
Ca = Concentration at year a
Nyrs = Number of years in the model run
This means that running multiple years in a single run will output an average annual average. However, some regulatory agencies ask that annual averages be reported as the maximum annual average extracted from a certain number of years. In these cases, AERMOD must be run separately for each year in order to find the maximum.
a = Index ranging from 1 to Nyrs
Ca = Concentration at year a
Nyrs = Number of years in the model run
This means that running multiple years in a single run will output an average annual average. However, some regulatory agencies ask that annual averages be reported as the maximum annual average extracted from a certain number of years. In these cases, AERMOD must be run separately for each year in order to find the maximum.
