The Premobl program processes hourly, gridded meteorology data to get them ready for use by MOBILE6 in the Emisfac program. The major purposes of the processing are to spatially and temporally aggregate the temperature, humidity, and pressure data and to align the data’s time steps with the local 6 AM to 5 AM hour range needed by MOBILE6.

The Premobl program must first determine whether the available meteorology data are sufficient for the calculations being requested. It compares the list of meteorology data files given in the METLIST file with the episode start date (EPI_STDATE) and episode duration (EPI_RUNLEN), calculating the required coverage period for the meteorology data. Since MOBILE6 expects hourly data to be provided from 6 AM to 5 AM local time, Premobl considers this in its assessment of the available meteorology data. The earliest time needed in the meteorology data will be 6 AM in the time zone closest to GMT (assuming that the meteorology data are in GMT). If this earliest time is earlier than the user-defined episode starting time, then the episode start date must be moved back one day. Correspondingly, the latest time needed in the meteorology data will be 5 AM in the time zone farthest from GMT (again assuming that the meteorology data are in GMT). If this latest time is later than the episode end time, the episode ending date is moved forward one day. Using these adjusted start and end dates and times, Premobl can then determine the total number of time periods that the meteorology data must cover.

Before computing the averaged temperatures and humidities, Premobl checks several aspects of the input files to ensure that the averaging will be able to work properly. It verifies that the meteorology files are valid by ensuring that the user-defined temperature, pressure, and mixing ratio variables to use in processing are actually in the files. Premobl also ensures that the grids of the meteorology files are the same and that the time span covered by all of the files is contiguous and that the files have the proper start and ends dates and times. If there are missing meteorology data files (missing days), Premobl generates an error message. Premobl also verifies that the ungridding matrix is consistent with the inventory (i.e., the number of sources match between the two) and that the grid matches the meteorology grid.

To compute the averaged meteorology data, Premobl first finds the *GROUP files created by Mbsetup to see which types of averaging are needed. Using the information from these files, Premobl assigns the correct type of averaging to each county in the domain and inventory. The program then reads the hourly meteorology data and applies the ungridding matrix to calculate county-based data, checking that the temperatures are within the minimum and maximum range allowed by MOBILE6. As each hour of data is read, Premobl shifts the data for each source to the appropriate local time based on the time zone of the source. For averaging types other than daily averaging, data corresponding to the same source and hour of the day (1-24) are summed across the days in the episode and a count is kept of the total days. Once all hours for the averaging period (week, month, or episode) have been processed, the summed value is divided by the total number of days in the averaging period. Additionally, the data for each hour are summed for all sources assigned to a county or reference county (depending on the spatial averaging used), then divided by the total number of sources summed.

After computing the average 24-hour temperature and humidity profiles, Premobl outputs the data to the appropriate file depending on averaging type. For daily profiles (no temporal averaging), each day is written to a separate file. Weekly profiles are written by week starting on Sunday and ending on Saturday. Monthly profiles follow the calendar dates, so a separate file is written for each month using the date for the first of the month in the file. The weekly or monthly files will never begin before the adjusted episode start date or end after the adjusted episode end date; for example, if the episode starts on a Tuesday, the first weekly file will also start on Tuesday (not the previous Sunday) and the subsequent files will start on a Sunday. Finally, for episode-based averaging, a single file is written that covers the entire episode.