Dealing with missing data

[Thiago Veloso]

 Greetings everybody,
 Currently I'm working on a 7-year long, sattelite precipitation estimates dataset. The data's temporal resolution is hourly. So, for each year of data, I should have 8760 binary files. Considering the files are named sequentially (year-month-day-hour), I've created a .ctl template GrADS file which intends to open the entire sequence at once and after that convert them from binary data to ascii (I wrote a script which is functional to this task).
  My problem is: Instead the hourly output, I need daily means, but some years present missing data. Instead of 8760 files, they
 have only 8733, or 8747, and so on. This is critical because if the series were complete, a simple "ave" function (starting from the first date and ending in the final date, incremented by 24) would produce the daily means of precipitation. But I'm sure that if I apply this function to a incomplete series, the results won't be reliable...
 So, I would like to listen your advices on how to deal with these missing files. What's the solution??
 Thanks in advance and best regards,
 Thiago Veloso.
Plants Disease LaboratoryDepartment of Phytopathology, Agronomy FacultyFederal University at Rio Grande do Sul, Brazil.  


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://gradsusr.org/pipermail/gradsusr/attachments/20091107/2674ae01/attachment.html