Hello,
MODIS-A PAR daily l3m question.
I am trying to make use of the PAR level-3, 9km daily MODIS-A data and see that the format for the data file changed from
A2014159.L3m_DAY_PAR_par_9km.nc and A2014160.L3m_DAY_PAR_par_9km.nc, such that the files sizes jumped from ~7mb to ~14mb. Has anyone else come on this?
John
PAR data set format
-
OB.DAACx - SeanBailey
- Posts: 1519
- Joined: Wed Sep 18, 2019 6:15 pm America/New_York
- Been thanked: 9 times
PAR data set format
John,
What changed was the data type. We put a new program in place to generate the SMI files, one that uses the same data type for products as in the L2 file.
The PAR product is now stored as a scaled integer rather than a float. 2 bytes per pixel vs 4 bytes, thus the factor of 2 reduction is size.
Sean
What changed was the data type. We put a new program in place to generate the SMI files, one that uses the same data type for products as in the L2 file.
The PAR product is now stored as a scaled integer rather than a float. 2 bytes per pixel vs 4 bytes, thus the factor of 2 reduction is size.
Sean
PAR data set format
Hi Sean,
Is the format going to be made consistent eventually to the entire time series at some point? And, is this the only format change or are there others out there in other products?
John
Is the format going to be made consistent eventually to the entire time series at some point? And, is this the only format change or are there others out there in other products?
John
-
OB.DAACx - SeanBailey
- Posts: 1519
- Joined: Wed Sep 18, 2019 6:15 pm America/New_York
- Been thanked: 9 times
PAR data set format
John,
There are similar "changes" to other products. Some were just changes in the scale factors used in the float-to-integer conversion, others went from float to int (like PAR).
When we reprocess, the files will all be generated with the new code. That said, the format is not that different, and most netCDF client software will treat the file differences transparently....kind of like in Oz...they ignore the man behind the curtain :wink: SeaDAS handles them just fine.
Sean
There are similar "changes" to other products. Some were just changes in the scale factors used in the float-to-integer conversion, others went from float to int (like PAR).
When we reprocess, the files will all be generated with the new code. That said, the format is not that different, and most netCDF client software will treat the file differences transparently....kind of like in Oz...they ignore the man behind the curtain :wink: SeaDAS handles them just fine.
Sean
PAR data set format
In my lab we run into changes in raw storage format and scaling with netcdf-cf data sets from non-NASA sources. In all the cases that were brought to my attention, explicit hard coding conversions and rescaling could be removed using current NetCDF libraries, resulting simpler and more robust code. As an added bonus it is generally easy to remove hard coded units strings and get the information from the netcdf-cf files.