File name follows ocean color naming convention:
Sensor.YearDay.L3m_DAY_EXP_SpatialResolution.algorithm.hdf.bz2
(1) Sensor: S=SeaWiFS, A=MODIS Aqua
(2) YearDay: YYYYDDD(begin)YYYYDDD(end). For example, January 2000 has the format of: 20000012000031
(3) L3m: we produces only level 3 mapped data at this time;
(4) DAY: monthly=MO, 8 day average = 8D;
(5) EXP: parameter is surface carbon export production;
(6) SpatialResolution: 9=9km*9km, 4=4km*4km;
(7) Algorithm: the algorithm used to produce the parameters. 
GP=equation derived from genetic programming in Li and Cassar [2016]; 
SVR=support vector regression derived in Li and Cassar [2016];
Siegel=Siegel et al. [2014]
Westberry=Westberry et al. [2012];
Laws2011 = Laws et al. [2011]
Dunne=Dunne et al. [2005]
Laws2000=Laws et al. [2000]
Baines=Baines et al. [1994]
Betzer=Betzer et al. [1984]
Eppley=Eppley and Peterson [1979]
Siegel=Siegel et al. [2014]
(8) hdf: data format;
(9) bz2: the compression format.
Example: “S20000012000031.L3m_MO_EXP_9.GP.hdf.bz2” is January average export production in 2000 derived by our algorithm GP (genetic programming) from SeaWiFS observations. It is a level 3 mapped file, with a spatial resolution of 9km * 9km.
------------------------------------------------------------------------------------------
The organization of folders follows ocean color website.
The top folder is “cassar_export_satellite” followed by sensor folders (“MODISA” and “SeaWiFS”).
In each sensor folder, there is a folder for each algorithm with data with different spatial resolution in separate folders. Data with different temporal resolutions are saved in separated folders.
Example:
“cassar_export_satellite\SeaWiFS\GP\9km\Annual” is the path of annual products produced by our algorithm GP (genetic programming) from 9km SeaWiFS observations.

Product units:
mg C m-2 day-1
Projection: 
Equidistant cylindrical
Spatial range:
Latitude: From 90o N to 90o S
Longitude: From 180o W to 180o E
These metadata have been included in each separated file.