GAME-AAN File Format

Dec. 13, 1999 decided.

GAME-AAN should collect the data observed by every region group and plan to compile data of common format. Acquired data is aimed to open to many and unspecified persons through GAIN (GAME Archive Information Network). Following is the explanations about the data form that is presented from each observation group to GAME-AAN.

1. Equalized time and time stump

The data for GAME-AAN is to be a mean for a half-hour in principle, such as 0000-0030UTC and 0030-0100UTC and time stump is to be the last time (UTC) in the thirty minutes. So, the time stump of data starts at 0030UTC and end at 2400UTC in each day.
In case of unsuitable items for equalization, the same instant figure with time stump is also available.

Each data should be attached flags which show the quality of data. There are five variations, "0","1","2","8","9", and the followings show each meaning.

0: There is no problem in data.
1: There are some problems in data. The problem data had already revised by PI (Principal Investigator).  To give the detailed information in data set documentation.
2:There are some problem in data. It hasn't yet revised. (To give the detailed information in data set  documentation.)
8: PI defines this meaning. (To give the detailed information in data set documentation.)
9: Missing data. (If necessary, the details is described in data set documentation)

A file of data is to be made by each region and every month. All items are to be compiled in one file, so files are not separated into each observation items.

Each file is named as  "Place- Years-Versions. Dat"  .For example, in case of the first version data at Tiksi in August 1998, the name is showed as "Tiksi-9808-a. dat".

Alphabets are used as the order of versions , so ,for instance, the beginning is "a" and followings are b,c,d,e,$B!&(B.

A header, which shows a data item is put in the first line of data. Each line below the header starts at the year of data, days and months, and time. In short, it starts at a time stump, and an observation item and a flag is lined in a row.

The form of each line is CVS (Comma Separated Value format), so each item is divided with a comma. Observation figures must be the originals, so figures after additional calculations, dividing the original into 10 or mutipling it by 10 and so on, are unavailable.

Arrangement order of the observation items depends on each region group, but once the order is decided, it would not be changed.

In case of a giving up item, you put figures for lack of observation in place of the vacant space, so you don't delete the item itself.

New items are put into the end of each line as an additional style.

The name of observation item and available figures in data will be united if those items can be united between regions and particularly another proportions are to be offered.

In case of items that can't be united, decisions by each region group are accepted.

In case of missing data, they are all input NaN.

year,date,UTC,T_air_1.5m,T_air_1.5m_F,T_air_3.0m,T_air_3.0_F,Press_1m,Press_1m_F,P_0.7m,P_0.7m,RS_down_1.5m,RS_down_1.5m_F...

1998,0701,0030,27.54,0,26.89,0,NaN,9,0,0,103.2,2,.......

1998,0701,0100,28.20,0,27.03,0,NaN,9,1,0,20.9,2,.......

1998,0701,0130,28.41,0,27.99,0,NaN,9,0,0,120.5,2,.......

1998,0701,2400,13.01,0,12.88,0,1002.3,2,0,0,-12.0,2,.......

1998,0702,0030,12.58,0,12.01,0,1001.9,2,0,0,-4.2,2,.......

1998,0731,2400,28.41,0,27.99,0,NaN,9,0,0,120.9,2,.......

7.The attached file for explanation

Each month data need to attach a data set documentation, which explains the order of observation items, details of lack of observation and unit of figures in data.

The deadline is within one year after an acquirement of data.

Requirement is an observation data in a part of GAME-ANN, such as PAM III, ACOS.
A decision what should be "a part of AAN" is planed to discuss and propose in each local group.