This section describes the general structure of a magnetic tape containing Level 1b data. Section 2.1.1 describes the record and file arrangement of Level 1b data tapes. For information regarding the data format of a specific type of Level 1b data, see Section 3 for AVHRR data and Section 4 for TOVS data.
SSB offers the user data written on several different types of magnetic media. These media include IBM 3480 cartridges, 4 mm DAT tapes and 8 mm Exabyte tapes. (Nine-track 800, 1600 or 6250 BPI tapes are no longer available.)
There are also several selection options available to the user. It is possible to obtain a full data set copy, or just a portion of a data set (specified by an area and/or time). In addition, channels can also be selected and the data can be unpacked. (Selection of data by channels automatically results in data which is in the 16-bit unpacked format.)
SSB uses specialized software on all Level 1b data (except for SBUV/2 data). This software creates a small header record at the beginning of the file which is known as the TBM header and is described in detail in Section 2.1.1. Level 1b tapes in SSB's archive do not contain a TBM record, although any Level 1b tape sent to a user by SSB will contain a TBM record.
Data sets may be stacked sequentially on tape. If a data set extends across two tapes (split data set), there will be an end of file (EOF) on the first tape and a double EOF after the last data set on the second tape. Note: there will not be any headers before the data on the second tape. User's software, when detecting end of medium while reading data records, should request the next tape and resume reading the records of the current data set. If a user cannot handle split data sets, the area of interest must be reduced through some selection process.
Unless otherwise requested, all Level 1b tapes generated by SSB will be written with TBM Header, packed data, and no split data sets. For example, if a user wants three full data set copies of LAC data on IBM 3480 cartridges, the first two data sets will be placed on Tape 1 and the third will be placed on Tape 2 (rather than splitting the third data set across two tapes). See Table 2.3-2 for the volume of data available on IBM 3480 cartridges, respectively.
Each copy of a Level 1b data set will be written on tape as a separate file (a series of records followed by an EOF). Regardless of whether a full copy or area, time, or channel selection is made, the following format applies.
The record structure of each file (data set) except LAC/HRPT is as shown below:
|TBM (Terabit memory) Header record (122 bytes)|
|Data Set Header record|
However, the above record structure does not hold true for any LAC/HRPT data sets. These particular data sets contain two data records for each scan. Each record has 7,400 bytes; therefore, the Data Set Header Record is also contained in two 7,400-byte records (the first record contains the Data Set Header and the second record is meaningless). The record structure of each file for LAC/HRPT data sets is as follows:
|Record 1:||TBM Header record (122 bytes)|
|Record 2:||Data Set Header record (7,400 bytes)|
|Record 3:||Dummy (7,400 bytes)|
|Records 4-n:||Data records (7,400 bytes)|
The above tape format also applies to data received from SSB prior to April 1985 when the TBM system was shut down. These data always included a TBM Header Record since they were accessed using the TBM. Currently, SSB's software creates a TBM Header record directly from operator input of the data set name and selection criteria. Users can verify which selection criteria were used by checking the TBM Header record.
The TBM Header contains information as to the type of data, selection parameters used, etc. The TBM Header record is 122 bytes in length and all fields are ASCII characters except the Channels Selected field which is in binary. The format for the TBM Header record is contained in Table 2.1.1-1.
|Byte #||# of Bytes in Field||Contents|
|31-74||44||Data Set Name|
|75||1||Total/Selective Copy ("T" or "S")|
|94-96||3||Number of Minutes|
|97||1||Appended Data Selection ("Y" or "N")|
|98-117||20||Channels Selected (in binary)|
|118-119||2||Sensor Data Word Size (ASCII)|
The data set name is in the same format as described in Section 2.0.3. The total/selective copy field is one byte long and contains either a "T" or an "S" for total or selective copy, respectively.
The beginning and ending latitude and longitude fields give the range of latitude and longitude if either or both were selected. If the latitude or longitude select option is not used, these fields will have "ALL" coded in them to indicate all latitudes or longitudes.
The start hour and minute fields indicate the beginning time of the selected data set. The number of minutes field indicates the number of minutes contained in the selected data set. "AL" indicates there was no time selection, and all times were included.
The appended data selection field indicates whether or not Earth location was appended to the data. Appended data are always included with any Level 1b data obtained from SSB. This is indicated by a "Y" in the field for Yes.
The channels selected field contains 20 bytes corresponding to 20 possible channels. The value of each byte can be either 0 or 1. If the data have been selected by channel, then byte 75 will contain an "S" for Select and there will be a 1 in each appropriate channel position in the channels selected field. If the total dataset has been chosen, byte 75 will contain a "T" for Total and there will be a 0 in each channel position. The channel number is indicated by the position of the byte (e.g., Channel 4 would be found in byte 4 of the field). However, channel select HIRS data are not ordered chronologically from 1 to 20. They are in the order indicated in Section 18.104.22.168 (i.e., byte 2 would indicate Channel 17, etc.).
The sensor data word size field contains 2 bytes in ASCII which indicate the size of the sensor data words in the dataset. Valid values are 08, 10 or 16; which corresponds to 8 bit (unpacked), 10 bit (packed) and 16 bit (unpacked) formats, respectively.
Table 2.1.1-2 contains the character code conversion from hexadecimal to ASCII code. This may be useful when reading the TBM Header Record.
The format of the Data Set Header is described in Section 2.0.4. Note that the Data Set Header Record will always contain the same number of bytes as a full channel data record. For full data set copies, the format of each type of data record is described in Sections 3 and 4 for AVHRR and TOVS, respectively.