GET_AVAILABLE Function:

Definition of the Greenland Depth Sounder Data File Structure

The University of Kansas, Radar Systems and Remote Sensing Lab (RSL) has been conducting experiments in Greenland to measure the thickness of the ice sheet. The radar depth sounder is an instrument that transmits a sequence of pulses from an aircraft flying over the ice sheet and coherently processes the received signal. The signal received is demodulated into its inphase (I) and quaduture (Q) components, which are sampled at a 18.75 MHz rate. After a pulse is transmitted, the system begins to sample after a delay. The amount of delay is set by the operator and is determined by the amount of time it takes the transmitted pulse to propagate to the ice surface and back. This delay is called the Sampling Window Delay. After this delay, the system begins sampling the received signal. The Number of Samples is also set by the operator and is determined by the anticipated thickness of the ice. Once a return is sampled, it is represented in the digital system as two sample vectors, one for the I channel and one for the Q channel. To reduce the data rates, several pulses are averaged together. The number of pulses averaged together is denoted as the Number of Coherent Integrations. Example:

I Channel of Pulse 1		I Channel of Pulse 2		I Channel of Pulse ...		I Channel of Pulse N*		I Channel Averaged
Sample 1	+	Sample 1	+	Sample 1	+	Sample 1	=	Sample 1
Sample 2	+	Sample 2	+	Sample 2	+	Sample 2	=	Sample 2
Sample ......	+	Sample ......	+	Sample ......	+	Sample ......	=	Sample ......
Sample ......	+	Sample ......	+	Sample ......	+	Sample ......	=	Sample ......
Sample M*-1	+	Sample M-1	+	Sample M-1	+	Sample M-1	=	Sample M-1
Sample M*	+	Sample M	+	Sample M	+	Sample M	=	Sample M

Q Channel of Pulse 1		Q Channel of Pulse 2		Q Channel of Pulse ...		Q Channel of Pulse N*		Q Channel Averaged
Sample 1	+	Sample 1	+	Sample 1	+	Sample 1	=	Sample 1
Sample 2	+	Sample 2	+	Sample 2	+	Sample 2	=	Sample 2
Sample ......	+	Sample ......	+	Sample ......	+	Sample ......	=	Sample ......
Sample ......	+	Sample ......	+	Sample ......	+	Sample ......	=	Sample ......
Sample M-1	+	Sample M-1	+	Sample M-1	+	Sample M-1	=	Sample M-1
Sample M	+	Sample M	+	Sample M	+	Sample M	=	Sample M

*Where M is the Number of Samples.

*Where N is the Number of Coherent Integrations.

When the data system is operating in coherent mode, it saves the averaged I and Q channel data to hard disk. When the data system is operating in incoherent mode, it squares both averaged I and the Q channel data and adds them together to determine the magnitude of each sample before saving these data to hard disk. Example:

I Channel Averaged		Q Channel Averaged		Incoherent Data
(Sample 1)²	+	(Sample 1)²	=	Sample 1
(Sample 2)²	+	(Sample 2)²	=	Sample 2
(Sample ..... )²	+	(Sample ..... )²	=	Sample ......
(Sample ..... )²	+	(Sample ..... )²	=	Sample ......
(Sample M-1)²	+	(Sample M-1)²	=	Sample M-1
(Sample M)²	+	(Sample M)²	=	Sample M

Several sets of incoherent data can be averaged also. The number of samples averaged is set by the operator and is determined by the Number of Incoherent Integrations. The data system also receives GPS data from the aircraft. The GPS string is stored in the data file with every new set of averaged data (coherent or incoherent).

After we return from the field, the raw data files (which can be fairly large) are broken up into smaller files which have the following format:

(float32) Pulse Repetition Frequency (PRF) value (Hz)
(float32) Sample Window Delay (seconds)
(uint32) DSP Mode (0=Coherent; 1=Incoherent)
(uint32) No. of Samples
(uint32) No. of Coherent Integations
(uint32) No. of Incoherent Integrations
(uint32) No. of Receiver cards (Interferometric Mode)
(uint32) Data format
(uint32) BLANK
(uint32) BLANK
(uint32) BLANK
(uint32) BLANK
(uint32) BLANK
(uint32) BLANK
(uint32) BLANK
(uint32) BLANK
(int32) Datatype in Position 1	(int32) Datasize in Position 1	(int32) No. of Records in Position 1
Position 1 data (of "Datatype" which is "Datasize" bytes long and "No. of Records" wide)
(int32) Datatype in Position 2	(int32) Datasize in Position 2	(int32) No. of Records in Position 2
Position 2 data (of "Datatype" which is "Datasize" bytes long and "No. of Records" wide)
(int32) Datatype in Position 3	(int32) Datasize in Position 3	(int32) No. of Records in Position 3
Position 3 data (of "Datatype" which is "Datasize" bytes long and "No. of Records" wide)
(int32) Datatype in Position …	(int32) Datasize in Position …	(int32) No. of Records in Position …
Position … data (of "Datatype" which is "Datasize" bytes long and "No. of Records" wide)

The first 64 bytes describe the radar settings when the data were collected. The next three integers (Datatype, Datasize, No. of Records) describe the Position 1 data. Next comes the Position 1 data. The next three integers describe the Position 2 data. This pattern continues until the end of the file. The following different datatypes are currently defined:

Datatype	Description	Format
1	Incoherent Raw Data	uint16
2	I Channel Raw Data (Coherent)	int16
3	Q Channel Raw Data (Coherent)	int16
4	GPS String	uint8
5	Computer Time	uint8
7..19	RESERVED
20	Top Curve	float32
21	Bottom Curve	float32

1998&1999 data were recorded with 16-bit resolution, while all previous year had only 8-bit resolution. The "Data format" word in the header indicates whether 8- or 16-bit resolution was used. When "Data format" = 0, the coherent and incoherent data are stored with 16-bit resolution with incoherent data stored as voltage (square root is applied). When "Data format" = 1, the coherent and incoherent data are stored with 8-bit resolution and incoherent data are stored as power (square root is not applied).