GLA02 is a level 1a product that contains the low level instrument corrections to the raw data such as background subtraction, laser energy normalization, dead time correction (532 channel), droop correction (1064 channel) and range square correction.
Lidar Profile Format
The format of the lidar profiles remains the same as for the raw (level 0) instrument data. For the 532 channel, there are 3 vertical 'layers' of data, each with different horizontal resolutions. The first layer extends from 20.5 to 41 km (268 bins) where the data are averaged to 1 second. The product variable name for this layer is 'i1_g_lid' at byte offset 26356. From 10.5 to 20.5 km, the data are averaged to 5 Hz (product variable name: "i5_g_lid"), and below 10.5 km the data is at the full 40 Hz resolution of the instrument (product variable name: i40_g_lid). The 1064 channel is the same as 532, except that data are not acquired above 20.5 km. The 1064 layer between 10.5 and 20.5 km is stored in product variable "i5_ir_lid" and the lower layer is stored in product variable "i40_ir_lid". The width of each return bin, and thus the vertical resolution, is 76.8 m for both channels. The GLA02 product consists of 1 second records with the time and position information corresponding to the first shot (of 40) of that second. Linear interpolation between seconds should be used to obtain location and time information for each of the 40 Hz profiles within a second.
Note that the height of the top most bin for each of the layers mentioned above is not constant. The height (above the ellipsoid) at which GLAS begins the lidar data acquisition depends on the height of the local topography below the nadir point of the satellite. GLAS is designed to start the acquisition of lidar data such that the end of the profile (the last bin of the lowest layer) is always roughly 1.5 km below the ground. Obviously, doing so over changing topography will cause the height of the top most bin (and all subsequent bins of the profile) to change as well. This 'vertical shifting' only occurs on second boundaries. Within a second, all 40 shots are acquired starting at the same height. To determine the height of the top bin, examine the "I_Rng2PCProf" variable (byte offset 28476) for the 532 channel and "I_rng2CDProf" variable (byte offset 29500) for the 1064 channel. Subtracting these values from the height of the spacecraft (I_Hsat, byte offset 56636) will yield the height (above the geoid) of the top most bin.
Note that all lidar data are stored in the product as 4 byte integers and the appropriate scaling factor must be applied to the data in order to obtain the backscatter cross section (please see the product format spreadsheet for GLA02 for the proper scaling factor).
Calibration
GLA07 contains the calibrated, attenuated backscatter for both the 1064 and 532 channels. The 532 data are calibrated using the molecular return from about 30 km altitude. The 1064 channel is not sensitive enough to measure a molecular return and is calibrated using the instrument parameters and airborne validation measurements. The molecular backscatter cross section is computed from either standard atmosphere data (temperature and pressure as a function of height) or, when available, the NCEP gridded analysis fields of temperature and pressure interpolated to the spacecraft position and time. The 1 byte flag "I_metFlg" on the GLA07 product gives the source of the meteorological data that was used to compute the molecular backscatter profiles. The complete profile of molecular backscatter cross section is stored on the GLA07 product for both channels. The molecular backscatter profiles are calculated and stored once per second. The 532 molecular profile spans from 41 to -1 km with vertical resolution of 76.8 m (product variable "I_g_mbscs"). The 1064 molecular profiles extend from 20.5 km to -1 km with the same vertical resolution (product variable "I_ir_mbscs"). The calibration constants used to calibrate the data are also stored on the product (I_g_cal_cof and I_ir_cal_cof for the 532 and 1064 channels, respectively). The 532 calibration is a 3 word array, where the first word is the calibration constant calculated from about 30 km altitude and the second word contains the calibration constant calculated from an altitude of roughly 10 km. The third word contains the value that was actually used to compute the calibrated backscatter. The 1064 calibration constant is a 2 word array where the first word contains the calculated calibration from the atmosphere at about the 10 km height, and the second word is the value that was actually used to compute the calibrated backscatter (which in the case of the 1064 channel will likely be a constant derived from instrument specs or validation measurements).
Lidar Profile Format
Note that all lidar data are stored in the product as 4 byte integers and the appropriate scaling factor must be applied to the data in order to obtain the backscatter cross section (please see the product format spreadsheet for GLA07 for the proper scaling factor)
The 1064 channel data are stored as 5 Hz profiles from 20.5 km (bin 1) to -1 km (bin 280) above the geoid and 40 Hz profiles from 10.3 km (bin 1) to -1.0 km (bin 148). The fundamental collection frequency of the instrument is 40 Hz for data below 10.3 km and 5 Hz for data above 10.3 km. Each vertical bin is 76.8 meters wide. Note that there are no 1064 data recorded by the instrument above 20.5 km, but data are collected up to 41 km for the 532 channel (see below).
The 532 channel data are stored as 5 Hz profiles from 41.0 km (bin 1) to -1 km (bin 548) above the geoid and 40 Hz profiles from 10.3 km (bin 1) to -1.0 km (bin 148). The fundamental collection frequency of the raw data is 40 Hz for data below 10.3 km, 5 Hz for data between 10.3 - 20.5 km, and 1 Hz for data above 20.5 km. As a result, for the 5 Hz profiles, the portion above 20.5 is repeated 5 times (in other words that portion of the profile does not change within a given second, since the fundamental data collection frequency of the instrument is 1 Hz above 20.5 km). The vertical bin size (resolution) is 76.8 meters for both channels.
GLA07 is designed to vertically align the data to a set altitude range (with respect to the geoid) of -1.0 km to 20.5 km (1064) and -1.0 km to 41.0 km (532). Originally, it was planned to set the range gate setting bias (the range gate governs the height of the start of data ) onboard GLAS to a value that would result in lidar profiles ending 1 km below ground. After launch we decided to set the range gate bias somewhat lower (1.75 km) in order to obtain more background information. This changes the altitude range that the raw data span. The 1064 data actually start at about 19.75 km (40.25 km for 532) above ground and extend to about -1.75 km below ground. Since GLA07 was designed to align the data such that the highest bin is 20.5 km (41.0 for 532), the 1064 profile is padded with 'invalids' from 20.5 km to 19.75 km (40.25 to 41.0 km for 532). This is strictly true only over regions at or close to sea level. For areas above sea level, there will be valid data above 19.75 km. For instance, if the GLAS 1064 profile were taken over a region with a surface elevation of 500 m above mean sea level, the data would begin at 20.25 km (40.75 km for 532) and GLA07 would insert invalids only between 20.25 and 20.50 km (40.75 and 41.0 km for 532). This behavior is a result of the fact that the start range gate for data acquisition is computed for each second onboard GLAS in such a way as to keep the end of the lidar profile about 1 km below the ground.
For the 532 channel there are saturation profiles that correspond bin for bin to the 40 Hz and 5 Hz atmospheric profiles described above. The saturation profiles are a zero/one flag that indicate the associated 532 bin is saturated (1=saturation). The 532 channel will saturate at a level of about 8-10 photons per detector. This is expected to occur mainly from dense water clouds and the ground return. The 532 saturation profiles are 'packed' on the data product (meaning each bin of the profiles is represented by a bit). Please see the product format spreadsheet to obtain further information on the saturation profile format.
Note that on the product format description, the 532 channel lidar profiles are labeled "532 nm merged attenuated backscatter". Whenever the 532 channel is saturated, the values can be replaced with a measurement of the cross section from the 1064 channel (which won't be saturated). Hence the label "merged attenuated backscatter". Whether or not this substitution takes place is controlled by a flag in the processing software. The value of the flag can be determined by examining the "I_532AttBS_Flag" (byte offset 70276). Please see the breakout description of this flag in the GLA07 product format documentation.
Elevated Aerosol Layers
The GLA08 product contains elevated aerosol layer height for layers below 20 km at 4 second horizontal resolution for a maximum of 5 layers. It also provides height for layers above 20 km at 20 second horizontal resolution for a maximum of 3 layers. If a layer top is found above 20 km, but the bottom is below 20 km, it is treated as a layer above 20 km. The layers are determined from the 532 channel data. It is very important for the user to examine the layer height flag (I_LayHgt_Flag, bytes 301-332) values to obtain information on the number, quality and type of layers found. The type of layer can be either normal aerosol or polar stratospheric cloud (PSC). The PSC flag is set only when the layer meets certain requirements regarding its average temperature, latitude and height. The quality flag (i20_aer_qf and i4_aer_qf) assigns a layer quality based on the ratio of the average signal within the layer divided by the average signal level above the layer. Values of 1 or 2 indicate a very tenuous layer that has a higher chance of being a false positive. The maximum value for this flag is 13 which indicates a layer containing very strong backscatter that is not likely to be a false positive.
Planetary Boundary Layer Height
Planetary Boundary Layer (PBL) height is determined at 4 second and 5 Hz resolution from the 532 channel data only. The PBL height is the lowest layer that can be resolved that has as its bottom the ground. The PBL thickness (top minus ground height) cannot exceed 6 km. If the layer top is greater than 6 km above the local ground height, it is not identified as the PBL top. The PBL top is first searched for using a 4 second average profile. If it is found at that resolution, the PBL top is located from each of the 20, 5 Hz shots that make up that 4 second period. Each PBL top retrieval is given a confidence rating which is similar to the elevated aerosol layer flag. The values run from 1 (lowest) to 13 (highest) confidence of a good PBL height retrieval. This number is computed by taking the ratio of the average signal within the PBL to the signal level above the PBL. The quality flag (among other things) is stored in the 'I_LayHgt_Flag' product variable. Please see the breakout description (available as part of the product description) of that variable for details.
The initial design of the GSAS atmospheric software used only the 532 data to determine cloud layer heights. As of version 3.4 of the GSAS code, cloud layer heights are also determined from the 1064 channel. Cloud layer heights from the 532 channel are determined at 4 different horizontal resolutions: 4 and 1 second, and 5 and 40 Hz. Cloud layer height from the 532 channel is first determined at a 4 second resolution. If a layer is found at that resolution, then a search at 1 second resolution is performed. If a layer is found at the 1 second resolution, then a search is performed at 5 Hz, and if a layer is found at 5 Hz and the layer is below 4 km, then the 40 Hz search is performed. The 1064 channel is treated similarly, where if a layer is found at 4 second resolution, then and only then is the search is done at 1 second. The layer heights from the 1064 channel are determined at 2 different resolutions: 4 and 1 second.
For each resolution there can be a maximum of 10 layers (that is a limitation of the GLA09 product format). The number of layers detected (the sum of the number of 532 layers plus the 1064 layers) at a given resolution is stored in the cloud layer flag for that resolution. Please see "I_LRCL_Flag", "I_MRCL_Flag", "I_HRCL_Flag", and "I_FRCL_Flag" product variables for more information. For each layer there is also a use flag (see the product format spreadsheet for GLA09). The cloud layers found from the 532 data have a use flag of 0. The layers found from the 1064 data have a use flag of 2. Thus, when looking at the 4 or 1 second cloud layers in GLA09, you must test the use flag to determine which channel was used to detect the layer. The layers detected from the 532 channel will always be listed first, followed by the layers detected from the 1064 channel. Note then, that there will often be duplication of layers, meaning that a given layer will be reported twice - once from the 532 channel and once from the 1064 channel (this of course assumes that both channels are operating). Note further that if this does occur, the top and bottom of the layer will most likely be slightly different (as two completely different algorithms are used to locate the layers). However, since the 532 channel is more sensitive than the 1064 channel, it will be able to detect optically thinner layers. This will result in some layers that were detected from the 532 channel to go undetected in the 1064 channel.
For each of the 4 horizontal resolutions, a ground return search is also performed. The presence or absence of a ground return must be used to determine if the bottom of the last (lowest) cloud layer found is valid.
Optical Properties Parameters Contained in Product
GLA10 contains the optical inversion results of the GLAS 532 nm atmospheric vertical profiles: backscatter cross section and extinction cross section separated into cloud and aerosol components. Polar stratospheric clouds are part of the aerosol category. In order to obtain the complete vertical optical structure, the user will need to merge the aerosol and cloud components. For every atmospheric layer that was detected by GLA09 from the 532 channel data, an extinction-to-backscatter ratio (S) value that reflects the current multiple scattering condition was calculated or estimated in order to complete the optical inversion. "True S" refers to this value under conditions of no multiple scattering. The "true S values use flag" refers to whether the calculated (flag=2) or estimated (flag=1) S value was used in the optical inversion process. GLA10 contains top and bottom heights of all the layers that were optically processed separated into cloud and aerosol. These are the same heights as reported in GLA09 and detected from the 532 channel and GLA08 (for aerosol layers or boundary layer). All cloud products (backscatter, extinction, S, layer top and bottoms) are reported at 1 second resolution while aerosol products are reported at 4 seconds only. The "ground detection" parameter refers to the height where the lidar sensed the ground. If the ground was not sensed, this parameter is set to invalid. This parameter is handy in determining whether the last layer bottom sensed was a true bottom or marks when the lidar signal was extinguished. All layer locations are referenced from the geoid. All extinction profiles have been corrected for multiple scattering. This multiple scattering correction factor is not stored on GLA10, but it is stored on GLA11. Each GLA10 record is a 4-second record that contains 1 group of aerosol products and 4 groups of cloud products (one for each second). The time stamp refers to the "J2000" time of the first shot of the record. J2000 time equates to the number of seconds since January 1, 2000 at 12 UTC.
For more information about the methodology used to created the products in GLA10 and details of the product descriptions, refer to the "GLAS Atmospheric Data Products" ATBD. This ATBD can be found at the following URL: www.csr.utexas.edu/glas/atbd.html Click on "Atmospheric Data Products" and go to Sections 3.5, 4.5, and specifically Section 4.5.3.
Note that all backscatter and extinction data are stored on the product as 4 byte integers and the appropriate scaling factor must be applied to the data in order to obtain the backscatter and extinction cross section (please see the product format spreadsheet for GLA10 for the proper scaling factor)
The 532 channel cloud cross section data are stored as 1 second average profiles from 20.4 km (bin 1) to -1 km (bin 280) above the geoid and the 532 channel aerosol cross section data are stored as 4 second average profiles from 41.0 km (bin 1) to -1.0 km (bin 548). Each vertical bin is 76.8 meters thick. The extinction-to-backscatter ratio (S) values are recorded one value per layer. There are up to 10 cloud layers in each cloud profile (obtained from the 1 second cloud layers detected by GLA09 using the 532 channel) and up to 9 aerosol layers in each aerosol profile (obtained from the 4 second aerosol layers detected by GLA08 using the 532 channel). Active layers are not necessarily packed toward the front of the array sequence. For aerosols, the top three layer positions are reserved for layers above 20.4 km and the last (9th) position is reserved for the Planetary Boundary Layer (PBL).
All the backscatter and extinction profiles have an associated composite flag variable that represents conditions in each layer that makes up the profile. The composite flag is made up of two main components: 1) layer quality flag and 2) layer use flag. For all cases, the quality flag is a categorized value of the percent error calculation. For backscatter cross section, the use flag gives saturation status. For extinction cross section, the use flag designates layer type category. More details of the values in these flags can be found in the ATBD, Section 4.5.4.
Optical Properties Parameters Contained in Product
GLA11 contains the layer by layer optical results of the GLAS 532 nm atmospheric vertical profiles: optical depth and multiple scattering factor separated into cloud and aerosol components. Polar stratospheric clouds are part of the aerosol category. In order to obtain the total column optical depth, the user will need to add the aerosol and cloud components. For every atmospheric layer, a multiple scattering factor value that reflects the current multiple scattering condition was estimated in order to complete the optical inversion. GLA11 contains top and bottom heights of all the layers that were optically processed separated into cloud and aerosol. All cloud products (optical depth, multiple scattering factor, layer top and bottoms) are reported at 1 second resolution (these are the same layers as detected using the 532 channel at 1 second resolution and stored in the GLA09 product. The aerosol products are reported at 4 seconds only (same layers as detected and reported in the GLA08 product). The "ground detection" parameter refers to the height where the lidar sensed the ground. If the ground was not sensed, this parameter is set to invalid. This parameter is handy in determining whether the last layer bottom sensed was a true bottom or marks when the lidar signal was extinguished. All layer locations are referenced from sea level. All layer optical depths have been corrected for multiple scattering. Each GLA11 record is a 4-second record that contains 1 group of aerosol products and 4 groups of cloud products (one for each second). The time stamp refers to the "J2000" time of the first shot of the record. J2000 time equates to the number of seconds since January 1, 2000 at 12 UTC.
For more information about the methodology used to created the products in GLA11 and details of the product descriptions, refer to the "GLAS Atmospheric Data Products" ATBD. This ATBD can be found at the following URL: www.csr.utexas.edu/glas/atbd.html Click on "Atmospheric Data Products" and go to Sections 3.5, 4.5, and specifically Section 4.5.3.
Note that all optical depth and multiple scattering factor data are stored on the product as 2 byte integers and the appropriate scaling factor must be applied to the data in order to obtain the optical depth and multiple scattering factor (please see the product format spreadsheet for GLA11 for the proper scaling factor)
The 532 channel cloud layer data are stored as 1 second average values and the 532 channel aerosol layer data are stored as 4 second average values. The optical depth and multiple scattering factor values are recorded one value per layer. There are up to 10 cloud layers in each cloud profile and up to 9 aerosol layers in each aerosol profile. However, for the optical depth product, the nine layers are separated into 8 elevated and 1 Planetary Boundary Layer (PBL). Active layers are not necessarily packed toward the front of the array sequence. For aerosols, the top three layer positions are reserved for layers above 20 km.
Optical depths for each layer have an associated composite flag variable that represents conditions in that layer. The composite flag is made up of two main components: 1) layer quality flag and 2) layer use flag. For all cases, the quality flag is a categorized value of the percent error calculation. The use flag designates layer type category. More details of the values in these flags can be found in the ATBD, Section 4.5.4. The "cloud multiple scattering warning flag" is a measure of multiple scattering intensity and is based on the total column optical depth. One value per 1-second profile is calculated. Refer to Section 3.6.2 in the ATBD.