NRB: The MPLNET Lidar Signal and Diagnostics Product
The Micro Pulse Lidar (MPL) and miniMPL used in MPLNET produces a calibrated L1 signal called the Normalized Relative Backscatter (NRB), as described by Campbell et al (2002) and Welton and Campbell (2002). See our publications page for full citations to these publications, and consult them for detailed information on the NRB product. Here we use the term "MPL" to refer to both the original MPL and the miniMPL, except were explicitly mentioned. Deadtime detector calibrations are provided by the manufacturer. Darkcount and afterpulse calibrations are performed on-site at least once per 1-2 months. Overlap calibrations are provided using a wide-field-of-view receiver (WFR) installed on the MPL (Berkoff et al 2003). Routine inspection and recalibrations of the MPL overlap and polarization are also performed on a monthly cycle.
The original MPL design from the early 1990s was unpolarized, providing only total signal power. A polarized MPL design was introduced by Flynn et al (2007) and commercial versions became available. The original design utilized a nematic liquid crystal (NLC) polarizer which was limited to slow switching between states. A modified design (Welton et al 2018) using a ferro-electric liquid crystal (FLC) was developed by NASA and the MPL company resulting in faster switching capability and less potential for data quality impacts caused by the slower NLC. MPLNET began using a new polarized MPL at the start of V3 development in 2013. See our Instruments section for more information. The network became fully polarized in 2017, and all NRB data product files going forward contain total signal power, as well as co and cross polarized signals, and the volume depolarization ratio. As presented in Welton et al. (2018), the unique design of the polarized MPL results in polarization impacts that are similar to the gain-ratio and crosstalk that arise in traditionally designed linearly polarized lidar instruments, but are due to vastly different causes that require specific polarization procedures. A gain-ratio-like impact is the only concern in the absence of significant manufacturing defects in the FLC or induced birefringence in the window, and provided the instrument temperatures are in bounds with the calibrations. This is considered the nominal polarization condition. Polarization calibrations can be achieved by approximating the angle offset between the MPL polarizing beam splitter and the FLC fast axes to achieve particle-free atmospheric depolarization ratios equal to the Cabannes baseline of 0.37%. The polarization calibrations are typically very stable over time, but the FLC fast axes are dependent upon temperature and sensitive to instrument movement. Thus polarization calibrations require routine inspection, and a re-calibration if the MPL is moved. Non-nominal polarized MPL conditions are problematic. If manufacturing defects exist in the FLC, then addition crosstalk-like impacts arise resulting in large bias in the volume depolarization ratios for molecular and spherical particle layers (bias decreases as the volume depolarization increases). More concerning, crosstalk results in an inability to accurately calculate the total signal power creating a low bias (bias decreases as the volume depolarization increases). Crosstalk impacts can be mitigated if laboratory calibrations of the FLC are available, but this is time consuming and unrealistic. Instead, we have procedures to detect the presence of crosstalk upon inspection of the polarization calibration (using thick elevated clouds and the WFR). If crosstalk is detected, the FLC is rejected and needs to be replaced. A more detailed paper on the new FLC polarized MPL and its full characterization, performance, and calibration will be forthcoming from MPLNET.
The NRB is a range corrected, energy normalized lidar signal with all instrument specific calibrations and corrections applied except the so-called lidar system constant (C). As such, the NRB is not an SI unit of measure but a relative signal strength that can vary based on the transmission of the instrument optics (including the window used to gather data). The NRB is analogous to the EOS Level: L1A. Conversion to fully calibrated L1B Attenuated Backscatter (SI unit) data requires the determination of the system constant C and subsequent normalization of signal.
The NRB product also contains all instrument diagnostics such as laser energy, solar background (count rate), and various important instrument temperatures (ambient, detector, laser, etc). In addition, the product contains all Quality Assurance (QA) flags necessary to interpret the data.
NRB Product Levels and QA
L1, L15, and L2 NRB products have identical file formats and content. The only difference between L1 and L15 NRB products are that data failing to meet the L15 QA criteria are screened and replaced with NaN in the files. The primary difference between L2 and L1/15 files is that L2 may have additional post-calibrations applied as well as corrections to instrument temperatures. These would be reflected in the appropriate temperature and calibration flags. In addition, any L2 NRB data that fail to meet L2 QA criteria would also be screened and replaced with NaN in the files. NRB products only have one QA confidence level: high.
NOTE: Some QA Flags are not currently used in V3 L1 or L15 NRB, but are placeholders for eventual use in later Version release or for custom/manual L2 QA screening. We are conducting ongoing research in these areas, and have not determined an adequate method or process to utilize the information in automated processing yet. The following flags are currently set to "ignore_flag" status in V3 L1 and L15 NRB: