How good are temperature and humidity measurements with lidar?

In this contribution, we will discuss the performance of state-of-the-art automatic temperature and humidity lidar (e.g., Wulfmeyer and Behrendt 2022). As example, we will investigate ARTHUS (Lange et al., 2019), a lidar system developed at University of Hohenheim. This automatic mobile instrument participated in recent years in a number of field campaigns.

ARTHUS technical configuration is the following: A strong diode-pumped Nd:YAG laser is used as transmitter. It produces 200 Hz laser pulses with up to 20 W average power at 355 nm. Only this UV light is sent after beam expansion into the atmosphere so that the system remains eye safe. The atmospheric backscatter signals are collected with a 40 cm telescope. A polychromator extracts the elastic backscatter signal and three inelastic signals, namely the vibrational Raman signal of water vapor, and two pure rotational Raman signals. The detection resolution of these backscatter signals are 1 to 10 s and 3.75 to 7.5 m. All four signals are simultaneously analyzed and stored in both photon-counting (PC) mode and voltage (so-called “analog” mode) in order to make optimum use of the large intensity range of the backscatter signals covering several orders of magnitude.

From these eight primary signals measured by ARTHUS, four independent atmospheric parameters are calculated merging the PC and analog signals: temperature, water vapor mixing ratio, particle backscatter coefficient, and particle extinction coefficient. The temporal resolution of these data is also 1 to 10 s, allowing studies of boundary layer turbulence (Behrendt et al, 2015) and - in combination with a vertical pointing Doppler lidar - sensible and latent heat fluxes (Behrendt et al, 2020).

From the measured number of photon counts in each range bin, the statistical uncertainty of the measured data due to so-called shot-noise can directly be calculated. This value, however, while determining the major part of the uncertainty, does not cover the total uncertainty because additional noise of the analog signals is not included. So the shot-noise uncertainty alone underestimates the uncertainties in the near range where the analog data is used. To solve with this problem, higher-order analyses of the turbulent fluctuations can be performed which allow to determine the total statistical uncertainty of the measurements (Behrendt et al, 2020).

Finally, to investigate the stability of the calibration and thus the accuracy of the measured data, we decided to compare averaged ARTHUS data with local radiosondes. In order to cope with the unavoidable sampling of different air masses between these different instruments, we are investigating the average of a larger number of profiles.  We found that the performance of the measured data of ARTHUS reaches even the stringent requirements of WMO.

The results will be presented at the conference.

 

References:

Behrendt et al. 2015, https://doi.org/10.5194/acp-15-5485-2015

Behrendt et al. 2020, https://doi.org/10.5194/amt-13-3221-2020

Lange et al. 2019, https://doi.org/10.1029/2019GL085774

Wulfmeyer and Behrendt 2022, https://doi.org/10.1007/978-3-030-52171-4_25