RETRIEVALS OF ICE-WATER CONTENT FROM AN AIRBORNE ELASTIC LIDAR IN TROPICAL CONVECTIVE CLOUDS

Ice water content (IWC) is one of the critical parameters in determining the cloud radiative impact. In this work lidar-based IWC retrievals obtained in tropical mesoscale convective systems are evaluated in the context of an extensive in-situ and remote sensing instrumentation suite. Based on a test case of May 27, 2015 lidar-derived IWC values at 50 m above the aircraft were on average within 25% of the in-situ IWC measurements obtained using an isokinetic probe.


INTRODUCTION
Ice and mixed-phase clouds cover roughly 25% of the Earth's surface and can have a significant and variable effect on the Earth's radiation budget [1].The ice water content (IWC), is one of the critical parameters in determining the cloud radiative impact and hydrological cycle.In addition, high IWC conditions are recognized by aviation authorities as potentially dangerous for operation of commercial aviation [2].A substantial effort has been expended to estimate IWC from various in-situ and remote sensing measurements.In particular, several works have previously shown the feasibility of IWC retrievals from lidar extinction profiles, especially in relation to spaceborne lidars.However, in many cases this technique is applied to mid-latitude clouds with IWC values on the order of mg/m 3 .We have found no works that evaluate the potential of lidar IWC retrievals in the high ice water content (HIWC) environment, with IWC values reaching as high as 2-3 g/m 3 .Furthermore, the quality of the remotely sensed IWC measurements is often evaluated through comparisons with in-situ data.This can turn out to be problematic in HIWC, because the standard measurement techniques of estimating IWC, either from 2D particle imagery or hot-wire probes, can suffer from numerous instrumental issues [3].
Recently developed isokinetic probes (IKP) [4] were designed to solve most of these issues and can provide accurate measurements at high IWC concentrations (up to 15 g/m    1 while the lidar data processing chain is briefly described below.First, the return signal is background-and range-corrected to compensate for parasite (internal) signal sources and the R 2 signal drop off with range, respectively.
Then the incomplete overlap correction is also applied to compensate for the effect of the laser beam not being entirely within the telescope FOV at the lowest ranges.We estimate that the full overlap is achieved at 150 m.Comparing the AECL results in the few lowest range bins against the in-situ (flight-level) data requires careful consideration of the incomplete overlap problem.In this work, we have used the method of [5] to correct the signal below 150 m by using calibration data in clear air.Despite the correction, we acknowledge that evaluating lidar data in the overlap region can be problematic and associated with a source of error.Once all the basic corrections have been applied, the lidar vertically resolved extinction coefficients are derived using a Klett technique with a boundary condition in the far range.This method requires an assumption of the particulate extinction-tobackscatter (or lidar) ratio, LR.In this work we used a constant vertically averaged value of LR=20 Sr consistent with climatological values from satellite statistics and other works.Only the lidar profiles in zenith direction were used for Klett inversions.Finally, once the extinction coefficient (κ) profiles have been computed, IWC profiles can be calculated using a power law parametrization: where we used the values of a=527 and b=1.32 as per [1].

In-situ instrumentation
During the campaign, Convair-580 carried a multitude of in-situ instruments designed for cloud studies.Among them were particle imaging probes (including SPEC 2D-S and DMT PIP) as well as IKP for bulk IWC measurements.The imaging probes provide a visual information on particle size and shape.Particle probes' data was also used to derive cloud extinction coefficient and particle median mass size and concentration.IWC information can also be derived from imaging probes using empirical size-to-mass parameterization.The IKP provides reliable measurements of IWC by measuring a change in absolute humidity from continuously aspirating and evaporating all encountered particles and ensuring that the air speed at the probe inlet matches the free stream air speed [3].

NAWX radar
The NAWX radar system [6] has polarimetric and Doppler capabilities at both frequencies (X-band:  While TP1 and TP3 are both associated with relatively high lidar extinction values (reaching values >20 km -1 ), in-situ measurements show that the extinction is caused by different types of particles.Referring to panes 5 and 6 of Figure 2, it can be seen that particle concentrations are an order of magnitude higher during TP1 than during TP3 (103 vs 10 4 L -1 ).At the same time, the particle median mass diameter, MMD, defined as the size that splits the mass size distribution into two parts equal by mass, gets as low as 50-60 µm during TP1 but stays at an average value of ~500 µm during TP3.These observations imply that TP1 is dominated by small liquid cloud droplets, while TP3 consists of larger glaciated particles.
The Cloud Imaging Probe (CPI) imagery (not shown) corroborates this analysis, showing small regular particles during TP1 and predominantly unrimed columns during TP3.TP3 is associated with the High Ice Water Content environment -a primary subject of investigation in the campaign.Previous results show (e.g.[7]) that MMD in HIWC varies between 200 and 500 µm and decreases with increasing TWC and/or decreasing temperature.These statistics are consistent with the results of Figure

CONCLUSIONS
In May 2015 AECL sampled regions of HIWC near Cayenne (French Guiana) with IWC values as high as 2-3 g/m 3 .Lidar IWC values derived from extinction using a power law parametrization were on average within 25% of the IKP measurements.The relatively good agreement is important for a number of reasons.First, a consistency between fundamentally different instruments yields more credence to both the absolute values and the relative variations of IWC in the HIWC environment.Second, it shows that the parametrization developed by [1] produces results consistent with in-situ measurements in HIWC environment.Third, it shows a potential for lidar-only IWC retrievals in the future once the inherent uncertainties in lidar data processing are accounted for.

Figure 1 .
Figure 1.AECL design with a rotatable beam bender AECL can estimate a degree of particle-induced depolarization of the emitted laser pulse by separately measuring parallel and perpendicular polarization components of the return signal, P || and P ⊥ .The (linear) depolarization ratio, LDR = P ⊥ / P || , can often provide an insight into the types of particles encountered by the laser beam with lower LDR values often associated with spherical (liquid) drops.

Figure 2
Figure 2 Lidar, radar and in-situ responses during the HAIC-HIWC on May 27, 2015.See text for details.

Table 1
Technical specifications of the AECL transmitter, receiver and data acquisition system IWC IKP ) and AECL-based IWC values taken at 50 m above aircraft (IWC lid ).During TP3 IWC lid values were on average within 25% of the IWC IKP and generally matched well the IWC IKP in terms of relative variations.