Time-resolved measurements of the densities of individual frozen hydrometeors and of fresh snowfall

Dhiraj K. Singh, Eric R. Pardyjak, and Timothy J. Garrett

Summary: The authors present a new way of estimating densities of frozen hydrometeors. The proposed method is based on Differential Emissivity Imaging Disdrometer measurements. This new instrument provides a more direct way of estimating mass of individual snowflakes. By combining mass observations with the rate of heat transfer during ice particle melting, the authors estimate particle density. This density is then converted to bulk-snow hydrometeor density, which acts as a proxy for snow-to-liquid ratio.

General comments:

This is a very good study that utilizes new instrument, DEID, and new method for estimation of the particle density. The paper is well written and presents the study in a logical way.

I only have one minor comment:  In Section 2.3 the authors present the equations that are used to estimate snowpack quantities, such as SWE and SLR. I think it would be good to explicitly point out that DEID measures falling snow properties and the observed liquid equivalent precipitation rate and bulk-snow hydrometeor density are proxies for the SWE and SLR. There are processes that take place at the surface that are neglected in the presented equations. You discuss this on page 20, but it would be good to also explicitly mention it here.

The manuscript describes measurements with the DEID, the Differential Emissivity Imaging Disdrometer. The instrument has been presented by Singh et al. 2021 who explain that it can “measure the mass, size, density and type of individual hydrometeors as well as their bulk properties.” Laboratory and field measurements have been described in that paper, which are very similar to the measurements presented in this manuscript. The authors also cite another study, Rees et al. 2021, which describes mass and density measurements of individual hydrometeors with the DEID in comparison to refined optical imagery data, again like what is presented in this manuscript.

The only new thing with the method is about density estimation. The determination of an effective thickness h allows a different estimation of the volume compared to the estimate based on area equivalent diameter used in Singh et al. 2021 and Rees et al. 2021 (spherical-particle method). Measurement of h is done during a calibration with certain ice particles under certain conditions determining P_0. This P_0 is then used for h, and thus density estimation.

The field campaign provides new data (compared to Singh et al. 2021 and Rees et al. 2021) and  comparison of densities in the field. In this campaign a different optical imaging system is used to provide complimentary data on the frozen hydrometeors. This imaging system uses a laser sheet that illuminates the falling hydrometeors from the side, which are then imaged by an SLR camera. Its optical resolution (judged from the images provided) seems similar or only slightly better than the optical resolution of the thermal camera of the DEID.

METHOD to determine mass m (as in Singh et al. 2021 and Rees et al. 2021):

The melted hydrometeor’s temperature and area are observed during evaporation, in the laboratory also during the initial faster melting.

From these observations, mass is determined using an equation derived from a heat transfer balance assuming that the heat transferred into the melted hydrometeor equals the sum of the heats required to a) increase the temperature of the hydrometeor and b) melt and c) evaporate the hydrometeor.

0) Determine κ (laboratory calibration).

1) Observe during evaporation: A(t), T_w(t). (T_p(t) about const)

2) Determine T_p,max.

3) Use Eq. (2) to determine m.

METHOD to determine density rho_MS (new in this manuscript):

The density is determined from m divided by V_s, the “volume of a snowflake”.

The volume V_s is estimated as Ap*h, where Ap is the “initial snowflake projected area” and h is the “particle’s effective thickness in the direction normal to the hotplate”. This effective thickness h is estimated as the product of a melting speed and the melting duration. The implied definition (or method of determination) of this melting speed v_melt is h divided by melting duration. v_melt can be determined from  T_melt (average temperature difference between hotplate and frozen part of melting particle) as the two variables are proportional. The proportionality factor includes, besides κ, the laboratory calibrated coefficient P_0.

For DEID measurements in the field it is easier to observe evaporation duration (instead of the much shorter melting duration), so Δ T_melt and melting duration can be replaced by Δ T_evap and evaporation duration. Thus, the volume, and with that the density, is estimated from Ap, Δ T_evap, and evaporation duration (Eq. 8).

0) Determine P_0 (laboratory calibration)

1) Observe during evaporation: T_w(t). (T_p(t) about const)

2) Determine Δ T_evap

3) Use Eq. (8) to determine density rho_MS

The manuscript continues with a method to determine the average density and derive from that the snow water equivalent precipitation rate and snow accumulation rate.

These methods are explained with a series of equations that are derived step by step. All these equations, however, suffer from unclear, implicit, badly motivated (or not at all), or wrong assumptions. This makes it extremely difficult to check the validity of the methods and to understand under which conditions and for which type of hydrometeors they can be applied. None of these assumptions is adequately discussed. That means that uncertainties related to these assumptions cannot be estimated.

Despite these many issues and unclear assumptions, the results are seemingly good. This is both surprising and interesting and needs better and further discussion. In particular it is interesting how DEID estimated well snowpack densities agree with manually measured density despite the fact that average hydrometeor density is wrongly consider to be the snow pack density.

Due to all the issues with the method and validation measurements using a limited variety of very simply-shaped hydrometeors, most of the conclusions should be re-evaluated after addressing these issues.

Before I can accept this manuscript for publication, I would like to see major changes taking into consideration my feedback (see Supplement!).