Research

Ice particle mass- and projected area-dimension (m-D and A-D) power laws are commonly used in the treatment of ice cloud microphysical and optical properties and the remote sensing of ice cloud properties. Although there has long been evidence that a single m-D or A-D power law is often not valid over all ice particle sizes, few studies have addressed this fact. Moreover, the cirrus m-D relationship is poorly constrained and this produces considerable uncertainty in both predicted and retrieved cloud properties.

To overcome these problems, self-consistent m-D and A-D expressions have been developed that are not power laws, but can easily be reduced to power laws for the ice particle size range of interest. In this way, there is no need for change in architecture of existing models. Also, such expressions are valid over a much larger size range than power laws. This was done by combining field measurements of individual ice particle m and D formed at temperature T < -20°C during Sierra Cooperative Pilot Project (SCPP) with 2-dimensional stereo (2D-S) and Cloud Particle Imager (CPI) probe measurements (or estimates) of D, A and m in synoptic and anvil ice clouds at similar temperatures during Small Particle in Cirrus Clouds (SPartICus). The resulting m-D and A-D expressions are functions of temperature and cloud type (synoptic vs. anvil), and are in good agreement with m-D power laws developed from recent field studies considering the same temperature range. In addition, by accounting for the non-linear expression (2nd-order polynomial fit), the uncertainties associated with m-D and A-D expressions was greatly reduced. Such uncertainties are quantified for use remote sensing. The results have been published in a peer-reviewed journal (Erfani and Mitchell, 2016). Moreover, this approach is employed by NCAR scientists to improve the new version of Community Atmosphere Model (CAM) (Eidhammer et al. 2016).

For more details, see:

Erfani, E., and D. Mitchell, 2016: Developing and Bounding Ice Particle Mass- and Area-dimension Expressions forUse in Atmospheric Models and Remote Sensing, Atmos. Chem. Phys., 16, 4379-4400, doi:10.5194/acp-16-4379-2016.

An understanding of the major governing processes of North American monsoon (NAM) is necessary to guide improvement in global and regional climate modeling of the NAM, as well as NAM's impacts on the summer circulation, precipitation, and drought over North America. NAM supplies more than half of the annual precipitation for NAM region. Also, wet summers in Arizona are associated with dry conditions in mid-west.

A mechanistic understanding of the NAM is suggested by incorporating local- and synoptic-scale processes. At the local scale, analysis of satellite observations and ship rawinsondes launched over the Gulf of California (GC) during North American Monsoon Experiment in 2004 (NAME 2004) show that the Marine Boundary Layer (MBL) temperature inversion over the GC controls the moisture transport from GC to NAM regions. The strong low-level inversion inhibits the exchange between the moist air in the MBL and the overlying dry tropospheric air. This inversion weakens with increasing sea surface temperatures (SSTs) in GC and generally disappears once SSTs exceed 29.5°C, allowing the moist air, trapped in the MBL, to mix with free tropospheric air. This leads to a deep, moist layer that can be transported by across-gulf (along-gulf) flow toward the NAM core region (southwestern U.S.) to form thunderstorms. Gulf surges also play a role in weakening or eroding this inversion since they are associated with cooling in the low-level free troposphere.

On the synoptic scale, limited soundings at the GC entrance (suggesting this local mechanism may also be active in that region), and climatologies based on satellite SST and reanalysis data of outgoing longwave radiation (OLR), horizontal wind, specific humidity and geopotential height from 1983 to 2010 exhibit a temporal correspondence between coastal warm tropical surface water, NAM deep convection, NAM anticyclone, and NAM-induced strong subsidence in the eastern Pacific. A hypothesis is proposed to explain this correspondence, as the coastal advection of tropical surface water might be responsible for the poleward advancement of deep convection, atmospheric heating and precipitation along the SMO. The NAM anticyclone can form as a result of atmospheric heating in the region of NAM deep convection. Thus, the poleward propagation of NAM convection may result in the poleward migration of the NAM anticyclone and the poleward movement of the anticyclone center appropriates mid-level moisture for convection further north along the Sierra Madre Occidental (SMO), complementing any low-level moisture from the Pacific coastal region. Strong subsidence occurs further northwest of the NAM convective system and moves northward from the Baja California peninsula in mid-May to southern California in mid-July, promoting drier conditions in these regions. This strong subsidence is generated as a result of Rossby wave response to the west of the NAM with midlatitude westerlies. The resulted paper has been published in a peer-reviewed journal (Erfani and Mitchell, 2014) that has won two awards.

Such findings can be critical in improving the numerical simulations of NAM, because most regional climate models under-predict Monsoon rainfall in this region. Therefore, I used Weather and Research Forecasting (WRF) model to simulate the 2004 NAM onset and the influence of SSTs. The observationally-based local scale mechanism is simulated by WRF, and it is shown that under-estimated rainfall is partially due to the biased simulations of boundary layer in this region (Erfani, 2016).  

For more details, see:

Erfani, E., and D. Mitchell, 2014: APartial Mechanistic Understanding of the North American Monsoon, J. Geophys. Res., 119, 13,096–13,115, doi:10.1002/2014JD022038.

Erfani, E., 2016: A partial mechanistic understanding of North American Monsoon and microphysical properties of ice particles. University of Nevada, Reno, Ph.D. Dissertation, 229 pp.

There is a long-standing challenge in cloud and climate models to simulate the process of ice particle riming realistically, partly due to the unrealistic parameterization of the growth of ice particle mass (m) and projected area (A) during riming. This study addresses this problem, utilizing ground-based measurements of m and ice particle maximum dimension (D) and also theory to formulate simple expressions describing the dependence of m and A on riming. It was observed that β in the m-D power law  appears independent of riming during the phase 1 (before the formation of graupel), with α accounting for the ice particle mass increase due to riming. This semi-empirical approach accounts for the degree of riming and renders a gradual and smooth ice particle growth process from unrimed ice particles to graupel, and thus avoids discontinuities in m and A during accretional growth. Once the graupel with quasi-spherical shape forms, D increases with an increase in m and A (phase 2 of riming). 

The treatment for riming is explicit, and includes the parameterization of the ice crystal-cloud droplet collision efficiency (Ec) for hexagonal columns and plates using hydrodynamic theory. Prior to this work, there was no practical method to calculate Ec in models. Moreover, most models use the Hall (1980) equation to calculate Ec for planar crystals, but this equation has important drawbacks inherited from the early numerical studies. To solve this problem, new equations for the calculation of Ec are developed based on the numerical study of Wang and Ji (2000) for both hexagonal plates and hexagonal columns that accounts for dependence of Ec on cloud droplet d and ice particle D in non-steady flow.  In particular, Ec for cloud droplet diameters less than 10 μm are estimated, and under some conditions observed in mixed phase clouds, these droplets can account for roughly half of the mass growth rate from riming. These physically-meaningful yet simple methods can be used in climate and cloud models to improve the riming process and cloud parameterizations. 

For more details, see:

Erfani, E., and D. Mitchell, 2017: Growth of ice particle mass and projected area during riming, Atmos. Chem. Phys., 17, 1241-1257, doi:10.5194/acp-17-1241-2017.

Microphysics parameterizations in regional and global models are computationally expensive and produce uncertainties in simulations. Such models mostly use microphysics parameterizations that employ multiple particle categories (such as ice crystal, snowflake, graupel) and they define arbitrary threshold for sudden change from ice crystal to aggregates and graupel, which is not realistic and does not happen in the nature. Another challenge exists in remote sensing of precipitation. Since the lowest radar reflectivity (Zw) over complex topography is often considerably above cloud base, radar quantitative precipitation estimates (QPE) often underestimate the precipitation at ground level. 

To find solutions, a snow growth model for rimed snowfall (SGMR) was developed based on the growth processes of vapor deposition, aggregation, and riming. The SGMR is initialized by radar reflectivity (Z) at cloud top, and thereafter simulates the vertical evolution of size spectra. The SGMR is based on the zeroth- and second- moment conservation equations with respect to mass, and thus conserves the number concentration and Z, respectively. New mass- and area-dimension expressions suitable for synoptic clouds are utilized in the model, and therefore the assumption of specific ice particle shapes is not required. In addition, the new approach to parameterize riming has the advantage of a smooth and gradual growth of mass and area by riming. In general, the processes of vapor deposition and aggregation lead to larger ice particles that fall faster and therefore, produce larger snowfall rate (rs). The rs and ice water content with the inclusion of riming are significantly greater than that produced by the vapor deposition and aggregation alone. Moreover, rs is sensitive to the cloud drop size distribution. The size spectra predicted by the SGMR were compared with those from two cases of Lagrangian spiral descent through frontal and cirrus clouds, and good agreement is seen between the vertical profiles of SGMR and observations. This analytical SGMR, due to its accuracy and short running time, can be used in climate models and remote sensing.

For more details, see:

Erfani, E., 2016: A partial mechanistic understanding of North American Monsoon and microphysical properties of ice particles. Ph.D. Dissertation, 229 pp.