Ice-Phase Precipitation Remote Sensing Using Combined Passive and Active Microwave Observations Benjamin T. Johnson UMBC/JCET & NASA/GSFC (Code 613.1) Gail Skofronick-Jackson NASA/GSFC (Code 613.1) IGARSS 2011 – Vancouver, Canada

Figure 1.: whiteout conditions during a snow storm. 2/22

Introduction Midlatitude/Winter precipitation is difficult to measure using radars or radiometers alone. Precipitating clouds consist of a wide range of particles with variable shape, size, number density, and composition, and microwave radiation is sensitive to these properties Furthermore, ice clouds, water clouds, and gases and attenuate/emit microwave radiation B. Johnson IGARSS /22

Physically-based microwave precipitation remote sensing methods require (at least): A physical description of the atmosphere and surface properties Physical descriptions of hydrometeors (PSD, shape(s), composition) Appropriate relationships between physical and scattering/extinction/backscattering properties An inversion method for retrieving the desired physical properties given observations B. Johnson IGARSS /22

Relevant Key Problems Uncertainties the physical description of the atmosphere: distribution of CLW, WV; particle composition, size distribution, and shape. No current method for validating MW scattering properties of ice-phase hydrometeors. Present Retrieval Approach Physical method using “consistency matching” -- adjust simulations until consistent with PMW and radar observations across multiple wavelengths (e.g., Meneghini, 1997). Pros: Simple to implement, works equally over land and water Cons: “matches” may not represent reality, geometric issues ignored (NUBF, beam matching) Important note: the uncertainty due to unknown particle shape is orders of magnitude greater than other known sources of uncertainties. B. Johnson IGARSS /22

Observed Reflectivities (Z ku, Z ka ) Observed Reflectivities (Z ku, Z ka ) Inversion Z-S, DWR, etc. Inversion Z-S, DWR, etc. Large set of Radar-Retrieved Vertical Profiles of PSD/IWC Large set of Radar-Retrieved Vertical Profiles of PSD/IWC Simulated Radiances ( TB sim ) Simulated Radiances ( TB sim ) PMW Retrieval Algorithm PMW Retrieval Algorithm Physical Model Precip. & Atmos. Physical Model Precip. & Atmos. Hydrometeor Model Ext., Scat., p( , Z Hydrometeor Model Ext., Scat., p( , Z Physical - Radiative Database Physical - Radiative Database (2) Forward Model TB Constrained PSD/IWC Profiles Retrieval Schematic Attenuation “Correction” (1) Radar-only Retrieval Observed Radiances ( TB obs ) Observed Radiances ( TB obs ) Radiative Transfer Model Radiative Transfer Model (3) Radar/Radiometer Retrieval 6/22

Observed Reflectivities and Passive Microwave TBs during the 2003 Wakasa Bay Experiment B. Johnson IGARSS /22

Observables: Z m,14, Z m,35, DWR Microphysics: Particle Density, Shape, PSD Type Retrieval Inputs at each vertical level Environment: Pressure, Temperature, Humidity, Cloud Water Content Forward Dual Wavelength Ratio Retrieval Method Z e,35 -IWC retrieval, infer D 0 / N 0 Match DWR with D 0 ( 3.67/  ) in Database; compute N 0 Update PIA for air, clouds, and precip. ( A 14, A 35 ) PIA-corrected Reflectivities Z e,14, Z e,35 Starting at storm top ( z top ) down to z=0 Is DWR  1 ? no yes (Const. Density Spheres) B. Johnson IGARSS /22

WBAY 03: Dual Wavelength Ratio, and retrieved N0, and D0 (assuming a single constant particle density) B. Johnson IGARSS /22

10/22

11/22

12/22

Part 1 comments: The basic retrieval works surprisingly well using only constant- density spheres approx. 5 K RMS error in precipitating regions, simply by adjusting the CLW and particle density. However, constant-density spheres likely are not representative of the true distribution of mass and sizes of particles within the observed volume of the atmosphere… Improvements: Inclusion of well-known size-density relationships for spheres (following Brown and Ruf, 2007), Include sets of non-spherical “realistically shaped” hydrometeors B. Johnson IGARSS /22

Magono and Nakamura (1965) Mitchell et al. (1990) Locatelli and Hobbs (1974) Barthazy (1998) UW-NMS (Tripoli, 1992) Constant Density Spheres Mass-Density Relationships (Fixed IWC = 1.0 g m -3 ) 14/22

Retrieved log 10 (IWC) [g m -3 ] using size-density relationships (Brown and Ruf, 2007) 15/22

B. Johnson IGARSS /22

Retrieved IWC [g m -3 ] :: “Realistic” particle shapes, exponential PSD B. Johnson IGARSS /22

18/22

19/22

Final comments: The present method is designed for testing advances in the physical-radiative properties of a physically based retrieval algorithm The choice of particle shape and size distribution appears to be the largest uncertainty in physically-based precipitation retrieval algorithms (most certainly renders them ill-posed) So, prior knowledge of the particle shapes and sizes should significantly constrain physically based retrievals However, this requires that one has already computed the necessary physical-radiative properties ahead of time! B. Johnson IGARSS /22

Next Steps for this work: (un-break my radiative transfer model… ) Create complete database of IWC as a function of reflectivity, dual-wavelength ratio, and particle shape. Add other non-spherical shapes (in progress, e.g., Kuo, G. Liu, others) Add melting particles (in progress) Apply retrieval to GPM satellite simulator data (T. Matsui, WK Tao, et al.) as a alg. dev. testbed. Incorporate database(s) into official GPM combined radar/radiometer algorithm currently assumes constant-density spheres(?) B. Johnson IGARSS /22

22/22