1. Page 1© Crown copyright 2006 Ice hydrometeor microphysical parameterisations in NWP Amy Doherty T. R. Sreerekha, Una O’Keeffe, Stephen English October 2006

2. Page 2© Crown copyright 2006 Outline  Motivation  Background  Model and data  Case study results  Summary and Future work

3. Page 3© Crown copyright 2006 Motivation  Currently precipitation and ice affected microwave radiances are not assimilated at the Met Office  Information in these conditions is sparse, so use of these data would be beneficial  For direct assimilation of radiances a scattering RTM is required – RTTOVSCATT  Testing of RTTOVSCATT before operational implementation revealed questions about ice microphysical assumptions

4. Page 4© Crown copyright 2006 Background  Ice scattering causes TB depression at AMSU-B frequencies  Strength of depression depends on microphysics of ice particles: size, shape, density  No prognostic a priori information is available about the microphysics so assumptions have to be made  Different methods of solving the scattering RTE perform to similar standard

5. Page 5© Crown copyright 2006 RTTOV 8.7  Simple two stream scattering solution (Eddington)  Fast geometric optics ocean surface emissivity model  Marshall-Palmer/Modified Gamma Drop Size Distribution  Ice particle diameter up to 100 microns, snow 100-20000 microns  Density of ice particles 0.9 g/cm 3  Density of snow particles 0.1 g/cm 3  Permittivity dependent on ice/water/air mixture of hydrometeors (Maxwell-Garnet mixing formula)

6. Page 6© Crown copyright 2006 Met Office Model Fields  Pressure, temperature and moisture profiles available from forecast model  Frozen hydrometeor, rain and liquid cloud content profiles available  Smooth transition between different types of frozen hydrometeor  Ice particle density inversely proportional to diameter and exponential size distribution dependant on temperature

7. Page 7© Crown copyright 2006 Case study simulations AMSU Ch 20 (183.3±7 GHz) NOAA-16 Observations RTTOV Simulated TBs

8. Page 8© Crown copyright 2006 Experiments ExperimentDensity Size distribution 10.1 g/cm 3 Modified Gamma 20.5 g/cm 3 Modified Gamma 30.5 g/cm 3 Field* 4 8.74 x 10 -4 exp{-0.625D 2 } + 4.5 x 10 -5 ( $ ) Modified Gamma 50.132 D -1 Modified Gamma 60.132 D -1 Field* *Field et al 2005 $ Jones 1995

9. Page 9© Crown copyright 2006 Same PSD different density Density = 0.5 g/cm 3 (exp 2) Density = 0.874 exp{-625*D 2 } + 0.045 (exp 4) Modified gamma distribution 183.3±7 GHz

10. Page 10© Crown copyright 2006 Same density different PSD Density = 0.132*D -1 Modified gamma distribution (exp 5) Paul Field size distribution (exp 6) 183.3±7 GHz

11. Page 11© Crown copyright 2006 Comparisons AMSU Channel 20: 183 ± 7 GHz

12. Page 12© Crown copyright 2006 Results for Experiment 6 Observation Experiment 6 183.3±7 GHz PSD = Field et al.,2005 (based on T and IWC) Density = 0.132 D -1 (Wilson and Ballard, 1999)

13. Page 13© Crown copyright 2006 Summary  Comparisons of TB observations with RTTOV8 simulations using Met Office forecast model inputs have highlighted strong sensitivity to ice microphysical assumptions at microwave frequencies affected by scattering  Interface between forecast and RT models is very important  Parameterisations of PSD based on T and IWC of cloud are better supported by simulations than more general ones  Parameterisations of density based on size of ice particles are better supported by TB simulations than constant density  Best parameterisation may depend on cloud type/latitude band, only tested so far with UK case studies

14. Page 14© Crown copyright 2006 Future Work  Option to use Experiment 6 microphysics will be available with RTTOV9  Test parameterisations in other conditions and areas  Investigate other available parameterisations  Implement best set of assumptions operationally at the Met Office

15. Page 15© Crown copyright 2006 References  Bauer et al., 2006, QJRMS, 132, 1259-1281  Doherty et al., 2006? Submitted to QJRMS  Field et al., 2005, QJRMS, 131, 1997-2017  Jones, 1995, PhD Thesis, University of Reading  Wilson and Ballard, 1999, QJRMS, 125, 1607- 1636

16. Page 16© Crown copyright 2006 Questions?

17. Page 17© Crown copyright 2006

18. Page 18© Crown copyright 2006  IPWG to work more closely with NWP centres  Scale matching – degrading the resolution of obs to make comparison with models agree better  Climatology, Hydrology, Nowcasting and Operational forecasts  Beam filling, justification for 3DVAR