REPRESENTING CLOUD AND PRECIPITATION IN NWP MODELS IN CANADA. Environment Canada [RPN], Dorval, Canada

Transcription

1 REPRESENTING CLOUD AND PRECIPITATION IN NWP MODELS IN CANADA. (Peter) M.K. Yau 1 and Jason Milbrandt 2 1 McGill University, Montreal, Canada 2 Environment Canada [RPN], Dorval, Canada

2 Environment Canada's forecast model GEM (Global Environmental Multiscale) Grid configurations: Global Uniform Global Variable Limited Area (LAM) medium-range (10-d) = 35 km 25 km t = 15 min Simple Cloud Scheme short-range (48-h) = 15 km 10 km t = 7.5 min eperimental short-range (24-h) = 2.5 km 1 km t=1 min (t=30s) Detailed Microphysics Scheme

3 The simple cloud scheme (Sundqvist) Cloud-cover fraction is diagnosed (function of RH) Condensation occurs when RH eceeds a threshold (80% near surface) Total condensate (cloud water/ice) is prognostic (advected) Precipitation falls instantly to the ground there is no advection of precipitation Global Uniform Global Variable

4 The detailed microphysics scheme NU vc, VD vc CLOUD selfcollection CN cr, CL cr selfcollection VAPOR VD vr VD vs NU vi, VD vi CL ci, ML ic, FZ ci CL cscnig ICE CN is, CL is VD vh Si hydrometeor categories: 2 liquid: cloud, rain 4 frozen: ice, snow, graupel, hail Multi-moment scheme VD vg CL cg GRAUPEL ML gr RAIN CL ch CN sg CL ir CL rh, ML hr, SH hr ML sr, CL sr CL rs CL sr CL sr-g CL ir-g CN gh SNOW CL sr-h CL ir-g CL ri CLih CL sh HAIL Milbrandt and Yau (JAS 2005 a,b) Milbrandt and Yau (JAS, 2006 a,b) Gultepe and Milbrandt (Pure App. Geoph.,2007) Milbrandt et al. (MWR, 2008) Milbrandt et al. (MWR, 2010) Dawson et al. (MWR, 2010) SEDIMENTATION Limited Area Model SEDIMENTATION Scheme implemented in GEM-LAM, Global variable (Canada) ARPS (U Oklahoma, US) WRF 3.2 (US)

5 Overview of the scheme Testing and improvement in IMPROVE-2 (GEM-LAM) Forecast in winter Olympics 2010 (GEM-LAM) Testing over Arctic (GEM-Global Variable)

6 Representing the size spectrum ANAYLTICAL FUNCTION N(D) [m -3 m -1 ] m D [ m] BULK METHOD

7 log N(D) log N(D) log N(D) Gamma Distribution Function: N( D) N 0 D D e Varying (slope) (N 0 and constant) Varying N 0 (intercept) ( and constant) Varying (shape) (Q * and N 0 constant) D [mm] D [mm] D [mm] INCREASING VALUES (of, N 0 and ) * Q = r q (mass content)

8 BULK METHOD Predict evolution of specific moment(s) e.g. q, N T,... Total number concentration, N T N N ( D) dd M (0) T 0 Mass miing ratio, q Implies prediction of evolution of parameters Size Distribution Function: N i.e. N 0,,... ( D ) N 0 D e D q c r where Radar reflectivity factor, Z Z D 0 0 m D 6 3 N N ( D) c ( D) dd M r c D ( D) dd 3, r M air density (6) (3), p th moment: M ( p) ( 1 p p D N ( D) dd N0 0 p1

9 BULK METHOD Predict evolution of specific moment(s) e.g. q, N T,... Implies prediction of evolution of parameters Size Distribution Function: N i.e. N 0,,... ( D ) N 0 p th moment: D e M D ( p) For every predicted moment, there is one prognostic parameter. The remaining parameters are prescribed or diagnosed. e.g. One-moment scheme: q is predicted; is prognosed (N 0 and are specified) Two-moment scheme: q and N T are predicted; and N 0 are prognosed; ( is specified) Three-moment scheme: q, N T and Z are predicted;, N 0 and is prognosed ( 1 p p D N ( D) dd N0 0 p1

10 CLOSURE OF SYSTEM c ( Solve for shape parameter α from NT Z ( 6)( 5)( G( ) 2 rq) ( 3)( 2)( 2 where N ( 3 cd, and r air density m D Solve for slope parameter λ from 0 cn T ( 4) rq( 1) 1 N T ( 1) 1 3 4), 1) Solve for intercept parameter N 0 from N T and q vary monotonically in a 1-moment scheme

11 Diagnostic closure for α in 2- moment scheme D m rq cn T 1 3, f ( D m )

12

13 r = M (p,α est ) / M(p,α corr ) r

14 Verification and improvement of Multimoment scheme in GEM- LAM (1 km) in IMPROVE-2

15 CASE STUDY November-December 2001: IMPROVE-2 Observational Campaign Improvement of Microphysical Parameterization through Observational Verification Eperiment Cresswell Sounding 150 km

16 CASE STUDY Dec 2001 case: chosen for study at W.M.O. International Cloud Modeling Workshop, Hamburg (July 2004) special issue of J. Atmos. Sci. (October 2005) dedicated to IMPROVE-2 GOES IR: 2239 UTC 13 Dec 2001 Characteristics: large-scale baroclinic system strong low-level cross-barrier flow Precipitation in IOP region: prefrontal showers; moderate to heavy stratiform rain (associated with mid-level baroclinic zone); surface frontal rain-band; transition to sporadic showers

17 CASE STUDY: MM5 Simulations Dec 2001 case: MM5 runs at 4-km and 1.3 km ehibited errors in surface precipitation attributed to problems associated with the microphysics (SM Reisner-2) UNDER-Predicted OVER-Predicted OBSERVED PRECIPITATION 1600 UTC 13 Dec 0800 UTC (18 h) BIAS SCORES 4-km MM5 Simulation Source: Garvert et al. (2005a) [J. Atmos. Sci.]

18 S-Pol 1.5 PPI R 150 km Portland 0.5 PPI R 200 km 4km-GEM 700 hpa

19 1-km GEM E. Reflectivity

20 No pronounced over prediction along lee side of Cascade

21

22 MICROPHYSICS: Observations Aircraft flight tracks ( UTC) NOAA P-3 Convair-580 Source: Stoelinga et al. (2003) [Bull. Amer. Meteor. Soc.]

23 NOAA P-3 Convair-580 MICROPHYSICS: Observations Z>4.5 km - single ice tal 3<Z<4 km dendrite 2<Z<3 km column & aggregate Source: Wood et al. (2005) [J. Atmos. Sci.] Mean size inc. with dec. height

24 MICROPHYSICS: Observations Combined Observations for UTC A Dec 2001 Case B Source: Garvert et al. (2005b) [J. Atmos. Sci.]

25

26 Cloud liquid water along P-3 flight legs Under prediction of vertical etent of cloud water

27 Ice/snow content along Corvair flight legs Over prediction of concentration of snow mass too large deposition and/or riming

28 IMPROVEMENTS OF SNOW CATEGORY Diffusional growth Growth by riming

29 Electrostatic Analogy for Diffusional Growth of Ice Crystals dm 4 C( S 1) i dt AB The electrostatic analogy of the capacitance theory of ice crystal growth is highly flawed and does not produce the observed growth rates of ice crystals. It severely overpredicts the growth rates in almost all cases [by a factor of 3 to 8+ for plates and 2 to 4 for columns] involving even simple heagonal shapes. i Bailey and Hallet (2006)

30 Add CORRECTION FACTOR to DIFFUSIONAL GROWTH EQUATION dm 4 C( Si 1) dt AB i dm 4 C fcorr ( Si 1) dt AB i where f corr must be < 1, with value justified by results

31 Sensitivity Tests for IMPROVE-2: f corr = 1.0 C = 0.5D g kg -1 q s q c With decreasing f corr, f corr = 0.50 C = 0.25D g kg -1 SNOW content (q s ) is reduced and CLOUD LWC (q c ) is increased f corr = 0.25 C = 0.125D g kg -1 Other evidence: Field et al. (2008) Westbrook et al. (2008)

32 RIMING of SNOW Stochastic collection equation: (for category collecting category y) CL y 1 V( D ) Vy ( Dy ) r ( D Dy my ( Dy ) Ey ( D, Dy ) N y ( Dy ) N ( D ) ddydd COLLECTION EFFICIENCY For the collection efficiency, E cs = 1 is often assumed (for collection of cloud by snow) If E cs < 1, the snow riming rate will be overestimated

33 150 m E cs (D c,d s ) 225 m 500 m 700 m 950 m 1700 m RIMING of SNOW D s Approimation: E cs min( Dc,30m) min( Ds,1000m ( Dc, Ds ) 30m 1000 m D c *Wang and Ji, 1992 Works for D c ~ m, and D s ~ m Reduces riming rate 10-80% (vs. E cs = 1)

34 Test of 2-moment microphysics in Vancouver Olympics 2010 in 1 km GEM-LAM

35 Nesting strategy for LAM-V10 system 3 nested LAM integrations twice daily from 0000 and 1200 UTC GEM-Regional forecasts: LAM-15 km 2.5 km 1 km Whistler 15 km 2.5 km Vancouver 1.0 km

36 Verification for LAM-V10 Olympic Autostation Network (OAN): appro. 40 standard and special surface observing sites (hourly or synop available on GTS) large number (relatively) of surface stations concentrated in small region

37 Verification Eamples Observations courtesy of George Isaac

38 Eperimental field: Solid-to-Liquid ratio Observed:* FLUFFY SNOWFLAKES SNOW PELLETS *Forecaster: Michael Gélinas

39 Testing of 2-moment microphysics in Global GEM variable 15 km over the Arctic

40 30 day simulation July 2008 over Arctic Polar-GEM: = 15 km

41 GPCP merged obs Sundqvist One-Moment Two-Moment

42 PRECIPITATION GPCP merged obs Cloud Scheme: Sundqvist One-Moment Two-Moment

43 SENSITIVITY TO TIME STEP Cloud Scheme: Sundqvist One-Moment GPCP merged obs Two-Moment t = 60s t = 120s t = 225s t = 450s 60-h Simulation

44 SENSITIVITY TO TIME STEP Cloud Scheme: Sundqvist One-Moment Two-Moment GPCP merged obs 60-h Simulation (t = 60 s)

45 SUMMARY 1) Multi-moment mied phase bulk cloud microphysical schemes have been developed and implemented in GEM-LAM and GEM-Global Variable 2) Comparison with in-situ field measurements allows improvements in the scheme 3) Implementation in GEM-Global Uniform is planned but still needs work to address a) time splitting for microphysics b) subgrid scale cloud fraction c) simplification to allow for a miture of higher and lower moment hydrometeor categories

46

47 SEDIMENTATION: Bulk scheme rq t SEDI ( rq V z q SM V = mass-weighted fall velocity q DM N t SEDI ( N V z N TM V N = number-weighted fall velocity Z t Z SEDI ( Z V z Z V = reflectivity-weighted fall velocity For a given size distribution, V Z V q V N

48 Effects on sedimentation terms (Q = r q) z [km] ANA Q [g m -3 ] SM DM0 TM TM better than DM0 better than SM DIFFERENCE RELATED TO SIZE SORTING

49 Disadvantages of 1-moment scheme a) Inconsistency in modeling physical processes From closure relation, N T and q vary monotonically N T increases or decreases with q, but in breakup, N T increases but q = constant, and in diffusional growth, q increases but N T = constant. c) Inconsistency in modeling size sorting in sedimentation mean size increases with decreasing height, but not necessarily true in 1-moment as mean diameter is D m rq cn T 1 3

50 Disadvantages of 2-moment fied α scheme in sedimentation Rate of change of D m (size sorting) proportional to fallspeed ratio

51 Diagnosed α sedimentation results in larger mean size (larger D m ) but narrower spectrum (larger α ) From TRIPLE- MOMENT sedimentation profiles:

52 How well do the various bulk scheme predict sources/sinks? dq dt S dq dt proc1 dq dt proc2... MICROPHYSICAL PROCESSES e.g. dq dt CL 0 dm( D) dt CL N( D) dd CONTINUOUS COLLECTION OF CLOUD WATER dm( D) dt CL D 4 2 V ( D) E c r q c E 4 c rq c D 2 b dq dt CL E 4 c rq c 0 D 2b N( D) dd M p th moment: p ( p) D N ( D) dd 0 dq dt CL M ( 2 b ) V ( D) a D b

53 How well do the various bulk scheme predict sedimentation and sources/sinks? TM and DIAG DM schemes better than SM AND FIXED DM schemes