Ro Rongqia ian Y Yang, J Jesse Meng Meng , , M Mich chael Ek - - PowerPoint PPT Presentation

Ro Rongqia ian Y Yang, J Jesse Meng Meng , , M Mich chael Ek Ek a and He Helin lin W Wei,

The EM EMC L Land/Hyd ydro T Team EM EMC/NCEP EP/NWS 5200 A Auth R Road, C Camp S Springs, M MD 2 D 20746, USA USA

 Assess skills of the NCEP CFS v2 in

predicting SST, Precipitation, and T2m anomalies.

 Examine impact of land surface

parameterizations on summer season predictions with the new CFS.

The Atmospheric Forecast Model is increased from T62, L , L28 to T126, L , L64 (~100 km) resolution and equipped with more advanced physics.

The Land Surface Model is upgraded from the 2-layer OSU U to the 4-layer Noah h LS LSM.

Introduction of a 3-layer global Sea Ic Ice M Model.

Fully Coupled Ocean-Land-Atmosphere System, implemented in March, 2011

Glo Global Ga l Gauge Di Distribution n Pr Precip cip d differenc nce ( (GF GFS-C

• CMAP)

• The

he c critical la l land nd s surface s ski kin t n temperature d depend nds o

• n r

n ratio o

• f a

and nd , , seasona nal G l GVF ( (with c h cons nstant nt L LAI) I), a , and nd o

• the

hers. .

• Bare s

soil a l and nd v vegetation a n are t treated t togethe her ( (one ne la layer), In N , In Noah, h, , a , a t tuna nable le p parame meter ( (varies i in d n different nt o

• perationa

nal mo l models ls), i , is used t to c compute , i , is p prescribed f for a all g ll grid c cells lls d depend nding ng o

• n v

n vegetation n typ ypes., V ., Von K n Karma man c n cons nstant nt k=0.4 .4, i , is t the he mo mole lecula lar v viscosity. . T The he p phys ysical c l cons nstraint nt s sho hould ld b be t the he c convergenc nce o

• f t

turbule lent nt f flu luxes a and nd t to bare s soil v l valu lues ( (i.e .e., , , a ., , , and nd d displa laceme ment nt he height ht) w whe hen t n the he a above b bioma mass approache hes z zero. .

 The Noah LSM has a cold bias of around 10 K in the early afternoon of summer over

semiarid regions.

 The previous efforts to reduce the bias were focused on the tunable parameter

by adjusting its value or taking as a function of vegetation height h, e.g. , = (Chen and Zhang, 2009).

 However, there is no vegetation height input to the Noah LSM. The derived from

the corresponding vegetation height would lead to an overestimation of , suggesting that the problem can’t be fixed by just tuning the parameter and the prescribed also needs to be adjusted, by explicitly applying the physical constraint. Following Zeng and Wang (2007), the bare soil roughness length is taken as 0.01, effective roughness length for momentum is , the maximum Green Fractional Cover is , and the prescribed roughness for momentum is .

 GP

GPCP P Pent ntad P Precipitation A n Ana nalys lysis f for pr precipit cipitat atio ion ( (Xi Xie e et a al.,2 l.,2003). .

 GH

GHCN/CAMS ( (la land nd) T T2m A m Ana nalys lysis f for T2m T2m ( (Fan a n and nd V Van d n den n Do Dool, 2 , 2008). .

 NOAA O

Optimu mum Int m Interpola lation ( n (OI) I) S SST f for SS SST ( (Reyno ynold lds, 1 , 1988). .

 Ano

noma maly c ly correla lation i n is u used a as a a me measure o

• f t

the he s ski kills lls f for mo mont nths hs o

• f M

May a y and nd June ne. .

References: Fan, Y., and H. van den Dool (2008), A global monthly land surface air temperature analysis for 1948- present, J. Geophys. Res., 113, D01103, doi: 10.1029/2007JD008470. Xie, P. and Coauthors, 2003: GPCP Pentad Precipitation Analyses: An Experimental Dataset Based on Gauge Observations and Satellite Estimates. J. Climate, 16, 2197–2214. doi: 10.1175/2769.1.

High skill globally for lead 0, decreases with lead 1 (mid-latitude), still maintains good performance over most of the globe, especially over the Nino regions

Cont ntrol Experime ment ntal l Ma May June ne

No surprise, small difference in lead 0, initial ocean conditions is the main control and land impact is very small Slightly better over the Pacific mid- latitudes and equatorial Atlantic

• cean, still small
• ver the tropics

Experime ment ntal - c l - cont ntrol l Ma May June ne

Higher skill and similar patterns in lead 0, decreases substantially in lead 1. As expected, the decrease is relatively small in the Southern Hemisphere (cold)

Cont ntrol Experime ment ntal l Ma May June ne

Mixed picture, varies with regions in the N.H., small changes in the patterns with both leads in the S.H.

Experime ment ntal - c l - cont ntrol l Ma May June ne

Cont ntrol

Patterns similar to the global, no big difference in lead 0. Skill gain/loss varies with different climate regimes in lead 1

Ma May June ne Experime ment ntal l CONUS CONUS

Better over the northwest Pacific states in lead 1, worse

• ver the

east (New England region) in both leads

Experime ment ntal - c l - cont ntrol l Ma May June ne CONUS CONUS

Higher skill than precipitation and close to each other in lead 0 decreases substantially in lead 1 over both hemispheres

Cont ntrol Experime ment ntal l Ma May June ne

Similar to the global skill, the difference varies with regions Main difference in Mid to high latitudes

Experime ment ntal - c l - cont ntrol l Ma May June ne

Higher skill in lead 0, decreases substantially in lead 1,

Cont ntrol Experime ment ntal ( l (Z0) Ma May June ne CONUS CONUS

Better over most of the CoNUS, especially over central great plains in lead 0. Higher skill over the Northwest Pacific and mid-Atlantic regions in lead 1 May come from better ocean

Experime ment ntal - c l - cont ntrol l Ma May June ne CONUS CONUS

The skill gain mainly comes from the disagreements between the two configurations

Ma May June ne CONUS CONUS

 The new formulations generally lead to a better skill in

predicting T2m over the CONUS in the first month and the skill gain/loss varies with different climate regimes for the second month globally.

 The changes made to the roughness lengths have a

relatively small impact on the precipitation skill, suggesting that the ocean and atmosphere are still the dominant controls over warm season precipitation for relatively short leads.

 The impact is also affected by the land-atmosphere

coupling strength. The differences mainly show up in the second month due to the coupled nature. An examination of the atmospheric circulation could be very useful.

 A careful treatment to land surface parameterization is

important to mid-range/seasonal predictions.

 More years may be needed to confirm the patterns.