Non-physical Kain-Fritsch Precipitation Bands

The Problem

As seen in the lower left quadrant of the lefthand image above, the accumulated non-convective precipitation at hour 60 in the forecast from 0Z July 19, 2011 has a non-physical linear appearance in the model's u direction. In this case, the winds are blowing west to east in this southern portion of the East Pacific High. The case studies described below indicate that this problem is due to the new Kain-Fritsch cumulus scheme as found in WRF versions 3.1.1 and 3.3.

As seen in the image above right, standing waves of precipitation develop in this case from 2012101912. For more on this case, see below.

Case Study Plots

Product Forecast Hour Case

The Cases

Name	Abbreviation	WRF Version	WRF cu_physics	Description
1. Kain-Fritsch	kf	3.1.1	1	Real-time, version 3.1.1 with standard, new Kain-Fritsch Scheme
2. Old Kain-Fritsch	kf99	3.1.1	99	Older Kain-Fritsch scheme does not have the problem.
3. Grell	grell	3.1.1	3	Grell-Devenyi ensemble scheme does not have the problem.
4. Grell5	grell5	3.1.1	5	New Grell 3d ensemble cumulus scheme does not have the problem.
5. Kain-Fritsch in WRF3.3	kfv33	3.3	1	Kain-Fritsch scheme as found in WRF version 3.3 still has the problem.
6. Old Kain-Fritsch WRFV3.3	kf99v33	3.3	99	Old Kain-Fritsch in WRFV3.3 does not have the problem.
7. Kain-Fritsch No FDDA Ptop50	kfptop50nofdda	3.1.1	1	Kain-Fritsch without FDDA analysis nudging and with the model top at 50 mb instead of 100 mb.
8. WRF v3.4.1 Kain-Fritsch Ptop50	wrfv341kf	3.4.1	1	Kain-Fritsch with the model top at 50 mb instead of 100 mb.
9. WRF v3.4.1 Grell Ptop50	wrfv341grell	3.4.1	1	Grell with the model top at 50 mb instead of 100 mb.

Table 2. Additional 12-km and 4-km Cases
Initialization	Domain	Product	Hour	Cases (control click to select multiple)
Choose a date:	Choose a domain 36-km 12-km 4-km	All products ------ Moisture --------- Low clouds sfc to 3000 feet Low clouds 3 - 10Kft Middle clouds 10-20Kft High clouds about 20Kft 700 mb Vertical Velocity OLR Integrated Cloud 1-hour Precip 3-hour Precip 24-hour Precip Western Washington 24-hour Precip 500mb Vorticity --- Vertical Cross Sections (Nov 30 case only) -- N-S at 126W, total cloud mr N-S at 126W, total precip mr N-S at 126W, rain water mr N-S at 126W, vertical velocity N-S at 127W, total cloud mr N-S at 127W, total precip mr N-S at 127W, rain water mr N-S at 127W, vertical velocity	Choose an hour Hour 0 Hour 6 Hour 12 Hour 18 Hour 21 Hour 24 Hour 30 Hour 33 Hour 36 Hour 48 All hours

All Cases
1. Std (aka wrfv341kf99)
1.1 Std but no cumulus scheme on 4-km (wrfv341kf99no4k)
1.2 Std but with 47 vertical levels and model top at 25 mb (wrfv341kf99z47)
2. Betts-Miller-Janjic (wrfv341bmj)
2.1 Betts-Miller-Janjic without 4-km cumulus scheme (wrfv341bmjno4k)
3. Old GFS simplified Arakawa-Schubert (wrfv341cu4)
3.1 Old GFS simplified Arakawa-Schubert without 4-km cumulus scheme (wrfv341cu4no4k)
3.2 New GFS simplified Arakawa-Schubert (wrfv341cu14)
4. Modified Tiedtke scheme (wrfv341cu6)
4.1 Modified Tiedtke scheme without 4-km cumulus scheme (wrfv341cu6no4k)
5. Grell-Devenyi ensemble scheme (wrfv341grell)
5.1 Grell-Devenyi ensemble scheme without 4-km cumulus scheme (wrfv341grellno4k)
5.2 wrfv341grell2way Grell with 2-way nests (wrfv341grell2way)
5.3 Grell with reduced boundary zone (wrfv341grellzonediff)
6. Grell 3D ensemble scheme (wrfv341grell5)
6.1 Grell 3D ensemble scheme with shallow cumulus turned on (wrfv341grell5shcu)
6.2 Grell 3D ensemble scheme without 4-km cumulus scheme (wrfv341grell5no4k)
7. wrfv341kf -- Standard Kain-Fritsch (wrfv341kf)
7.1 wrfv341kf -- Standard Kain-Fritsch without 4-km cumulus scheme (wrfv341kfno4k)
8. Previous model setup, 3.1.1, old vertical levels (oct11)
9. Single, larger 4-km domain without cumulus scheme (wrfv341big4)
10. MM5-NAM (mm5)

Non-physical Kain-Fritsch Precipitation Bands

The Problem

Case Study Plots

The Cases

Table 2. Additional 12-km and 4-km Cases