Observation and Forecast Background Errors in Analysis for Rapid Update Cycle

Obs errors - in array OEQ

Background errors - in array FGQ
As of October 2003.


C *** This file contains tables of observation errors and 12-h forecast
C   errors for use in upper-air quality control and analysis routines.
 
C   Early 1986  K. Brewster     Original version
C   31 Dec 1986 S. Benjamin     Updated to include both observation
C                                and forecast errors.
C   1 Sep 1987  S. Benjamin     Subjectively reduced temp ob error
C                                from 1.0 to 0.7 deg C to force
C                                closer fit and reduced height
C                                ob errors by factor of 0.8 for
C                                same reason.
C   22 Dec 1989 S. Benjamin     Changed RH* to condensation pressure.
C                                Also, added SAO obs errors
C   30 May 1990 S. Benjamin     Added p to forecast errors, keeping
C                                same ratio to obs error as for temp.
C   04 Aug 2000 S. Weygandt     Change PC errors to lnQ errors


C   Units: heights (m), temp (deg C), condensation pressure (mb), 
C   winds (m/s)

C   The forecast error data is entered for summer and winter season.    

C     REQUIREMENTS
C    1) ATIME, the analysis time be
C       available in the routine into which it is included.
C    2) Also a dummy variable   integer*4 julday   must be declared.
C    3) 'CONFIG1' must also be included in the routine

C This version determines whether to use
C        dynamic season: 30 Sept to May 15
C  or convective season: 16 May to 29 Sept

C Dissect the date to determine julian day
C   by pulling the third through the 5th characters off of atime
C
      READ (ATIME(3:5),'(I3)') JULDAY

C Standard deviations for 12-h LFM verification 
C       from NMC Systems Evaluation Branch statistics.
C       North American 110 station network used for verification
C       These numbers are for forecast minus observation differences.
C                 /------------------------------
C       FE sd = \/ [(fcst-ob) var - ob err var]

      IF (JULDAY.GT.135  .AND.  JULDAY.LT.273) THEN  
C convective season

C       July 1985 statistics ***********************

C Heights
        FGQ(H_P,P850MB_P) = 15.0
        FGQ(H_P,P700MB_P) = 16.8  
        FGQ(H_P,P500MB_P) = 21.0 
        FGQ(H_P,P400MB_P) = 25.2  
        FGQ(H_P,P300MB_P) = 30.2  
        FGQ(H_P,P250MB_P) = 32.7
        FGQ(H_P,P200MB_P) = 35.8
        FGQ(H_P,P150MB_P) = 39.0  
C
C  Temp
        FGQ(T_P,P850MB_P) = 2.7
        FGQ(T_P,P700MB_P) = 1.8
        FGQ(T_P,P500MB_P) = 1.8
        FGQ(T_P,P400MB_P) = 1.8  
        FGQ(T_P,P300MB_P) = 1.8 
        FGQ(T_P,P250MB_P) = 2.5 
C       FGQ(T_P,P200MB_P) = 1.8 
        FGQ(T_P,P200MB_P) = 3.0 
C       FGQ(T_P,P150MB_P) = 1.7 
C increased to account for higher temp errors in stratosphere
C    28 May 96 - Stan B.
        FGQ(T_P,P150MB_P) = 3.5 
C

C  Pressure
        FGQ(P_P,P850MB_P) = 200.
        FGQ(P_P,P700MB_P) = 180.
        FGQ(P_P,P500MB_P) = 120.
        FGQ(P_P,P400MB_P) = 130.  
        FGQ(P_P,P300MB_P) = 170. 
        FGQ(P_P,P250MB_P) = 190. 
        FGQ(P_P,P200MB_P) = 180. 
        FGQ(P_P,P150MB_P) = 170. 
C
C  To estimate standard deviation errors for condensation pressure,
C    I took the RH* obs errors, figured that they were appropriate for
C    a scale of 0.0-1.0, and equated them to a scale of condensation
C    pressure deficit of 0-300 mb.
c-----SSW Aug 2000 change to reasonable values for lnQ
C  ln Q (units: none)
        FGQ(qv_P,P850MB_P) =0.50
        FGQ(qv_P,P700MB_P) =0.50
        FGQ(qv_P,P500MB_P) =0.50
        FGQ(qv_P,P400MB_P) =0.50 
        FGQ(qv_P,p300MB_P) =0.60
        FGQ(qv_P,P250MB_P) =0.70
        FGQ(qv_P,P200MB_P) =0.70
        FGQ(qv_P,P150MB_P) =0.70

C --- Set precipitable water forecast background error (mm)
        do k=1,nzp_p
        FGQ(pwob_P,k) = 5.0
        end do
C
C Winds
        FGQ(U_P,P850MB_P) = 3.3 
        FGQ(U_P,P700MB_P) = 4.0
        FGQ(U_P,P500MB_P) = 4.5 
        FGQ(U_P,P400MB_P) = 4.5 
        FGQ(U_P,P300MB_P) = 4.8 
        FGQ(U_P,P250MB_P) = 5.4 
        FGQ(U_P,P200MB_P) = 4.5 
        FGQ(U_P,P150MB_P) = 3.7 

        FGQ(V_P,P850MB_P) = FGQ(U_P,P850MB_P)
        FGQ(V_P,P700MB_P) = FGQ(U_P,P700MB_P)
        FGQ(V_P,P500MB_P) = FGQ(U_P,P500MB_P)
        FGQ(V_P,P400MB_P) = FGQ(U_P,P400MB_P)
        FGQ(V_P,P300MB_P) = FGQ(U_P,P300MB_P)
        FGQ(V_P,P250MB_P) = FGQ(U_P,P250MB_P)
        FGQ(V_P,P200MB_P) = FGQ(U_P,P200MB_P)
        FGQ(V_P,P150MB_P) = FGQ(U_P,P150MB_P)

      ELSE 
C fall,winter,spring
C       print *, ' Using dynamic season LFM error stats........'
C
C       Standard deviations for LFM (fcst-ob) differences for 12h fcsts
C       January 1985 statistics ***********************
C
C 
        FGQ(H_P,P850MB_P) = 21.6  
C15.0
        FGQ(H_P,P700MB_P) = 22.0  
C16.8  
        FGQ(H_P,P500MB_P) = 23.3  
C21.0 
        FGQ(H_P,P400MB_P) = 27.4  
C25.2  Heights
        FGQ(H_P,P300MB_P) = 31.6  
C30.2  
        FGQ(H_P,P250MB_P) = 33.2  
C32.7
        FGQ(H_P,P200MB_P) = 35.8  
C35.8
        FGQ(H_P,P150MB_P) = 38.4  
C39.0  
C
        FGQ(P_P,P850MB_P) = 200.
        FGQ(P_P,P700MB_P) = 180.
        FGQ(P_P,P500MB_P) = 120.
        FGQ(P_P,P400MB_P) = 130.  
C  Pressure
        FGQ(P_P,P300MB_P) = 170. 
        FGQ(P_P,P250MB_P) = 190. 
        FGQ(P_P,P200MB_P) = 180. 
        FGQ(P_P,P150MB_P) = 170. 

        FGQ(T_P,P850MB_P) = 2.7 
C2.7
        FGQ(T_P,P700MB_P) = 2.0 
C1.8
        FGQ(T_P,P500MB_P) = 2.0 
C1.2
        FGQ(T_P,P400MB_P) = 2.0 
C1.3  Temp
        FGQ(T_P,P300MB_P) = 2.4 
C1.7 
        FGQ(T_P,P250MB_P) = 2.6 
C1.9 
        FGQ(T_P,P200MB_P) = 3.0 
C1.8 
c       FGQ(T_P,P150MB_P) = 2.4 
C1.7 
C increased to account for higher temp errors in stratosphere
C    28 May 96 - Stan B.
        FGQ(T_P,P150MB_P) = 3.5 
C
c-----SSW Aug 2000 change to reasonable vaules for lnQ
C  ln Q (units: none)
        FGQ(qv_P,P850MB_P) = 0.50 
C 0.13
        FGQ(qv_P,P700MB_P) = 0.50 
C 0.14
        FGQ(qv_P,P500MB_P) = 0.50 
C 0.15
        FGQ(qv_P,P400MB_P) = 0.50 
C 0.15   RH*
        FGQ(qv_P,P300MB_P) = 0.60 
C 0.15
        FGQ(qv_P,P250MB_P) = 0.70 
C 0.15
        FGQ(qv_P,P200MB_P) = 0.70 
C 0.15
        FGQ(qv_P,P150MB_P) = 0.70 
C 0.15
C

C --- Set precipitable water forecast background error (mm)
        do k=1,nzp_p
        FGQ(pwob_P,k) = 5.0
        end do

        FGQ(U_P,P850MB_P) = 4.0 
C 3.3 
        FGQ(U_P,P700MB_P) = 4.5 
C 3.2 
        FGQ(U_P,P500MB_P) = 4.7 
C 3.5 
        FGQ(U_P,P400MB_P) = 5.7 
C 4.2 ! Winds
        FGQ(U_P,P300MB_P) = 6.5 
C 4.8 
        FGQ(U_P,P250MB_P) = 6.9 
C 5.4 
        FGQ(U_P,P200MB_P) = 6.5 
C 4.5 
        FGQ(U_P,P150MB_P) = 5.5 
C 3.7 

        FGQ(V_P,P850MB_P) = FGQ(U_P,P850MB_P)
        FGQ(V_P,P700MB_P) = FGQ(U_P,P700MB_P)
        FGQ(V_P,P500MB_P) = FGQ(U_P,P500MB_P)
        FGQ(V_P,P400MB_P) = FGQ(U_P,P400MB_P)
        FGQ(V_P,P300MB_P) = FGQ(U_P,P300MB_P)
        FGQ(V_P,P250MB_P) = FGQ(U_P,P250MB_P)
        FGQ(V_P,P200MB_P) = FGQ(U_P,P200MB_P)
        FGQ(V_P,P150MB_P) = FGQ(U_P,P150MB_P)

       ENDIF  
Cseason determination

C  THE OBSERVATIONAL ERRORS ARE ASSUMED TO BE INDEPENDENT OF THE SEASON

C   Units: heights (m), temp (deg C), RH* (nd), winds (m/s)
C
C Set std. deviation of height and wind observation error. 
C   These figures
C        (for RAOBs) are taken from Dey and Morone (1985, MWR).
C    They are also used for the Regional OI at NMC.
C ---------------------------------------------------------------------
C-
C Heights
        OEQ(H_P,P850MB_P,RAOB_P) = 8.0  * 0.8
        OEQ(H_P,P700MB_P,RAOB_P) = 8.6  * 0.8 
        OEQ(H_P,P500MB_P,RAOB_P) = 12.1 * 0.8
        OEQ(H_P,P400MB_P,RAOB_P) = 14.9 * 0.8   
        OEQ(H_P,P300MB_P,RAOB_P) = 18.8 * 0.8
        OEQ(H_P,P250MB_P,RAOB_P) = 25.4 * 0.8
        OEQ(H_P,P200MB_P,RAOB_P) = 27.7 * 0.8
        OEQ(H_P,P150MB_P,RAOB_P) = 32.4 * 0.8
C
C *** WIND OBSERVATION ERRORS FOR RAOBS FROM DEY AND MORONE ARE
C       AS FOLLOWS FOR MANDATORY LEVELS 850-150 /1.8,2.1,3.4,4.2,5.2
C    5.9,5.9,5.9/

C       THE VALUES USED HERE ARE MUCH SMALLER AND ARE EQUAL TO
C    8% OF THE IMPLICIT FIRST GUESS WIND ERROR USING THE
C       SOAR AND LFM HEIGHT ERRORS.
C       --------------------------------------------------------------

C *** RAOBS
C Winds
        OEQ(U_P,P850MB_P,RAOB_P) = 0.83
        OEQ(U_P,P700MB_P,RAOB_P) = 0.87
        OEQ(U_P,P500MB_P,RAOB_P) = 0.92
        OEQ(U_P,P400MB_P,RAOB_P) = 1.01         
        OEQ(U_P,P300MB_P,RAOB_P) = 1.15
        OEQ(U_P,P250MB_P,RAOB_P) = 1.20
        OEQ(U_P,P200MB_P,RAOB_P) = 1.24
        OEQ(U_P,P150MB_P,RAOB_P) = 1.38

c-----SSW Aug 2000 change to reasonable vaules for lnQ
c-----SGB Oct 2000 revision - higher values, esp. aloft
C  ln Q (units: none)
        OEQ(QV_P,P850MB_P,RAOB_P) = 0.10
        OEQ(QV_P,P700MB_P,RAOB_P) = 0.10
        OEQ(QV_P,P500MB_P,RAOB_P) = 0.10
        OEQ(QV_P,P400MB_P,RAOB_P) = 0.15         
        OEQ(QV_P,P300MB_P,RAOB_P) = 0.15
        OEQ(QV_P,P250MB_P,RAOB_P) = 0.18
        OEQ(QV_P,P200MB_P,RAOB_P) = 0.18
        OEQ(QV_P,P150MB_P,RAOB_P) = 0.18
C                                
         DO 31415 K   =1,NZP_P
           OEQ(V_P,K,RAOB_P) = OEQ(U_P,K,RAOB_P) 
           OEQ(T_P,K,RAOB_P) = 0.4        
C              DEG C
C *** changed from 0.7 to 0.4 - 28 Oct 87 - Stan B.
C       because of relatively poor fit to temp obs
           OEQ(P_P,K,RAOB_P) = 50.         
C                                mb
31415   CONTINUE

C *** PROFILER
         DO 31416 K   =1,NZP_P
          OEQ(H_P,K,PROFILER_P) = OEQ(H_P,K,RAOB_P)
          OEQ(T_P,K,PROFILER_P) = 0.8   
C                                DEG C
c-----SSW Aug 2000 change to reasonable vaules for lnQ
C  ln Q (units: none)
          OEQ(qv_P,K,PROFILER_P) = 0.40  
C                                 
          OEQ(U_P,K,PROFILER_P) = 1.0   
C                                 Profiler wind error
          OEQ(V_P,K,PROFILER_P) = 1.0
          OEQ(P_P,K,PROFILER_P) = 100.  
C                                 mb

C *** AIRCRAFT REPORTS
C --- When automated aircraft reports are received, they will
C       have to be treated as a whole new data type by the MAPS
C       system because of their different error characteristics.
C     ----------------------------------------------------------
          OEQ(H_P,K,PIREPS_P) = 50.0    
C                                 m
          OEQ(T_P,K,PIREPS_P) = 1.0     
C                                 DEG C
c-----SSW Aug 2000 change to reasonable vaules for lnQ
C  ln Q (units: none)
          OEQ(qv_P,K,PIREPS_P) = 0.40
C                                 
          OEQ(U_P,K,PIREPS_P) = 1.5    
          OEQ(V_P,K,PIREPS_P) = 1.5
          OEQ(P_P,K,PIREPS_P) = 100.    
C                                 mb
C *** SAO
          OEQ(P_P,K,SAO_P) = 20.
          OEQ(H_P,K,SAO_P) = 6.
          OEQ(T_P,K,SAO_P) = 0.8
c-----SSW Aug 2000 change to reasonable vaules for lnQ
C  ln Q (units: none)
          OEQ(qv_P,K,SAO_P) =0.08
          OEQ(U_P,K,SAO_P) = 0.5
          OEQ(V_P,K,SAO_P) = 0.5

C *** Satellite
          OEQ(T_P,K,SAT_P) = 2.0        
C                                DEG C
c-----SSW Aug 2000 change to reasonable vaules for lnQ
C  ln Q (units: none)
          OEQ(qv_P,K,SAT_P) = 0.40
C                                
          OEQ(H_P,K,SAT_P) =30.0   
          OEQ(U_P,K,SAT_P) = 2.0   
          OEQ(V_P,K,SAT_P) = 2.0
C                                m/s
C *** Precipitable water observation errors - mm
          OEQ(pwob_P,k,SAT_P) = 3.0
          OEQ(pwob_P,k,GPS_P) = 1.5
          OEQ(pwob_P,k,RAOB_P) = 1.0
31416   CONTINUE