GitHub Issue NOAA-EMC/GSI#339: Modify setupsst.f90 to include mixed s…

…urface in situ observations, modify global_convinfo.txt to loose O-B QC test for in situ observations, and modify exglobal_atmos_analysis.sh for smaller dmesh for surface sensetive radiances for GSI/NSST

scripts/exglobal_atmos_analysis.sh

@@ -804,7 +804,7 @@ cat > gsiparm.anl << EOF
   $OBSQC
 /
 &OBS_INPUT
-  dmesh(1)=145.0,dmesh(2)=150.0,dmesh(3)=100.0,dmesh(4)=70.0,time_window_max=3.0,
+  dmesh(1)=145.0,dmesh(2)=150.0,dmesh(3)=100.0,dmesh(4)=25.0,time_window_max=3.0,
   $OBSINPUT
 /
 OBS_INPUT::

src/gsi/setupsst.f90

@@ -61,6 +61,8 @@ subroutine setupsst(obsLL,odiagLL,lunin,mype,bwork,awork,nele,nobs,is,conv_diags
 !   2017-02-09  guo     - Remove m_alloc, n_alloc.
 !                       . Remove my_node with corrected typecast().
 !   2019-11-12  li      - add 4 nsst variables to netcdf sst diag file
+!   2022-04-04  li      - modify to use mixed surface in situ observations
+!                         if ( isli == 0 ) then =>  if ( owpct > 0.05_r_kind ) then
 !
 !   input argument list:
 !     lunin    - unit from which to read observations
@@ -309,7 +311,7 @@ subroutine setupsst(obsLL,odiagLL,lunin,mype,bwork,awork,nele,nobs,is,conv_diags
      if(.not.in_curbin) cycle
 
 ! Interpolate to get sst at obs location/time
-     if ( isli == 0 ) then
+     if ( owpct > 0.05_r_kind ) then
        nobs_qc = nobs_qc + 1
        call intrp2a11(dsfct(1,1,ntguessfc),dsfct_obx,dlat,dlon,mype)
      else