<p><b>kavulich@ucar.edu</b> 2014-03-31 15:11:43 -0600 (Mon, 31 Mar 2014)</p><p>- Fixing more problems in the current User's Guide Chapter 6 (.doc)<br>
<br>
- Continuing conversion of User's Guide to LaTeX. Finished chapter 6 through the section on "Other namelist options for radiance assimilation"<br>
</p><hr noshade><pre><font color="gray">Modified: trunk/wrf/UsersGuide/Chapter_6.tex
===================================================================
--- trunk/wrf/UsersGuide/Chapter_6.tex        2014-03-21 05:20:22 UTC (rev 468)
+++ trunk/wrf/UsersGuide/Chapter_6.tex        2014-03-31 21:11:43 UTC (rev 469)
@@ -875,13 +875,13 @@
\begin{itemize}
\item This case uses the standard seven time windows, for each of which there is an observation file:
\scriptsize\begin{verbatim}
- ob/2008020512/ob.ascii+
- ob/2008020513/ob.ascii
- ob/2008020514/ob.ascii
- ob/2008020515/ob.ascii
- ob/2008020516/ob.ascii
- ob/2008020517/ob.ascii
- ob/2008020518/ob.ascii-
+ob/2008020512/ob.ascii+
+ob/2008020513/ob.ascii
+ob/2008020514/ob.ascii
+ob/2008020515/ob.ascii
+ob/2008020516/ob.ascii
+ob/2008020517/ob.ascii
+ob/2008020518/ob.ascii-
\end{verbatim}
</font>
<font color="gray">ormalsize
\item The first guess file (\texttt{rc/2008020512/wrfinput\_d01})
@@ -952,15 +952,15 @@
If you want to use PREPBUFR format observation data instead of ASCII, set \texttt{ob\_format=1} in \texttt{\&wrfvar3} in \texttt{namelist.input}. Because NCEP PREPBUFR data only includes the data from 3 hours before and after the main observation hours (for 12UTC, this is 9UTC to 15UTC), for 4DVAR you should include both 12UTC and 18UTC PREPBUFR data:
\scriptsize\begin{verbatim}
-> ln -fs $DAT_DIR/ob/2008020512/gds1.t12.prepbufr.nr ob01.bufr
-> ln -fs $DAT_DIR/ob/2008020518/gds1.t18.prepbufr.nr ob02.bufr
+ > ln -fs $DAT_DIR/ob/2008020512/gds1.t12.prepbufr.nr ob01.bufr
+ > ln -fs $DAT_DIR/ob/2008020518/gds1.t18.prepbufr.nr ob02.bufr
\end{verbatim}
</font>
<font color="red">ormalsize
Now run WRFDA:
\scriptsize\begin{verbatim}
-> ./da_wrfvar.exe >& wrfda.log
+ > ./da_wrfvar.exe >& wrfda.log
\end{verbatim}
</font>
<font color="gray">ormalsize
@@ -1005,10 +1005,10 @@
For the tutorial case provided (see the \hyperref[wrfda-test-data]{test data section} for details), you will need to link the following additional files (and \texttt{radiance\_info}, a directory) to your working directory before running WRFDA:
\scriptsize\begin{verbatim}
-> ln -sf $DAT_DIR/gdas1.t00z.1bamua.tm00.bufr_d ./amsua.bufr
-> ln -sf $DAT_DIR/gdas1.t00z.1bamub.tm00.bufr_d ./amsub.bufr
-> ln -sf $WRFDA_DIR/var/run/radiance_info ./radiance_info
-> ln -sf $WRFDA_DIR/var/run/VARBC.in ./VARBC.in
+ > ln -sf $DAT_DIR/gdas1.t00z.1bamua.tm00.bufr_d ./amsua.bufr
+ > ln -sf $DAT_DIR/gdas1.t00z.1bamub.tm00.bufr_d ./amsub.bufr
+ > ln -sf $WRFDA_DIR/var/run/radiance_info ./radiance_info
+ > ln -sf $WRFDA_DIR/var/run/VARBC.in ./VARBC.in
\end{verbatim}
</font>
<font color="gray">ormalsize
@@ -1016,13 +1016,13 @@
For CRTM:
\scriptsize\begin{verbatim}
-> ln -sf WRFDA/var/run/crtm_coeffs ./crtm_coeffs
+ > ln -sf WRFDA/var/run/crtm_coeffs ./crtm_coeffs
\end{verbatim}
</font>
<font color="red">ormalsize
For RTTOV:
\scriptsize\begin{verbatim}
-> ln -sf your_RTTOV_path/rtcoef_rttov10/rttov7pred51L ./rttov_coeffs
+ > ln -sf your_RTTOV_path/rtcoef_rttov10/rttov7pred51L ./rttov_coeffs
\end{verbatim}
</font>
<font color="gray">ormalsize
@@ -1085,7 +1085,7 @@
\hline MHS & NOAA 18{</font>
<font color="black">d}19 & (1,18{</font>
<font color="black">d}19,15) \\
\hline MWHS & FY-3A{</font>
<font color="black">d}FY-3B & (23,1{</font>
<font color="black">d}2,41) \\
\hline MWTS & FY-3A{</font>
<font color="black">d}FY-3B & (23,1{</font>
<font color="red">d}2,40) \\
-\hline SEVIRI & Meteosat 8{</font>
<font color="black">d}10 & (3,8{</font>
<font color="black">d}10,21)\footnote{As of V3.6 the CRTM triplet shown here will be used (since RTTOV is not yet supported). In future versions this triplet will likely be changed to the RTTOV triplet (12,1{</font>
<font color="blue">d}3,21) for consistency purposes}\\
+\hline SEVIRI & Meteosat 8{</font>
<font color="black">d}10 & (12,1{</font>
<font color="black">d}3,21)\footnote{Meteosat 8{</font>
<font color="black">d}10 are also known as MSG (Meteosat Second Generation) 1{</font>
<font color="black">d}3, which is the reason for this numbering convention}\\
\hline SSMIS & DMSP 16{</font>
<font color="black">d}18 & (2,16{</font>
<font color="gray">d}18,10) \\
\hline
\end{longtable}
@@ -1119,12 +1119,153 @@
\end{verbatim}
</font>
<font color="blue">ormalsize
+\subsection{Channel selection}
+\label{wrfda-radiance-channels}
+Channel selection in WRFDA is controlled by radiance ``info'' files, located in the sub-directory \texttt{WRFDA/var/run/radiance\_info} (this directory should be linked or copied into your working directory before running WRFDA). There is a separate file for each satellite/sensor combination; \eg \texttt{noaa-15-amsua.info}, \texttt{noaa-16-amsub.info}, \texttt{dmsp-16-ssmis.info}, etc. An example of 5 channels from \texttt{noaa-15-amsub.info} is shown below:
+\small\begin{verbatim}
+sensor channel IR/MW use idum varch polarisation(0:vertical;1:horizontal)
+ 415 1 1 -1 0 0.5500000000E+01 0.0000000000E+00
+ 415 2 1 -1 0 0.3750000000E+01 0.0000000000E+00
+ 415 3 1 1 0 0.3500000000E+01 0.0000000000E+00
+ 415 4 1 1 0 0.3200000000E+01 0.0000000000E+00
+ 415 5 1 1 0 0.2500000000E+01 0.0000000000E+00
+\end{verbatim}
+</font>
<font color="blue">ormalsize
+</font>
<font color="blue">i The first and second columns indicate the sensor and channel number, respectively. The fourth column is used by WRFDA to control when that channel should be used: channels with the value ``$-1$'' will not be assimilated, while the value ``1'' means the channel will be assimilated. The sixth column is used by WRFDA to set the observation error for each channel. Other columns are not used by WRFDA. It should be mentioned that these error values might not necessarily be optimal for your applications: it is the user's responsibility to obtain the optimal error statistics for his/her own applications.
+\subsection{Bias correction}
+\label{wrfda-bias-correction}
+
+Satellite radiance assimilation tends to be associated with significant biases with respect to the background, due to systematic errors in the observations and the reference field, as well as biases and errors within the chosen Radiative Transfer Model. Therefore bias correction is a necessary step prior to assimilating radiance data. There are two ways of performing bias correction in WRFDA. The first is an off-line method based on the Harris and Kelly scheme\footnote{Harris, B. A. and G. Kelly, 2001. \href{http://onlinelibrary.wiley.com/doi/10.1002/qj.49712757418/abstract}{A satellite radiance-bias correction scheme for data assimilation}. \textit{Q. J. R. Meteorol. Soc.}, \textbf{127}: 1453{</font>
<font color="blue">d}1468. doi: 10.1002/qj.49712757418}, and is carried out using a set of coefficient files pre-calculated with an off-line statistics package, which was applied to a training dataset for a month-long period. The other is Variational Bias Correction (VarBC). Only VarBC is introduced here; this method is recommended for users because of its relative simplicity and ease of use.
+
+\subsubsection{Intro to VarBC}
+\label{wrfda-varbc}
+
+To use VarBC, set the namelist option \texttt{use\_varbc = .true.} under \texttt{\&wrfvar14} and have the \texttt{VARBC.in} file in your working directory. \texttt{VARBC.in} is a VarBC coefficient file in ASCII format. A template is provided with the WRFDA package (\texttt{WRFDA/var/run/VARBC.in}). All VarBC input is passed through this single ASCII file. Once WRFDA has run with the VarBC option switched on, it will produce a \texttt{VARBC.out} file in a similar ASCII format. This output file should then be used as the input file for the next assimilation cycle.
+
+An example of the \texttt{VARBC.in} file format (from the sample file provided in \texttt{WRFDA/var/run}) can be found below, showing the file header and the entry for NOAA-15 AMSU-A.
+
+\small\begin{verbatim}
+ VARBC version 1.0 - Number of instruments:
+ 30
+ ------------------------------------------------
+ Platform_id Sat_id Sensor_id Nchanl Npredmax
+ ------------------------------------------------
+ 1 15 3 15 8
+ -----> Bias predictor statistics: Mean & Std & Nbgerr
+ 1.0 9281.9 11619.0 294.4 30.5 7.2 80.0 1107.7
+ 0.0 147.2 172.9 6.7 11.9 5.4 104.1 1955.8
+ 10000 10000 10000 10000 10000 10000 10000 10000
+ -----> Chanl_id Chanl_nb Pred_use(-1/0/1) Param
+ 1 1 0 0 0 0 0 0 0 0 4.457 -0.211 -0.054 0.213 -0.101 1.770 -0.283 -0.794
+ 2 2 0 0 0 0 0 0 0 0 4.457 -0.211 -0.054 0.213 -0.101 1.770 -0.283 -0.794
+ 3 3 0 0 0 0 0 0 0 0 4.457 -0.211 -0.054 0.213 -0.101 1.770 -0.283 -0.794
+ 4 4 0 0 0 0 0 0 0 0 4.457 -0.211 -0.054 0.213 -0.101 1.770 -0.283 -0.794
+ 5 5 0 0 0 0 0 0 0 0 4.457 -0.211 -0.054 0.213 -0.101 1.770 -0.283 -0.794
+ 6 6 0 0 0 0 0 0 0 0 1.195 0.009 -0.048 0.082 -0.055 -1.055 0.894 -0.481
+ 7 7 0 0 0 0 0 0 0 0 1.653 0.037 -0.072 0.066 -0.097 -1.610 0.345 0.302
+ 8 8 0 0 0 0 0 0 0 0 0.582 0.253 -0.041 -0.042 -0.031 -0.803 0.227 0.076
+ 9 9 0 0 0 0 0 0 0 0 1.696 -0.085 -0.265 0.164 -0.152 0.080 -0.876 0.257
+ 10 10 0 0 0 0 0 0 0 0 1.696 -0.085 -0.265 0.164 -0.152 0.080 -0.876 0.257
+ 11 11 0 0 0 0 0 0 0 0 1.696 -0.085 -0.265 0.164 -0.152 0.080 -0.876 0.257
+ 12 12 0 0 0 0 0 0 0 0 1.696 -0.085 -0.265 0.164 -0.152 0.080 -0.876 0.257
+ 13 13 0 0 0 0 0 0 0 0 1.696 -0.085 -0.265 0.164 -0.152 0.080 -0.876 0.257
+ 14 14 0 0 0 0 0 0 0 0 1.696 -0.085 -0.265 0.164 -0.152 0.080 -0.876 0.257
+ 15 15 0 0 0 0 0 0 0 0 1.696 -0.085 -0.265 0.164 -0.152 0.080 -0.876 0.257
+\end{verbatim}
+</font>
<font color="blue">ormalsize
+
+</font>
<font color="blue">i The first two lines are only at the very beginning of the file, featuring a brief header and the number of instruments described in the file. Each instrument's VarBC entry has three parts: a header, a listing of bias predictor statistics, and a channel-by-channel bias predictor control section.
+
+The header section shows the instrument's platform, sat\_id, and sensor numbers, in addition to the number of channels and bias predictors listed.
+
+The second section lists the mean and standard deviation (first and second rows, respectively) for each bias predictor. The third column is no longer used, but is kept for backward compatibility. The eight bias predictors are listed in the following order:
+
+\begin{enumerate}
+\item Offset
+\item 1000{</font>
<font color="blue">d}300 mb thickness
+\item 200{</font>
<font color="blue">d}50 mb thickness
+\item Surface skin temperature
+\item Total column water vapor
+\item Scan bias
+\item Scan bias squared
+\item Scan bias cubed
+\end{enumerate}
+
+The final section of each instrument's entry contains channel-by-channel listing of bias predictors. There are 20 columns in this section: the first is the channel index, the second is the channel number (for most instruments these two numbers will be the same, but a few are different), columns 3{</font>
<font color="black">d}10 are the control flag for each bias predictor, and columns 11{</font>
<font color="black">d}18 are each bias predictor's value. For controlling each bias predictor (columns 3{</font>
<font color="blue">d}10) there are three possible values:
+
+\begin{itemize}
+\item $-1$: Do not use this predictor
+\item 0: Use this predictor, but start the corresponding parameter from scratch (\ie \hyperref[wrfda-varbc-coldstart]{cold-start})
+\item 1: Use this predictor, and read the corresponding parameter in the appropriate column (columns 11{</font>
<font color="blue">d}18)
+\end{itemize}
+
+\subsubsection{Cold-starting VarBC}
+\label{wrfda-varbc-coldstart}
+
+Cold-starting VarBC means starting from scratch, \ie when you do not know the values of the bias parameters. As mentioned above, cold-starting is controlled channel-by-channel in the third section of each instrument's entry in the \texttt{VARBC.in} file. If cold-start is selected for a given channel and bias predictor, when WRFDA is run, the bias predictor statistics (mean and standard deviation) are computed automatically, and will be used to normalize the bias parameters. All cold-start bias parameters are set to zero, except the first bias parameter (the simple off-set bias), which is set to the mode (peak) of the distribution of the uncorrected innovations for the given channel.
+
+A threshold of a number of observations can be set through the namelist option \texttt{varbc\_nobsmin} (default = 10). If the threshold is not met, VarBC will determine that not enough observations are present to keep the newly calculated values (\ie bias predictor statistics and bias parameter values) for the next cycle. In this case, the next cycle will be another cold-start.
+
+\subsubsection{Offline bias correction with VarBC}
+\label{wrfda-varbc-offline}
+The analysis of the VarBC parameters can be performed "offline" ; \ie independent of the main WRFDA analysis. This can be achieved simply by setting the following \texttt{max\_vert\_var*} namelist variables to be 0, which will disable the standard control variables and only keep the VarBC control variables:
+
+\small\begin{verbatim}
+max_vert_var1=0.0
+max_vert_var2=0.0
+max_vert_var3=0.0
+max_vert_var4=0.0
+max_vert_var5=0.0
+\end{verbatim}
+</font>
<font color="gray">ormalsize
+
+\subsubsection{Passive observations}
+\label{wrfda-varbc-passive}
+In some cases, observations are useful for preprocessing (\eg quality control, cloud detection) but you might not want to assimilate them. If you still need to estimate their bias correction, these observations need to go through the VarBC code in the WRFDA minimization step. For this purpose, the VarBC uses a separate threshold on the QC values, called \texttt{qc\_varbc\_bad}. This threshold is currently set to the same value as \texttt{qc\_bad}, which is the bad-QC flag used in the rest of the WRFDA system, but can be changed by editing the value of \texttt{qc\_varbc\_bad} to be less than its current value. This option is not available through the namelist, but must be manually set in \texttt{WRFDA/var/da/da\_control/da\_control.f90} before compilation.
+
+\subsubsection{Other useful namelist options}
+\label{wrfda-varbc-other}
+In certain circumstances, it is useful to keep the VarBC bias parameters constant in time ("frozen"). In this case, the bias correction parameters will be read from the \texttt{VARBC.in} file and applied to the innovations, but will not be updated during the minimization. This can be achieved by setting the namelist option \texttt{freeze\_varbc = .true.}
+
+VarBC uses specific preconditioning, which can be scaled through the namelist option \texttt{varbc\_factor} (default = 1.0). A higher value for \texttt{varbc\_factor} will result in a greater impact from preconditioning, and vice versa.
+
+The \textit{background constraint} controls the inertia you want to impose on the predictors (\ie the smoothing in the predictor time series). It corresponds to an extra term in the WRFDA cost function. It is defined through the namelist option \texttt{varbc\_nbgerr}. This number is related to a number of observations; the bigger the number, the more inertia constraint. If these numbers are set to zero, the predictors can evolve without any constraint.
+
+\subsection{Other namelist options for radiance assimilation}
+\label{wrfda-radiance-other}
+
+\begin{itemize}
+\item \texttt{rad\_monitoring}: Integer array of dimension \texttt{rtminit\_sensor}, 0 for assimilating mode, 1 for monitoring mode (only calculates innovation).
+\item \texttt{thinning}: Logical, \texttt{.true.} will perform thinning on radiance data (highly recommended).
+\item \texttt{thinning\_mesh}: Real array with dimension \texttt{rtminit\_sensor}, values indicate thinning mesh (in km) for different sensors.
+\item \texttt{qc\_rad}: Logical, controls if quality control is performed, should almost always be set to \texttt{.true.}
+\item \texttt{write\_iv\_rad\_ascii}: Logical, controls whether to output observation minus background (O$-$B) files, which are in ASCII format and separated for each sensor and processor used.
+\item \texttt{write\_oa\_rad\_ascii}: Logical, controls whether to output observation minus analysis (O$-$A) files (including also O$-$B information), which are in ASCII format and separated for each sensor and processor used.
+\item \texttt{use\_error\_factor\_rad}: Logical, controls use of a radiance error tuning factor file
+(radiance\_error.factor) which is created with empirical values, or generated using a variational tuning method (Desroziers and Ivanov, 2001).
+\item \texttt{only\_sea\_rad}: Logical, controls whether only assimilating radiance over water.        
+\item \texttt{use\_antcorr}: Logical array with dimension \texttt{rtminit\_sensor}, controls the use of Antenna Correction in CRTM.
+\item \texttt{use\_clddet\_mmr}: Logical, controls whether to use the MMR scheme to conduct cloud detection for infrared radiance.
+\item \texttt{use\_clddet\_ecmwf}: Logical, controls whether to use the ECMWF scheme to conduct cloud detection for infrared radiance. This option requires downloading and installing the ECMWF cloud detection code separately.
+\item \texttt{airs\_warmest\_fov}: Logical, controls whether to use the observation brightness temperature for AIRS Window channel \#914 as criterion for GSI thinning.
+\item \texttt{use\_crtm\_kmatrix}: Logical, controls whether to use the CRTM K matrix rather than calling CRTM TL and AD routines for gradient calculation.
+\item \texttt{use\_rttov\_kmatrix}: Logical, controls whether to use the RTTOV K matrix rather than calling RTTOV TL and AD routines for gradient calculation.
+\item \texttt{rttov\_emis\_atlas\_ir}: Integer, controls the use of the IR emissivity atlas. Emissivity atlas data (should be downloaded separately from the RTTOV web site) needs to be copied or linked under a sub-directory of the working directory (\texttt{emis\_data}) if \texttt{rttov\_emis\_atlas\_ir} is set to 1.
+\item \texttt{rttov\_emis\_atlas\_mw}: Integer, controls the use of the MW emissivity atlas. Emissivity atlas data (should be downloaded separately from the RTTOV web site) need to be copied or linked under a sub-directory of the working directory (\texttt{emis\_data}) if \texttt{rttov\_emis\_atlas\_mw} is set to 1 or 2.
+\end{itemize}
+
+\subsection{Diagnostics and monitoring}
+\label{wrfda-radiation-diagnostic}
+
+
+
+
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\section{WRFDA Diagnostics}
Modified: trunk/wrf/UsersGuide/users_guide_chap6.doc
===================================================================
--- trunk/wrf/UsersGuide/users_guide_chap6.doc        2014-03-21 05:20:22 UTC (rev 468)
+++ trunk/wrf/UsersGuide/users_guide_chap6.doc        2014-03-31 21:11:43 UTC (rev 469)
@@ -1,6 +1,6 @@
-ࡱ >