1.8. Build CROCO river forcing

Unlike data from global ocean reanalyses, river data are heterogeneous and do not follow the same formalism. The make_rivers.py is a script that attempts to compensate for this lack of formalism by being able to read input data from river reanalysis (e.g. CEMS), or observations (e.g. in Europe Marinesitu).

Warning

Unlike the matlab tool, this script does not allow you to build a river forcing file from a climatology. However, it is possible possible to create a one-year, 365.25-day cycle.

Note

The script is currently unable to manage variations in river temperature and salinity. The cpp key to be used during simulation is PSOURCE_NCFILE. This allows us to take into account only a flow with a default temperature and salinity set in croco.in.

1.8.1. Input list

Unlike previous scripts, make_rivers does not use a reader to decrypt data. This would make the task more complicated, since on a zone, several types of data can co-exist. The list of rivers to be taken into account is given to the script in a text file following this formalism:

# input_file     |    Lon    |  Lat/Qmin  |
river.nc            -2.05        47.30
loire.dat           -2.05        47.30
Rhone.txt            4.82        43.37
2ddata.nc             X            10

Each column represents, respectively, the river file, the longitude, and the latitude of the river mouth. The script can also process text data, where each line represents the river flow at a specific date such as:

1950/01/01 12:00:00 730.000000
1950/01/02 12:00:00 695.000000
1950/01/03 12:00:00 675.000000
1950/01/04 12:00:00 645.000000
1950/01/05 12:00:00 610.000000
1950/01/06 12:00:00 600.000000

Each line must follow the format YYYY/MM/DD hh:mm:ss Qsrc where Qsrc is the river flow at the time YYYY/MM/DD hh:mm:ss

Note

In the case of a netcdf file, the variable containing the river flow must have an element from this list as a unit attribute. ['m3 s-1','m**3 s**-1','m3/s'].

If the input file is a 2-D reanalysis of a region, and no longitude is specified, the code will automatically determine the position of rivers meeting the minimum flow condition. In the example above, in the 2ddata.nc line we therefore indicate X for longitude data and 10 m3/s for the minimum flow to be considered.

1.8.2. Using make_rivers

Before running make_rivers.py, USER CHANGES section needs to be filled. There are several parts in it:

# Dates
Yorig = 2013 # year origin of time : days since Yorig-01-01
Ystart, Mstart = 2013,1   # Starting month
Yend,   Mend   = 2013,3   # Ending month

# Input data informations
input_dir = './' # where .txt file can be found
input_file = 'river_list.txt' # txt file name
Crange = 3 # only used in 2D reanalysis, number of iterations to broaden the coast mask

# CROCO path and filename informations
croco_dir = '../../CROCO_FILES/'
croco_grd = 'croco_grd.nc'

# Rivers file informations
river_filename = 'croco_runoff.nc' # output will be put in croco_dir by default
river_file_format = "MONTHLY" # How outputs are split (MONTHLY,YEARLY,FULL)
rivers_cyl = 0. # if cycle is needed
rivers_output_frequency = 'DAILY' # output frequency for rivers (HOURLY,DAILY,MONTHLY depending on your data)

Origin year, along with starting and ending date are defined.

Informations on the location and name of the text file containing rivers to consider.

Note

The Crange variable is only used in the case of two-dimensional data and allows the user to widen the land/sea interface zone to take into account more rivers (some of which are not connected to the sea, as they should be, due to the resolution of the data).

Informations on where to find CROCO grid.

Informations about river output file. Several format (river_file_format) are available to specify how to separate data (monthly, yearly files). As river data may come from several sources, they may not have the same temporal distribution and frequency. To co-locate temporally, we need to choose the data output frequency by setting rivers_output_frequency.

To use make_rivers.py do:

python make_rivers.py

Note

The river flow components are located on U/V grid points.

The script can open a graphic interface for helping you positionning the rivers after a first “guess” made using the input parameters. To use it, simply answer yes to the question:

Do you want to manually edit river positions using interactive GUI ?: y,[n]

The following interface will show up:

A loop will be performed on all rivers to adjust the position. Several informations are avaible on the plot:

• Green cross: position in the input text file

• Red dot with black arrow: first guess position at land/ocean interface and direction

The basic commands of the window are:

Zoom in	Magnifying glass on top row
Show/hide mesh	Button in right column
Change point position	Right click on the desired cell (Cyan dot)
Change flow direction	Lower row
Add a new point for the same river	Button in the right column

On the graphical interface, the active point is colored in cyan, while the others are colored in green. When adding a new point for the same river, the flow rate will be divided by the total number of points, so that each point has the same flow rate.

To pass to the next river click on Save chosen position and leave.

After finishing editing river positions, a summury of rivers in the domain can be shown by answering yes to the question

Do you want a summary of the position of the rivers ?: y,[n]

At the end of the script you will have:

• croco_runoff.nc containing river inputs

• for_croco_in.txt a text file to copy in psource section of croco.in