June 26, 2016 in Hindcast 202007F¶

on 83rd time step goes NaN (physics) 84th (biology)

In [1]:

import xarray as xr

In [2]:

gT = xr.open_dataset('/home/sallen/202007/202007F/SalishSea_1h_20160626_20160630_grid_T_20160626-20160626.nc')

In [3]:

gT.votemper[0, 0].plot();

Out[3]:

<matplotlib.collections.QuadMesh at 0x7f69404db760>

In [4]:

gU = xr.open_dataset('/home/sallen/202007/202007F/SalishSea_1h_20160626_20160630_grid_U_20160626-20160626.nc')

In [5]:

gU.vozocrtx[0, 0].plot();

In [8]:

gT.vosaline[0, 4, 700:900, 0:250].plot();

So we have NaN about 100 grid points from the boundary. That is 100 barotropic grid points or about 5 baroclinic time steps. We'd expect that on time step 90 we would see 7*20: about 140 grid cells impacts.

In [13]:

gT.sossheig[0].plot(vmax=1, vmin=-1, cmap='plasma');

However,¶

We note that the lowR run went unstable on the first time step for 21Jul15 and that the maximum velocity at the first time step here is 15.2 m/s. Things appear to have calmed down before the instability above, but a smooth restart might easily solve this problem.

Note that I originally thought the dates matched but they are 11 months out. Also lowR is western boundary, this is northern boundary. So not clear that this is a starting boundary problem.

Rivers¶

In [16]:

rivers = xr.open_dataset('/home/sallen/MEOPAR/Tools/I_ForcingFiles/Rivers/ncfiles/R202102Dailies_y2016m06d26.nc')

In [17]:

rivers

Out[17]:

<xarray.Dataset>
Dimensions:       (time_counter: 1, x: 398, y: 898)
Coordinates:
  * x             (x) int64 0 1 2 3 4 5 6 7 ... 390 391 392 393 394 395 396 397
  * y             (y) int64 0 1 2 3 4 5 6 7 ... 890 891 892 893 894 895 896 897
  * time_counter  (time_counter) int64 0
Data variables:
    rorunoff      (time_counter, y, x) float64 ...
Attributes:
    acknowledgements:      Based on river fit
    creator_email:         sallen@eoas.ubc.ca
    creator_name:          Salish Sea MEOPAR Project Contributors
    creator_url:           https://salishsea-meopar-docs.readthedocs.org/
    institution:           UBC EOAS
    institution_fullname:  Earth, Ocean & Atmospheric Sciences, University of...
    title:                 Rivers for y2016m06d26
    notebook:              MakeDailyRiverNCfiles.ipynb
    rivers_base:           river_202101
    summary:               Daily Runoff
    history:               [2021-02-02] File creation.

In [18]:

rivers.rorunoff.max()

Out[18]:

<xarray.DataArray 'rorunoff' ()>
array(24.70158577)

In [19]:

rivers.rorunoff.min()

Out[19]:

<xarray.DataArray 'rorunoff' ()>
array(0.)

In [23]:

rivers.rorunoff[0, 800:, 50:150].plot(cmap='gist_ncar_r');

In [28]:

sshb = xr.open_dataset('/results/forcing/sshNeahBay/obs/ssh_y2016m06d25.nc')
ssh = xr.open_dataset('/results/forcing/sshNeahBay/obs/ssh_y2016m06d26.nc')

In [25]:

ssh

In [29]:

ssh.sossheig[:, 0, 40].plot()
sshb.sossheig[:, 0, 40].plot()

Out[29]:

[<matplotlib.lines.Line2D at 0x7f69305d6760>]

Whatever, smooth the boundary conditions and see if that works (if not add the NaN code)¶

In [30]:

import matplotlib.pyplot as plt
import numpy as np

%matplotlib inline

In [31]:

restart_file = xr.open_dataset('/data/sallen/results/MEOPAR/202007/202007F/SalishSea_04324320_restart.nc')
restart_file

In [32]:

ib, ie = 40, 60
uveln19 = restart_file.un[0, 25-1]
uvelb19 = restart_file.ub[0, 25-1]
uvelb19[896, ib:ie].plot(label="Before 1", linewidth=5)
uveln19[896, ib:ie].plot(label="Next 1")
uvelb19[894, ib:ie].plot(label="Before 3", linewidth=5)
uveln19[894, ib:ie].plot(label="Next 3")
plt.plot([45-ib, 45-ib], [-0.5, 2], color='k');
plt.legend();

In [33]:

restart_file.un[0, :, 896, :] = restart_file.un[0, :, 895, :]
restart_file.ub[0, :, 896, :] = restart_file.ub[0, :, 895, :]
restart_file.vn[0, :, 895, :] = restart_file.vn[0, :, 894, :]
restart_file.vb[0, :, 895, :] = restart_file.vb[0, :, 894, :]

In [34]:

restart_file.attrs['Modified'] = 'copied velocity one away from north boundary to boundary'
restart_file.attrs

Out[34]:

{'file_name': 'SalishSea_04324320_restart.nc',
 'TimeStamp': '21/03/2021 02:40:06 -0700',
 'Modified': 'copied velocity one away from north boundary to boundary'}

In [35]:

restart_file.to_netcdf('SalishSea_04324320_restart.nc')

Looks like its running okay now.

In [ ]: