make_input_files -- E_Whidbey¶
Notebook to make an array of fgmax points based on a RuledRectangle and parameters set in params.py, and a RuledRectangle covering these points for AMR flagging.
Also make the allow_initially_wet array for this area, if needed. This part of the notebook may need to be customized.
In the next cell use %matplotlib notebook for live interactive plots or %matplotlib inline for non-interactive plots.
%matplotlib inline
from pylab import *
import os,sys
from imp import reload
from clawpack.geoclaw import topotools, dtopotools
from clawpack.visclaw import colormaps
Set some values needed below:¶
loc = 'E_Whidbey' ## SET THIS to name to be used for this study region
save_figs = True # make png files for figures?
try:
rootdir = os.environ['WA_EMD_2019']
except:
print('*** Need to set environment variable WA_EMD_2019')
print('Set rootdir to ', rootdir)
# directory for topo files:
topodir = rootdir + '/topo'
# directories where some python modules can be found:
sys.path.insert(0, rootdir + '/new_code')
sys.path.insert(0, rootdir + '/topo')
import fgmax_tools, fgmax_routines, kmltools # from new_code
import region_tools, marching_front # from topo
#reload(marching_front) # reload if debugging module
def savefigp(fname):
global save_figs
if save_figs:
savefig(fname, bbox_inches='tight')
print('Created ', fname)
else:
print('save_figs = False')
Load Google Earth image, if available¶
GEdir = rootdir + '/info/fgmax_regions'
try:
GEfile = GEdir + '/%s_bbox_GE.png' % loc
GEmap = imread(GEfile)
GEextent_file = GEdir + '/%s_bbox_GE_extent.txt' % loc
GEextent_tokens = open(GEextent_file,'r').read().strip().split(',')
GEextent = [float(s) for s in GEextent_tokens]
print('GE extent: ', GEextent)
except:
print('Problem reading Google Earth image file or extent')
GEmap = None
if GEmap is not None:
figure(figsize=(8,8))
imshow(GEmap, extent=GEextent)
gca().set_aspect(1./cos(48*pi/180.))
ticklabel_format(useOffset=False)
xticks(rotation=20);
zmin = -60.
zmax = 40.
land_cmap = colormaps.make_colormap({ 0.0:[0.1,0.4,0.0],
0.25:[0.0,1.0,0.0],
0.5:[0.8,1.0,0.5],
1.0:[0.8,0.5,0.2]})
sea_cmap = colormaps.make_colormap({ 0.0:[0,0,1], 1.:[.8,.8,1]})
cmap, norm = colormaps.add_colormaps((land_cmap, sea_cmap),
data_limits=(zmin,zmax),
data_break=0.)
sea_cmap_dry = colormaps.make_colormap({ 0.0:[1.0,0.7,0.7], 1.:[1.0,0.7,0.7]})
cmap_dry, norm_dry = colormaps.add_colormaps((land_cmap, sea_cmap_dry),
data_limits=(zmin,zmax),
data_break=0.)
PTPSm = topotools.read_netcdf(topodir + '/topofiles/PTPS_merged_13s.nc')
Ruled rectangle used as boundary polygon for fgmax points¶
We will select fgmax points that lie within this polygon, with other conditions also imposed later.
slu = \
array([[-122.750, 48.220, 48.240],
[-122.623, 48.187, 48.320],
[-122.581, 48.120, 48.320],
[-122.575, 48.040, 48.320],
[-122.550, 48.005, 48.290],
[-122.515, 48.005, 48.250],
[-122.500, 48.030, 48.170],
[-122.450, 48.050, 48.120]])
# [ 48.036, -122.578, -122.448],
# [ 48.12 , -122.581, -122.448],
# [ 48.187, -122.623, -122.5 ],
# [ 48.191, -122.684, -122.5 ],
# [ 48.221, -122.755, -122.5 ],
# [ 48.33, -122.66, -122.58 ]])
fg_poly = region_tools.RuledRectangle(slu=slu)
fg_poly.ixy = 1
fg_poly.method = 1
# Cropped topo file:
#extent = [-122.54, -122.41, 47.9, 48.04] # custom extent if desired
x1,x2,y1,y2 = fg_poly.bounding_box()
padding = 5./(3*3600.) # add a few grid cells around bounding box
extent = [x1-padding, x2+padding, y1-padding, y2+padding]
print('Cropping PTPSm topo to: ', extent)
topo = PTPSm.crop(extent)
extent = topo.extent
print('New extent of topo: ', extent)
fg_poly.write('fg_poly_%s.slu' % loc)
def plot_topo(Z, show_cb=True):
#figure(figsize=(8,8))
pc = pcolormesh(topo.X, topo.Y, Z, cmap=cmap, norm=norm)
if show_cb:
cb = colorbar(pc)
cb.set_label('meters')
axis(extent)
gca().set_aspect(1./cos(48*pi/180.))
ticklabel_format(useOffset=False)
xticks(rotation=20)
figure(figsize=(8,8))
plot_topo(topo.Z)
xpoly,ypoly = fg_poly.vertices()
plot(xpoly,ypoly,'w')
title('Topography and fg_poly')
grid(True)
savefigp('fg_poly_%s.png' % loc)
Make kml files¶
Make a kml file showing the polygon used above, and one showing the bounding box as a rectangle. The latter is useful for capturing a Google Earth image of the fgmax region in order to plot or topography on top of the image.
kmltools.poly2kml((xpoly,ypoly), name='%s_fgmax_poly' % loc, fname='%s_fgmax_poly.kml' % loc)
kmltools.box2kml(extent, name='%s_fgmax_bbox' % loc, fname='%s_fgmax_bbox.kml' % loc)
Select fgmax points within this polygon¶
First use the criterion that points are selected everywhere in the bounding box rectangle, as specified by the extent of the topo array, and if one of these conditions holds:
- the point is offshore at any depth of water, i.e.,
Z < 0, - the point lies within
onshore_buffergrid points of the shoreline, or - the topography satisfies
B < onshore_Z, and is connected to the shore by a chain of grid points also satisfying this condition.
The latter requirement avoids selecting points in a local depression that are separated from the coast by high ground that won't be overtopped.
This is accomplished by a marching front algorithm defined in rootdir + '/topo/marching_front.py'.
# mask the points not in fg_poly:
mask_out = fg_poly.mask_outside(topo.X, topo.Y)
#mask_fg = logical_or(mask_out, mask_near)
fgmax_pts_chosen = marching_front.select_by_flooding(topo.Z, mask=mask_out,
prev_pts_chosen=None,
Z1=0, Z2=1e6, max_iters=10)
fgmax_pts_chosen = marching_front.select_by_flooding(topo.Z, mask=mask_out,
prev_pts_chosen=fgmax_pts_chosen,
Z1=0, Z2=15., max_iters=None)
# mask the points not selected above:
mask_fg = logical_not(fgmax_pts_chosen)
# topography as masked array:
Z_fg = ma.masked_array(topo.Z, mask_fg)
print('Selected %i fgmax points' % Z_fg.count())
figure(figsize=(8,8))
plot_topo(Z_fg)
plot(xpoly,ypoly,'r')
title('%i fgmax points in specified polygon' % Z_fg.count())
grid(True)
savefigp('fgmax_points_%s.png' % loc)
Write out the masked array indicating fgmax points¶
This is written in the style of a topo_type=3 topography file, with a header followed point values at all points on a uniform grid. The values are simply the integer 1 for points that should be used as fgmax points and 0 for other points. Note that format %1i is used for compactness.
fname_fgmax_mask = 'fgmax_pts_topostyle_%s.data' % loc
topo_fgmax_mask = topotools.Topography()
topo_fgmax_mask._x = topo.x
topo_fgmax_mask._y = topo.y
topo_fgmax_mask._Z = where(Z_fg.mask, 0, 1)
topo_fgmax_mask.generate_2d_coordinates()
topo_fgmax_mask.write(fname_fgmax_mask, topo_type=3, Z_format='%1i')
print('Created %s' % fname_fgmax_mask)
Write out topo file with DEM Z values¶
Don't need this since we'll put topo in netCDF file.
if 0:
fname_fgmax_topo = 'fgmax_topo_%s.nc' % loc
topo.write(fname_fgmax_topo, topo_type=4)
print('Created %s' % fname_fgmax_topo)
Make RuledRectangle for AMR flagging to the finest level¶
We wish to force GeoClaw to refine the region around the fgmax points to the finest level (generally 1/3" to agree with the resolution of the fgmax points). Rather than trying to specify rectangular regions that cover the fgmax points efficiently, we can use a RuledRectangle.
We can make a RuledRectangle that includes all the fgmax points in two different ways, either with s = x so the rules go from ylower to yupper on each column of grid points, or with s = y so that the rules go from xlower to xupper along each row of grid points.
Which approach is best depends on the geometry of the fgmax points. We would like the RuledRectangle to cover as few points as possible that are not fgmax points.
Here we compute and display both options.
x = topo.x
y = topo.y
# Ruled rectangle with s = x:
rr1 = region_tools.RuledRectangle()
rr1.ixy = 1
s = []
lower = []
upper = []
for i in range(len(x)):
if fgmax_pts_chosen[:,i].sum() > 0:
j = where(fgmax_pts_chosen[:,i]==1)[0]
j1 = j.min()
j2 = j.max()
s.append(x[i])
lower.append(y[j1])
upper.append(y[j2])
rr1.s = array(s)
rr1.lower = array(lower)
rr1.upper = array(upper)
rr1.method = 0
rr1.ds = s[1] - s[0]
rr1_npts = int(ceil(sum(rr1.upper - rr1.lower) *3*3600))
print('RuledRectangle rr1 covers %s grid points' % rr1_npts)
# Ruled rectangle with s = y:
rr2 = region_tools.RuledRectangle()
rr2.ixy = 2
s = []
lower = []
upper = []
for j in range(len(y)):
if fgmax_pts_chosen[j,:].sum() > 0:
i = where(fgmax_pts_chosen[j,:]==1)[0]
i1 = i.min()
i2 = i.max()
s.append(y[j])
lower.append(x[i1])
upper.append(x[i2])
rr2.s = array(s)
rr2.lower = array(lower)
rr2.upper = array(upper)
rr2.method = 0
rr2.ds = s[1] - s[0]
rr2_npts = int(ceil(sum(rr2.upper - rr2.lower)*3*3600))
print('RuledRectangle rr2 covers %s grid points' % rr2_npts)
# Create files for AMR flag regions:
fname1 = 'RuledRectangle_%s_ixy1.data' % loc
rr1.write(fname1)
print('Created ',fname1)
fname2 = 'RuledRectangle_%s_ixy2.data' % loc
rr2.write(fname2)
print('Created ',fname2)
# Plot comparison of two approaches:
figure(figsize=(10,8))
subplot(121)
plot_topo(Z_fg, show_cb=False)
x2,y2 = rr1.vertices()
plot(x2,y2,'r')
title(fname1 + '\n%i points covered' % rr1_npts)
subplot(122)
plot_topo(Z_fg, show_cb=False)
x2,y2 = rr2.vertices()
plot(x2,y2,'r')
title(fname2 + '\n%i points covered' % rr2_npts)
savefigp('RuledRectangles_%s.png' % loc)
Deal with dry land below MHW¶
If this region does not have any land below MHW, can simply set:
#allow_wet_init = None
Otherwise, determine which points should be intially wet. This is only necessary if there are regions near the coast where the topography value Z is below 0 but the initial water depth should be 0 because the region is behind a dike or levy.
The basic approach is to process the DEM to determine what points are both below MHW and connected to deep water by a path of grid points that is entirely below MHW.
Normally the approach below is sufficient: Start by selecting points with Z < -5 and then iterate to flood all points up to Z = 0 (MHW):
wet_points = marching_front.select_by_flooding(topo.Z, mask=None,
prev_pts_chosen=None,
Z1=-5., Z2=0., max_iters=None)
mask_dry = logical_not(wet_points)
mm = logical_or(mask_fg, wet_points)
Z_dry = ma.masked_array(topo.Z, mm)
mask_dry_onshore = logical_and(mask_dry, topo.Z<0.)
mm = logical_or(mask_fg, mask_dry_onshore)
Z_fg2= ma.masked_array(topo.Z, mm)
figure(figsize=(10,8))
subplot(121)
pc = pcolormesh(topo.X, topo.Y, Z_dry, cmap=cmap_dry, norm=norm_dry)
pc = pcolormesh(topo.X, topo.Y, Z_fg2, cmap=cmap, norm=norm)
#cb = colorbar(pc, shrink=0.5, extend='both')
#cb.set_label('meters')
gca().set_aspect(1./cos(48*pi/180.))
ticklabel_format(useOffset=False)
xticks(rotation=20);
title('%s: %i fgmax points\n(dry regions below MHW pink)' % (loc, Z_fg.count()))
subplot(122)
if GEmap is not None:
imshow(GEmap, extent=GEextent)
gca().set_aspect(1./cos(48*pi/180.))
ticklabel_format(useOffset=False)
xticks(rotation=20);
else:
print('Could not make plot with image')
savefigp('allow_wet_init_%s.png' % loc)
Zooming in on various areas shows there is only one small region where we actually need to worry about this, so restrict to that region:
x1,x2,y1,y2 = [-122.67, -122.65, 48.276, 48.288]
I = where(logical_and(topo.x >= x1, topo.x <= x2))[0]
J = where(logical_and(topo.y >= y1, topo.y <= y2))[0]
Zw = topo.Z[J,:][:,I]
Xw = topo.X[J,:][:,I]
Yw = topo.Y[J,:][:,I]
print('Restricted topo.Z to Zw with shape: ', Zw.shape)
#topo_for_wet = topo.crop([-122.67, -122.65, 48.276, 48.288])
wet_points = marching_front.select_by_flooding(Zw, mask=None,
prev_pts_chosen=None,
Z1=-5., Z2=0., max_iters=None)
mask_dry = logical_not(wet_points)
Z_wet = ma.masked_array(Zw, mask_dry)
figure(figsize=(6,6))
pc = pcolormesh(Xw, Yw, Zw, cmap=cmap_dry, norm=norm_dry)
pc = pcolormesh(Xw, Yw, Z_wet, cmap=cmap, norm=norm)
gca().set_aspect(1./cos(48*pi/180.))
ticklabel_format(useOffset=False)
xticks(rotation=20);
Make allow_wet_init array after padding¶
The wet_points identified above are points that the DEM indicates should be initially wet.
In GeoClaw we don't use the DEM topography directly, instead the B values are cell averages over finite volume cells, that ideally are centered at the DEM points with the same resolution.
In GeoClaw we want to force a point to be dry only if wet_points[i,j] == 0 and also the 8 neighboring points have this property. This picks up all the locations well separated from the coast. We want to avoid forcing a GeoClaw cell to be dry if it is at the coastline, since the B value might be slightly negative even if the DEM Z value is positive. Forcing it to be dry while the neighbor offshore is wet will cause immediate flow near the coast.
We accomplish this by setting a new array allow_wet_init We will allow GeoClaw to initialize as wet or dry depending on the value of B in cells where allow_wet_init[i,j] = 1 and only force cells to be dry if allow_wet_init[i,j] = 0.
Summing wet_points over 3x3 blocks of point centered at each [i,j] allows quickly determining if any of the 8 neighboring points are potentially wet.
wet_points_sum = wet_points[1:-1,1:-1] + wet_points[0:-2,1:-1] + wet_points[2:,1:-1] + \
wet_points[1:-1,0:-2] + wet_points[0:-2,0:-2] + wet_points[2:,0:-2] + \
wet_points[1:-1,2:] + wet_points[0:-2,2:] + wet_points[2:,2:]
allow_wet_init = wet_points
allow_wet_init[1:-1,1:-1] = wet_points_sum
allow_wet_init = where(allow_wet_init > 0, 1, 0) # 1 if point or any neighbor is wet
Confirm that allow_wet_init contains more points near the shore than wet_points:
print('wet_points has %i nonzeros, allow_wet_init has %i nonzeros'\
% (wet_points.sum(), allow_wet_init.sum()))
Create the allow_wet_init file used by GeoClaw¶
allow_wet_init_topo = topotools.Topography()
allow_wet_init_topo._x = Xw[0,:]
allow_wet_init_topo._y = Yw[:,0]
allow_wet_init_topo._Z = allow_wet_init
allow_wet_init_topo.generate_2d_coordinates()
fname_allow_wet_init = 'allow_wet_init_%s.data' % loc
allow_wet_init_topo.write(fname_allow_wet_init, topo_type=3, Z_format='%1i')
print('Created %s' % fname_allow_wet_init)
print('Shape: %i points in longitude, %i in latitude' % (Xw.shape[1],Xw.shape[0]))
Expand to cover full fgmax array for use in netCDF file¶
By default we want to allow GeoClaw to fill any cell with $B<0$ up to MHW, so set allow_wet_init_full to 1 everywhere except in this smaller region we have defined allow_wet_init.
allow_wet_init_full = ones(topo.Z.shape)
allow_wet_init_full[J[0]:J[-1]+1, I[0]:I[-1]+1] = allow_wet_init
allow_wet_init_full.shape, topo.Z.shape
contourf(topo.X,topo.Y,allow_wet_init_full, [-2,0,2], colors=[[1,.5,.5], [0.5,0.5,1]])
Make figure for viewing as kml¶
reload(kmltools)
Z_dry_above0 = ma.masked_where(Z_dry<0., Z_dry)
fig,ax,png_extent,kml_dpi = kmltools.pcolorcells_for_kml(topo.x, topo.y, Z_dry_above0,
png_filename='dry_above0_for_kml.png',
dpc=2, cmap=cmap_dry, norm=norm_dry)
Z_dry_below0 = ma.masked_where(Z_dry>0., Z_dry)
Z_dry_below0 = ma.masked_where(logical_or(Z_dry_below0.mask,allow_wet_init_full),
Z_dry_below0)
fig,ax,png_extent,kml_dpi = kmltools.pcolorcells_for_kml(topo.x, topo.y, Z_dry_below0,
png_filename='dry_below0_for_kml.png',
dpc=2, cmap=cmap_dry, norm=norm_dry)
mask_dry = logical_or(logical_not(allow_wet_init_full), topo.Z>0.)
mm = logical_or(mask_fg, mask_dry)
Z_wet = ma.masked_array(topo.Z, mm)
fig,ax,png_extent,kml_dpi= kmltools.pcolorcells_for_kml(topo.x, topo.y, Z_wet,
png_filename='wet_for_kml.png',
dpc=2, cmap=cmap, norm=norm)
kmltools.png2kml(topo.extent, png_files=['dry_above0_for_kml.png', 'dry_below0_for_kml.png',
'wet_for_kml.png'],
png_names=['dry above MHW','dry below MHW','allow wet'],
name='fgmax_points_%s' % loc,
fname='fgmax_points_%s.kmz' % loc)
print('original shape of X,Y,Z: ', Z_fg.shape)
iindex = array([i for i in range(Z_fg.shape[0]) \
if ma.notmasked_edges(Z_fg[i,:]) is not None])
jindex = array([j for j in range(Z_fg.shape[1]) \
if ma.notmasked_edges(Z_fg[:,j]) is not None])
i1 = iindex.min(); i2 = iindex.max()
j1 = jindex.min(); j2 = jindex.max()
print('First and last iindex: %7i, %7i' % (i1,i2))
print('First and last jindex: %7i, %7i' % (j1,j2))
Zcrop = Z_fg[i1:i2+1, j1:j2+1]
print('new shape of X,Y,Z: ', Zcrop.shape)
Xcrop = topo.X[i1:i2+1, j1:j2+1]
Ycrop = topo.Y[i1:i2+1, j1:j2+1]
Plot to make sure it looks right¶
figure(figsize=(8,8))
pc = pcolormesh(Xcrop, Ycrop, Zcrop, cmap=cmap, norm=norm)
gca().set_aspect(1./cos(48*pi/180.))
ticklabel_format(useOffset=False)
xticks(rotation=20);
#axis(extent)
#plot(xpoly,ypoly,'r');
fgm = fgmax_tools.FGmaxMaskedGrid()
fgm.id = loc
fgm.X = Xcrop
fgm.Y = Ycrop
fgm.Z = Zcrop
fgm.fgmax_point = where(Zcrop.mask, 0, 1)
print('Created fgm object')
print('Check that this has the expected number of fgmax points: ',\
fgm.fgmax_point.sum())
if allow_wet_init is not None:
#fgm.allow_wet_init = allow_wet_init_topo.Z #[i1:i2+1, j1:j2+1]
fgm.allow_wet_init = allow_wet_init_full[i1:i2+1, j1:j2+1]
fname_nc = '%s_input.nc' % loc
fgmax_routines.make_nc_input(fname_nc, fgm, force=True)
print('Created ',fname_nc)
Save resulting input files¶
Save the *.data, *.nc (and optionally *.png and *.kml) files created by this notebook to a (new) directory input_files if you don't already have them saved in this directory, since this is where GeoClaw will look for them. Back up an old version first if you want to, for safety.