SEAWAT is a model developed by the USGS for the simulation of three-dimensional variable density groundwater flow with solute and heat transport. The software is based on MODFLOW-2000 and MT3DMS and on its latest version it can simulate viscosity variations and provide faster execution times. SEAWAT is implemented on Flopy, the Python library to build, run and represent MODFLOW models. This tutorial has the complete workflow to create and represent a basic example of saline instrusion with SEAWAT and Flopy on a Jupyter Notebook.

Model Geometry

The model has 100 columns, 50 layers and 1 row. Total dimension in X, Y and Z axis are 2, 1 and 1 meter repectively. There is a injecting wel on the left side and a fixed head on the right.

Tutorial

Python code

Import the required libraries

This tutorial requires some Python core libraries, Scipy libraries (numpy and matplotlib) as well as the Flopy library. Since we work on the interactive enviroment of Jupyter Notebook, we will use the Jupyter option to have inline output representation of the generated Matplotlib graphs. We will use some IPython widgets to create an interactive visualization of the model concentration distribution.

%matplotlib inline
import os
import sys
import numpy as np
import flopy
import flopy.utils.binaryfile as bf
import matplotlib.pyplot as plt
from matplotlib.colors import BoundaryNorm
from matplotlib.ticker import MaxNLocator
import ipywidgets as widgets

Create a basic MODFLOW model object

As well as any other model generated by Flopy, first we have to setup the model name and the working directory. We strongly recommend to follow the folder estructure provided on the "Input Files" of this tutorial.

# Create the basic MODFLOW model structure
modelname = 'model1a'
workspace = '../Model'
swt = flopy.seawat.Seawat(modelname, exe_name='../Exe/swt_v4x64.exe', model_ws=workspace)
print(swt.namefile)

model1a.nam

Define model dimensions, spatial discretization and hydraulic parameters

This model is almost a 2D models of 100 columns, 50 layers and 1 row. Total dimension in X, Y and Z axis are 2, 1 and 1 meter repectively. According to the Flopy input data procedure, there is only value or array for model_top (surface) and a list of elevations for every layer bottom elevation.

# Model dimensions
Lx = 2.
Ly = 1.
Lz = 1.

# Spatial discretization
nlay = 50
nrow = 1
ncol = 100
delr = Lx / ncol
delc = Ly
delv = Lz / nlay

# Elevation for model surface and layer bottom elevation
model_top = 1.
model_botm = np.linspace(model_top - delv, 0., nlay)

model_botm

array([0.98, 0.96, 0.94, 0.92, 0.9 , 0.88, 0.86, 0.84, 0.82, 0.8 , 0.78,
       0.76, 0.74, 0.72, 0.7 , 0.68, 0.66, 0.64, 0.62, 0.6 , 0.58, 0.56,
       0.54, 0.52, 0.5 , 0.48, 0.46, 0.44, 0.42, 0.4 , 0.38, 0.36, 0.34,
       0.32, 0.3 , 0.28, 0.26, 0.24, 0.22, 0.2 , 0.18, 0.16, 0.14, 0.12,
       0.1 , 0.08, 0.06, 0.04, 0.02, 0.  ])

# Another parameters
qinflow = 6.6E-5 # m3/s equiv to 5.702 m3/day
dmcoef = 6.6E-6 # m2/s equiv to 0.57024 m2/day  Could also try 1.62925 as another case of the Henry problem
hk = 0.01 # m/s equivalent to 864m/day

Definition of the flow packages for the SEAWAT model

In this part we define the packages related to groundwater flow on the SEAWAT model. First we define the DIS package that has the geometry as well as the simulation type (steady / transient). The model run on steady conditions on the head distribution, but transient on the saline distribution.

perlen = 86400 # 1 day in seconds 

dis = flopy.modflow.ModflowDis(swt, nlay, nrow, ncol, nper=1, delr=delr,
                               delc=delc, laycbd=0, top=model_top,
                               botm=model_botm, perlen=perlen, nstp=15)

Then we define another SEAWAT packages as:

the BAS package for setting the constant head cells,
the LPF that defines the vertical / horizontal hydraulic conductivity,
the PCG packages that solves the model matrix
the OC packages for the output record

# Variables for the BAS package
ibound = np.ones((nlay, nrow, ncol), dtype=np.int32)
ibound[:, :, -1] = -1
bas = flopy.modflow.ModflowBas(swt, ibound, 0)

# Add LPF package to the MODFLOW model
lpf = flopy.modflow.ModflowLpf(swt, hk=hk, vka=hk, ipakcb=53)

# Add PCG Package to the MODFLOW model
pcg = flopy.modflow.ModflowPcg(swt, hclose=1.e-8)

# Add OC package to the MODFLOW model
oc = flopy.modflow.ModflowOc(swt, 
                             stress_period_data={(0, 0): ['save head', 'save budget']},
                             compact=True)

Definition of the flow and transport packages for the SEAWAT model

A well is created with the WEL package and defined as a source of fresh water with concentration of 0. The left boundary is defined as sea water with concentration of 35. The well inflow is distributed over the model layers.

# Create WEL and SSM data
itype = flopy.mt3d.Mt3dSsm.itype_dict()
wel_data = {}
ssm_data = {}
wel_sp1 = []
ssm_sp1 = []
for k in range(nlay):
    wel_sp1.append([k, 0, 0, qinflow / nlay])
    ssm_sp1.append([k, 0, 0, 0., itype['WEL']])
    ssm_sp1.append([k, 0, ncol - 1, 35., itype['BAS6']])
wel_data[0] = wel_sp1
ssm_data[0] = ssm_sp1
wel = flopy.modflow.ModflowWel(swt, stress_period_data=wel_data, ipakcb=53)

Setup of the MT3DMS models and the SEAWAT variable density flow package

According with the SEAWAT setup the paramaters for the transport model. Values variable density flow package of Seawat are also declared.

# Create the basic MT3DMS model structure
btn = flopy.mt3d.Mt3dBtn(swt, nprs=-5, prsity=0.35, sconc=35., ifmtcn=0,
                         chkmas=False, nprobs=10, nprmas=10, dt0=300)
adv = flopy.mt3d.Mt3dAdv(swt, mixelm=0)
dsp = flopy.mt3d.Mt3dDsp(swt, al=0., trpt=1., trpv=1., dmcoef=dmcoef)
gcg = flopy.mt3d.Mt3dGcg(swt, iter1=500, mxiter=1, isolve=1, cclose=1e-7)
ssm = flopy.mt3d.Mt3dSsm(swt, stress_period_data=ssm_data)

# Create the SEAWAT model structure
vdf = flopy.seawat.SeawatVdf(swt, iwtable=0, densemin=0, densemax=0,
                             denseref=1000., denseslp=0.7143, firstdt=1e-3)

Write files of the SEAWAT model and run simulation

Write the .nam file and the files declared on the .nam file. Then runs the simulation and shows the simulation information on the screen

# Write the input files
swt.write_input()
swt.run_model()

FloPy is using the following  executable to run the model: ../Exe/swt_v4x64.exe

                                  SEAWAT Version 4
    U.S. GEOLOGICAL SURVEY MODULAR FINITE-DIFFERENCE GROUND-WATER FLOW MODEL
                             Version 4.00.05 10/19/2012                      

Incorporated MODFLOW Version: 1.18.01 06/20/2008                      
Incorporated MT3DMS  Version: 5.20    10/30/2006                      


 This program is public domain and is released on the
 condition that neither the U.S. Geological Survey nor
 the United States Government may be held liable for any
 damages resulting from their authorized or unauthorized
 use.


 Using NAME file: model1a.nam                                                                                                                                                                                             
 Run start date and time (yyyy/mm/dd hh:mm:ss): 2019/02/18 10:30:57


 STRESS PERIOD NO.    1

  STRESS PERIOD    1 TIME STEP    1 FROM TIME =   0.0000     TO    5760.0    

 Transport Step:    1   Step Size:   300.0     Total Elapsed Time:   300.00    
 Outer Iter.  1  Inner Iter.  1:  Max. DC =  0.4179      [K,I,J]   50    1    1
 Outer Iter.  1  Inner Iter.  2:  Max. DC =  0.3685      [K,I,J]    1    1    2
 Outer Iter.  1  Inner Iter.  3:  Max. DC =  0.2005      [K,I,J]    1    1    3
 Outer Iter.  1  Inner Iter.  4:  Max. DC =  0.2249      [K,I,J]   49    1    3
 .....
 Run end date and time (yyyy/mm/dd hh:mm:ss): 2019/02/18 10:31:02
 Elapsed run time:  5.310 Seconds

 Normal termination of SEAWAT





(True, [])

Model results post-processing

This tutorial has three representations of the flow and transport model with SEAWAT:

The first representation is a flow direction vector over the concentrations at the end of the model
Second representation is the head distribution on the model extension
Third representation is a interactive graph of concentration development along the model time steps.

1. Flow directions and concentrations

# Load data
ucnobj = bf.UcnFile('../Model/MT3D001.UCN', model=swt)
times = ucnobj.get_times()
concentration = ucnobj.get_data(totim=times[-1])
cbbobj = bf.CellBudgetFile('../Model/model1a.cbc')
times = cbbobj.get_times()
qx = cbbobj.get_data(text='flow right face', totim=times[-1])[0]
qz = cbbobj.get_data(text='flow lower face', totim=times[-1])[0]

# Average flows to cell centers
qx_avg = np.empty(qx.shape, dtype=qx.dtype)
qx_avg[:, :, 1:] = 0.5 * (qx[:, :, 0:ncol-1] + qx[:, :, 1:ncol])
qx_avg[:, :, 0] = 0.5 * qx[:, :, 0]
qz_avg = np.empty(qz.shape, dtype=qz.dtype)
qz_avg[1:, :, :] = 0.5 * (qz[0:nlay-1, :, :] + qz[1:nlay, :, :])
qz_avg[0, :, :] = 0.5 * qz[0, :, :]

# parameters for the colorbar
levels = MaxNLocator(nbins=15).tick_values(concentration.min(), concentration.max())
cmap = plt.get_cmap('PiYG')
norm = BoundaryNorm(levels, ncolors=cmap.N, clip=True)

# Make the plot
fig = plt.figure(figsize=(20, 10))
ax = fig.add_subplot(1, 1, 1, aspect='equal')
im = ax.imshow(concentration[:, 0, :], interpolation='nearest',
           extent=(0, Lx, 0, Lz))
fig.colorbar(im, ax=ax, fraction=0.05, label="Concentration (g/l)")

y, x, z = dis.get_node_coordinates()
X, Z = np.meshgrid(x, z[:, 0, 0])
iskip = 3
ax.quiver(X[::iskip, ::iskip], Z[::iskip, ::iskip],
           qx_avg[::iskip, 0, ::iskip]*1E5, -qz_avg[::iskip, 0, ::iskip]*1E5,
           color='w', scale=3, headwidth=3, headlength=2,
           headaxislength=2, width=0.0025)

ax.set_title('Flow Direction and Concentration')
plt.savefig('../Output/ConcentrationDistribution.png')

2. Head distribution on model extension

# Extract the heads
fname = '../Model/model1a.hds'
headobj = bf.HeadFile(fname)
times = headobj.get_times()
head = headobj.get_data(totim=times[-1])
head[0:3,0,0:3]

array([[-0.00049478, -0.00062826, -0.00076412],
       [-0.00049201, -0.00062547, -0.00076129],
       [-0.00048649, -0.0006199 , -0.00075567]], dtype=float32)

levels = MaxNLocator(nbins=15).tick_values(head.min(), head.max())
cmap = plt.get_cmap('Blues')
norm = BoundaryNorm(levels, ncolors=cmap.N, clip=True)

# Make a simple head plot
fig = plt.figure(figsize=(20, 10))
ax = fig.add_subplot(1, 1, 1, aspect='equal')
im = ax.imshow(head[:, 0, :], interpolation='nearest',
               extent=(0, Lx, 0, Lz))
fig.colorbar(im, ax=ax, fraction=0.05, label="Head (m)")
ax.set_title('Simulated Heads')
plt.savefig('../Output/HeadDistribution.png')

3. Interactive representation of concentration development

times = np.asarray(ucnobj.get_times())
times[:5]

array([1500., 3000., 4500., 5760., 7260.], dtype=float32)

def plotConcentration(time):
    # Load data
    ucnobj = bf.UcnFile('../Model/MT3D001.UCN', model=swt)

    concentration = ucnobj.get_data(totim=time)

    # For colormap
    levels = MaxNLocator(nbins=15).tick_values(concentration.min(), concentration.max())
    cmap = plt.get_cmap('PiYG')
    norm = BoundaryNorm(levels, ncolors=cmap.N, clip=True)

    #Figure definition
    fig = plt.figure(figsize=(20, 10))
    ax = fig.add_subplot(1, 1, 1, aspect='equal')
    im = ax.imshow(concentration[:, 0, :], interpolation='nearest',
               extent=(0, Lx, 0, Lz))

    cbaxes = fig.add_axes([0.05, 0.15, 0.02, 0.7])

    fig.colorbar(im, ax=ax, fraction=0.05, label="Concentration (g/l)", cax = cbaxes)

# create a model interactive representation
widgets.interact(plotConcentration, time=widgets.SelectionSlider(options=times, value=times[0], disabled=False))