This page was generated from mf6_mnw2_tutorial01.py. It's also available as a notebook.

Working with the Multi-node Well (MNW2) Package

[1]:

import os

[2]:

import sys
from tempfile import TemporaryDirectory

import numpy as np
import pandas as pd

import flopy

print(sys.version)
print(f"numpy version: {np.__version__}")
print(f"pandas version: {pd.__version__}")
print(f"flopy version: {flopy.__version__}")

3.9.19 (main, Mar 20 2024, 16:40:02)
[GCC 11.4.0]
numpy version: 1.26.4
pandas version: 2.2.2
flopy version: 3.7.0.dev0

Make an MNW2 package from scratch

[3]:

# temporary directory
temp_dir = TemporaryDirectory()
model_ws = temp_dir.name

m = flopy.modflow.Modflow("mnw2example", model_ws=model_ws)
dis = flopy.modflow.ModflowDis(
    nrow=5, ncol=5, nlay=3, nper=3, top=10, botm=0, model=m
)

MNW2 information by node

(this could be prepared externally from well reconds and read in from a csv or excel file) * this table has two multi-node wells, the first (well1) consisting of two nodes that are manually specified (where the variable rw is specified by node) * node that some variables that are constant for the whole well are also included (losstype, zpump, etc.)

[4]:

node_data = pd.DataFrame(
    [
        [1, 1, 9.5, 7.1, "well1", "skin", -1, 0, 0, 0, 1.0, 2.0, 5.0, 6.2],
        [1, 1, 7.1, 5.1, "well1", "skin", -1, 0, 0, 0, 0.5, 2.0, 5.0, 6.2],
        [3, 3, 9.1, 3.7, "well2", "skin", -1, 0, 0, 0, 1.0, 2.0, 5.0, 4.1],
    ],
    columns=[
        "i",
        "j",
        "ztop",
        "zbotm",
        "wellid",
        "losstype",
        "pumploc",
        "qlimit",
        "ppflag",
        "pumpcap",
        "rw",
        "rskin",
        "kskin",
        "zpump",
    ],
)
node_data

[4]:

	i	j	ztop	zbotm	wellid	losstype	pumploc	rw	rskin	kskin	zpump
0	1	1	9.5	7.1	well1	skin	-1	1.0	2.0	5.0	6.2
1	1	1	7.1	5.1	well1	skin	-1	0.5	2.0	5.0	6.2
2	3	3	9.1	3.7	well2	skin	-1	1.0	2.0	5.0	4.1

Stress period information

(could also be developed externally)

[5]:

stress_period_data = pd.DataFrame(
    [
        [0, "well1", 0],
        [1, "well1", 100.0],
        [0, "well2", 0],
        [1, "well2", 1000.0],
    ],
    columns=["per", "wellid", "qdes"],
)
stress_period_data

[5]:

	per	wellid	qdes
0	0	well1	0.0
1	1	well1	100.0
2	0	well2	0.0
3	1	well2	1000.0

[6]:

pers = stress_period_data.groupby("per")
stress_period_data = {i: pers.get_group(i) for i in [0, 1]}
stress_period_data

[6]:

{0:    per wellid  qdes
 0    0  well1   0.0
 2    0  well2   0.0,
 1:    per wellid    qdes
 1    1  well1   100.0
 3    1  well2  1000.0}

Make `ModflowMnw2` package object

note that extraneous columns in node_data and stress_period_data are ignored
if itmp is positive, it must equal the number of active wells being specified in stress_period_data, otherwise the package class will raise an error.

[7]:

mnw2 = flopy.modflow.ModflowMnw2(
    model=m,
    mnwmax=2,
    node_data=node_data,
    stress_period_data=stress_period_data,
    itmp=[2, 2, -1],  # reuse second per pumping for last stress period
)

[8]:

# "nodtot" is computed automatically
mnw2.nodtot

[8]:

[9]:

pd.DataFrame(mnw2.node_data)

[9]:

	i	j	ztop	zbotm	wellid	losstype	pumploc	...
0	1	1	9.5	7.1	well1	skin	-1	...
1	1	1	7.1	5.1	well1	skin	-1	...
2	3	3	9.1	3.7	well2	skin	-1	...

3 rows × 33 columns

[10]:

pd.DataFrame(mnw2.stress_period_data[0])

[10]:

	k	i	j	wellid	qdes	capmult	cprime	hlim	qcut	qfrcmn	qfrcmx
0	0	1	1	well1	0.0	0	0.0	0.0	0	0.0	0.0
1	0	3	3	well2	0.0	0	0.0	0.0	0	0.0	0.0

[11]:

pd.DataFrame(mnw2.stress_period_data[1])

[11]:

	k	i	j	wellid	qdes	capmult	cprime	hlim	qcut	qfrcmn	qfrcmx
0	0	1	1	well1	100.0	0	0.0	0.0	0	0.0	0.0
1	0	3	3	well2	1000.0	0	0.0	0.0	0	0.0	0.0

[12]:

tmp = flopy.modflow.ModflowMnw2(
    model=m,
    itmp=[1, 1, -1],  # reuse second per pumping for last stress period
)

/opt/hostedtoolcache/Python/3.9.19/x64/lib/python3.9/site-packages/flopy/mbase.py:659: UserWarning: Unit 34 of package MNW2 already in use.
  warn(
/opt/hostedtoolcache/Python/3.9.19/x64/lib/python3.9/site-packages/flopy/mbase.py:668: UserWarning: Two packages of the same type, Replacing existing 'MNW2' package.
  warn(

empty `node_data` and `stress_period_data` tables can also be generated by the package class, and then filled

[13]:

node_data = tmp.get_empty_node_data(3)
node_data

[13]:

rec.array([(0, 0, 0, 0., 0., 0, 0, 0, 0, 0, 0, 0., 0., 0., 0., 0., 0., 0., 0., 0, 0, 0, 0., 0., 0, 0., 0., 0., 0., 0., 0., 0., 0.),
           (0, 0, 0, 0., 0., 0, 0, 0, 0, 0, 0, 0., 0., 0., 0., 0., 0., 0., 0., 0, 0, 0, 0., 0., 0, 0., 0., 0., 0., 0., 0., 0., 0.),
           (0, 0, 0, 0., 0., 0, 0, 0, 0, 0, 0, 0., 0., 0., 0., 0., 0., 0., 0., 0, 0, 0, 0., 0., 0, 0., 0., 0., 0., 0., 0., 0., 0.)],
          dtype=[('k', '<i8'), ('i', '<i8'), ('j', '<i8'), ('ztop', '<f4'), ('zbotm', '<f4'), ('wellid', 'O'), ('losstype', 'O'), ('pumploc', '<i8'), ('qlimit', '<i8'), ('ppflag', '<i8'), ('pumpcap', '<i8'), ('rw', '<f4'), ('rskin', '<f4'), ('kskin', '<f4'), ('B', '<f4'), ('C', '<f4'), ('P', '<f4'), ('cwc', '<f4'), ('pp', '<f4'), ('pumplay', '<i8'), ('pumprow', '<i8'), ('pumpcol', '<i8'), ('zpump', '<f4'), ('hlim', '<f4'), ('qcut', '<i8'), ('qfrcmn', '<f4'), ('qfrcmx', '<f4'), ('hlift', '<f4'), ('liftq0', '<f4'), ('liftqmax', '<f4'), ('hwtol', '<f4'), ('liftn', '<f4'), ('qn', '<f4')])

Mnw objects

at the base of the flopy mnw2 module is the Mnw object class, which describes a single multi-node well. A list or dict of Mnw objects can be used to build a package (using the example above):

flopy.modflow.ModflowMnw2(model=m, mnwmax=2,
                 mnw=<dict or list of Mnw objects>,
                 itmp=[1, 1, -1], # reuse second per pumping for last stress period
                 )

or if node_data and stress_period_data are supplied, the Mnw objects are created on initialization of the ModflowMnw2 class instance, and assigned to the .mnw attribute, as items in a dictionary keyed by wellid.

[14]:

mnw2.mnw

[14]:

{'well1': <flopy.modflow.mfmnw2.Mnw at 0x7f1cac1992b0>,
 'well2': <flopy.modflow.mfmnw2.Mnw at 0x7f1c7eb1e0a0>}

[15]:

mnw2.mnw["well1"].rw

[15]:

[1.0, 0.5]

Note that Mnw object attributes for variables that vary by node are lists (e.g. rw above)

Each Mnw object has its own `node_data` and `stress_period_data`

[16]:

pd.DataFrame(mnw2.mnw["well1"].node_data)

[16]:

	k	i	j	ztop	zbotm	wellid	losstype	pumploc	qlimit	ppflag	...	hlim	qcut	qfrcmn	qfrcmx	hlift	liftq0	liftqmax	hwtol	liftn	qn
0	0	1	1	9.5	7.1	well1	skin	-1	0	0	...	0.0	0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
1	0	1	1	7.1	5.1	well1	skin	-1	0	0	...	0.0	0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0

2 rows × 33 columns

Instead of a dict keyed by stress period, Mnw.stress_period_data is a recarray with pumping data listed by stress period for that well

note that data for period 2, where itmp < 1, is shown (was copied from s.p. 1 during construction of the Mnw object)

[17]:

pd.DataFrame(mnw2.mnw["well2"].stress_period_data)

[17]:

	i	j	per	qdes
0	3	3	0	0.0
1	3	3	1	1000.0
2	3	3	1	1000.0

Build the same package using only the `Mnw` objects

[18]:

mnw2fromobj = flopy.modflow.ModflowMnw2(
    model=m,
    mnwmax=2,
    mnw=mnw2.mnw,
    itmp=[2, 2, -1],  # reuse second per pumping for last stress period
)

/opt/hostedtoolcache/Python/3.9.19/x64/lib/python3.9/site-packages/flopy/mbase.py:659: UserWarning: Unit 34 of package MNW2 already in use.
  warn(
/opt/hostedtoolcache/Python/3.9.19/x64/lib/python3.9/site-packages/flopy/mbase.py:668: UserWarning: Two packages of the same type, Replacing existing 'MNW2' package.
  warn(

[19]:

pd.DataFrame(mnw2fromobj.node_data)

[19]:

	i	j	ztop	zbotm	wellid	losstype	pumploc	...
0	1	1	9.5	7.1	well1	skin	-1	...
1	1	1	7.1	5.1	well1	skin	-1	...
2	3	3	9.1	3.7	well2	skin	-1	...

3 rows × 33 columns

[20]:

pd.DataFrame(mnw2fromobj.stress_period_data[0])

[20]:

	k	i	j	wellid	qdes	capmult	cprime	hlim	qcut	qfrcmn	qfrcmx
0	0	1	1	well1	0.0	0	0.0	0.0	0	0.0	0.0
1	0	3	3	well2	0.0	0	0.0	0.0	0	0.0	0.0

[21]:

pd.DataFrame(mnw2fromobj.stress_period_data[1])

[21]:

	k	i	j	wellid	qdes	capmult	cprime	hlim	qcut	qfrcmn	qfrcmx
0	0	1	1	well1	100.0	0	0.0	0.0	0	0.0	0.0
1	0	3	3	well2	1000.0	0	0.0	0.0	0	0.0	0.0

By default, the `node_data` and `stress_period_data` tables attached to the `ModflowMnw2` package class are definitive

on writing of the package output (mnw2.write_file()), the Mnw objects are regenerated from the tables. This setting is controlled by the default argument use_tables=True. To write the package file using the Mnw objects (ignoring the tables), use mnw2.write_file(use_tables=False).

[22]:

per1 = flopy.modflow.ModflowMnw2.get_empty_stress_period_data(itmp=2)
per1

[22]:

rec.array([(0, 0, 0, 0, 0., 0, 0., 0., 0, 0., 0.),
           (0, 0, 0, 0, 0., 0, 0., 0., 0, 0., 0.)],
          dtype=[('k', '<i8'), ('i', '<i8'), ('j', '<i8'), ('wellid', 'O'), ('qdes', '<f4'), ('capmult', '<i8'), ('cprime', '<f4'), ('hlim', '<f4'), ('qcut', '<i8'), ('qfrcmn', '<f4'), ('qfrcmx', '<f4')])

Write an MNW2 package file and inspect the results

[23]:

mnw2.write_file(os.path.join(model_ws, "test.mnw2"))

[24]:

junk = [
    print(l.strip("\n"))
    for l in open(os.path.join(model_ws, "test.mnw2")).readlines()
]

# MNW2 package for MODFLOW-2005 generated by Flopy 3.7.0.dev0
2 0 0
well1 -2
            skin -1 0 0 0
              -1.0000000E+00    2.0000000E+00    5.0000000E+00
               9.5000000E+00    7.0999999E+00 2 2    1.0000000E+00
               7.0999999E+00    5.0999999E+00 2 2    5.0000000E-01
               6.1999998E+00
well2 -1
            skin -1 0 0 0
               1.0000000E+00    2.0000000E+00    5.0000000E+00
               9.1000004E+00    3.7000000E+00 4 4
               4.0999999E+00
2  Stress Period 1
well1    0.0000000E+00
well2    0.0000000E+00
2  Stress Period 2
well1    1.0000000E+02
well2    1.0000000E+03
-1  Stress Period 3

Load some example MNW2 packages

[25]:

path = os.path.join("..", "..", "examples", "data", "mnw2_examples")
m = flopy.modflow.Modflow("MNW2-Fig28", model_ws=model_ws)
dis = flopy.modflow.ModflowDis.load(os.path.join(path, "MNW2-Fig28.dis"), m)

[26]:

m.get_package_list()

[26]:

['DIS']

[27]:

mnw2pth = os.path.join(path, "MNW2-Fig28.mnw2")
mnw2 = flopy.modflow.ModflowMnw2.load(mnw2pth, m)

[28]:

pd.DataFrame(mnw2.node_data)

[28]:

	k	i	j	ztop	zbotm	wellid	losstype	pumploc	qlimit	ppflag	...	hlim	qcut	qfrcmn	qfrcmx	hlift	liftq0	liftqmax	hwtol	liftn	qn
0	0	29	40	-5.0	-65.0	well-a	skin	0	1	0	...	-7.5	-1	0.1	0.15	0.0	0.0	0.0	0.0	0.0	0.0

1 rows × 33 columns

[29]:

pd.DataFrame(mnw2.stress_period_data[0])

[29]:

	k	i	j	wellid	qdes	capmult	cprime	hlim	qcut	qfrcmn	qfrcmx
0	0	29	40	well-a	0.0	0	0.0	0.0	0	0.0	0.0

[30]:

mnw2.mnw

[30]:

{'well-a': <flopy.modflow.mfmnw2.Mnw at 0x7f1c7ea48d60>}

[31]:

pd.DataFrame(mnw2.mnw["well-a"].stress_period_data)

[31]:

	i	j	per	qdes
0	29	40	0	0.0
1	29	40	1	-10000.0
2	29	40	2	-10000.0

[32]:

path = os.path.join("..", "..", "examples", "data", "mnw2_examples")
m = flopy.modflow.Modflow("br", model_ws=model_ws)
mnw2 = flopy.modflow.ModflowMnw2.load(
    os.path.join(path, "BadRiver_cal.mnw2"), m
)

[33]:

df = pd.DataFrame(mnw2.node_data)
df.loc[:, df.sum(axis=0) != 0]

[33]:

	i	j	ztop	zbotm	wellid	losstype	pumploc	rw	rskin	kskin	zpump
0	294	503	181.630005	161.630005	br_birch1	skin	-1	1.0	2.0	10.0	162.630005
1	295	503	179.119995	159.119995	br_birch2	skin	-1	1.0	2.0	10.0	160.119995
2	175	342	400.220001	312.220001	br_diaperville1	skin	-1	1.0	2.0	10.0	313.220001
3	174	342	399.119995	312.119995	br_diaperville2	skin	-1	1.0	2.0	10.0	313.119995
4	248	454	565.200012	555.200012	br_franks1	skin	-1	1.0	2.0	10.0	556.200012
5	249	453	564.419983	554.419983	br_franks2	skin	-1	1.0	2.0	10.0	555.419983
6	180	396	453.959991	447.959991	br_odanah1	skin	-1	1.0	2.0	10.0	448.959991
7	181	395	450.559998	444.559998	br_odanah2	skin	-1	1.0	2.0	10.0	445.559998
8	181	395	380.489990	371.489990	br_odanah3	skin	-1	1.0	2.0	10.0	372.489990
9	180	396	450.739990	444.739990	br_odanah4	skin	-1	1.0	2.0	10.0	445.739990
10	170	350	475.410004	472.410004	br_oldschool	skin	-1	1.0	2.0	10.0	473.410004
11	172	312	377.410004	348.410004	br_recycle	skin	-1	1.0	2.0	10.0	349.410004
12	216	412	562.200012	542.200012	br_unspec	skin	-1	1.0	2.0	10.0	543.200012
13	516	424	1079.910034	1072.910034	cfsp_ncp1	skin	-1	1.0	2.0	10.0	1073.910034
14	515	424	1077.900024	1074.900024	cfsp_ncp2	skin	-1	1.0	2.0	10.0	1075.900024
15	539	415	1093.010010	1003.010010	cfsp_of	skin	-1	1.0	2.0	10.0	1004.010010
16	360	2	706.330017	699.330017	hayward_bait_n	skin	-1	1.0	2.0	10.0	700.330017
17	360	6	705.289978	698.289978	hayward_bait_ne	skin	-1	1.0	2.0	10.0	699.289978
18	362	4	696.979980	689.979980	hayward_bait_of	skin	-1	1.0	2.0	10.0	690.979980
19	456	699	1352.050049	1342.050049	ironbelt2	skin	-1	1.0	2.0	10.0	1343.050049
20	456	699	1350.439941	1343.439941	ironbelt3	skin	-1	1.0	2.0	10.0	1344.439941
21	599	414	1198.839966	1188.839966	mellen2	skin	-1	1.0	2.0	10.0	1189.839966
22	576	408	1151.599976	1131.599976	mellen3	skin	-1	1.0	2.0	10.0	1132.599976
23	253	149	622.179993	602.179993	milestone	skin	-1	1.0	2.0	10.0	603.179993
24	208	148	520.119995	500.119995	nsp	skin	-1	1.0	2.0	10.0	501.119995
25	80	162	535.299988	30.299999	washburn1	skin	-1	1.0	2.0	10.0	31.299999
26	91	143	541.280029	-67.720001	washburn2	skin	-1	1.0	2.0	10.0	-66.720001

[34]:

try:
    # ignore PermissionError on Windows
    temp_dir.cleanup()
except:
    pass

Working with the Multi-node Well (MNW2) Package

Make an MNW2 package from scratch

MNW2 information by node

Stress period information

Make ModflowMnw2 package object

empty node_data and stress_period_data tables can also be generated by the package class, and then filled

Mnw objects

Each Mnw object has its own node_data and stress_period_data

Instead of a dict keyed by stress period, Mnw.stress_period_data is a recarray with pumping data listed by stress period for that well

Build the same package using only the Mnw objects

By default, the node_data and stress_period_data tables attached to the ModflowMnw2 package class are definitive

Write an MNW2 package file and inspect the results

Load some example MNW2 packages

Make `ModflowMnw2` package object

empty `node_data` and `stress_period_data` tables can also be generated by the package class, and then filled

Each Mnw object has its own `node_data` and `stress_period_data`

Build the same package using only the `Mnw` objects

By default, the `node_data` and `stress_period_data` tables attached to the `ModflowMnw2` package class are definitive