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
import sys

[2]:

from pathlib import Path
from tempfile import TemporaryDirectory

import git
import numpy as np
import pandas as pd
import pooch

import flopy

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

3.12.10 | packaged by conda-forge | (main, Apr 10 2025, 22:21:13) [GCC 13.3.0]
numpy version: 2.2.5
pandas version: 2.2.3
flopy version: 3.9.3

Check if we are in the repository and define the data path.

[3]:

try:
    root = Path(git.Repo(".", search_parent_directories=True).working_dir)
except:
    root = None

[4]:

data_path = root / "examples" / "data" if root else Path.cwd()

Make an MNW2 package from scratch

[5]:

# 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.)

[6]:

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

[6]:

	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)

[7]:

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

[7]:

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

[8]:

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

[8]:

{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.

[9]:

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
)

[10]:

# "nodtot" is computed automatically
mnw2.nodtot

[10]:

[11]:

pd.DataFrame(mnw2.node_data)

[11]:

	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

[12]:

pd.DataFrame(mnw2.stress_period_data[0])

[12]:

	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

[13]:

pd.DataFrame(mnw2.stress_period_data[1])

[13]:

	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

[14]:

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

/home/runner/micromamba/envs/flopy/lib/python3.12/site-packages/flopy/mbase.py:622: UserWarning: Unit 34 of package MNW2 already in use.
  warn(
/home/runner/micromamba/envs/flopy/lib/python3.12/site-packages/flopy/mbase.py:631: 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

[15]:

node_data = tmp.get_empty_node_data(3)
node_data

[15]:

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.

[16]:

mnw2.mnw

[16]:

{'well1': <flopy.modflow.mfmnw2.Mnw at 0x7f959c429b80>,
 'well2': <flopy.modflow.mfmnw2.Mnw at 0x7f959c42ac60>}

[17]:

mnw2.mnw["well1"].rw

[17]:

[np.float32(1.0), np.float32(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`

[18]:

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

[18]:

	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)

[19]:

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

[19]:

	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

[20]:

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

/home/runner/micromamba/envs/flopy/lib/python3.12/site-packages/flopy/mbase.py:622: UserWarning: Unit 34 of package MNW2 already in use.
  warn(
/home/runner/micromamba/envs/flopy/lib/python3.12/site-packages/flopy/mbase.py:631: UserWarning: Two packages of the same type, Replacing existing 'MNW2' package.
  warn(

[21]:

pd.DataFrame(mnw2fromobj.node_data)

[21]:

	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

[22]:

pd.DataFrame(mnw2fromobj.stress_period_data[0])

[22]:

	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

[23]:

pd.DataFrame(mnw2fromobj.stress_period_data[1])

[23]:

	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).

[24]:

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

[24]:

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

[25]:

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

[26]:

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 3.9.3
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

[27]:

folder_name = "mnw2_examples"

[28]:

file_names = {
    "BadRiver_cal.mnw2": None,
    "MNW2-Fig28.bas": None,
    "MNW2-Fig28.dis": None,
    "MNW2-Fig28.lpf": None,
    "MNW2-Fig28.mnw2": None,
    "MNW2-Fig28.mnwi": None,
    "MNW2-Fig28.nam": None,
    "MNW2-Fig28.oc": None,
    "MNW2-Fig28.pcg": None,
    "MNW2-Fig28.rch": None,
    "MNW2-Fig28.wel": None,
}
for fname, fhash in file_names.items():
    pooch.retrieve(
        url=f"https://github.com/modflowpy/flopy/raw/develop/examples/data/{folder_name}/{fname}",
        fname=fname,
        path=data_path / folder_name,
        known_hash=fhash,
    )

[29]:

path = data_path / folder_name
m = flopy.modflow.Modflow("MNW2-Fig28", model_ws=model_ws)
dis = flopy.modflow.ModflowDis.load(os.path.join(path, "MNW2-Fig28.dis"), m)

[30]:

m.get_package_list()

[30]:

['DIS']

[31]:

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

[32]:

pd.DataFrame(mnw2.node_data)

[32]:

	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

[33]:

pd.DataFrame(mnw2.stress_period_data[0])

[33]:

	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

[34]:

mnw2.mnw

[34]:

{'well-a': <flopy.modflow.mfmnw2.Mnw at 0x7f959c27a360>}

[35]:

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

[35]:

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

[36]:

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

[37]:

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

[37]:

	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

[38]:

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