Surface water network

A Python package to create and analyze surface water networks.

Python packages

Python 3.10+ is required.

Required

• geopandas - process spatial data similar to pandas
• packaging - used to check package versions
• pandas - tabular data analysis
• pyproj - spatial projection support
• rtree - spatial index support

Optional

• flopy >=3.3.6 - read/write MODFLOW models
• netCDF4 - used to read TopNet files

Testing

Run pytest -v or python3 -m pytest -v

For faster multi-core pytest -v -n 2 (with pytest-xdist)

To run doctests pytest -v swn --doctest-modules

Examples

import geopandas
import pandas as pd
import swn

Read from Shapefile:

shp_srs = "tests/data/DN2_Coastal_strahler1z_stream_vf.shp"
lines = geopandas.read_file(shp_srs)
lines.set_index("nzsegment", inplace=True)  # optional

Or, read from PostGIS:

from sqlalchemy import create_engine, engine

con_url = engine.url.URL(drivername="postgresql", database="scigen")
con = create_engine(con_url)
sql = "SELECT * FROM wrc.rec2_riverlines_coastal"
lines = geopandas.read_postgis(sql, con)
lines.set_index("nzsegment", inplace=True)  # optional

Initialise and create network:

n = swn.SurfaceWaterNetwork.from_lines(lines.geometry)
print(n)
# <SurfaceWaterNetwork: with Z coordinates
#   304 segments: [3046409, 3046455, ..., 3050338, 3050418]
#   154 headwater: [3046409, 3046542, ..., 3050338, 3050418]
#   3 outlets: [3046700, 3046737, 3046736]
#   no diversions />

Plot the network, write a Shapefile, write and read a SurfaceWaterNetwork file:

n.plot()

swn.file.gdf_to_shapefile(n.segments, "segments.shp")

n.to_pickle("network.pkl")
n = swn.SurfaceWaterNetwork.from_pickle("network.pkl")

Remove segments that meet a condition (stream order), or that are upstream/downstream from certain locations:

n.remove(
    n.segments.stream_order == 1,
    segnums=n.gather_segnums(upstream=3047927))

Read flow data from a TopNet netCDF file, convert from m3/s to m3/day:

nc_path = "tests/data/streamq_20170115_20170128_topnet_03046727_strahler1.nc"
flow = swn.file.topnet2ts(nc_path, "mod_flow", 86400)
# remove time and truncate to closest day
flow.index = flow.index.floor("d")

# 7-day mean
flow7d = flow.resample("7D").mean()

# full mean
flow_m = pd.DataFrame(flow.mean(0)).T

Process a MODFLOW/flopy model:

import flopy

m = flopy.modflow.Modflow.load("h.nam", model_ws="tests/data", check=False)
nm = swn.SwnModflow.from_swn_flopy(n, m)
nm.default_segment_data()
nm.set_segment_data_inflow(flow_m)
nm.plot()
nm.to_pickle("sfr_network.pkl")
nm = swn.SwnModflow.from_pickle("sfr_network.pkl", n, m)
nm.set_sfr_obj()
m.sfr.write_file("file.sfr")
nm.grid_cells.to_file("grid_cells.shp")
nm.reaches.to_file("reaches.shp")

Citation

Toews, M. W.; Hemmings, B. 2019. A surface water network method for generalising streams and rapid groundwater model development. In: New Zealand Hydrological Society Conference, Rotorua, 3-6 December, 2019. p. 166-169.