w2rpy

w2rpy

GitHub

PyPI

To install or update:

pip install w2rpy
pip install --upgrade w2rpy

PyTorch and Meta's SegmentAnything model must be installed separately to run the pebble_count function. Download a pytorch wheel specific to your GPU here. Download SegmentAnything model checkpoints here. And install the SegmentAnything python package:

pip install https://github.com/facebookresearch/segment-anything/archive/master.zip

To create a REM and estimate channel extent:

import w2rpy as w2r
import matplotlib.pyplot as plt
import rasterio as rio
from rasterio.plot import show

dem = '/dem.tif'
valley_centerline = '/VCL.shp'

xs = w2r.xs(valley_centerline,1500,100)

haws = r'C:/users/lrussell/downloads/haws.tif'
w2r.create_REM(dem,xs,haws)

channel = w2r.inundate(haws,[3],remove_holes=True,largest_only=True)


fig,ax = plt.subplots(1,1)

with rio.open(haws) as src:
    show(src,ax=ax,cmap='viridis',vmin=-1,vmax=10,alpha=0.75)
    
xs.plot(ax=ax,color='k',lw=0.5)
channel.plot(ax=ax,ec='b',fc='gray',alpha=0.5)

To estimate a ratinge curve:

import w2rpy as w2r
import geopandas as gpd
import matplotlib.pyplot as plt

import geopandas as gpd

# Get inundation polygons over a range of relative elevations
inundation_polygons = w2r.inundate(haws,range(3,11),largest_only=True)

# clip polygons to area of hydraulic calcs
aoi = gpd.read_file(r'C:/Users/lrussell/Downloads/aoi_clip.shp')
inundation_polygons.geometry = inundation_polygons.intersection(aoi.union_all())
channel.geometry = channel.intersection(aoi.union_all())

# run htab calcs
htab = w2r.htab_2D(haws,
            extents=inundation_polygons,
            vcl=valley_centerline,
            slope=0.01,
            roughness=0.1,
            channel_area=channel,
            channel_roughness=0.032)


fig,ax = plt.subplots(1,2)

inundation_polygons.plot(ax=ax[0],cmap='Blues_r',ec='k',alpha=0.1)

ax[1].plot(htab.Q,htab.WSE,color='k')
ax[1].scatter(htab.Q,htab.WSE,color='k')

ax[1].set_xlabel('Discharge (cfs)')
ax[1].set_ylabel('Relative WSE (ft)')

ax[1].grid()

To run a photo-based pebble count:

w2r.pebble_count(['/pebble_photo.png'],obj_size_mm=180)

To run a habitat suitability (HSI) analysis:

dep = r'/depth_raster.tif'
vel = r'/velocity_raster.tif'
w2r.HSI(dep,vel,'Adult Coho Spawning',r'/HSI_output.tif')

To run a tree delineation analysis for trees between 50 and 100 feet tall:

ch = '/canopy_height.tif'
output = r'/trees.shp'
w2r.delineate_trees(ch, output)

Documentation

terrain(dem_file)

• Returns Terrain object. Terrain object contains the following objects:
  • DEM, flow accumulation, flow direction, pysheds Grid, and coordinate system
• Input options:
  • dem_file: file path to a DEM raster (.tif, .vrt, etc…)

streamlines(terrain, pour_point, threshold=None, snap_threshold=None, save=None)

• Returns a geodataframe of delineated stream lines.
• Input options:
  • terrain: Terrain object
  • pour_point: file path to or geodataframe of points at the bottom of the desired watershed. Will only use the first point.
  • threshold: threshold for stream delineation in number of flow accumulated pixels. Defaults to number of pixels/100.
  • snap_threshold: threshold for snapping pour point to the flow accumulation raster. Defaults to number of pixels/10.
  • save: optional file path to save streamlines as shapefile rather than returning a geodataframe object.

catchment(terrain, pour_point, snap_threshold=None, save=None)

• Returns a geodataframe of delineated catchment.
• Input options:
  • terrain: Terrain object
  • pour_point: file path to or geodataframe of points at the bottom of the desired watershed. Will only use the first point.
  • snap_threshold: threshold for snapping pour point to the flow accumulation raster. Defaults to number of pixels/10.
  • save: optional file path to save catchment as shapefile rather than returning a geodataframe object.

get_xs(cl, xs_length, spacing, save=None)

• Returns a geodataframe of cross-sections.
• Input options:
  • cl: centerline, or any lines, to generate cross-sections along.
  • xs_length: length of cross-section from cl. Full XS length is xs_length*2
  • spacing: distance between cross-sections on cl.
  • save: optional file path to save cross-sections as shapefile rather than returning a geodataframe object.

get_points(lines, spacing, ep=True, save=None)

• Returns a geodataframe of cross-sections.
• Input options:
  • lines: centerline, or any lines, to generate cross-sections along.
  • spacing: distance between points on lines.
  • ep: option to include end point on line. Either True or False.
  • save: optional file path to save points as shapefile rather than returning a geodataframe object.

inundate(raster, rel_wse_list, largest_only=False, invert=False, remove_holes=False, save=None)

• Returns a geodataframe of polygons showing inundation extent for given relative WSEs.
• Input options:
  • raster: file path to DEM or REM to perform inundation on.
  • rel_wse_list: list for WSEs to create inundation polygons for. (Example: [1,3.5,7,10])
  • largest_only: True/False to return only the largest inundated polygon. Helpful for channel delineation.
  • invert: True/False to select the are below or above a given WSE.
  • remove_holes: True/False option to remove all holes within inundation extent polygons.
  • save: optional file path to save the inundation extents as shapefile rather than returning a geodataframe object.

sample_raster(raster, points, buff=None, metric="min", multiband=False, crs=None)

• Returns list of values sampled from raster at location of points.
• Input options:
  • raster: file path to DEM or imagery to sample data.
  • points: geodatabase or file path of points to sample.
  • buff: optional buffer distance to around each point to sample.
  • metric: suite of options to sample for including "min", "mean", "max", and "mode"
  • multiband: True/False option to include samples from all bands, otherwise it only returns the first band.
  • crs: option to convert points to specific coordinate system, only use if raster has no set coordinate system.

zonal_stats(df,raster,metric="mean",band=1)

• Returns list of values sampled from raster within extent of given polygons.
• Input options:
  • raster: file path to DEM or imagery to sample data.
  • df: geodatabase or file path of polygons to sample.
  • metric: suite of options including "min", "mean", "max", "sum", and "nonzero"
  • band: integer value of which raster band to sample data from.

edit_raster(raster, output, crs=None, resample=None, resample_method="bilinear", clip=None, nodata=None)

• All in one function to edit or warp rasters in various ways. Saves output to a new tif file.
• Input options:
  • raster: file path to DEM or imagery to warp/edit.
  • output: file path to save new raster.
  • crs: coordinate system to reproject raster to.
  • resample: pixel size to resample raster to.
  • resample_method: options for resampling ("bilinear","nearest","mean")
  • clip: geodataframe or path to polygons to clip raster to.
  • nodata: will set the nodata value for the output raster.

merge_rasters(rasters, output, method="first", compression=None, nodata=None)

• Merges rasters into a single tif file. This does not build a mosaic, be wary of resulting file sizes.
• Input options:
  • rasters: list of file paths to DEM or imagery to merge.
  • output: file path to save new raster.
  • method: options for merging rasters including "first", "last", "max" and "min". See rasterio documentation for more info.
  • compression: option to compress merged raster. Best to use are "JPEG" or "LZW".
  • nodata: will set the nodata value for the output raster.

difference_rasters(r1, r2, output, match_affine="first", method="nearest")

• Will match the extent and resolution of both rasters then difference. Output = r2 – r1.
• Input options:
  • r1: file path to DEM or imagery.
  • r2: file path to DEM or imagery.
  • output: file path to save new raster.
  • match_affine: If "first" then will warp r2 to match r1 extent and resolution. If "last" then will warp r1 to match r2 extent and resolution.
  • method: method for raster transformation includes "nearest" and "bilinear"

create_REM(dem, xs, output, sample_dist=3, smooth_window=5, buffer=1000, limits=[-50,50], method="min", vb=None, ret_xs=False, wse_path=None)

• Will create a REM from a given DEM and set of cross-sections.
• Input options:
  • dem: file path to DEM.
  • xs: file path or geodataframe of cross-sections to use.
  • output: file path to save new REM raster.
  • sample_dist: Length along cross-sections to sample elevation data.
  • smooth_window: number of cross-sections to average using rolling window.
  • buffer: clips new raster to include buffer around total bounding box of cross-sections.
  • limits: sets all data above or below limits to nodata.
  • method: Option to use "min" or "mean" sampled elevation per cross-section
  • vb: geodataframe or path to valley bottom polygon to clip cross-sections.
  • ret_xs: will return XS if set to true, no need to use this.
  • wse_path: file path to save WSE/GGL surface raster.

htab_1D(station, elevation, slope, D50, max_depth=10, breaks=None, save=None)

• Returns dataframe of hydraulic table (htab) calcs for a given cross-section. Htab table returns either the sum (for discharge, area, and perimeter) or discharge-weighted average (for roughness, hydraulic radius, and velocity) across all flowing channels.
• Input options:
  • station: list , array, or series of station data.
  • elevation: list , array, or series of elevation data.
  • slope: dimensionless slope for hydraulic calcs. Constant across all flows/depths.
  • D50: representative grain size used for bathhurst calcs to vary roughness with depth.
  • max_depth: depth from thalweg to perform calcs from. Splits this range in 25 WSEs to build rating curve.
  • breaks: station values to split hydraulic calcs by. Helpful for separating channel from floodplain.
  • save: optional file path to save htab as excel rather than returning a dataframe object.

htab_2D(rem,extents,vcl,slope,roughness,channel_area=None,channel_roughness=0.035,save=None)

• Returns dataframe of reach-averaged hydraulic table (htab) calcs for a given 2D area. Htab table returns either the sum (for discharge, area, and perimeter) or discharge-weighted average (for roughness, hydraulic radius, and velocity) across all flowing channels.
• Input options:
  • rem: Relative elevation model raster file. Stages in the rating curve will be referencing this REM.
  • extents: Shapefile of area to run hydraulic calcs.
  • vcl: Shapefile of the valley center line, used to calculate reach length.
  • slope: dimensionless slope for hydraulic calcs. Constant across all flows/depths.
  • roughness: Averge or assumed floodplain roughness. Will apply to channel as well without channel_area provided.
  • channel_area: Optional shapefile of active channel area for separate hydraulic calcs from floodplain.
  • channel_roughness: Manning's roughness for channel, only used if channel_area is provided
  • save: optional file path to save htab as excel rather than returning a dataframe object.

pebble_count(photos, obj_size_mm, method='rapid')

• Creates classified image with grain size distribution and CSV file of raw data in the same folder as your image.
• Input options:
  • photos: list of file paths to PNG/JPG of gravel bar with clear identifying object in the photo.
  • obj_size_mm: length of object along major axis in mm.
  • method: "rapid" for a faster Wolman pebble count or "detailed" for an area-weighted count of most visible grains

delineate_trees(ch_file, output, canopy_floor=10, min_ht=60, max_ht=120, min_area=20, combine_dist=5)

• Saves a set of polygons representing delineated trees from a canopy height model.
• Input options:
  • ch_file: Path to the canopy height model raster file.
    • output: Path to save the resulting tree locations shapefile.
    • canopy_floor: Height threshold to consider the ground level of the canopy (default 15).
    • min_ht: Minimum height to consider a detection as a tree (default 50).
    • max_ht: Maximum height to still consider the detection as a tree (default 100).
    • min_area: Minimum area for a tree detection to be considered valid (default 20).
    • combine_dist: Number of pixels within which to combine tree points (default 5).

def HSI(dep_raster,vel_raster,curve,output)

• Determines habitat suitability index given depth and velocity rasters. Values range from 0-1. Curves based on WDFW Instream Flow Guidelines.
• Input options:
  • dep_raster: Path to the depth raster file.
  • vel_raster: Path to the velocity raster file.
  • curve: Species and lifestage curve, options are: -"Adult Chinook Spawning Large River" -"Adult Chinook Spawning Small River" -"Juvenile Chinook Rearing" -"Adult Coho Spawning" -"Juvenile Coho Rearing" -"Adult Sockeye Spawning" -"Juvenile/Adult Rainbow Trout Rearing" -"Spring Chinook Holding" -"O. mykiss Juvenile"
    • output: Path to save the resulting HSI raster.

Dcrit(shear_raster,output,tc=0.06,psed=162.3128,pwater=62.428)

• Determines critical grain size (Dcrit) using Shields' equation. .
• Input options:
  • shear_raster: Path to the shear stress (lbs/ft^2) raster file.
  • tc: dimensionless critical shear, use values between 0.045 and 0.06.
  • psed: density of water (lbs/ft^3)
  • pwater: density of water (lbs/ft^3)
    • output: Path to save the resulting Dcrit raster.