Bankfull Regression Equations for CONUS

The dataset generated by the methodology presented in this repository can be directly downloaded from this repository (bankfull_physio.parquet) or HydroShare using this link.

For example, in Python, after installing geopandas and pyarrow, you can download and read this GeoParquet file as follows:

import geopandas as gpd
import urllib.request

url = "https://www.hydroshare.org/resource/5d3ab5cc521344c881c38621fa3769f1/data/contents/bankfull_phyiso.parquet"
file_path = "bankfull_phyiso.parquet"
_ = urllib.request.urlretrieve(url, file_path)

bf = gpd.read_parquet(file_path)

Motivation

For hydrological and hydraulic modeling, the bankfull width and depth are important parameters that are challenging to measure in the field and/or estimate a priori. Many studies suggested different methodologies for estimating these parameters based on various geomorphological and hydrological characteristics. One such study is by Bieger et al. (2015) which proposed a methodology to estimate bankfull width and depth for the conterminous United States (CONUS) using regression equations. They gathered hydraulic geometry data from hundreds of sites across the CONUS and developed regional regression equations for the Physiographic Regions of CONUS. The general form of the regression equations is:

$$ Y = a \times A^b, $$

where $Y$ is the bankfull width or depth, $A$ is the upstream drainage area, and $a$ and $b$ are the coefficients of the regression equation. These coefficients are specific to each physiographic region. Thus, bankfull width and depth can be estimated for any location in the CONUS using upstream drainage area and the coefficients of the corresponding physiographic region.

However, the dataset provided by Bieger et al. (2015) is not readily available for use in hydrological and hydraulic modeling. This repository aims to provide a methodology to generate a geospatial dataset for the bankfull width and depth of the CONUS using the regression equations provided by Bieger et al. (2015). So, users can easily query the bankfull width and depth for any location in the CONUS using the drainage area and the coefficients of the corresponding physiographic.

Methodology

You can run bankfull_physio.py to reproduce the dataset. The script depends on pandas, geopandas, pyogrio, pyarrow, and openpyxl packages that can be installed with pip or mamba:

pip install pandas geopandas pyogrio pyarrow openpyxl

or

mamba install pandas geopandas-base pyogrio pyarrow openpyxl

Here is a brief overview of the methodology to generate the bankfull width and depth dataset:

1. From the supplementary material of Bieger et al. (2015), which is a Microsoft Word file, convert SUPPLEMENTARY TABLE 1 to an Excel file. Save the results to bankfull_equations.xlsx.
2. Extract the coefficients of the regression equations for each physiographic region.
3. Download the Physiographic Regions of CONUS shapefile from the ScienceBase.
4. Ensure the keywords for the regions in the shapefile match the keywords in the Excel file.
5. Use the shapefile and the coefficients to generate the bankfull width and depth dataset for the CONUS.
6. Save the dataset to a GeoParquet file.