Rainfall intensity duration frequency (IDF) curve calculation and plotting
- STATUS: Active
- LATEST RELEASE: v.0.4.3
- LAST UPDATED: 2019-11-28
- LICENSE: Public Domain (except where otherwise noted)
- URL: https://github.com/dt-woods/idf
The intensity-duration-frequency (IDF) curve for a given region is often used as a design criteria for hydrologic structures: from rain barrels to dams. Computing an IDF curve allows the designer to quickly calculate expected rainfall quantities at different return periods; answering questions such as "What storm characteristics should I expect in the next five or ten years and how much water will it bring?" In order to more accurately predict future storm characteristics and their return periods, long-term rainfall observations are critical (>10 years). Because different regions will likely have different rainfall characteristics that change over time, a method for quickly computing the IDF curve is advantageous.
This script reads rainfall data, identifies individual rainfall events, calculates each event's duration and intensity, computes the return period probabilities for all storm events, and plots the IDF curve.
Results will be highly dependent on the quality, resolution, and length of the rain data set. Some assumptions have been made and can be edited within the code. Notably is the minimum inter-event time (MIT), which is set to five (5) hours; this may not be suitable for all regions: consult the literature for advice. This script also assumes that storm starting and ending times are immediately before and after they are recorded, which may not be accurate for data sets with long time intervals (e.g., hourly data). The durations and the return periods have been hard-coded; if they are not what you need for your analysis, they are clearly identified in the code for IDF Analysis and Computing the IDF Curve; make adjustments as needed.
Please note that this code is not intended for use with designs that are life-saving or life-threatening.
Although all reasonable efforts have been taken to ensure the accuracy and reliability of this code, the author does not and cannot warrant the performance or results that may be obtained by its use. The author disclaims all warranties, express or implied, including warranties of performance, merchantability or fitness for any particular purpose.
For comments or concerns, please use GitHub's Issues.
idf.py
- Python script for reading precipitation data, identifying rainfall events, and computing/plotting the IDF curve
usage: idf.py [-h] [--usgs] [--make_regular] [--save_plot] [--verbose] file
IDF.py - Calculate IDF curves from rainfall data.
positional arguments:
file input rainfall file; format should be two-column (datetime
and rainfall amount) comma-separated plain text
optional arguments:
-h, --help show this help message and exit
--usgs input file format is based on USGS raingage station; the
script will format the file for you
--make_regular make regular irregular time stamped rainfall.
--save_plot save IDF curve to PNG file
--verbose print out all rainfall events
EXAMPLE 1 - USGS RAINGAGE DATA
Converts tab-separated file to csv and saves the IDF curve as a PNG image file.
python idf.py --usgs --save_plot nwis.waterdata.usgs.gov
EXAMPLE 2 - TWO-COLUMN PLAIN TEXT FORMAT
Plots the IDF curve.
python idf.py rainfall.txt
This script reads one of two types of rainfall data: USGS raingage tab-separated plain text file or a two-column comma-separated plain text file.
EXAMPLE DATA
- URL: https://waterdata.usgs.gov/nc/nwis/inventory/?site_no=354057080362545
- description: Contains 5-minute resolution rainfall data in CSV format for USGS rain gage at RO-149 PIEDMONT RS 1 NR BARBER, NC
- USGS site: 354057080362545
- Retrieved: 2019-04-10
- Parameter: Precipitation, total, inches
- Date range: 2009-10-01 01:00 EDT to 2019-04-10 09:55 EDT
- Rows of data: 960109
DATA ACCESS
- Browse to URL
- Click Current / Historical Observations
- Check 00045 Precipitation
- Output format: Tab-separated
- Begin date: 2009-10-01
- End date: 2019-04-10
- Click Go
- Right-click on page: Save page as "nwis.waterdata.usgs.gov"
DATA FORMAT
# ---------------------------------- WARNING ----------------------------------------
# Some of the data that you have obtained from this U.S. Geological Survey database
# may not have received Director's approval. Any such data values are qualified
# as provisional and are subject to revision. Provisional data are released on the
# condition that neither the USGS nor the United States Government may be held liable
# for any damages resulting from its use.
#
# Additional info: https://help.waterdata.usgs.gov/policies/provisional-data-statement
#
# File-format description: https://help.waterdata.usgs.gov/faq/about-tab-delimited-output
# Automated-retrieval info: https://help.waterdata.usgs.gov/faq/automated-retrievals
#
# Contact: gs-w_support_nwisweb@usgs.gov
# retrieved: 2019-11-24 12:55:21 EST (nadww01)
#
# Data for the following 1 site(s) are contained in this file
# USGS 354057080362545 RAINGAGE AT RO-149 (NC193) PIEDMONT RS 1 NR BARBER
# -----------------------------------------------------------------------------------
#
# Data provided for site 354057080362545
# TS parameter Description
# 90643 00045 Precipitation, total, inches
#
# Data-value qualification codes included in this output:
# A Approved for publication -- Processing and review completed.
#
agency_cd site_no datetime tz_cd 90643_00045 90643_00045_cd
5s 15s 20d 6s 14n 10s
USGS 354057080362545 2010-10-01 00:00 EDT 0.00 A
USGS 354057080362545 2010-10-01 00:05 EDT 0.00 A
USGS 354057080362545 2010-10-01 00:10 EDT 0.00 A
...
Use the --usgs flag to auto-format this file.
You may create your own rainfall data file, so long as it meets the formatting guidelines of:
- single header row (e.g., "Datetime,Rainfall")
- ISO formatted timestamps (i.e., "YYYY-MM-DD HH:MM:SS")
EXAMPLE DATA
datetime,precip_inches
2019-02-06 13:00,0
2019-02-06 13:05,0
2019-02-06 13:10,0.1
...
If your rainfall data comes from a personal weather station or other source that has irregular time stamp intervals, you may use the flag --make_regular to calculate a regular time series (if needed).
- 2019-11-28: v0.4.3
- changed usgs_to_csv rain_file naming (use os.path.splitext)
- 2019-11-27: v0.4.2
- fixed typo in usgs_to_csv
- fixed potential problem in make_regular_ts with variable
my_data - added command line arguments for easier workflows with other rainfall files
- 2019-11-25: v0.4.1
- added check for formatted csv to avoid running conversion twice
- changed usgs_to_csv to accept two string arguments (input/output file names) and return nothing
- added boolean argument to make_plot for saving figure to PNG
- 2019-11-24: v.0.4
- added three new functions: usgs_to_csv, writeout, writeline
- removed file path variables (assumes rainfall file is in local dir)
- new workflow: reads tab-separated usgs rain gage file, processes to csv, runs idf
- 2019-03-23: v.0.3.1
- 2019-03-22: v.0.3.0-dev
This script was tested using Python 3.6.7 and requires the installation of the following third-party packages (version numbers have been successfully tested).
- numpy (1.16.2; 1.17.4)
- scipy (1.2.1; 1.3.3)
- matplotlib (3.0.3)
-
Download the latest version of Python 3 for Windows (https://www.python.org/downloads/windows/).
-
Install locally with pip and defaults (it should result in you having
pycommand line tool). -
Test that the installation works.
- Click the Windows start, search "cmd," and open the Command Prompt app
- Type
py -V
-
Download necessary wheel files for the following three packages:
- NumPy (https://www.lfd.uci.edu/~gohlke/pythonlibs/#numpy)
- SciPy (https://www.lfd.uci.edu/~gohlke/pythonlibs/#scipy)
- Matplotlib (https://www.lfd.uci.edu/~gohlke/pythonlibs/#matplotlib)
Note that the wheel file (.whl) you want should match the Python version you downloaded (for example -cp27- for Python 2.7 or -cp37- for Python 3.7) and its bitness (for example win32 for 32-bit installations or amd64 for 64-bit installations).
-
Install the wheel files using pip on the command line.
- Open the Command Prompt
- Type
cd %USERPROFILE%\Downloadsto move into your Downloads folder - Type
py -m pip install "numpy‑1.16.2+mkl‑cp37‑cp37m‑win32.whl"replacing the file name with whichever version you downloaded - Type
py -m pip install scipy‑1.2.1‑cp37‑cp37m‑win32.whl - Type
py -m pip install matplotlib‑3.0.3‑cp37‑cp37m‑win32.whl
- Download the latest version of Python 3 for Mac (https://www.python.org/downloads/mac-osx/)
- Open the Terminal app and test that the installation works by typing
python3 -Vandpip3 -V - Use pip to install numpy, scipy and matplotlib
- Open the Terminal
- Type
pip3 install numpy - Type
pip3 install scipy - Type
pip3 install matplotlib