GUI using python for extrapolation of a Hydrograph

_{The Brahmaputra River in Majuli island, Assam, India (source: Akash Jyoti Handique 2019).}

---

Background: Gumbel distribution is used to model the maximum or minimum of samples of various distribution. Gumbel distribution predicts the chances of an extreme events like floods, earthquakes and other natural calamity. The ability of the Gumbel distribution to represent the maxima can be used to predict future discharges in a river.

Goal: To predict 50, 100, 150, 200 and 1000 year flood discharges using Gumbel distribution and showcasing the results in a GUI

GitHub Repository URL

https://github.com/aswathCodeX/HASA-Python-Project.git

Theory

Gumbel Distribution

The Gumbel distribution is named after Emil Julius Gumbel. Gumbel distribution is a special case of generalized extreme value distribution.

The gumbel distribution formula is

Q = Q_m.(1+K.C_v)
where

• Q = Probable discharge with a return period of T years

• T = 50, 100, 150, 200, 500, 1000 years

• Q_m = Mean discharge

• K = Frequency factor

  • K = (Y_t - Y_n) / S_d
  • Y_t = -ln (ln(T / T-1)
  • Y_n = Expected Reduce mean of the gumbel distribution
  • S_d = Expected Reduce standard deviation of gumbel distribution

• C_v = Coefficient of variation

  • C_v = S_d / Q_m

In order to determine the value of K, from the gumbel reduce table Yn and Sd needs to be extracted

Data

Gumbel reduce table

In the gumbel reduce table possible values of Expected mean and Expected standard deviation is given with respect to the amount of data available.

The gumbel reduce data referred to is continuous from the 10 to 100 years. After this, data for 150 and 200 years is provided respectively in the table. The reduced data is available in .csv format

Number of data	Reduce mean	Reduce std
10	0.4952	0.9497
11	0.4996	0.9676
12	0.5035	0.9833
...	...	...
100	0.56	1.2065
150	0.5646	1.2253
200	0.5672	1.236

Hydrological data

We have considered Hydrological data for River Neckar using the LUBW Geo-Portal Hydrological data is from Obendorf Neckar station measuring point 409 in Neckar.

Data is available in .csv format

DD-MM-YYYY	Q (m3/s)
01-11-1929	4.086
02-11-1929	6.326
...	...
31-12-2021	11.79

Code-Description

Framework

Functions and Custom Classes were created as mentioned in following UML Diagram

Brief Overview of the Components

• main.py contains 4 functions- verify_gumbel(),get_reduce_mean(),get_reduce_std(), gumbel_distribution()
• discharge_data.pycontains a custom class DischargeDataHandler. This class contains 2 methods and a Magic method
• plot_discharge.pycontains a custom class PlotDischarge. This class contains 2 methods and 4 Magic methods
• gui.py
• config.pyrequired packages and modules are imported and the input file paths that are required for the project
• log.pycontains logging configuration and initiates the logger
• gumbel_reduce.pycontains gumbel_reduce() function

Extracting hydrological and gumbel reduce data

river_discharge_data.csv is the input discharge data file that should discharge data. Currently, the Neckar River data used as input is available from the years 1929 to 2021, separated by ;

Please use this specified format to input different data. Date: DD-MM-YYY and Discharge: Q in (m³/s)

discharge_data.py

This file consists of DisachargeDataHandlerclass which has Methods for handling and printing the Discharge Data, Extracting data from user inputted ".csv" file as an argument data_csv_file and sorting the data.

get_discharge_data()

method sorts the data using pandas DataFrame and resampled to annual maxima using resample(rule='A', kind='period').max() Each year is indexed and returned as pandas DataFrame, which includes Discharge [CMS] and Year columns with maximum annual discharge values in the Discharge [CMS] column.

print_discharge_data()

To print the sorted pandas dataframe this method is used.

---

gumbel_reduce.py

###gumbel_reduce()

function takes an argument csv_file which is the file path for the gumbel reduce data. pandas DataFrame is used to sort the data into 3 columns and assign index column to the Data Index

plot_discharge.py

This script contains a custom class PlotDischarge and 6 methods. Among the methods 4 are magic methods including __init__

Aim of the PlotDischarge class is to plot the hydrograph for the River Discharge data (Neckar) and the extrapolated discharges for 50, 100, 150, 200, 500, 1000 years.

plot_discharge()

Method takes 4 arguments time_series, q_series, title and color

time_series and q_series have to be pandas DataFrames. color and title should be strings.

plot_discharge() is generating a scatter plot for the annual_max hydrological data. The flexibility on the size of the scatter points are controlled by the magic methods __mul__(), __gt__() and __lt__().

###__mul_() Magic method multiplies the scatter points size by a factor of 10. __mul__() method can be accessed by * operator

###__gt__() Magic Method warns the user that scatter points are too large and might overlap and generates a print statement The Hydrograph scatter points are too big. __gt__() can be accessed by >

###__lt__() Magic method warns the user if the scatter points are too small to discern by generating a WARNING msg The Hydrograph scatter points are too large. __lt__() can be accessed by <

###gumbel_plotting() Method takes 4 arguments t_series, q_series which are numpy arrays and title and color. The method generates a line graph for the extrapolated discharge values.

main.py

In the main script verify_gumbel(), get_reduce_mean(), get_reduce_std() and gumbel_distribution() functions are present.

###get_reduce_mean() Function takes an argument and index the gumbel.csv file to extract the reduced mean.

The argument is the number of years of hydrological data available.

###get_reduce_std() takes an argument and index the gumbel.csv file to extract to reduce standard deviation.

The argument is the number of years of hydrological data available

If the number of years of hydrological data that is available is greater than 100 years and less than 150 years, reduce mean and standard deviation of 150 year will be considered. For greater than 150 years data, 200 year value will be taken into account. As the table doesn't contain statistical parameters for values above this range, User will be warned (limitation of the code in it's current state).

###gumbel_distribution() Function calculates the extrapolated discharges for 50, 100, 150, 200, 500 and 1000 years using gumbel distribution and provide a list with discharge values.

The inbuilt pandas library statistical methods shape(), mean(), std() are used to get the discharge data size, mean and standard deviation.

get_reduce_mean() and get_reduce_std() are called to determine the reduced mean and reduced standard deviation.

Inside the for loop all the anticipated parameters of the Gumbel Distribution are calculated and extrapolated discharge is appended to an empty list.

###verify_gumbel()

Function is a wrapper function which wraps the get_reduce_mean() and get_reduce_std() functions to make sure that the indexing value is within the limits of the Gumbel Reduce Data provided in the file gumbel.csv. Otherwise, the function shows an Error.

main()

DischargeDataHandler class is instantiated and get_discharge_data() is called to retrieve and sort the discharge data.

Discharge data is put as an argument in the gumbel_distribution() function and the output is list.

PlotDischarge class is instantiated, Magic methods are exercised.

A dictionary is generated containing the years as keys and the extrapolated discharges as values.

*results *are returned as a Dictionary

GUI

gui.py script contains class OurApp. OurApp class has 7 methods.

Aim of the class is to create Graphic user interface and showing the extrapolated discharges and hydograph, extrapolated graph.

GUI allows the user to choose the discharge data path and the size of the scatter points to generate extrapolated discharge and all the plots. ###__init__()

The method creates a tk frame geometry and 4 buttons along with 2 entries. Buttons are provided with necessary command and user input from two entries are stored.

###call_estimator()

This method gets the user input of return period and shows a message containing the extrapolated discharge by putting arguments in the method estimate_u(). ###sel_discharge_data() this method calls select_file method to get the user selected discharge path. ###select_file() this method open the discharge data file that the user selects.

###estimate_u() This method takes an argument(value of the user return period) and return the extrapolated discharge by indexing the dictionary generated by the Main(). ###plot_hydrograph() the method instantiated DischargeDataHandler with user inputed directory of the discharge and method get_discharge_data() is called to generate pandasDataFrame to plot the scaterred hydrograph. otherwise the method returns np.nan . the method is exercising all the magic methods for scatter size flexibility.

###plot_extrapolation() this method is instantiating DischargeDataHandler class and calling method get_discharge_data to generate the plot for extrapolated discharge. otherwise the method returns np.nan

log.py

The file starts the logger and set up its StreamHandler. Log configurations are also set up.

config.py

This script contains all the necessary imports statements of libraries and packages.

# necessary basic python libraries
    import logging
    import os
    import math
    
# global python libraries used
    
    import numpy as np
    import pandas as pd
    import matplotlib.pyplot as plt
    from matplotlib.backends.backend_tkagg import FigureCanvasTkAgg
    import tkinter as tk
    from tkinter import ttk
    from tkinter.messagebox import showinfo, showerror

If the above packages are not install on the host computer, please do so to run the code without any hassle.

# gumbel_reduce_data.csv file path
gumbel_reduce_path = os.path.abspath('') + '/gumbel_reduced/gumbel.csv'

Please specify the Gumbel Reduce data file path here