Runoff Estimation in Ungauged Catchments Using Regionalization

Technique, Blue Nile River Basin

Seminar II
By
Temesgen Tsehayneh (PhD. Student)

Tuesday, Aug 16,2022

 Outline of Presentation

 Introduction
 Statement of problem
 Objective of the study
 Material and methodology
 Time Schedule
2
 Introduction

Runoff simulations have a significant influence on sustainable water resources

management and engineering design

Flood protection water supply Real time flood water resources

design forecasting forecasting management

Accurate estimates of continuous streamflow in the given catchments are crucial

Used Rain fall Runoff Model

They all need stream flow data to compare the predictions

3
 Cont.…

It is now recognized that the availability of stream flow data world-wide

is not improving (Takeuchi, 2002, Lavers et al., 2019…)

It is likely that the majority of catchments world-wide are ungauged (Sivapalan, 2003)
It is far from adequate in a lot of countries especially in developing countries like Ethiopia
The transfer of information from
This transfer of information is called
gauged catchments to ungauged
Regionalization (Sivapalan, 1995)
catchments is necessary
This is achieved by extrapolating the model parameters from gauged to ungauged
sites belonging to homogeneous regions. 4
 Cont.…

Regionalization approaches can be classified into hydrologic model-dependent

and hydrologic model-independent groups.

The methods employed by the first group directly transfer rainfall-

runoff model parameters between basins.

The second group does not transfer hydrologic model parameters

instead the structure and parameter equation are transferred.
5
 Statement of problem

 Most sub basins of the Blue Nile River Basin are not gauged
 Even the gauged one had limited hydrologic data
 Little attention has been given to the estimation of runoff at ungauged basins

Hydrologic responses of the catchment in the Due to this the Basin experiences
basin not well understood recurrent drought and famine

This limitation motivates the need to develop a reliable Regionalization model

Thus the question as to whether it will be possible to extend and/or generalize model parameters obtain
through calibration of gauged watersheds to ungauged catchments within the same region. 6
 Cont.…
A number of methods have been attempted to estimate the runoff from ungauged catchments
some of selected previous studies

 According to the literature review, to the best knowledge

of the authors, watersheds have been classified based on
the average characteristics of each sub-watershed.

PCC,s
 However, in Blue Nile basin it is generally difficult to
determine and understand the hydrological behavior of
the ungauged watersheds due to the lack of sufficient and
available data.

7
 Cont.…

 Furthermore, watersheds with the same physical properties or with spatial proximity do
not always present the same hydrological behavior (Patil, Stieglitz, & Sciences, 2011).

 This issue has the potential to be addressed using remote sensing techniques that are
widely used for studying earth events and hydro-climate patterns (Blöschl et al., 2013).

 Therefore, this study will present a pixel-based classification for identification of

homogeneous areas in addition to classification using watershed physical properties and
hydrological behaviors.

One of the most important objectives and interests of International Association

of Hydrological Sciences (IAHS) (Sivapalan et al. 2003). 8
 Objectives of the research
 The main objectives of this study is to test the rainfall-runoff model for the catchment
under study, and estimating daily runoff for the ungauged catchments.

To achieve this general objective, the following specific objectives will addressed
1. To identify hydrologically homogenous regions and propose a watershed classification
methodology prior to regionalization
 Determine Hydrologically Homogenous Regions in the basin
 Evaluate the efficiency of a systematic watershed classification

2. To simulate streamflow using hydrological model on daily time interval for gauged catchments
 To determine model parameters required to estimate streamflow for ungauged catchments
 Selection of catchment characteristics which is used for characterization of streamflow at ungauged catchments
9
 Cont.…

3. To identify appropriate regionalization method that can be used in extending the observed
hydrologic regimes to ungauged basins of different spatial scales in the study area.

 Develop regionalization model

 Selection of the best regionalization technique for Blue Nile River Basin

4. Evaluate and propose an efficient streamflow regionalization methodology with

uncertainty estimate.

10
 Study Area
Covers an area of 324,530 km2.

Provides 62% of flow to the Nile River (Conway, 2005)

Characterized by a highly rugged topography and

The Blue Nile Altitude ranging from 350 m at Khartoum to over 4,250 m Ethiopian highlands.
River Basin
The livelihood of the study area depends largely on rain-fed agriculture and
small-scale irrigation
Mean annual temperature 18.3 C (Kim et.al, 2007)

Originates from Lake Tana Ethiopia

There are few reliable hydrologic stations throughout the entire basin
11
Cont.…

12
 Material and Methodology
Phase 1: Data collection The required input data will adequately collect

Metrological data Hydrological Data Catchment Attributes

Rainfall average annual flow Area
 Soil type
Temperature minimum monthly flow Elevation
 Land Use
Sunshine hour maximum monthly flow Slope

Radiation Stream length

Wind speed Aridity index

National Meteorological Ministry of Water, Irrigation

Agency and Electricity DEM

13
 Selection of representative catchments and catchment
characteristics

 Based on the availability of daily flow records and reliability in the modelling results I will select
the representative catchment characteristics

 In this study I will select a total of 23 PCCs from SRTM DEM analysis, geological maps, land
cover maps and soil maps and 18 sub-basins will be considered in the regionalization procedure.

 SAVI
 LAI
In addition, I will extract remote sensing indices including
 NDMI and
 NDVI from MODIS

14
 Cont.…
Catchment area Reflects volume of water that can be generated from rainfall

Length of longest flow path Distance from the catchment’s outlet to the most distant source on the catchment boundary

Hypsometric integral Describes the distribution of elevation across the catchment area.

Average altitude Average elevation of the catchment from SRTM DEM

Average slope of catchment Calculated from digital elevation model SRTM DEM pixel by pixel
Drainage density Total stream length for the basin divided by catchment area
%Level: Level to undulation/level/, dominant slopes ranging between 0 and 8%
Slope class as per FAO %Hilly: Rolling to hilly/hilly/, dominant slopes ranging between 8 and 30%
%Steeply: Steeply dissected to mountainous/steeply/, dominant slopes over 30%
Circularity index: is the ratio of perimeter square to the area of the catchment.
Basin shape
Elongation ratio: is the ratio of length of longest drainage to diameter of a circle which has the same area as the basin
Land cover (%) forest, grassland, cropland, bare land, urban and built-up, woody savannah
Soil (%) luvisols, leptosols, nitisols, vertisols and fluvisols
 Selected catchment characteristics 15
 Cont.…
Phase 2: Input Variable Reduction (selection) techniques

 Principal component analysis (PCA)

 Missing Value Ratio used to identify the most relevant
watershed characteristics
 Random Forest

Phase 3: Homogeneous areas determination techniques

1. Connectivity-based Clustering (Hierarchical clustering)
2. Centroids-based Clustering (Partitioning methods)     
3. Distribution-based Clustering
Cluster (homogeneous)
4. Density-based Clustering (Model-based methods)
5. Fuzzy Clustering
6. Constraint-based (Supervised Clustering) 16
 Cont.…
Study Area

Watershed attribute
Stream flow Indices
SAVI, LAI, NDMI Variable reduction Techniques
and NDVI from
MODIS Hierarchical clustering  Density-based Clustering
Reference Clustering Centroids-based Clustering        Fuzzy Clustering
Distribution-based Clustering  Supervised Clustering

Identify best classification techniques based on similarity index by comparing Reference clustering

Take bast clustering

Techniques 17
 Cont.…

Phase 4: Individual model calibration using observed runoff at donor catchments)

Phase 5: Develop regionalization model.

Phase 6: Leave-one-out cross-validation (validation). The performance of the proposed

regionalization approach will validate.

Phase 7: Enhanced regionalization (Regional model calibration). Functional

relationships between model parameters and physical-catchment attributes were
established
18
 Cont.…
Doner/ Acceptor (Parameter optimization)

Regionalization Techniques

IDW SPT SBT RBT

Identify best regionalization

Techniques

Hydrological model parameter Develop

Gauged Ungauged regionalization model
Watersheds Watersheds

Continuous streamflow
regionalization
pseudo-ungauged Catchment
Model Performance Evaluation 19
 Time Schedule
Time in Months (Jun 2022 to Jun 2025)
Main Activities
6 12 18 24 30 36 42 48
Proposal Development
Data collection and processing
Literature Review
Input Data Preparation
Modeling Simulation
(Conceptual)
Regionalization model
Model validation work
20
 nd !!
E u !
he o
T k y
a n
T h
21