Page 1 of 92 - Cover Page Submission ID trn:oid:::1:3081449489

Nuzhat Huma
thesis
Assignment 8

Dont Add class, Dont add assignment, Use View

HCCCI LMS LTI 1.3

Document Details

Submission ID

trn:oid:::1:3081449489 90 Pages

Submission Date 19,235 Words

Nov 17, 2024, 12:53 AM GMT+8

108,551 Characters

Download Date

Nov 17, 2024, 1:07 AM GMT+8

File Name

Mohammed__plagiarized.docx

File Size

5.1 MB

Page 1 of 92 - Cover Page Submission ID trn:oid:::1:3081449489

 Page 2 of 92 - AI Writing Overview Submission ID trn:oid:::1:3081449489

22% detected as AI Caution: Review required.

The percentage indicates the combined amount of likely AI-generated text as It is essential to understand the limitations of AI detection before making decisions
well as likely AI-generated text that was also likely AI-paraphrased. about a student’s work. We encourage you to learn more about Turnitin’s AI detection
capabilities before using the tool.

Detection Groups
1 AI-generated only 11%
Likely AI-generated text from a large-language model.

2 AI-generated text that was AI-paraphrased 11%

Likely AI-generated text that was likely revised using an AI-paraphrase tool
or word spinner.

Disclaimer
Our AI writing assessment is designed to help educators identify text that might be prepared by a generative AI tool. Our AI writing assessment may not always be accurate (it may misidentify
writing that is likely AI generated as AI generated and AI paraphrased or likely AI generated and AI paraphrased writing as only AI generated) so it should not be used as the sole basis for
adverse actions against a student. It takes further scrutiny and human judgment in conjunction with an organization's application of its specific academic policies to determine whether any
academic misconduct has occurred.

Frequently Asked Questions

How should I interpret Turnitin's AI writing percentage and false positives?

The percentage shown in the AI writing report is the amount of qualifying text within the submission that Turnitin’s AI writing
detection model determines was either likely AI-generated text from a large-language model or likely AI-generated text that was
likely revised using an AI-paraphrase tool or word spinner.

False positives (incorrectly flagging human-written text as AI-generated) are a possibility in AI models.

AI detection scores under 20%, which we do not surface in new reports, have a higher likelihood of false positives. To reduce the
likelihood of misinterpretation, no score or highlights are attributed and are indicated with an asterisk in the report (*%).

The AI writing percentage should not be the sole basis to determine whether misconduct has occurred. The reviewer/instructor
should use the percentage as a means to start a formative conversation with their student and/or use it to examine the submitted
assignment in accordance with their school's policies.

What does 'qualifying text' mean?

Our model only processes qualifying text in the form of long-form writing. Long-form writing means individual sentences contained in paragraphs that make up a
longer piece of written work, such as an essay, a dissertation, or an article, etc. Qualifying text that has been determined to be likely AI-generated will be
highlighted in cyan in the submission, and likely AI-generated and then likely AI-paraphrased will be highlighted purple.

Non-qualifying text, such as bullet points, annotated bibliographies, etc., will not be processed and can create disparity between the submission highlights and the
percentage shown.

Page 2 of 92 - AI Writing Overview Submission ID trn:oid:::1:3081449489

 Page 3 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

JIGJIGA UNIVERSITY

PERFORMANCE EVALUATION OF CHIRPS SATELITE

RAINFALL PRODUCT FOR PROBABLE MAXIMUM
PRECIPITATION(PMP) IN DATA SCARE REGION OF UPPER
FAFAN SUB-BASIN, WABI-SHEBELE RIVER BASIN, ETHIOPIA

M.Sc THESIS

MOHAMMED FARAH AHMED

JIGJIGA, ETHIOPIA
JUNE, 2024

Page 3 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 4 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

PERFORMANCE EVALUATION OF CHIRPS SATELITE RAINFALL

PRODUCT FOR PROBABLE MAXIMUM PRECIPITATION(PMP) IN
DATA SCARE REGION OF UPPER FAFAN SUB-BASIN, WABI-
SHEBELE RIVER BASIN, ETHIOPIA

MOHAMMED FARAH AHMED

A THESIS SUBMITTED TO THE DEPARTMENT OF WATER RESOURCES AND

IRRIGATION ENGINEERING, SCHOOL OF WATER RESOURCES ENGINEERING,
JIGJIGA INSTITUTE OF TECHNOLOGY, THE SCHOOL OF GRADUATE STUDIES OF
JIGJIGA UNIVERSITY IN PARTIAL FULFILLMENT OF THE DEGREE OF MASTER OF
SCIENCE IN IRRIGATION AND DRAINAGE ENGINEERING

MAJOR ADVISOR: Dr. DEMELASH WONDIMAGEGNEHU (PhD,

ASSISTANT PROFESSOR)
CO-ADVISOR: TESFU ABEBE (PhD STUDENT)

JIGJIGA, ETHIOPIA
JUNE, 2024

Page 4 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 5 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Declaration

I, Do research, Irrigation and Drainage Engineering Department, Jigjiga University, Somali

Region, Ethiopia, and hereby declare that this research, entitled “PERFORMANCE
EVALUATION OF CHIRPS SATELITE RAINFALL PRODUCT FOR PROBABLE
MAXIMUM PRECIPITATION(PMP) IN DATA SCARE REGION OF UPPER FAAFAN SUB-
BASIN, WABI-SHEBELE RIVER BASIN, ETHIOPIA”, was done by me, and this thesis was
composed by myself. The work contained here is my own, and this research work has neither been
submitted nor presented for any reason, neither by me nor any other student, in this university or
any other institution.

Name: Mohamed Farah Signature :

Page 5 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 6 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Approval Page

PERFORMANCE EVALUATION OF CHIRPS SATELITE RAINFALL

PRODUCT FOR PROBABLE MAXIMUM PRECIPITATION(PMP) IN
DATA SCARE REGION OF UPPER FAFAN SUB-BASIN, WABI-
SHEBELE RIVER BASIN, ETHIOPIA

Submitted by:
Mohammed Farah
Name of student Signature Date

Approved by:
1. Dr. Demelash Wondimagegnehu (Assistant Professor)
Major Advisor Signature Date

2. Tesfu Abebe (PhD Student)

Co-Advisor Signature Date

3. __________________
External Examiner Signature Date

4. __________________
Internal Examiner Signature Date

5.
Department Head Signature Date

6.
Dean,SGS Signature Date

7.
Program Coordinator, SGC Signature Date

I
Page 6 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489
 Page 7 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

ACKNOWLEDGMENTS

First of all, I would like to thank the Allah for his grace upon me during all my works and in all
my life. My sincere and special thanks indebted to my supervisor Dr. Demelash Wondimagegnehu
(Assistant Professor) for his supervision, encouragement, critical comment and helpful guidance
since early age of the thesis work. Without him, this research work wouldn’t have been realized. I
would like to express my sincere gratitude to my employer Jigjiga University (JJU) for providing
me to learn my MSc. Program and research finance. I gratefully acknowledge to all offices and
personalities who have given me the necessary data for my study: Particulary, National
Metrological Agency of Ethiopia. I would also like to express my great appreciation to JJU School
of Graduate Study staff members and libraries for their provision of the necessary reference books
and assistance for my study. I would like to extend my appreciation to all members of JJU staff
member especially Water resources and Irrigation Engineering staff members and my class mate
for their support, socialization and help each other. My special thanks also should reach to Tesfu
Abebe who provided all the necessary information for my thesis work. I am also highly indebted
to my special friends, Mohamed Ismail & Firew Kasahun for their giving moral, and support for
me. Last but not least, I would like to thank all my family members especially my mother, for her
love, care, support and pray for me though out my life; my wife for her continoues support to
complete this thesis work.

Mohammed Farah

II
Page 7 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489
Page 8 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

ABBREVIATIONS

CHIRPS Climate Hazards Group Infrared Precipitation with Stations

EV1 Extreme Value Distribution one
EV2 Extreme Value Distribution two
EV3 Extreme Value Distribution three
FBS Frequency Based Storm
FOS Factor of Safety
ITCZ Inter Tropical Coverage Zone
K-S test Kolmogorove-Smirnove test
m.a.s.l Meters above sea level
MNSA National Meteorological Agency
NOAA National Oceanic and Atmospheric Administration
NRC National Research Council
NWS National Weather System
PE Probability of exceedance
PMF Probable Maximum Flood
PMP Probable Maximum Precipitation
USACE United State of America Corps Engineers
USBR Unites State Bureau of Reclamation
WMO World Meteorological Organization

III
Page 8 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489
 Page 9 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

TABLE OF CONTENTS

Contents page
Declaration.................................................................................................................................................... I
Approval Page .............................................................................................................................................. I
ACKNOWLEDGMENTS .......................................................................................................................... II
ABBREVIATIONS .................................................................................................................................... III
TABLE OF CONTENTS .......................................................................................................................... IV
LIST OF TABLES .................................................................................................................................... VII
LIST OF FIGURES ................................................................................................................................. VIII
ABSTRACT ................................................................................................................................................ IX
1. INTRODUCTION............................................................................................................................... 1
1.1 Background ................................................................................................................................. 1
1.2 Problem statement ...................................................................................................................... 2
1.3 Objectives of the Study ............................................................................................................... 2
1.3.1 Specific Objectives .............................................................................................................. 3
1.4 Research questions ...................................................................................................................... 3
1.5 Significance of the study ............................................................................................................. 3
1.6 Scope and Limitation of the study ............................................................................................. 4
1.7 Organization of the thesis ........................................................................................................... 4
2 LITERATURE REVIEW .................................................................................................................. 5
2.1 Probable Maximum Precipitation (PMP) ................................................................................. 5
2.1.1 Estimation of PMP .............................................................................................................. 5
2.1.2 Statistical Methods for PMP ............................................................................................. 5
2.1.3 Extraction of annual maximum Precipitation (GBP) ...................................................... 6
2.2 Data reconstruction..................................................................................................................... 6
2.3 PMP estimation using Hershfield Approach ............................................................................ 7
2.3.1 Advantages of Hershfield Method ..................................................................................... 7
2.3.2 Estimation Precipitation Ratio .......................................................................................... 8
2.4 Satellite Precipitation Products ................................................................................................. 8
2.5 Probability Distribution Function ........................................................................................... 10
2.5.1 Data fitting ......................................................................................................................... 11
2.5.2 Testing the goodness of fit ................................................................................................ 11

IV
Page 9 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489
 Page 10 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

2.5.3 Derivations for Return Periods ........................................................................................ 12

2.6 Performance of CHIRPS .......................................................................................................... 13
3 Materials and Methods ..................................................................................................................... 14
3.1 Description of study area .......................................................................................................... 14
3.1.1 Climate ............................................................................................................................... 15
3.1.2 Topography ....................................................................................................................... 16
3.1.3 Soil Type ............................................................................................................................ 17
3.1.4 Land use/Land cover ........................................................................................................ 19
3.2 Data and sources ....................................................................................................................... 21
3.3 Data collection ........................................................................................................................... 22
3.4 Data quality and analysis ......................................................................................................... 22
3.4.1 Test of consistency, Homogeneity and trend of ground-based precipitation ............... 23
3.4.2 Test of consistency, homogeneity and trend of CHIRPS ............................................... 27
3.5 Bias Correction of CHIRPS ..................................................................................................... 31
3.6 PMP Estimation Procedure...................................................................................................... 32
3.6.1 Hershfield Approach for PMP Analysis ......................................................................... 33
3.6.2 Advantage of Using Hershfield Method .......................................................................... 34
3.6.3 The magnitude of PMP to maximum observation rainfall Ration . Error! Bookmark not
defined.
3.7 Fitting Data to the Probability Distribution Functions ......................................................... 35
3.8 Testing the Goodness of Fit of Data to Probability Distribution .......................................... 36
3.9 Intensity Duration Frequency (IDF) ....................................................................................... 37
3.10 Estimation of PMP, PDF and CDF.......................................................................................... 38
3.10.1 Hershfield method ............................................................................................................. 38
3.10.2 Frequency analysis and Best-fit ....................................................................................... 40
3.11 Estimation of Return Period Values for PMP ........................................................................ 42
3.12 Performance Evaluation of CHIRPS ...................................................................................... 43
4 RESULTS AND DISCUSSION ....................................................................................................... 45
4.1 Probability Exceedence (PE) of PMP of Fafan Sub-basin .................................................... 45
4.1.1 PE of PMP of ground-based Precipitation...................................................................... 45
4.1.2 PE of PMP of CHIRPS Precipitation .............................................................................. 46
4.2 Flow Duration Curves of GBP ................................................................................................. 47
4.3 Flow Duration of CHIRPS ....................................................................................................... 49

V
Page 10 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489
 Page 11 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

4.4 Maximum Frequency Factor (Km)........................................................................................... 51

4.5 Ratio of PMP to Highest Observed Precipitation (HOP) ...................................................... 53
4.6 Comparison of the Probability Distribution Functions ......................................................... 56
4.6.1 Probability Density Functions (PDFs)............................................................................. 57
4.6.2 Cumulative Distribution Functions (CDFs).................................................................... 58
4.7 Performance of CHIRPS .......................................................................................................... 62
5 CONCLUSIONS AND RECOMMENDATION ............................................................................ 64
5.1 Conclusions ................................................................................................................................ 64
5.2 Recommendation....................................................................................................................... 64
REFERENCE ............................................................................................................................................ 66
ANNEX ...................................................................................................................................................... 74

VI
Page 11 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489
 Page 12 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

LIST OF TABLES

Table 2-1. Performance statistics description ............................................................................................. 13

Table 3-1: CHIRPS satellite rainfall stations available in and around Fafan Sub-basin with area coverage
and recording period of 1990-2020............................................................................................................. 21
Table 3-2. Statistical Estimated value of Fafan Sub-basin Precipitation for GB of selected metrological
stations ........................................................................................................................................................ 26
Table 3-3. Homogeneity test methods used in Fafan Sub-basin Precipitation for GB of selected eight
metrological stations ................................................................................................................................... 27
Table 3-4. Statistical Estimated value of Fafan Sub-basin Precipitation for CHIRPS of selected
metrological stations ................................................................................................................................... 28
Table 3-5. Homogeneity test methods used in Fafan Sub-basin Precipitation for CHIRPS of selected eight
metrological stations ................................................................................................................................... 29
Table 3-6: Values of Standard Normal Deviate for the Cumulative Normal Distribution ......................... 35
Table 4-1. Maximum KT of ground-based Fafan sub-basin ....................................................................... 52
Table 4-2. Maximum KT of CHIRPS of Fafan sub-basin .......................................................................... 52
Table 4-3: The maximum frequency factors for different basins or regions compared with Fafan sub-basin
maximum frequency factors ....................................................................................................................... 53
Table 4-4. Probable Maximum Precipitation (mm) PMP to HOP Ratio of Fafan sub-basin using ground-
based precipitation ...................................................................................................................................... 54
Table 4-5. Probable Maximum Precipitation (mm) PMP to HOP Ratio of Fafan sub-basin using CHIRPS
precipitation ................................................................................................................................................ 54
Table 4-7. Performance statistics of CHIRPS rainfall estimates of Upper Fafan Sub-basin (1990-2020)
annual time scale based ............................................................................................................................... 63

VII
Page 12 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489
 Page 13 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

LIST OF FIGURES

Figure 3-1: Map of Study sub basin.............................................................................................. 15

Figure 3-2. Mean monthly Precipitation, mean maximum and minimum temperature of Fafan
Sub-basin....................................................................................................................................... 16
Figure 3-3. Upper Fafan sub-basin Elevation map ....................................................................... 17
Figure 3-4. FAO Soil Map of Upper Fafan Sub-Basin ................................................................. 18
Figure 3-5. Land use land cover (2018) of Upper Fafan Sub-basin ............................................. 20
Figure 3-6. Image by Mohamed Farah Ahmed for Cactus (A) , Lantana camara (B) and Acacia
(C) found in Upper Fafan Sub-basin and surrounding.................................................................. 21
Figure 3-7: Thiessen Polygon map of Fafan Sub-basin ................................................................ 23
Figure 3-8. Long-Term maximum precipitation of Ground Based of Fafan Sub-basin ............... 24
Figure 3-9: Homogeneity test of Fafan Sub-basin Precipitation for GB of selected eight
metrological stations using Pettit test ........................................................................................... 26
Figure 3-10: Long-Term maximum precipitation of CHIRPS of Fafan Sub-basin ...................... 28
Figure 3-11: Consistency test of precipitation of CHIRPS of Fafan Sub-basin of eight stations . 28
Figure 3-12. Homogeneity test of Fafan Sub-basin Precipitation for CHIRPS of selected stations
using Pettit test .............................................................................................................................. 31
Figure 4-1. Probability of Exceedance of ground-based Precipitation of Gursum, Degehabour,
Jigjiga and Gode ........................................................................................................................... 45
Figure 4-2. Probability of Exceedance of ground-based Precipitation of Hartisheik, Hadew,
Chinaksen and Kebri Dehar .......................................................................................................... 46
Figure 4-3. Probability of Exceedance of CHIRPS Precipitation of Gode, Gursum, Chinaksen
and Degehabour ............................................................................................................................ 47
Figure 4-4. Probability of Exceedance of CHIRPS Precipitation Hadew, Hartishek, Kebri Dehar
and Jigjiga ..................................................................................................................................... 47
Figure 4-5. Flow duration curves of ground based precipitation .................................................. 49
Figure 4-6. Flow duration curves of CHIRPS precipitation ......................................................... 51
Figure 4-7. Relation between CHIRPS and estimated PMP using Gumbel distribution function 55
Figure 4-8. Relation between GB and estimated PMP using Gumbel distribution function ........ 56
Figure 4-9. The PDF of ground-based precipitation all stations using normal distribution function
of Fafan sub-basin ......................................................................................................................... 57
Figure 4-10. The PDF of ground-based precipitation all stations using Log-normal distribution
function of Fafan sub-basin .......................................................................................................... 58
Figure 4-11. The PDF of ground-based precipitation all stations using Weibul distribution
function of Fafan sub-basin .......................................................................................................... 58
Figure 4-12. The CDF of ground-based precipitation of all stations using normal probability
distribution function of Fafan sub-basin ....................................................................................... 60
Figure 4-13. The CDF of CHIRPS precipitation of all stations using normal probability
distribution function of Fafan sub-basin ....................................................................................... 62

VIII
Page 13 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489
 Page 14 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

ABSTRACT

Probable Maximum Precipitation (PMP) holds significant importance in determining the optimal
precipitation depth for a specific location at any given time of the year. PMP serves as a vital
parameter for designing road crossing structures and planning agricultural activities in particular
areas. The development of PMP relies on various factors such as ground-based and satellite
precipitation data, frequency distribution analysis, flow duration curves, and probability functions,
all of which illustrate the variability of precipitation over the study period.This study aimed to
assess Probable Maximum Precipitation (PMP) using both ground-based and satellite precipitation
datasets in the data-limited Fafan Sub-Basin within the Wabi-Shebele River Basin, Ethiopia. A
31-year dataset was utilized for this evaluation. Data quality was ensured through Theissen
Polygon analysis, double mass curve examination for consistency, and Pettit test for homogeneity.
Subsequently, a frequency factor (Km) was derived from the refined dataset, revealing notable
variations across stations within the Fafan sub-basin. The highest frequency factor was observed
at the Chinaksen gauging station, while the lowest was recorded at the Degehabour station. This
investigation underscores the spatial variability of PMP across different monitoring stations. The
comparison between Probable Maximum Precipitation (PMP) and the highest observed
precipitation revealed nearly identical values. To determine the most accurate model for each
station, various probability distribution functions, including normal probability distribution
(NPD), log-normal probability distribution (Log-NPD), log Pearson type-III (LPT-III PD), and
Gumbel Extreme Value 1 (Gumbel-EV1), were employed. Predicted values were then evaluated
using the coefficient of determination (R2). The findings indicated that the log Pearson type-III
model performed the best, followed by the Log-normal and Gumbel-EV1 models. Furthermore,
the study analyzed the probability density function (PDF) and cumulative density function (CDF),
revealing deviations of precipitation data from the mean and standard deviation. These variations
underscored the precipitation variability within the study sub-basin. The results suggest that for
more reliable findings, it is advisable to prioritize uniformly distributed stations and utilize more
extensive and updated datasets due to the dynamic climate patterns within the study sub-basin.
Overall, understanding the PMP of the Fafan sub-basin is crucial for designing road crossing
structures, managing floods, implementing agricultural water management practices, and
effectively developing and managing water resources within the sub-basin, thereby aiding in the

IX
Page 14 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489
 Page 15 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

sustainable management of natural resources. Furthermore, the study analyzed the probability
density function (PDF) and cumulative density function (CDF), revealing deviations of
precipitation data from the mean and standard deviation. These variations underscored the
precipitation variability within the study sub-basin. The results suggest that for more reliable
findings, it is advisable to prioritize uniformly distributed stations and utilize more extensive and
updated datasets due to the dynamic climate patterns within the study sub-basin.

Key words: PMP, Ground-based, CHIRPS, FDCcurves, (PMP), , CHIRPS, Probable

distribution, Data scarce, Upper Fafan Sub-basin

X
Page 15 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489
 Page 16 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

1. INTRODUCTION

1.1 Background
Precipitation, the water that falls to the Earth's surface in liquid or solid form, plays a crucial role
in shaping our planet's environment. Variations in precipitation patterns, both in terms of intensity
and frequency, have a significant impact on various environmental and societal factors. These
variations can influence hydrological processes such as runoff, streamflow, and the occurrence of
floods and droughts, which in turn affect communities worldwide.

Accurate estimation of Probable Maximum Precipitation (PMP) is essential for the design of water
resource infrastructure. PMP represents the theoretical maximum amount of precipitation that
could occur at a specific location over a given duration. It is used in various applications, including
weather forecasting, hydrological modeling, and structural design.

While traditional methods for estimating PMP rely on long-term rainfall records from rain gauges,
many regions, especially in developing countries, lack sufficient and reliable data. This poses a
significant challenge in accurately assessing PMP and understanding its implications for water
resource management.

Satellite Precipitation Products (SPPs) offer a valuable alternative for estimating precipitation in
data-scarce regions. These products provide global coverage and can help fill gaps in ground-based
observations. However, it is important to carefully evaluate the accuracy and reliability of SPPs
before using them for hydrological applications.

In this study, we focus on the Climate Hazards Group Infrared Precipitation with Stations
(CHIRPS) satellite rainfall product. CHIRPS has been widely used in various hydrological studies,
including PMP estimation. By assessing the performance of CHIRPS in the Faafan Sub-basin, a
region with limited ground-based data, we aim to contribute to the understanding of its
applicability in data-scarce areas. This information can be valuable for water resource management
and decision-making in the region.

Page 16 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 17 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

1.2 Problem statement

Wabi-Shebelle River basin is experienced by Agricultural and Agri-pastoral activities for a long
period of time. Wabi-Shebelle River passes through the basin that serves the community by
providing water for drinking, hydropower at the upper of the river basin and irrigation specifically
at the lower of the river basin. The downstream of the river basin is influenced by flood inundation
at Mustahil, Kelafo and Gode area. Flood design at this area must be important; to determine the
flood design, the PMP estimation will be significant input for flood modeling, prediction, as well
as the development of flood duration curves (FDC) based on various years of return periods.

Flood design will be useful for infrastructure expansion such as highway and hydraulic structures
like dams and spillways, irrigation structures such as canals, diversion structures and irrigation
agriculture fields. For all these types of structures the PMP value will be valuable due to it proves
the highest precipitation for the river basins. Evaluation of PMP using GBD and SPP will be a
solution for ungauged watersheds to implement the SPP data during the precipitation scarcity and
failure of rain gauges. The selected SPP data providers will be used to study climate change,
streamflow, flood frequency analysis, water availability identification, rainfall trend analysis,
groundwater investigation and water resources development and management at a large-scale
level.

Surface irrigation modeling relies on precipitation, which is further used to determine runoff
analysis for surface water harvesting structure building at water scarcity sites in the river basin. In
the lower Wabi-Shebel river basin, there is a problem with gauging stations and those existing are
far apart by long kilometers which of out of standards of gauging stations installment. Hence, SPP
data providers will solve the precipitation data scarcity for water resources related studies.
Therefore, the main purpose of this study focuses on the evaluation and selection of the best of the
SPP types by cross-checking with GBP based on the output that will be obtained for the PMP.

1.3 Objectives of the Study

The main objective of this study is to evaluate Probable Maximum Precipitation (PMP) Using
Ground-Based and Satellite Precipitation Products Data in Faafan Sub-basin.

Page 17 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 18 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

1.3.1 Specific Objectives

To achieve the general objective, the following specific objectives were drawn: -

 To evaluate the performance of CHIRPS satellite rainfall, estimate over Faafan Sub-basin
basin.
 To evaluate the applicability of CHIRPS for point SPP and PMP in the Faafan Sub-basin
using selected probability functions.
 To estimate Flow Duration Curve (FDC) using PMP in different set

1.4 Research questions

• Why evaluate the performance of CHIRPS satellite rainfall and estimate over Faafan Sub-
basin basin?
• How to evaluate the applicability of CHIRPS for point SPP and PMP in the Faafan Sub-
basin using selected probability functions?
• Why estimate the FDC using the PMP and SPP based on different returning periods?

1.5 Significance of the study

Availability of evaluated and designed storms using the PMP and SPP will be useful for water
harvesting’s structure designs such as dams, irrigation command areas, and agricultural and agro-
pastoral activities consulting. The SPP evaluation and selection will solve the rainfall data shortage
in the river basin. For any study, project development, and decision making the SPP data suppliers
will be used to accomplish successful work. Hence, studying the relations between PMP and SPP
can be used fully to know the amount of precipitation that occurs in the Wabi-Shebele River Basin.
Selection of the SPP is an option for precipitation data shortage replacement to study agricultural,
engineering, and natural disasters like flood and drought in the river basin in the particular lower
Wabi-Shebele river basin and further for the Somali Regional State level. The development of the
Wabi-Shebele River Basin PMP and SPP Isohyetal map can be an indicator of the rainfall
condition of the river basin and its spatial distribution of precipitation. Generally, agricultural
experts, designers of water harvesting structures, irrigation experts, and meteorological experts
will be benefited from this research work to substitute SPP instead of PMP ungauged areas in the
river basin.
3

Page 18 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 19 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

1.6 Scope and Limitation of the study

This study is limited in scope by relying on weather data from the National Meteorological Agency
of Ethiopia and downloaded precipitation data from selected SPP websites for the Faafan sub-
basin. While the study proposes to estimate point PMP and SPP for rain gauge stations in the sub-
basin using various probability functions for different return periods, it is limited to developing an
Isohyetal map of the Faafan sub-basin. The research proposal's broader scope, which included
exploring the impact of precipitation on climate, water yield, hydrological water balance, drought,
and flood prediction, was not addressed in this study.

1.7 Organization of the thesis

This thesis consists of five chapters. Chapter one consists background, problem statement,
objectives, research questions, significance of the study, scope and limitation of the study and
organization of the thesis. The second chapter focused on literature review. The third chapter is
about study area description, data collection and analysis and methods. The fourth chapter is results
and discussion in general while it focuses on PMP estion based on sateleite and ground based
precipitation data. The final section of this thesis section is chapter five which has conclusions and
recommendations.

Page 19 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 20 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

2 LITERATURE REVIEW

2.1 Probable Maximum Precipitation (PMP)

The Probable Maximum Precipitation (PMP) has long been employed in the design of flood hazard
reduction structures, such as flood retention basins and levees along riverbanks. Although these
techniques have been used since the 1950s, they remain efficient in managing stormwater events
(Pearce and Kennedy, 1993). The World Meteorological Organization (WMO) defines PMP as the
theoretically maximum depth of precipitation that can occur over a given river basin at a particular
time of year (WMO, 1973). This definition incorporates advancements in the understanding of
atmospheric physics. Hansen et al. (1982) note that PMP was initially formulated without
considering climatic trends due to their slower impact relative to other uncertainties. In 1982, the
definition of PMP was updated to reflect the maximum depth of precipitation possible for a given
geographic area and time, regardless of climatic variability. Hansen (1994) later added that storm
size also influences precipitation, challenging earlier assumptions about PMP's dependence on
basin area. Consequently, PMP is now viewed as the maximum estimated precipitation depth over
a specific period, though actual measured precipitation has not exceeded PMP estimates
(Kourtsoyiannis, 1999).

2.1.1 Estimation of PMP

Accurate estimates of extreme precipitation, such as PMP, are critical for hydrological design.
Many countries, including the United States, India, Australia, the United Kingdom, and China,
utilize PMP in flood management and related infrastructure design (Svensson and Rakhecha,
1998). Several methodologies exist for estimating PMP, including storm modeling, maximization
and transposition of actual storms, and statistical approaches (Koutsoyiannis, 1999). For example,
the WMO (1986) and Hansen et al. (1988) recommend generalized data for basin areas larger than
1000 km² to adjust point values of precipitation.

2.1.2 Statistical Methods for PMP

Implementation of a statistical method of analysis on PMP researcher should consider that the
adequacy of sample size, quality such as consistency and homogeneity of the sample. Based an
adequate sample and homogeneous data, are required for enhanced precision and accuracy of
5

Page 20 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 21 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

results in statistical methods of analysis. Such problems of entirety and homogeneity of the sample
in extreme precipitation analysis can be filled first by extracting daily annual precipitation time
series and identifying the missed year’s data; testing the consistency of records and finally
reconstructing the data for any gaps and inconsistency of records (Chow et, al., 1988).

In Statistical Methods on PMP computation, the sample size is central and has a significant role in
extreme precipitation analysis for the estimation of PMP. Theoretically, a sample size as small as
ten has no significant difference in the estimation of the probability of exceedance of a variety as
compared to the parent distribution, however, in practice, this small size implies a few tens of years
cannot provide sufficient empirical approach for preferring the best fit distribution (Koutsoyiannis,
2004).

Although the small sample in a few tens of years is not adequate, recently this size has been used
for the derivations of point extreme rainfalls for larger return periods and PMP estimations. For
instance, Rezacova et al (2005) have derived point PMP for river basins in the Check Republic for
rain gauge stations having 10 and more years of daily annual maximum Precipitation record.
Similarly, Rakhecha et al. (2006) in Malaysia and Alemayew and Samu (2010) in the Ethiopian
river basin have estimated point PMP using precipitation records of 15.

2.1.3 Extraction of annual maximum Precipitation (GBP)

Daily extreme precipitation data and time series are scrutinized for irregularities. Stations with
over 30% missing data or consecutive data gaps exceeding three years are unsuitable for statistical
analysis. Time series with more than 90% data completeness are ideal for analyzing trends in
extreme precipitation (Peralta-Hernandez et al., 2009).

2.2 Data reconstruction

Daily extreme precipitation data and time series are scrutinized for irregularities. Stations with
over 30% missing data or consecutive data gaps exceeding three years are unsuitable for statistical
analysis. Time series with more than 90% data completeness are ideal for analyzing trends in
extreme precipitation (Peralta-Hernandez et al., 2009).

Page 21 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 22 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Numerous methods have been developed to fill gaps in precipitation data, including the station
average method and isohyetal methods (Chow et al., 1988). The Hershfield approach is frequently
used for calculating PMP values statistically (WMO, 1986; Koutsoyiannis, 1999). For basins larger
than 1000 km², area reduction curves are applied to adjust point values (WMO, 1973). Studies in
India have demonstrated that PMP estimates closely match the highest recorded rainfall values in
some regions (Dhar et al., 1969, 1971; Rakhecha et al., 1992). In Ethiopia’s Blue Nile Basin, PMP
was found to be 1.9 times greater than the highest observed 1-day rainfall recorded over 10,000
years (Alemayew and Semu, 2010).

2.3 PMP estimation using Hershfield Approach

Hershfield approach has been used worldwide for statistically calculating PMP magnitudes
(WMO, 1986; Koutsoyiannis, 1999). It requires a series of maximum annual daily precipitation
values for a particular location at a particular time in the study period. The WMO (1973) and
WMO (1986), suggested that basin area larger than 1000 km2 requires area reduction curves for
adjusting the point values to various magnitudes of precipitation. The method is widely applied in
India (Dhar and Kamte, 1969 & 1971) and Rakhecha et al. (1992) works are some of the several
studies that have carried out an extensive investigation into the estimation of point PMP for one
day for stations throughout the country.

Where, X1 is the highest observed annual maximum rainfall in the series, Xn-1 is the mean of the
annual maximum, excluding the highest value and σ , is the standard deviation of the annual

maximum, excluding the highest value.

2.3.1 Advantages of Hershfield Method

Collier and Haradaker (1996) have defended the inadequacy of storm models in estimating PMP
by pointing to their sequential approach to statistical techniques. Rainstorm analysis using hydro-
meteorological approaches is ineffective and necessitates a thorough examination of historical
rainfall events (WMO, 2009). As demonstrated in the USA (Myers, 1967) and more recently in
the Netherlands (Koutsoyiannis, 1999), the statistical approach is much quicker, easier to use, and
freely available. It also enables the design engineer to obtain the average maximum precipitation
estimates for an area (WMO, 2009). The PMP estimates obtained by this approach are also closely

Page 22 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 23 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

comparable to those obtained by other estimates, such as hydro-meteorological methods. In their

numerous technical publications and manuals (WMO, 1969, 1970, 1986), the World
Meteorological Organization (WMO) has also suggested this approach for estimating point PMP
for river basins with daily rainfall data that have been available for a considerable amount of time.

2.3.2 Estimation Precipitation Ratio

PMP ratios can also serve as a factor of safety in design, with values comparable to safety factors
used in structural and geotechnical engineering (Al-mamu and Hashim, 2004). In regions with
sparse rainfall data, satellite precipitation products have become an essential tool. Algorithms
combining microwave (MW) and infrared (IR) data have improved the accuracy of satellite-based
precipitation estimates (Huffman et al., 2007). Additionally, the Tropical Rainfall Measuring
Mission (TRMM) and Precipitation Estimation from Remotely Sensed Information using Artificial
Neural Networks (PERSIANN) provide critical data for hydrological applications (Sorooshian et
al., 2000).

2.4 Satellite Precipitation Products

The availability of high-resolution satellite precipitation products has made them especially
attractive for hydrological applications in regions with less dense and unreliable ground-based
observations. Due to their (near) real-time accessibility, some of these items can be utilized for
flood predictions. The idea behind these high-resolution satellite precipitation algorithms is that
we can take advantage of the complementing capabilities of both the more frequent (but indirect)
infrared (IR) and the more accurate (but less frequent) microwave (MW) data. Combining can be
done in a variety of ways. The Tropical Rainfall Measuring Mission (TRMM) Multi-satellite
Precipitation Analysis (TMPA) 3B42RT method (Huffman et al. 2007) produces estimates that
utilize the calibrated IR estimates in situations where MW data is unavailable, as well as MW-
derived rainfall estimates when and where MW data are available. A neural-network methodology
is used in the Precipitation Estimation from Remotely Sensed Information using Artificial Neural
Networks (PERSIAN) method (Sorooshian et al. 2000) to identify correlations between IR and
MW data, which are then applied to the IR data to provide rainfall predictions. In order to transmit
raining pixels, the National Oceanic and Atmospheric Administration's (NOAA) Climate
Prediction Center morphing technique (CMORPH) method (Joyce et al. 2004) uses a tracking

Page 23 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 24 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

strategy in which infrared data are only employed to create a cloud motion field. solely rainfall
estimates derived from MW data are used in CMORPH, whereas PERSIANN solely uses estimates
derived from IR data. In 3B42RT, rainfall estimates generated from MW data are used when they
are available; in the absence of MW data, estimates generated from IR are used. The resolutions
of CMORPH and PERSIANN are 0.25° and 3-hourly, respectively, however they also provide
finer resolutions.

Satellite precipitation products could be particularly useful in hilly conditions where it is difficult
to monitor large areas with a dense rain gauge network. Rain gauge sampling is, at best, subpar
even in technologically advanced nations, while in less developed regions, the gauges are widely
distributed and usually seen as unreliable. However, it is challenging to comprehend the accuracy
of satellite products in mountainous regions due to a lack of reliable ground-based observations.
The IR rainfall retrieval algorithms may miss light-precipitation events (Hong et al. 2007) or
underestimate total rainfall in mountainous regions due to the relatively warm clouds in these
regions (e.g., Bitew and Gebremichael 2010) due to the thermal IR thresholds commonly used to
differentiate between raining and non-raining clouds.

Integrated Multi-satellitE Retrievals for GPM (IMERG) (Huffman et al., 2020, 2019) and four
other well-known SPPs were selected in order to integrate multiple satellite sensor types, including
infrared (IR) and passive microwave (PMW), as well as gauge-adjusted and non-adjusted outputs.
This has made measuring the impact of these factors on SPP skills easier. The strengths and
limitations of the SPPs across Australia's many climate areas were also shown by this analysis,
since the competency of these datasets differs according to precipitation type, climate, and
topography (Pipunic et al., 2015). The current investigation looks at the application and usage of
the SPPs using one-day annual time periods.

Another SPP for estimating climatic variability in any region is CHIRPS. Recently, Funk et al.
(2015) used historical data from 1981 to the present to generate the long-period, high-resolution
CHIRPS satellite precipitation product (0.05°). The World Meteorological Organization (WMO)
states that more than 30 years of historical data are needed for climate studies, even though the
CHIRPS offers 40 years of precipitation data. This is one of the main advantages of the CHIRPS
program for hydrological and climatological modeling. To evaluate this product from the

Page 24 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 25 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

viewpoints of drought, water, and weather, numerous research have been carried out (Zambrano
et al. 2016; Guo et al. 2017).

2.5 Probability Distribution Function

In statistical analysis, the probability distribution function expresses the likelihood of different
possible outcomes for a given set of variables. Key concepts such as the probability density
function (PDF) and cumulative distribution function (CDF) are foundational when working with
various distribution models, including the normal distribution (ND), Log-Pearson type III (LP-III),
Gumbel Distribution (Extreme Value Type I, EV I), and Weibull distribution.

Rainfall patterns are often unpredictable and variable, but design rainfall for specific return periods
can still be estimated using several probability distributions. Bhaka et al. (2006) demonstrated how
hydrological variables, such as rainfall, need independent observations that represent an unbiased
and broad population. These observations typically fall into four main data categories: 1) complete
duration series, 2) annual series, 3) partial duration series, and 4) extreme value series. Among
these, the partial duration series often results in higher rainfall estimates for lower return periods,
as it generally has a higher mean and lower variance than the annual series. However, research by
Hershfield (1961) indicated that estimates for return periods greater than ten years tend to be
similar for both data series (Pathak, 2001).

Several probability distributions are widely used for rainfall frequency analysis, including the
Gumbel, Normal, Log-Normal, Gamma, and Weibull distributions (Pathak, 2001). Investigating
extreme precipitation depth is a critical step in understanding such occurrences (Hansonl and
Vogel, 2008). According to extreme value theory, there are three types of extreme value
distributions (EV1, EV2, and EV3), with EV1 (Gumbel) being the most frequently used for
assessing hydrological and meteorological extremes (Eliasson, 1994). These distributions help
quantify risks associated with extreme rainfall events, as shown in studies by Coles and Pericchi
(2003).

Distributions often applied in hydrology and meteorology, such as exponential, gamma, Weibull,
normal, and log-normal, generally belong to the exponential type. These distributions are typically
aligned with the EV1, or Gumbel, distribution, as noted by Melice and Reason (2007). The EV1
distribution is used when the initial variable distribution is of exponential type and has no upper

10

Page 25 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 26 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

limit for extreme values. This model has become standard for assessing risk related to extreme
rainfall (Koutsoyiannis, 2004).

2.5.1 Data fitting

Distribution fitting can be done using methods such as the method of moments or the method of
maximum likelihood, with the former often favored in hydrological analysis for its stability (Chow
et al., 1988). Goodness-of-fit tests, such as the Kolmogorov-Smirnov and Chi-square tests, are
used to assess how well a distribution fits observed data (Pathak, 2001).

2.5.2 Testing the goodness of fit

Using criteria to evaluate the goodness of fit is crucial when fitting several probability distributions
to a dataset. The Kolmogorov-Smirnov test and the Chi-Square test are two popular statistical tests
for this purpose. According to Pathak (2001), these tests assess the hypothesis that a specific data
series adheres to a specific probability distribution. For instance, by looking at the cumulative
frequency function, the Chi-Square and Kolmogorov-Smirnov tests can compare the empirical
distribution of sample data with a theoretical EV1 distribution (Stephens, 1977).

When assessing whether a sample originates from a population that adheres to a given distribution,
the Kolmogorov-Smirnov test is especially helpful (Pathak, 2001). By contrasting the empirical
cumulative distribution function (CDF) of the sample with the CDF of the theoretical distribution,
it assesses the quality of fit. At the value of each ordered data point in this test, the step function
rises by 1/N. The Kolmogorov-Smirnov test has the benefit of having a test statistic distribution
independent of the distribution under test. Furthermore, it is an exact test as opposed to the Chi-
Square test, which necessitates a substantial sample size to guarantee the validity of the
approximation (Stephens, 1977).

By contrasting the sample's empirical distribution with the theoretical one, the Kolmogorov-
Smirnov test hypothesis—specifically, for the EV1 distribution—is established. By grouping data
into several groups, the Chi-Square test, on the other hand, determines if a sample fits a particular
population distribution. However, the way the data is categorized may have an impact on the test's
results. The Chi-Square test has a drawback in that it is usually better suited for discrete
distributions and necessitates a sizable sample size for its approximation to hold (Stephens, 1977).

11

Page 26 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 27 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

2.5.3 Derivations for Return Periods

Efficiency, economy, and safety are crucial considerations while designing engineering buildings.
Therefore, over the course of their anticipated lifespan—which might be anywhere from a few
decades to several centuries—structures must be built to withstand the forces they may encounter.
For water control systems, which need to be able to withstand periods of heavy rainfall, this is
particularly crucial. Estimating the frequency of rainfall of a specific duration and intensity is
required in order to evaluate the possible costs and advantages of installing suitable measures.
These approximations are referred to as return periods or recurrence intervals. Return period values
of 10, 25, 50, 100, 500, and 1000 years for durations of 1, 2, and 3 days are widely utilized globally
for hydraulic structure design and planning, as well as for appropriate water resource management
and flood control research.

The average time interval (measured in years) between occurrences of a rainfall event of a specific
or higher magnitude is referred to as the return period. A recurrence interval is indicated by the
return period. It is a statistical indicator of the likelihood that a rainfall event of a specific size will
occur. (Melice and Reason, 2007) The return period is stated as follows:

where F is the value cumulative distribution function and T is the return time.

Fitting a known probability distribution to the yearly maximum rainfall series and estimating its
parameters statistically is a popular method for determining the likelihood or return period of
extreme rainfall occurrences for hydrological design. Theoretically, the PMP is an event that will
never be surpassed, but empirical data indicates that this is not the case. As a result, a number of
studies have tried to measure the risk involved with PMP estimates. The return period of PMP in
the US was predicted by the NRC (1994) to be between 10^5 and 10^9 years. The return period
(T) for PMP can be represented by the frequency factor K for a particular probability distribution.
For instance, using the Gumbel model, frequency factors of 10, 20, and 40 correspond to return
periods of approximately 6.710^5 and 2.510^22 years, respectively. The Hershfield largest
frequency factor (Km=15) has return periods between 10^5 and 10^11, which are significantly
higher than those estimated by Hershfield (1981). However, return periods on the order of 10^3,
which are far from Hershfield's estimates, have been extrapolated for PMP by fitting the Gumbel

12

Page 27 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 28 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

model to annual extreme rainfall series of 1-3 day durations at stations located within the Indus
Basin in India (Deshpande et al., 2008). For the generalized extreme value (GEV) distribution, the
Hershfield largest frequency factor corresponds to a return period of approximately 6000 years,
which falls at the lower end of the NRC range (Koutsoyiannis, 1999). Recent studies in the Blue
Nile River Basin (Alemayew and Semu, 2010) have also derived return periods for PMP that are
closer to the NRC range.

2.6 Performance of CHIRPS

Table 2-1. Performance statistics description

Statistic Description Performance Indication

CC (Pearson Correlation Measures the linear Higher values (closer to 1) indicate

Coefficient) correlation between CHIRPS better agreement in capturing trends.
estimates and rain gauge
observations. Ranges from -1
to 1.

PBIAS (Percentage Expresses the average Positive values indicate

Relative Bias) difference between CHIRPS overestimation, negative values
estimates and rain gauge indicate underestimation. Ideally
observations as a percentage close to 0%.
of the observed mean.

ME (Mean Error) Average difference between Positive values indicate

CHIRPS estimates and rain overestimation, negative values
gauge observations. indicate underestimation. Units of
precipitation (mm).

MAE (Mean Absolute Average of the absolute Lower values indicate better
Error) differences between CHIRPS correspondence. Units of
estimates and rain gauge precipitation (mm).
observations.

13

Page 28 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 29 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

RMSE (Root Mean Square root of the squared Lower values indicate better
Square Error) differences between CHIRPS correspondence. Units of
estimates and rain gauge precipitation (mm).
observations, averaged over
all observations.

3 Materials and Methods

3.1 Description of study area

Somali Regional state is one of the constituents of the Regional states in which Ethiopia is
comprised of that has been established in 1994 etc. to create the current federal system of the
government. It is located in the eastern part of the country, bordering with Kenya to the south-
west, Djibouti in the north, Somalia to the north-east and south, to the west Oromia, Afar and Dire
Dawa. The capital city is Jijiga. Also, Upper Fafan sub-basin where this study focus is located
most of its part is in Somali Region. The sub-basin characterized by its long flow length as
compared to other sub-basins found in Wabi-Shebele river basin (Figure 3-1).

14

Page 29 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 30 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Figure 3-1: Map of Study sub basin

The Fafan sub-basin has 12 districts Namely Jigjia, Harawa, Aw-bare, Kebribeyah, Tuli, Babili,
Goljano, Harorey, Harshin, Ararso, and it is considered the most densely populated sub-basin of
the whole Somali region as more than 2 million people resides. it is characterized by bi-modal
rainfall and pure agro-pastoralist mode of life.

3.1.1 Climate

The primary summer rains (June to September) in the Fafan sub-basin are preceded by a minor
spring rainfall surge. During the growing season, the mean temperature in these places is the
greatest. There are two temperature ranges that define it. Areas with mean maximum temperatures
between 270 and 290 degrees Celsius during the growing season and mean lowest temperatures
between 130 and 16.50 degrees Celsius are the first. It is generally accepted that the Fafan sub-
basin is semi-arid and devoid of any permanent rivers (MoA, 2004). The sub-basin experiences
variable mean annual rainfall, with a maximum of 600 mm (NSMA, 2006).

15

Page 30 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 31 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Figure 3-2. Mean monthly Precipitation, mean maximum and minimum temperature of Fafan
Sub-basin

3.1.2 Topography

Rainfall distribution is affected by the topography of the land surface. Topography defines the
elevation difference of the watershed, basin and other related characteristics. The highest elevation
has the maximum rainfall while the lower elevation has minimum rainfall. Based on this fact,
Upper Fafan sub-basin has maximum elevation of 3016 m and 356 m lower elevation (Figure .

16

Page 31 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 32 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Figure 3-3. Upper Fafan sub-basin Elevation map

3.1.3 Soil Type

The inclusion of soil parameter as a land quality is very important both for the plants which are
the source of animal feed. The word soil type is used to indicate a land quality under the context
of this study which may not fully have the definition of land quality per say. However, it has
embodied both the mixture of physical and chemical properties that have an important value for
livestock production. It is also one of the most important parameters used by the pastoral

17

Page 32 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 33 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

community to evaluate the suitability of the rangelands for different livestock as a source of feed
and livestock disease prevalence.

Based on the land resources of Ethiopia Map Agency (1984) and on the nature of slope,
existing parent materials and the prevailing climatic condition, the major soils of the study area
have been identified. The soil types in the plains are eutric cambisols, chromic vertisols,
calcic cambisols, and luvisols. Whereas the dominant soils in the hills are chromic luvisols.
Fluvisols which are soils developing from recent alluvial deposits or materials washed down
the hills and overlying other soils are also available.

Locally the communities classify the various soils by their colour; these are Vertisols as “aro
medaw’’ or black soil, Cambisols as “labeeb’’ or mixed soil, Luvisols as “ara cassan” or red soil,
Fluvisols as “hashewa” and Lithsols as “white soil” (Figure 3-4).

Figure 3-4. FAO Soil Map of Upper Fafan Sub-Basin

18

Page 33 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 34 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

3.1.4 Land use/Land cover

This study describes the land use/land cover of the Upper Fafan Sub-basin. Regarding land use,
agricultural activities dominate the upstream parts. Both annual and perennial crop production are
prevalent. Downstream, the area is covered with open and closed shrubland. Dense, moderate, and
wooded forests are found along dry stream rivers. Conversely, the upstream area is dominated by
close and open grasslands, with very minimal grasslands downstream. Water bodies and bare soil
are also present, though less common.

On the other hand, agricultural practices are common in the Upper Fafan sub-basin. The
livelihoods of the people there are characterized as purely agro-pastoral. Sorghum, maize, barley,
wheat, peanuts, and beans are the most important agricultural crops in the sub-basin. These crops
serve as staple foods for the rural communities. Downstream, however, pastoral activities are more
predominant.

The type of vegetation and land cover that would exist in a particular area is usually
influenced by the prevailing climatic and soil conditions of the area. Open shrub land
dominates the land covers of Upper Fafan Sub-Basin. In spite of natural resource
degradation, the area is vegetated with open shrub, bushes and grass species of very low
productivity. Because of continues over-use sub-basind lack of soil moisture to support growth of
vegetation, large areas of land in the area have turned into bare lands. The degradation of the
whole ecosystem in the sub basin has a fundamental influence on the living standard and
serious limitation for future development.

The study sites have bushy vegetation, predominantly consisting of Acacia (Figure 3.5C), Cactus
(Figure 3.5A), and Lantana camara (Figure 3.5B), except for rocky areas and farmland.
Ecologically, the medium-altitude plains of the Jijiga and Kebri Beyah areas are classified as
upper semi-arid. The major vegetation types include medium-height grasslands and open
shrubland dominated by Acacia etbaica, which forms a multi-strata formation. This vegetation
type covers a considerable portion of the plains at an altitude between 1,600 and 1,800 meters
above sea level (masl).

19

Page 34 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 35 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

The area generally appears overgrazed and degraded in parts of the vast plains, where most of
the perennial vegetation has been replaced by annual grasses, forbs, as well as xerophytic and
spiny bushes.

The available tree and shrub species include acacias, balanites, commiphora, Euphorbia ("Eyin"),
and grewia. The medium-height grasslands are dominated by Chrysopogon and Andropogon
(though degraded). Cenchrus, Aristida, and Cynodon are some of the other important grass species.
Le Houerou and Haywood (1979) reported Sonchus, Lactuca, and Crotalaria species as some of
the dominant forbs in the sites. The main grass species observed and found in the meadows of the
Fafan Valley are Cynodon aethiopicus and Cynodon plectostachus.

Figure 3-5. Land use land cover (2018) of Upper Fafan Sub-basin

20

Page 35 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 36 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Figure 3-6. Image by Mohamed Farah Ahmed for Cactus (A) , Lantana camara (B) and Acacia
(C) found in Upper Fafan Sub-basin and surrounding

3.2 Data and sources

A total of eight rain gauge ground-based stations and satellite-based CHIRPS precipitations were
used for this study in Fafan Sub-basin. The ground based and satellite precipitations were classified
based on the longitude and latitude geographical station locations (Table 3.1). The details of the
rain gauge stations are included in Tables 3.1 whereas this distribution of rain gauge stations and
their location in the sub-basin are indicated in Figure 3.1. In the Figure 3.2 and Table 3.1the missed
and recorded data and years provided with the area percentage coverage.
Table 3-1: CHIRPS satellite rainfall stations available in and around Fafan Sub-basin with area
coverage and recording period of 1990-2020

Long Area
ID Station Lat (N) Ground-Based (Recorded Days)
(E) (%)
1 Jigjiga 9.33 42.78 11200 5.71
2 Hadew 9.36 42.66 10552 2.47
3 Hartishek 9.15 43.35 10305 2.28
4 Gursum 9.35 42.38 11086 1.33
5 Chinaksen 9.5 42.6 10905 0.99
6 Degahabour 8.24 43.55 10978 23.89
7 Gode 5.91 43.58 10196 6.09
8 Kebri Dehar 6.73 44.3 10896 57.24

21

Page 36 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 37 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

3.3 Data collection

Precipitation and geography data sets were among the data sets required for the investigation. The
National Meteorological Service Agency in Addis Ababa, Ethiopia, provided the precipitation data
sets, which included daily annual extreme rainfalls and annual totals along with the latitude and
longitude of each point gauging station's geographic location. The data sets were downloaded from
the CHIRPS satellite precipitation websites (http://chg. geog.ucsb.edu/data/chirps/). Stations with
daily rainfall data spanning more than 21 years provided the data. Point severe rainfalls for longer
return periods and PMP predictions have been derived using this volume of rainfall data.

3.4 Data quality and analysis

The first step in analyzing rainfall data was to carefully review any records that were missing
throughout the research period. In hydrological data, missing values can seriously impair modeling
and analysis. As a result, estimating these missing values is frequently required before starting data
analysis. Years that had too few daily records to choose the annual maximum were marked as
missing in this study. Visual examination and Excel filtering options were used to gather missing
data for the study period. No missing data was discovered during the study period, despite a
thorough review. The yearly maximum rainfall figures were taken from the annual records after
this preliminary data quality assessment. The yearly maximum for each year was determined by
taking the highest daily rainfall value in that year.

22

Page 37 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 38 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Figure 3-7: Thiessen Polygon map of Fafan Sub-basin

3.4.1 Test of consistency, Homogeneity and trend of ground-based precipitation

The consistency and homogeneity of the rainfall record data must be verified before use (Figure
3-9 and Table 3.2). The double mass-curve approach was used to assess the consistency of the data
set of the specified stations. Each station's yearly cumulative total rainfall was plotted against the
average annual cumulative of all stations to create the double mass curve (Figure). Excel was used
to create a scatter plot of the data, and its consistency was examined for additional analysis. The
homogeneity test of precipitation within the study range was the other data test. The uniformity of
the precipitation at the research site was examined using the Pettit test in this investigation (Figure).
Based on the test results, the study data has been approved because the p-value is greater than

23

Page 38 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 39 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

0.005, indicating that the data is significant. A break in homogeneity was indicated by a significant
shift in the slope or trend of the resulting line (Shahn, 2002).

Figure 3-8 shows the Fafan subbasin's long-term precipitation trend. As can be seen from the
graph, Kebri Dahar displayed a maximum precipitation in 1999, while Chinaksen showed a
maximum in 1994. During the research period, Hartishek experienced very little precipitation.

Figure 3-8. Long-Term maximum precipitation of Ground Based of Fafan Sub-basin

The Double Mass Curve (DMC) is a powerful tool in hydrological analysis, especially for
evaluating the temporal trends of hydro-meteorological data over long periods. Its strength lies in
its ability to effectively track changes in precipitation, across sub-basins, and quantify the evolving
trends. By analyzing slope changes in the DMC, researchers can identify significant increases or
decreases in the variables of interest, and even pinpoint the timing of these changes. Furthermore,
by combining DMC analysis with local environmental data, it's possible to quantitatively
determine the impact of land management and climate on these hydrological trends. Despite its
widespread application, however, a lack of standardized derivation and breakpoint identification
methods has limited the DMC's full potential. This calls for a robust and repeatable approach that
considers hydrological characteristics, for instance, precipitation data, ultimately leading to more
accurate and reliable insights into the complex interplay between environmental factors and
hydrological processes.

24

Page 39 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Page 40 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

25

Page 40 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 41 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Figure 3-9: Homogeneity test of Fafan Sub-basin Precipitation for GB of selected eight
metrological stations using Pettit test
Table 3-2. Statistical Estimated value of Fafan Sub-basin Precipitation for GB of selected
metrological stations

Station Minimum Maximum µ σ CV (%)

Chinaksen 13.15 85.05 29.22 15.03 51.44
Gode 7.69 88.9 28.20 17.06 60.51
Kebri Dehar 13.97 154.84 43.88 37.26 84.90
Gursum 14.9 65.37 24.87 10.69 42.98
Hadew 14.9 65.37 24.87 10.69 42.98
Hartisheik 10.83 62.27 22.35 10.83 48.47
Degehabur 11.39 70.67 23.46 10.95 46.66
Jigjiga 13.38 47.16 23.36 8.21 35.13

26

Page 41 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 42 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Table 3-3. Homogeneity test methods used in Fafan Sub-basin Precipitation for GB of selected
eight metrological stations

Station Pettitt SNHT test Buishand von Neumann

Chinaksen 0.90 0.54 0.31 0.78
Gode 0.85 0.04 0.14 0.05
Kebri Dehar 0.51 0.14 0.07 0.25
Gursum 0.40 0.99 1.00 0.97
Hadew 0.40 0.99 1.00 0.97
Hartisheik 0.10 0.98 0.94 0.97
Degehabur 0.25 0.39 0.28 0.47
Jigjiga 0.81 0.83 0.59 0.59

3.4.2 Test of consistency, homogeneity and trend of CHIRPS

CHIRPS satellite rainfall data offers a valuable tool for analyzing rainfall trends in semi-arid
regions, especially where ground-based rain gauge data is limited (Bouizrou et al., 2023). In addition
studies have shown that CHIRPS data can be accurate, particularly in capturing monthly rainfall
patterns (Dinku et al., 2018).
Provides rainfall estimates in areas with sparse rain gauge networks offers good agreement with
observed rainfall data in many semi-arid regions (Boluwade, 2020).
Overall, CHIRPS data is a viable resource for analyzing rainfall trends (Figure 3-10) in semi-arid
zones. However, it's important to acknowledge potential limitations and consider validation with
ground-based data when possible.
Consistency tests (Figure 3-11) are a valuable tool for assessing the reliability of satellite-derived
rainfall data like CHIRPS. Two common approaches to consistency testing CHIRPS rainfall exist:
Comparison with Ground Observations (Dinku et al., 2018) , and the other is Internal Consistency
Checks (Dinku et al., 2018). However, this study employed Internal Consistency Checks or tests.
The test used, the Double Mass Curve method (Figure 3.11), can be used to identify inconsistencies
within the CHIRPS data itself. This method relies on the assumption that total rainfall over a region
should increase as the area being considered increases. Deviations from this trend might indicate
errors in the data for specific locations or periods. Consistency tests can reveal biases in CHIRPS

27

Page 42 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 43 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

data (Funk et al., 2015), like underestimation of low rainfall events or overestimation of dry days.

Figure 3-10: Long-Term maximum precipitation of CHIRPS of Fafan Sub-basin

21200 Chinaksen Degehabur

Commulative rainfall of each station

Gode Hadew
18200 Gursum Hartisheik
Jigjiga Kebri Dehar
15200

12200
(mm)

9200

6200

3200

200
400 50400 100400 150400 200400
Commulative mean annual rainfall of all stations (mm)
Figure 3-11: Consistency test of precipitation of CHIRPS of Fafan Sub-basin of eight stations

Tell us what is mean by the Table, how this is performed and for what purpose.

Table 3-4. Statistical Estimated value of Fafan Sub-basin Precipitation for CHIRPS of selected
metrological stations

28

Page 43 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 44 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Station Minimum Maximum µ σ CV (%)

Chinaksen 22.26 73.86 36.17 11.59 32.03
Degahabour 21.27 74.74 37.75 13.68 36.23
Gode 17.16 63.36 38.58 11.50 29.81
Gursum 19.67 62.61 36.97 11.63 31.46
Hadew 16.16 116.66 35.95 16.68 46.40
Hartishek 18.24 60.32 30.52 8.67 28.41
Jijiga 19.11 84.87 34.29 12.84 37.44
Kebri Dehar 33.34 115.77 56.64 16.20 28.59

Table 3-5. Homogeneity test methods used in Fafan Sub-basin Precipitation for CHIRPS of
selected eight metrological stations
Station Pettitt SNHT Buishand von Neumann
Chinaksen 0.333 0.778 0.664 0.574
Degahabour 0.179 0.277 0.088 0.287
Gode 0.141 0.562 0.918 0.968
Gursum 0.189 0.291 0.199 0.331
Hadew 0.222 0.340 0.126 0.145
Hartishek 0.343 0.975 0.941 0.600
Jijiga 0.344 0.498 0.371 0.757
Kebri Dehar 0.405 0.415 0.274 0.398

In Table 3.5, the homogeneity tests are crucial for assessing the usability of rainfall data from
satellite products like CHIRPS. These tests ensure data consistency over time, meaning there are
no artificial shifts or trends caused by factors unrelated to actual rainfall patterns.

Reliable trend analysis of rainfall data with inhomogeneity can lead to misleading interpretations
of trends over time. Homogeneity tests identify any inconsistencies, allowing for data corrections
or separation into homogeneous periods for accurate trend analysis.

For improved comparisons of CHIRPS data with ground observations or other

datasets, homogeneity ensures both datasets are comparable and represent actual rainfall patterns.

29

Page 44 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 45 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

There are several statistical tests used to assess homogeneity in CHIRPS rainfall data. Some
common methods include the Standard Normal Homogeneity Test that compares the means and
variances of rainfall data across different periods to detect abrupt shifts (Figure 3-12 A & B). The
other method is rank-based tests. These are non-parametric tests that analyze the order or ranking
of rainfall values instead of raw magnitudes, making them less sensitive to outliers. The one used
in this study is the Pettitt test. This test identifies abrupt changes in the median of rainfall time
series.

30

Page 45 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 46 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Figure 3-12. Homogeneity test of Fafan Sub-basin Precipitation for CHIRPS of selected stations
using Pettit test

3.5 Bias Correction of CHIRPS

The study focused on bias correction of satellite-based rainfall estimates (SPPs) CHIRPS for PMP
evaluations. Rainfall data has more influence on PMP and other hydrological estimations. To
improve the efficacy of bias correction methods, researchers performed bias correction on raw
SPPs after removing the "light rain effect" and grouping daily records by day, month or annual to
remove errors. There are two bias correction methods has been used including Quantile Mapping
Bias Correction (QMBC) and Principal Components Bias Correction (PCBC. CHIRPS dataset was
used as the reference for bias correction due to its accuracy for Africa is the Principal Components
(PC) analysis transforms data into a new coordinate system. It is commonly used to reduce
dimensionality by retaining significant components that explain most variance. In this study, PC
analysis was used to develop PC bias correction. Similar to QMBC, PCBC assumes that statistical

31

Page 46 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 47 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

properties between reference data and SPPs can be transferred to correct raw estimates. The
process involves reshaping CHIRPS data into two dimensional matrices and standardizing them
along the temporal dimension. Then a covariance matrix is calculated. To compute corrected
values, Singular Value Decomposition (SVD) is used to decompose the original matrix of
CHIRPS. The bias-corrected estimates are calculated by combining the eigenvectors from raw
SPPs with the singular values calculated for CHIRPS. Unlike traditional PC analysis, PCBC retains
total variance to correct raw estimates. However, dimensionality reduction can be applied to
potentially improve the performance of this method and the new data for PMP estimations and
evaluations.

∗ ∗ ∗ ∗ ∗ ∗

where ’( × ) is the matrix containing the bias-corrected daily satellite-based rainfall estimates.
is the matrix containing the PC of calculated for the SPPk . is the matrix containing the
Empirical Orthogonal Function of calculated for the SPPk. is the diagonal matrix containing the
singular values of calculated for CHIRPS. In definations and equations n, m and k are space by
space m and time respectively.

3.6 PMP Estimation Procedure

An array of annual maximum values was created for each station by choosing the annual maximum
series of one-day maximum precipitation. The month with the highest daily rainfall was
determined for each station by selecting the daily maximum for every month. For analysis and
interpretation, the annual total rainfall as well as the annual daily maximum rainfall were computed
and arranged in Excel. The Hershfield approach, a modified version of the Chow method for
frequency analysis of rainfall data, was used to estimate the PMP. To derive station-based PMP
estimates, one-day yearly maximum rainfall measurements from every station were examined. The
coefficient of variability was found by calculating the mean and standard deviation for the yearly
maximum rainfall dataset, eliminating the greatest value. A frequency table was made once each
station's frequency factor (Km) was determined using the relevant formula. The most likely Km
value was chosen. PMP values for a one-day period were calculated for each station using this Km
value, the mean and standard deviation for each station, and the relevant equation. To determine
the largest rainfall event and the ratio of estimated PMP to observed rainfall values for comparison,

32

Page 47 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 48 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

the estimated PMP values were compiled. According to Hershfield (1962), the point PMP
magnitude at a single station usually doesn't surpass three times the heaviest rainfall that has been
recorded for an extended period of time.

3.6.1 Hershfield Approach for PMP Analysis

The Hershfield methodology is one of the most popular techniques for statistically evaluating PMP
estimations (WMO, 1986; Koutsoyiannis, 1999). It requires a range of maximum annual daily
rainfall values for a particular site and only generates point values of PMP. Therefore, area
reduction curves are required to adjust the point values to various magnitudes of area larger than
1000 km2 (WMO, 1973; WMO, 1986). In India, the method is widely employed. Numerous
studies have carried out in-depth evaluations on the estimation of point PMP for a single day for
stations across the country, such as those by Dhar and Kamte (1969, 1971) and Rakhecha et al.
(1992).
According to Hershfiled (1969), Ven T. Chow showed that the comb=n statistical variable, K
(frequency factor), or standard variance in the following equation, is the only distinction between
the various distributions that lend themselves to the study of PMP using the following equation.

In years P_x ̅ and P_σ are the mean and standard deviation for a sequence of N yearly maxima,
respectively, and PT is the rainfall for the return period T.

According to Hershfied (1961, 1965), there is a frequency factor number, say Km, that should
not be exceeded for PMP estimate. He then changed Chow's formula as follows:

Xpmp = Xn + SnKm

In this case, Km is the maximum frequency factor, Xn is the mean of the annual extreme series,
Sn is the standard deviation of the annual extreme series, and Xpmp is the PMP estimation for a
station.

One way to calculate the sample mean (X) and standard deviation (S) is to:

X= ∑ Xi

33

Page 48 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 49 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Xi
S= 1

where S is the sample standard deviation, X is the random variable's ith value, and (x) ̅ is the
random variable's mean.

The following formula can be used to determine the maximum frequency factor (Km) (Hershfiled,
1961, 1965):

Where σ_(n-1) is the standard deviation of the annual maximum, excluding the highest value, Xn-
1 is the mean of the annual maximum, excluding the highest value, and X1 is the highest observed
annual maximum rainfall in the series.

3.6.2 Advantage of Using Hershfield Method

In recent years, statistical methods have gradually replaced storm models as the primary means of
estimating PMP (Collier and Haradaker, 1996). In the USA (Myers, 1967) and most recently in
the Netherlands (Koutsoyiannis, 1999), it has been shown that the PMP estimates produced by this
method are more in line with natural behavior, offer stronger foundations for estimation, and are
nearly identical to those produced by the traditional moisture maximization and storm
transposition techniques. On the other hand, design engineers can get the average maximum
rainfall estimates for a region using the statistical method, which is much simpler to apply and
takes a lot less time. In their many manuals and technical publications, the World Meteorological
Organization (WMO) has also recommended this method for point PMP estimation for river basins
whose daily rainfall data are available for a long time (WMO, 1969, 1970, 1986).

3.6.3 The ratio of PMP to the maximum rainfall observation rate

Hershfield's research indicate that over an extended length of time, the PMP at a particular area
typically does not surpass three times the maximum recorded rainfall. According to research by
Durbude and Dhar et al., the highest one-day rainfall levels ever reported in some parts of India
were either slightly higher or comparable to the estimated PMP. In Malaysia, Desa and Rakhecha

34

Page 49 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 50 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

found the PMP to be twice as high as the maximum recorded rainfall, Alemayew and Semu,
however, found that it was 1.9 times greater than the 1-day rainfall that takes place in the Ethiopian
Blue Nile basin, which is 180–420 mm every ten millinium. A pertinent metric for evaluating the
factor of safety (FOS), which is frequently employed in engineering design, is the ratio of PMP to
the greatest rainfall that has been observed or to the design rainfall for particular return periods. In
other words, the reasonableness of the projected PMP values for design purposes can be assessed
by comparing them to the corresponding factor of safety (Al-mamu and Hashim, 2004).

3.7 Fitting Data to the Probability Distribution Functions

The yearly maximum rainfall data was analyzed by fitting theoretical probability distributions to
the observed data using frequency analysis techniques. The data for the normal, Gumbel EVI, Log
Pearson III, and Log normal distribution models were ranked using Weibull's plotting position.
The extreme value (XT) and standard normal (Z) values for the normal distribution were taken
from Table 3.6, and the plotting probability was determined using the Weibull method. The
matching Z value was interpolated if Table 3.6 did not include the desired exceedance probability.

Table 3-6: Values of Standard Normal Deviate for the Cumulative Normal Distribution
Exceedance probability (%) Return period Z
50 2 0.0000
20 5 0.8416
10 10 1.2816
4 25 1.7507
2 50 2.0538
1 100 2.3264
0.2 500 2.8782

The probability distribution function was shown against the yearly maximum rainfall data for the
Log Normal distribution. The highest value was given the rank m after the values were arranged
in descending order. Using the relevant formulae, Z and W values were determined. A similar
process to that used for the normal and log-normal distributions was used to fit the data to the Log
Pearson Type III distribution. The precise procedures described by Raghunath (2006) were
adhered to.

35

Page 50 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 51 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

1. Every flow in the series underwent a logarithmic modification (Yi = log Xi).
2. Equations 2.21 and 2.22 were used to calculate the logarithms' mean (¯Y), standard deviation
(Sy), and standardized skew (Sk).
3. Weibull estimated the probability plotting places for every event formula
4. Equations 2.31 and 2.32 were used to calculate KT and K.

Plotting the ranked yearly maximum rainfall values against the exceedance probability allowed for
the fitting of the Gumbel Extreme Value Type I distribution to the data. To determine the extreme
value, the following procedures were followed.
1. The reduced variate was calculated by equation 2.34
2. The value of the return period was obtained by taking the inverse of the probability
plotting position which was obtained -by using Weibull formula table 1.
3. Frequency factor KT was derived from the equation 2.33, where Y and S
were obtained from the reduced variate table
4. Finally XT = X + KTS

Return period KT
2 -0.1552
5 0.8376
10 1.49474
25 2.32526
50 2.94146
100 3.55304
500 4.9661

3.8 Testing the Goodness of Fit of Data to Probability Distribution

By contrasting the theoretical and sample values of the relative frequency or cumulative frequency
function, one can determine whether a probability distribution fits well. To determine whether the
chosen probability distribution fit the theoretical frequency distribution, the chi-square goodness
of fit test was used to the annual maximum rainfall value for each station (Chandra, 1981; Law
and Kelton, 1991; Bhakar et al., 2006). Finally, a probability distribution statistical test, which is
tabulated in many statistics texts, was used to compare the resulting χ2†values (Haan, 1977). The

36

Page 51 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 52 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

degree of freedom (v) in the test is determined by ν = m-p-1, where m is the number of interval
classes in the data set and p is the number of parameters (in this example, mean and standard
deviation) used to fit the suggested distribution. For the test, a confidence level was selected, which
is commonly represented as 1-α, where α is known as the significance level. 95 percent is a
standard confidence level value. The test's null hypothesis is that the suggested probability
distribution provides a sufficient fit to the data. If the value of χc2 exceeds the limiting value,
which can be found using the χ2 distribution with v degrees of freedom, then this hypothesis must
be rejected.

3.9 Intensity Duration Frequency (IDF)

IDF techniques use this data to calculate the frequency of rainfall occurrences of different durations
and intensities.

Statistical techniques are used to determine the likelihood of surpassing a particular rainfall amount
within a specified timeframe, like a 100-year flood event, in order to estimate IDF connections.
After these procedures, the outcomes are shown as IDF curves that illustrate the correlation
between the frequency, duration, and intensity of rainfall.

These IDF curves are used by hydrologists and engineers for a number of purposes, such as
planning infrastructure projects, flood control strategies, and drainage systems. As a result,
CHIRPS and ground-based precipitation data can both be useful resources for IDF techniques to
comprehend rainfall trends.

The empirical reduction method created by the Indian Meteorological Department (IMD) was used
to calculate the relationship between the depth of rainfall and the development of IDF over a
variety of time periods shorter than 24 hours. The IMD equation yields the best estimations of IDF,
as shown below, according Rizwan and Ali (2017).

/
24
where P24, the length of time t in an hour, is the depth of the maximum daily rainfall, and Pt is the
maximum hourly rainfall depth.
The development of IDF curves began by fitting processes derived from sample data with
equations from selected probability distribution functions, such as the Gumbel distribution. While
37

Page 52 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 53 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

this study utilizes the Gumbel equation provided below, several studies worldwide have also
recommended its use for developing IDF curves (Ahmed et al., 2012; Raiford et al., 2007). In
particular, the Gumbel distribution function is recommended for annual maximum rainfall data
spanning more than 20 years, as has been suggested in many past studies (Chikabvumbwa et al., 2019;
Koutsoyiannis, 2007). Rainfall depths for return periods of 2, 5, 10, 25, 50, and 100 years are
computed using this method. The following formula can be used to determine the rainfall depth
for any duration with a given return period.
F(x) = exp (-e^-( )

In this case, α and μ are parameters of the extreme value type I (Gumbel) distribution that may be
inferred using the method of moment from the observed yearly extreme rainfall series.

, = + (-ln.ln )

Where, ln natural logarithm and

3.10 Estimation of PMP, PDF and CDF

The statistical or Hershfield’s method has been widely applied to estimate the PMP from the
extreme precipitation using the GBP data at meteorological stations and from high-resolution
SPP. The statistical method is useful for estimating PMP because it is easy and quick to apply
once a statistical model is constructed (Afzali-Gorouh et al. 2018; Hershfield 1961; Sarkar and
Maity 2020).

3.10.1 Hershfield method

The Hershfield (1965) method, which was modified from the Chow (1952) version for the
frequency analysis of precipitation, was used to estimate PMP. This method was used to calculate
the one-day annual maximum rainfall values of all stations available in the particular area of
watersheds or large river basins. According to Chow (1951) suggestions, PMP for each station in
the Upper Fafan Sub-basin was estimated using the following equations both for ground based
precipitation (GBP) and satellite precipitation product (SPP) for CHIRPS.

38

Page 53 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 54 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Statistical computations of precipitation for each of the stations have a significant value to estimate
the PMP. Included in the statistics are estimates of the annual maximum precipitation mean (X ̅n),
the annual maximum precipitation mean excluding the highest value (X ̅_(n-1)), the annual
maximum precipitation standard deviation (σ_n), and the annual maximum precipitation standard
deviation excluding the highest value (σ_(n-1)). and the following formula was used to determine
the Hershfield frequency factor (Km) from these statistical values.

̅ ̅
_

The Hershfield approach, which may provide accurate PMP values for stations located in the
Wabi-Shebele river basin, was used to estimate the proper frequency factor (Km). However, in
this case, the statistical analysis was based on the annual maximum precipitation amounts series
of the stations located within the sub-basin.

The mean and standard deviation of the annual maximum precipitation levels for both SPP and
GBP data were modified. Extreme precipitation events that may occur at some point throughout a
much shorter period of record allow for this. The mean and standard deviation of the annual time
series may be visibly impacted by such a rare occurrence, known as an outlier. Similarly, the
standard deviation adjustment graph was created using the annual series calculated after the largest
item in the time series was removed, σ_(n-m) (y-axis) and σ_(n-m)/σ_n (x-axis) (Hershfield,
1961b). The frequency factor and the Km graph for each station vs the duration of records were
developed using the mean and standard deviation corrections.

Because of the skewed frequency distribution of rainfall extremes to the right, which increases the
likelihood of a large extreme throughout the length of records, the mean and standard deviation of
the yearly time series tend to rise with record length.To make up for the mistakes, the mean and
standard deviation are adjusted for the length of the records. For example, there aren't many
precipitation records that go back more than 50 years, which makes it difficult to assess the impact
of sample size. However, the few lengthier records that are available show only minor adjustments
compared to the 50-year records.

39

Page 54 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 55 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

3.10.2 Frequency analysis and Best-fit

Frequency analysis can be used to characterize the likelihood and behavior of hydrological
phenomena, including floods, droughts, and precipitation. This type of analysis is crucial to the
economic and unique assessment of a project and is utilized for future water resource development
and infrastructure construction, such as dams (Kong et al., 2020; Farooq et al., 2018). The most
often used statistical frequency analysis techniques are the probability distribution functions, such
as the normal distribution (ND), Log-Pearson type III (LP-III), Gumbel Distribution of Extreme
Value Type I (EV I), and Weibull. The PMP of the Wabi-Shebele river basin was estimated in this
study by analyzing and assessing the GBP and SPP data using all of the probability distribution
functions mentioned above.

Normal distribution: The Gaussian or N distribution is commonly employed in rainfall

distribution assessment (Markovic, 1965). The variance (σ) and mean (µ) are the ND variables.
The following equations give the probability density function (PDF) and cumulative distribution
function (CDF) for a normal random variable x.

√2

1
√2

In the normal distribution, the value of PMP ( ) for return period (T) expressed by the following
equation:
̅ 1

Where ̅ the average, is the coefficient of variation, is the frequency factor and it can be
calculated using the following equation:

Log-Pearson type III (LP-III): A statistical method for fitting frequency distribution data to
forecast the PMP for the stations in the river basin is the log-Pearson type III (LP-III) distribution.
A frequency distribution for both SPP and GBP data can be created after the statistical data for
each station has been determined. The benefit of this specific method is that values for occurrences

40

Page 55 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 56 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

with return periods significantly longer than the observed precipitation events can be extrapolated.
Equations listed below were used to determine the LP-III distribution's probability density function
(PDF) and cumulative distribution function (CDF).

1 ln ln
exp
| |Γ

ln

Γ
Where, α, β and γ are shape, scale and location parameters, respectively. In the LP-III distribution,
the maximum value of expected rainfall (XT) corresponding to any return period (T) can be
calculated using equations given below.

log

2
_ 1 1
6 6

Where, , and Cs are the mean, standard deviation and coefficient of skewness of precipitation
data, respectively.
Log-normal distribution: A distribution of random variables with a normally distributed
logarithm is known as the log-normal distribution. A random variable y is part of the log-normal
distribution model, and Log(y) has a normal distribution. Equations given below were used to
compute the log-normal distribution's probability density function (PDF) and cumulative
distribution function (CDF).

1 ln
exp
2
√2
ln 1 ln
2 √2

where "Φ" is the Laplace Integral, "γ" is the location parameter, "σ" is the scale parameter, and
"µ" is the shape parameter. The highest value of predicted precipitation (XT) corresponding to any

41

Page 56 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 57 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

returning period (T) was determined using the equation since the log-normal distribution assumes
that Y=ln(X).
exp
1

Where and Cvy are the mean and coefficient of variation of Y, respectively. KT is the frequency
factor.
Gumbel (EV I) distribution: The PMP for the SPP and GBP data in the Wabi-Shebele river basin
was calculated using the Gumbel distribution, sometimes referred to as the Extreme Value Type I
(EV I) distribution. This distribution is a continuous probability distribution. The model maximum
or minimum values (extreme values) of a random variable are present in this distribution function.
Equations listed below were used to determine the Gumbel distribution's probability density
function (PDF) and cumulative distribution function (CDF).

Where the scale and location parameters are denoted by "σ" and "µ," respectively. Using the
following formula, the maximum value of predicted precipitation (XT) corresponding to the return
period (T) can be determined using the Gumbel distribution.

X 1

√6
_ 0.5772
1
where KT is the frequency factor, Cv is the coefficient of variation, and X ̅ is the mean. These
variables are dependent on the probability distribution of the precipitation events and the return
period (T).

3.11 Estimation of Return Period Values for PMP

The location and scale parameters were estimated and used for the computation of return period
values corresponding to estimated PMP value for durations of one day PMP for all stations.

42

Page 57 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 58 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

F(x) = exp (-e^-( )

√ ∗

̅ 0.5772 ∗

Where, are parameters of the extreme value type I (Gumbel) distribution that can be
estimated from the observed annual extreme rainfall series using the method of moment.

∗
1

Where, ln natural logarithm and

0.5772
K
√6

PMP ,

3.12 Performance Evaluation of CHIRPS

This study evaluated the performance of CHIRP satellite rainfall data using statistical methods and
graphical comparison charts. Five performance metrics were used to objectively compare rain
gauge data with the rainfall estimations from the CHIRPS satellite. The Pearson correlation
coefficient (CC), mean error (ME), mean absolute error (MAE), root mean square error (RMSE),
and percentage relative bias (PBIAS) were among the measures that were used. The degree of
agreement between the satellite estimates and rain gauge readings about the temporal dynamics is
shown by the CC (%). The percentage systematic bias of satellite rainfall as opposed to rain gauge
readings is represented by the PBIAS (%). The average and magnitude of errors are indicated by
the ME and MAE (mm), respectively. A tighter agreement between the two datasets is indicated
by reduced values of the RMSE (mm), which calculates the average absolute errors of the satellite
rainfall. The following formulas were used to assess these statistical measures on an annual basis.

43

Page 58 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 59 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

∑ G G S S
CC
∑ G G ∑ S S

∑ S G
PBIAS 100
∑ G

1
E S G
n

1
| |

Note: Gi denotes gauged rainfall, Si denotes satellite rainfall, and n denotes the number of samples
of rainfall data pair time series. G ̅ and S ̅ are the means of the gauge and satellite rainfall datasets,
respectively.

44

Page 59 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 60 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

4 RESULTS AND DISCUSSION

4.1 Probability Exceedence (PE) of PMP of Fafan Sub-basin

4.1.1 PE of PMP of ground-based Precipitation

The Figure 4.1 shows four exceedance probability graphs of Degehabur, Gursum, Jigjiga, and
Gode and the Figure 4.2 showed four stations including Hartisheik, Hadew, Chinaksen and Kebri
Dehar. The x-axis of each graph shows the amount of precipitation (mm), and the y-axis shows
the probability of exceeding that amount of precipitation. For example, the Degehabur graph shows
that there is a 99% chance of exceeding 10 mm of precipitation, a 90% chance of exceeding 20
mm of precipitation, and a 50% chance of exceeding 50 mm of precipitation. On the other hand,
in Figure 4.2, the probability of exceedance for Chinaksen station around 10 % has 100 mm
precipitation while this 10 % for Hadew is 80 mm precipitation. Hence, in Chinaksen, there is a
10% coincidental of exceeding 100 mm of precipitation.
The elucidation of these graphs both reported in Figures 4.1 and 4.2 depends on the specific activity
in which they are proposed. For example, they could be used to help farmers decide when to start
agriculture activity to produce crops, while engineers used to design infrastructure that can
withstand maximum flood that has been occurred due to heavy rainfall (Du et al., 2019; Nathan et
al., 2016).

Figure 4-1. Probability of Exceedance of ground-based Precipitation of Gursum, Degehabour,

Jigjiga and Gode

45

Page 60 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 61 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Figure 4-2. Probability of Exceedance of ground-based Precipitation of Hartisheik, Hadew,

Chinaksen and Kebri Dehar

4.1.2 PE of PMP of CHIRPS Precipitation

The probability of exceedance of CHIRPS precipitation refers to the likelihood of a specific

precipitation amount being surpassed; using satellite precipitation products such as CHIRPS that
has been used in the area of precipitation gauge scarcity as well as complex topography (Ramos et
al., 2022; Sharannya et al., 2020). In this probability estimation the ranking precipitation values
from highest to lowest from the annual maximum precipitation has been conducted. For instance,
in this study, 31 data values of precipitation measurements were converted to probability
exceedance (Figures 4.3 and 4.4). The probability of exceedance of precipitation evidence is
valuable for various purposes (Du et al., 2019; Quijano et al., 2015), including risk assessment in
agriculture water management, water resource management, and natural disaster preparedness like
flood risk management.

46

Page 61 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 62 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Figure 4-3. Probability of Exceedance of CHIRPS Precipitation of Gode, Gursum, Chinaksen

and Degehabour

Figure 4-4. Probability of Exceedance of CHIRPS Precipitation Hadew, Hartishek, Kebri Dehar
and Jigjiga

4.2 Flow Duration Curves of GBP

The graphs in Figure 4.4 show that the flow duration in the Fafan sub-basin is often brief, occurring
less than 50% of the time. The graphs created from the ground-based precipitation data show that,
using the Gumbel probability distribution function, the flow duration of stations like Gode is less

47

Page 62 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 63 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

than 10% for the majority of the year, while Jigjiga's flow duration is less than 20% for the majority
of the year.

PMP of ground-based precipitation was the main focus of this study report. It is crucial to
remember that these are only broad conclusions drawn from the graphs; rainfall, evaporation, and
water use are just a few of the variables that can affect the actual flow duration (Trenberth et al.,
2003; Berghuijs et al., 2016).

One useful tool in hydrology and water resource management is a flow-duration curve. The
cumulative frequency of discharges (streamflow) over a certain time period is shown by a flow-
duration curve. It displays the proportion of time that the designated discharges were met or
surpassed. In essence, it records the full discharge range without taking the order of events into
account1. Stream and basin characteristics can be inferred from the flow-duration curve's form.
The curve illustrates the kind of flood regime that the basin is expected to encounter in the high-
flow area. The basin's capacity to support low flows during dry seasons is characterized by the
low-flow area.

Numerous applications have made use of it, including water management, which uses flow-
duration curves to forecast future flow distributions for irrigated agriculture, hydropower, and
water supply. The low-flow endpoints of flow-duration curves are influenced by the geology in
geological evaluation, which makes them helpful for evaluating the geologic features of drainage
basins. Information on the relative duration of time that flows surpass particular levels is essential
for developing structures on streams. In general, flow-duration curves offer a thorough
understanding of streamflow behavior, supporting engineering design, environmental evaluation,
and water resource planning.

48

Page 63 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 64 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Figure 4-5. Flow duration curves of ground based precipitation

4.3 Flow Duration of CHIRPS

Flow duration curves are essential tools for understanding the distribution of streamflow or river
discharge over time as stated under section 4.2. Flow duration curves provide insights into the
frequency and duration of different flow levels. This study present under section 4.2 flow duration
49

Page 64 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 65 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

carvers of various stations from ground-based precipitation; however, to observe the behavior of
flow duration from satellite precipitation it has used CHIRPS dataset that estimates precipitation
(Figure 4.6). And, some similarities and variations were observed from both data sources. Hence,
in the area of precipitation scarcity, satellite precipitation products was used as reports from this
study suggested. There are similar studies done on flow duration curves such as on ungauged basin
(Müller & Thompson, 2016; Anagnostou & Mei, 2018). These studies has compared flow duration
curves produced from satellite precipitation and flow recorded some gauged river within the
watershed, but in this current study, Fafan sub-basin has no recorded flow for comparison.

Flow duration curves from satellite precipitation data are valuable tools in hydrological analysis,
but understanding their relationship requires considering their individual functionalities and
limitations.

As shown in Figure 4.6, it is a representation of the probability of exceeding a specific streamflow

value over a given time period; even though in this study estimated from precipitation.

The plotting showed the percentage of time it was equaled or exceeded during that period and used
for various water resource management applications, including flood analysis for identifying high
flow events and their frequency while low flow analysis used to determine water availability
during dry periods. Moreover, it has been used for designing infrastructure like dams and
reservoirs.

While Flow duration curves directly reflect streamflow, they are not directly tied to precipitation
data. However, satellite precipitation data can be used as an input for hydrological models or
empirical equation including excel that simulate streamflow generation. By considering various
factors like land cover: Vegetation type and soil characteristics influence infiltration and runoff,
topography: Slope and drainage patterns affect water flow and evaporation and
transpiration: Losses of water to the atmosphere.

It's crucial to remember that satellite precipitation data may not be perfect and hydrological models
have inherent uncertainties. Therefore, combining these tools with ground-based observations and
careful estimation from satellite precipitation like CHIRPS is essential for drawing reliable
assumptions.

50

Page 65 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 66 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Figure 4-6. Flow duration curves of CHIRPS precipitation

4.4 Maximum Frequency Factor (Km)

With an average CV of 2.63 and 2.52%, the ground-based precipitation maximum frequency factor
values for the stations ranged from 3.28 Chinaksen station to 1.63 Degehabour station (Table 4.1).
The Km result, which shows the data's variability, is greater than 20%. Variability in record

51

Page 66 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 67 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

duration or variations in the microclimatic conditions of the rain gauge stations might be the cause
of this fluctuation.

To see the trend, the calculated Km was compared to the station mean annual maximum rainfall
depths. As can be seen from Table 4.1, the trend indicates an inverse relationship, meaning that as
the mean annual total rainfall and the mean annual daily maximum increased, the value of Km
typically decreased.

Table 4-1. Maximum KT of ground-based Fafan sub-basin

Station μ α CV Km
Chinaksen 30.33 11.53 2.63 3.28
Degahabour 35.22 13.95 2.52 1.63
Gode 44.34 22.94 1.93 2.34
Gursum 41.53 17.92 2.32 2.52
Hadew 36.72 14.00 2.62 2.05
Hartishek 26.47 10.02 2.64 2.65
Jijiga 28.46 7.83 3.64 1.92
Kebri Dehar 44.67 25.74 1.74 3.08

Table 4.2 presents the CHIRPS precipitation maximum frequency factor values for the stations,
which ranged from 3.20 at Chinaksen station to 1.43 at Hadwe station. The average coefficient of
variation (CV) was found to be 2.63 and 2.52% (Table 4.1). Data variability was indicated by the
Km results, which were greater than 20%. Variations in record duration or microclimatic
conditions among the rain gauge sites may be the cause of this discrepancy. To find patterns, the
predicted Km values were contrasted with the station mean annual maximum rainfall depths.
According to Table 4.1, the trend shows an inverse relationship, with Km values often declining
as average annual total rainfall and mean annual daily maxima rise. This result is in line with
earlier studies that found that wet locations had greater Hershfield frequency factors than arid
regions.

Table 4-2. Maximum KT of CHIRPS of Fafan sub-basin

Station μ α CV Km
Chinaksen 34.91 9.09 3.84 3.20
Degahabour 36.51 11.64 3.14 3.22
Gode 37.76 10.37 3.64 2.02
Gursum 36.12 10.44 3.46 2.26
Hadew 33.26 7.22 4.61 1.43

52

Page 67 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 68 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Hartishek 29.52 6.57 4.50 2.32

Jijiga 32.60 8.61 3.79 3.00
Kebri Dehar 54.67 11.72 4.67 2.59

The observed enveloping Km (3.13) has estimated 79.10 % of the largest Hershfield frequency.
For Atrak Watershed, Iran this value has decreased to 67.35 % and 63.97 % in the Humid Regions
Malaysia. In Ethiopia such as Blue Nile Basin, Benishangul Gumuz region, West Shewa Zone
Oromia region, Arsi Zone and Tigray Region the value ranges from 65.17%, 61.78%, 53.91%,
47% and 47.04% respectively (Table).

Table 4-3: The maximum frequency factors for different basins or regions compared with Fafan
sub-basin maximum frequency factors

(K envelop-4.39)/K
Basin/Regions K envelop
envelop Source
Largest Hershfield frequency 15 70.73 Hershfield (1961)
Atrak Watershed, Iran 9.6 54.27 Ghamhraman (2008)
49.54 Desa and Rakhecha
Humid Regions Malaysia 8.7
(2006)
51.22 Alemayehu and Semu
Blue Nile Basin, Ethiopia 9
(2010)
Benishangul Gumuz region, 46.46
8.2 Mulualem (2010)
Ethiopia
West Shewa Zone Oromia 35.44
6.8 Mulugeta (2012)
region, Ethiopia
Arsi Zone, Ethiopia 5.9 25.59 Tilahun (2015)
Tigray Region, Ethiopia 5.91 25.72 Yohannes (2013)

4.5 Ratio of PMP to Highest Observed Precipitation (HOP)

The PMP estimates were contrasted with PMP estimations for both dry and humid locations, as
well as maximum recorded rainfall records. Table 4.4 shows the expected PMP to one-day HOP
ratio. The range of this ratio was 1.039 for Hartisheik and 1.038 for Gursum. Consistent with
Hershfield's (1962) finding, the magnitude ratio at a single site did not surpass three times the
maximum recorded rainfall depth. As a result, neither overestimation nor underestimation was
found in the study's anticipated PMP values. It is crucial to remember that new recordings of severe
storms may cause the PMP value to fluctuate over time. The estimated PMP values are the best
estimations that can be made using the information, methods, and expertise that are currently

53

Page 68 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 69 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

available. Disparities in record lengths and microclimates across stations could account for
variations in PMP estimations (Hershfield, 1961).

Table 4-4. Probable Maximum Precipitation (mm) PMP to HOP Ratio of Fafan sub-basin using
ground-based precipitation

Station PMP HOP Ratio

Chinaksen 76.22 73.86 36.17 11.40 1.032
Degahabour 77.10 74.74 37.75 13.45 1.032
Gode 65.72 63.36 38.58 11.32 1.037
Gursum 64.97 62.61 36.97 11.44 1.038
Hadew 119.02 116.66 35.95 16.41 1.020
Hartishek 62.68 60.32 30.52 8.53 1.039
Jijiga 87.23 84.87 34.29 12.63 1.028
Kebri Dehar 118.13 115.77 56.64 15.93 1.020
Table 4-5. Probable Maximum Precipitation (mm) PMP to HOP Ratio of Fafan sub-basin using
CHIRPS precipitation
Station PMP HOP Ratio
Chinaksen 74.75 71 34.96 10.19 1.053
Degahabour 79.17 75.42 36.54 12.25 1.050
Gode 72.75 69.00 37.38 10.11 1.054
Gursum 71.895 68.15 35.77 10.24 1.055
Hadew 96.985 93.235 34.74 15.21 1.040
Hartishek 69.873 66.123 29.32 7.32 1.057
Jijiga 94.75 91.00 33.09 11.43 1.041
Kebri Dehar 108.75 105.00 55.44 14.73 1.036

The relationship between PMP estimated and HOP observed precipitation has present in Figure
4.7 and except Hadew station all values are correlated well for CHIRPS satellite precipitation. In
addition, Figure 4.8, presents the ground-based precipitation relationship between PMP and HOP
using the graphical estimation of the correlation values.

54

Page 69 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 70 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Figure 4-7. Relation between CHIRPS and estimated PMP using Gumbel distribution function

55

Page 70 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 71 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Figure 4-8. Relation between GB and estimated PMP using Gumbel distribution function

4.6 Comparison of the Probability Distribution Functions

For the Gode station's yearly maximum rainfall data (mm), the reduced variate values for
exceedance probabilities were computed and displayed in Annex 1. As the plotting probability
56

Page 71 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 72 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

rises (i.e., as the recurrence interval falls), the results demonstrate that the reduced variate values
for all station records drop. Additionally, the extreme values produced exhibit a linear connection
with the reduced variate. The plotting probability (recurrence interval) increases as the variate
values for the station records decrease, and the extreme values obtained exhibit a linear relationship
with the standard normal variable. This is consistent with a comparison of the probability
distribution functions for the stations.

4.6.1 Probability Density Functions (PDFs)

Cumulative Distribution Functions (CDFs) and Probability Density Functions (PDFs) are effective
tools for evaluating rainfall data and comprehending how it relates to the mean and standard
deviation.

The likelihood that a given amount of rainfall will occur is explained in the PDF. The PDF's form
can be analyzed using the normal distribution function for all stations, as shown in Figure 4.9. The
rainfall is skewed toward heavier right- or lighter-to-left-skewed rainfall episodes, and tends to
cluster around the mean. The "average" quantity of rainfall is indicated by the mean, which is the
center of the PDF. The standard deviation, which is represented by the PDF's spread, informs us
how much the rainfall usually deviates from the mean.

Figure 4-9. The PDF of ground-based precipitation all stations using normal distribution function
of Fafan sub-basin

57

Page 72 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 73 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Figure 4-10. The PDF of ground-based precipitation all stations using Log-normal distribution
function of Fafan sub-basin

Figure 4-11. The PDF of ground-based precipitation all stations using Weibul distribution
function of Fafan sub-basin

4.6.2 Cumulative Distribution Functions (CDFs)

The CDF displays the likelihood that rainfall will fall below or equal a specific threshold. This is
useful for figuring out how likely it is to see rainfall below a certain threshold, like the amount

58

Page 73 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 74 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

required for farming or the amount that can cause flooding. This study offers the benefit of a
thorough grasp of rainfall patterns and their possible effects on the study area's natural resources
by examining both PDFs (Section 4.6.1) and CDFs together with the mean and standard deviation.

As a result, the CDF of ground-based precipitations in the Fafan sub-basin is shown in Figure 4.12;
in this study, which used the PMP, the CDF demonstrated the variability of precipitation amounts
from station to station. The CDF of precipitation at multiple locations in the Fafan sub-basin is
displayed in Figure 4.12. The likelihood that precipitation variability will be less than or equal to
a given value is displayed by the CDF. The CDF for Jigjiga, for instance, indicates that there is a
10% chance of precipitation that is less than or equal to 11 mm.

Each graph's vertical axis displays the likelihood of the CDF, while the horizontal axis displays
precipitation in millimeters. Confidence limits, which show the range of values that the true CDF
is likely to lie within, are also displayed on the graphs. The degree of certainty regarding the
genuine CDF decreases with increasing confidence bounds.

The Fafan sub-basin exhibits a broad range of precipitation values, as the graphs demonstrate.
Precipitation of less than or equal to 11 mm is 10% likely to have occurred in Jigjiga, whilst
precipitation of less than or equal to 20 mm is 10% likely to have occurred in Degehabour. We are
less positive of the true CDF for lower precipitation amounts, as the figures clearly demonstrate
that the confidence bounds are wider for these values.

59

Page 74 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 75 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Figure 4-12. The CDF of ground-based precipitation of all stations using normal probability
distribution function of Fafan sub-basin
Figure 4.13 displays the likelihood that rainfall will be less than or equal to a certain quantity based
on the CHIRPS CDF. This aids in determining the probability of precipitation falling below critical
levels for flood risk or agriculture. As a result, CDF mean and standard deviation PDFs offer a
thorough grasp of rainfall trends and their possible effects on the natural resources in the studied
area.

60

Page 75 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 76 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Figure 4.13 depicts the CDF of CHIRPS satellite precipitation in the Upper Fafan sub-basin. Using
the PMP, the figures for stations states variations in precipitation amounts across stations. It shows
the probability that precipitation falls below a certain level at several stations. For example, the
Jigjiga CDF indicates a 10% chance of receiving 11 mm or less of precipitation.

Overall, Figure 4.13 suggests substantial spatial variability in precipitation within the Upper Fafan
sub-basin. This information is valuable for water resource management, agriculture, and
meteorological analysis.

Ayehu et al. (2018) and Prakash (2019) both validate CHIRPS satellite precipitation in the upper
Nile river basin using the CHIRPS satellite precipitation CDF analysis.

However, in the same way as CHIRPS does, a ground-based precipitation Cumulative Distribution
Function (CDF) indicates the likelihood of witnessing rain amounts that are less than or equal to a
specific value. By comparing CDF and PDF, this useful tool for comprehending historical
precipitation patterns like that of CHIRPS is created using rainfall data from ground gauges (Belay
et al., 2019).

61

Page 76 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 77 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Figure 4-13. The CDF of CHIRPS precipitation of all stations using normal probability
distribution function of Fafan sub-basin

4.7 Performance of CHIRPS

CHIRPS rainfall estimates generally well agreed with gauged rainfall data at temporal scales used
in studies, like annual maximum rainfall data. This study compared CHIRPS with rain gauge
measurements in the Upper Fafan Sub-basin. Significant agreement was observed for Degahabour
and Jigjiga based on the statistical ME results. However, Gode, Kebri Dehar, and Gursum showed
very poor agreement with gauge-based precipitation measurements (GBPMP) estimations.
Evaluations revealed statistically significant correlations between CHIRPS and rain gauge data
(Table 4.6). Gode and Kebri Dehar are characterized by scarce rainfall, particularly in recent years.
Gursum is located in mountainous terrain; CHIRPS may have difficulty with rain detection in such
areas, as previous studies have indicated (Aksu & Akgül, 2020; Saeidizand et al., 2018).

Therefore, while CHIRPS is a valuable tool for rainfall estimation, it's important to understand its
strengths and limitations (Paredes et al., 2016). Users are advised to employ CHIRPS cautiously
in areas with sparse gauge networks or complex terrain, like the Upper Fafan Sub-basin, and to

62

Page 77 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 78 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

consult validation studies conducted in specific regions for a more comprehensive understanding
of CHIRPS accuracy (Ossa-Moreno et al., 2019; Du et al., 2024).

Table 4-6. Performance statistics of CHIRPS rainfall estimates of Upper Fafan Sub-basin (1990-
2020) annual time scale based

Station CC PBias RMSE ME MAE

Chinaksen 0.982 -5.340 3.102 4.156 0.134
Degahabour 0.939 0.340 7.445 0.493 0.515
Gode 0.559 10.256 30.285 43.243 1.267
Gursum 0.785 14.779 28.445 36.535 0.872
Hadew 0.937 8.613 19.664 1.052 0.704
Hartishek 0.963 3.219 12.604 3.959 0.524
Jijiga 0.979 6.410 13.405 0.725 0.159
Kebri Dehar 0.854 -4.017 21.690 39.073 1.255

63

Page 78 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 79 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

5 CONCLUSIONS AND RECOMMENDATION

5.1 Conclusions

The following are the main conclusions of the study that was carried out in the Fafan sub-basin
utilizing the PMP of satellite and ground-based CHIPRS precipitation:

Each station in the Fafan sub-basin has a different exceedance probability of precipitation.
Precipitation of less than or equal to 11 mm is 10% likely to have occurred in Jigjiga, whilst
precipitation of less than or equal to 20 mm is 10% likely to have occurred in Degehabour.

The Fafan sub-basin typically has short flow durations, with the majority of them lasting less than
50% of the time. Rainfall, evaporation, and water use are probably some of the contributing
variables.

Flow duration curves are useful resources for comprehending how river discharge or streamflow
is distributed over time. They can be applied to a number of tasks, including infrastructure design,
flood analysis, and water resource management.

The study's overall conclusions imply that the Fafan sub-basin exhibits notable fluctuation in
precipitation and flow duration. Road and water supply infrastructures, agriculture, and other uses
like water resource management may benefit from this knowledge.

5.2 Recommendation

The study on PMP using data from ground-based and satellite on the Fafan sub-basin with the
respective best fit probability distribution function to produce estimated PMP and FDC, the
following recommendations may be suggested:

 More evenly spaced gauging stations and longer-term rainfall records are needed to
confirm the following facts; therefore, a higher enveloping Km factor may be discovered,
leading to higher PMP values and a more validated best fit probability distribution function
for the area.
 The country's climate unpredictability necessitates that projects and study update the most
recent PMP data for the area.
 Since the resulting isohyetal map only offers point PMP calculations, areal reduction

64

Page 79 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 80 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

factors must be created in order to obtain more accurate and dependable area estimations.
The study recommends additional rain gauge stations that were utilized to compare gauge PMP
with satellite-based PMP since the gauging stations have an impact on the results derived from the
study area's annual maximum PMP.

65

Page 80 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 81 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

REFERENCE

Ahmed, Z.; Rao, D.; Reddy, K.; Raj, E. (2012). Rainfall intensity variation for observed data and
derived data: a case study of Imphal. ARPN J. Eng. Appl. Sci. 11(7), 1506–1513

Aksu, H., & Akgül, M. A. (2020). Performance evaluation of CHIRPS satellite precipitation
estimates over Turkey. Theoretical and Applied Climatology, 142(1), 71-84.

Al Mamoon, A., Rahman, A. (2017). Selection of the best fit probability distribution in rainfall
frequency analysis for Qatar. Nat. Hazards 86, 281–296. https://doi.org/10. 1007/s11069-
016-2687-0.

Alam, M., Emura, K., Farnham, C., Yuan, J. (2018). Best-fit probability distributions and return
periods for maximum monthly rainfall in Bangladesh. Climate 6, 9. https://
doi.org/10.3390/cli6010009.

Alemayew, D and Semu Ayalew, 2010. Development of Regional Probable maximum

precipitation for probable maximum flood development and application in design flood
hydrograph: case study of Blue Nile Basin, Ethiopia. The 210 Conference on water
Resoource in Ethiopia Addis Ababa.12-16 January 2010 Ministry of water Resource.

Anagnostou, E., & Mei, Y. (2018). Using high‐resolution satellite precipitation for flood frequency
analysis: Case study over the Connecticut River Basin. Journal of Flood Risk
Management, 11, S514-S526.

Ayehu, G. T., Tadesse, T., Gessesse, B., & Dinku, T. (2018). Validation of new satellite rainfall
products over the Upper Blue Nile Basin, Ethiopia. Atmospheric Measurement
Techniques, 11(4), 1921-1936.

Baker, F. B., & Kim, S. H. (Eds.). (2004). Item response theory: Parameter estimation techniques.
CRC press.

Barros, A. P., S. Chiao, T. J. Lang, D. Burbank, and J. Putkonen. (2006). From weather to
climate—Seasonal and interannual variability of storms and implications for erosion
process in the Himalaya. Geological Society of America Special Paper 398, Penrose
Conference Series, Boulder, CO, 17–38.

66

Page 81 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 82 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Behrangi, A., Khakbaz, B., Jaw, T. C., AghaKouchak, A., Hsu, K., and Sorooshian, S. (2011).
Hydrologic evaluation of satellite precipitation products over a mid-size basin, J. Hydrol.,
397, 225–237.

Belay, A. S., Fenta, A. A., Yenehun, A., Nigate, F., Tilahun, S. A., Moges, M. M., ... &
Walraevens, K. (2019). Evaluation and application of multi-source satellite rainfall product
CHIRPS to assess spatio-temporal rainfall variability on data-sparse western margins of
Ethiopian highlands. Remote Sensing, 11(22), 2688.

Berghuijs, W. R., Hartmann, A., & Woods, R. A. (2016). Streamflow sensitivity to water storage
changes across Europe. Geophysical Research Letters, 43(5), 1980-1987.

Bitew, M. M., and M. Gebremichael, 2010: Evaluation through independent measurements:

Complex terrain and humid tropical region in Ethiopia. Satellite Rainfall Applications for
Surface Hydrology, M. Gebremichael and F. Hossain, Eds., Springer-Verlag, 205–214.

Boluwade, A. (2020). Spatial-temporal assessment of satellite-based rainfall estimates in different

precipitation regimes in water-scarce and data-sparse regions. Atmosphere, 11(9), 901.

Bouizrou, I., Bouadila, A., Aqnouy, M., & Gourfi, A. (2023). Assessment of remotely sensed
precipitation products for climatic and hydrological studies in arid to semi-arid data-scarce
region, central-western Morocco. Remote Sensing Applications: Society and
Environment, 30, 100976.Upadhyay, S., Silwal, P., Prajapati, R., Talchabhadel, R.,
Shrestha, S., Duwal, S., & Lakhe, H. (2022). Evaluating magnitude agreement and
occurrence consistency of CHIRPS product with ground-based observations over medium-
sized river basins in Nepal. Hydrology, 9(8), 146.

Chikabvumbwa, S., Suribabu, C. R., & Worku, D. (2019). Rainfall frequency analysis using
Gumbel distribution.

Collier, C.G. and P.J. Hardaker, 1996. Estimating probable maximum precipitation using a storm
model approach journal of Hydrology 182:277-306.

Desa, M. M., & Rakhecha, P. R. (2006). Deriving the highest persisting monthly 24-hour dew
points in Malaysia for the estimation of PMP. IAHS PUBLICATION, 308, 287.

Desa, M.M., A.B. Noriah and R.R. (2001). Probable maximum precipitation for 24hr duration
over South east Asian Monson Region- Selagor Atomosheric Research j. 58:41-54
67

Page 82 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 83 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Deshpande, N. R., Kulkarni, B. D., Verma, A. K., & Mandal, B. N. (2008). Extreme rainfall
analysis and estimation of Probable Maximum Precipitation (PMP) by statistical methods
over the Indus river basin in India. Journal of spatial hydrology, 8(1).

Dhar, O. N., & Rakhecha, P. R. (1981). The effect of elevation on monsoon rainfall distribution in
the central. Monsoon dynamics, 253.

Dinku, T., Funk, C., Peterson, P., Maidment, R., Tadesse, T., Gadain, H., & Ceccato, P. (2018).
Validation of the CHIRPS satellite rainfall estimates over eastern Africa. Quarterly
Journal of the Royal Meteorological Society, 144, 292-312.

Dinku, T., Funk, C., Peterson, P., Maidment, R., Tadesse, T., Gadain, H., & Ceccato, P. (2018).
Validation of the CHIRPS satellite rainfall estimates over eastern Africa. Quarterly
Journal of the Royal Meteorological Society, 144, 292-312.

Dinku, T., S. J. Connor, and P. Ceccato, 2010: Comparison of CMORPH and TRMM-3B42 over
mountainous regions of Africa and South America. Satellite Rainfall Applications for
Surface Hydrology, M. Gebremichael and F. Hossain, Eds., Springer-Verlag, 193–204.

Du, H., Tan, M. L., Zhang, F., Chun, K. P., Li, L., & Kabir, M. H. (2024). Evaluating the
effectiveness of CHIRPS data for hydroclimatic studies. Theoretical and Applied
Climatology, 155(3), 1519-1539.

Du, H., Wang, Y., Liu, K., & Cheng, L. (2019). Exceedance probability of precipitation for the
Shuhe to Futuan Water Transfer Project in China. Environmental Earth Sciences, 78, 1-12.

Evans DL, Drew JH, Leemis LM. (2008). The distribution of the Kolmogorov-Smirnov, Cramer–
von Mises, and Anderson-Darling Test statistics for exponential populations with
estimated parameters. Commun Stat Simul Comput 37(7):1396–1421.

Farooq M, Shafique M, Khatta MS. (2018). Flood frequency analysis of river swat using Log
Pearson type 3, Generalized Extreme Value, Normal, and Gumbel Max distribution
methods. Arab J Geosci 11(9):1–10.

Funk C, Peterson P, Landsfeld M, Pedreros D, Verdin J, Shukla S, Husak G, Rowland J, Harrison

L, Hoell A, Michaelsen J. (2015). The climate hazards infrared precipitation with stations-
a new environmental record for monitoring extremes. Sci Data 2:150066.
https://doi.org/10.1038/sdata.2015.66
68

Page 83 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 84 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Funk, C., Peterson, P., Landsfeld, M., Pedreros, D., Verdin, J., Shukla, S., ... & Michaelsen, J.
(2015). The climate hazards infrared precipitation with stations—a new environmental
record for monitoring extremes. Scientific data, 2(1), 1-21.

Funk, C., Peterson, P., Landsfeld, M., Pedreros, D., Verdin, J., Shukla, S., Husak, G., Rowland, J.,
Harrison, L., Hoell, A., and Michaelsen, J. (2015a). The climate hazards infrared
precipitation with stations – a new environmental record for monitoring extremes,
Scientific Data, 2, 150066, https://doi.org/10.1038/sdata.2015.66, 2015a.

Gebremedhin, Y. G., Quraishi, S., & Itefa, H. (2017). Development of one day probable maximum
precipitation (PMP) and isohyetal map for Tigray Region, Ethiopia. Global Journal of
Science Frontier Research: H Environment & Earth Science, 17(2).

Guo H, Bao A, Liu T, Ndayisaba F, He D, Kurban A, de Maeyer P. (2017). Meteorological drought

analysis in the lower mekong basin using satellite-based long-term CHIRPS product.
Sustainability:9. https://doi.org/10.3390/su9060901

Hansen, E. M. (1994). Probable maximum precipitation, Pacific Northwest states: Columbia River
(including portions of Canada), Snake River and Pacific coastal drainages.

Hansen, E.M., L.C. Schreiner and J.F. Miller, 1982. Application of probable maximum
precipitation estimates, United States east of the 105th meridian Hydro meteorological
Report No.52, National Weather Service, Washington D.C.

Hong, Y., D. Gochis, J. T. Cheng, K. L. Hsu, and S. Sorooshian. (2007). Evaluation of

PERSIANN-CCS rainfall measurement using the NAME Event Rain Gauge Network. J.
Hydrometeor., 8 , 469–482.

Huffman, G. J., and Coauthors. (2007). The TRMM Multi satellite Precipitation Analysis (TMPA):
Quasi-global, multilayer, combined-sensor, precipitation estimates at fine scales. J.
Hydrometeor., 8, 38–55.

Huffman, G.J., et al. (2020). Integrated Multi-satellite Retrievals for the Global Precipitation
Measurement (GPM) Mission (IMERG). In: Levizzani, V., et al. (Eds.), Satellite
Precipitation Measurement, vol. 1. Springer International Publishing, Cha, pp. 343–353.
doi:10.1007/978-3-030-24568-9_19.

69

Page 84 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 85 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Huffman, G.J.; Bolvin, D.T.; Nelkin, E.J.; Wolff, D.B.; Adler, R.F.; Gu, G.; Hong, Y.; Bowman,
K.P.; Stocker, E.F. (2007). The TRMM multisatellite precipitation analysis (TMPA):
Quasi-global, multiyear, combined-sensor precipitation estimates at fine scales. J.
Hydrometeorol. 2007, 8, 38–55.

Joyce, R. J., J. E. Janowiak, P. A. Arkin, and P. Xie. (2004). CMORPH: A method that produces
global precipitation estimation from passive microwave and infrared data at high spatial
and temporal resolution. J. Hydrometeor., 5, 487–503.

Kennedy, M.R., H.J. Pearce, R.P. Canterford and L.J. Minty. (1988). The estimation of generalized
probable maximum precipitation in Australia. Proceeding of the Workshop on Spillway
Design Floods, AnU, Canberra, Australia, 4 February, An International Commission on
Large Dams, Bulletin no.79.

Kong X, Hao Z, Zhu Y. (2020). Entropy theory and Pearson type-3 distribution for rainfall
frequency analysis in semi-arid region. IOP Conf Ser Earth Environ Sci 495:012042.

Koutsoyiannis D., 1999. A probabilistic view of Hershfield method for estimating probable
maximum precipitation. Water Resources Research 35(4): 1313-1322.

Koutsoyiannis, D. (2004). Statistics of extremes and estimation of extreme rainfall: II. Empirical
investigation of long rainfall records/Statistiques de valeurs extrêmes et estimation de
précipitations extrêmes: II. Recherche empirique sur de longues séries de précipitations.
Hydrological Sciences Journal, 49(4).

Koutsoyiannis, D. (2007). A critical review of probability of extreme rainfall: principles and

models. Advances in urban flood management, 151-178.

Kozu, T., Iguchi, T., Kubota, T., Yoshida, N., Seto, S., Kwiatkowski, J., & Takayabu, Y. N. (2009).
Feasibility of raindrop size distribution parameter estimation with TRMM precipitation
radar. 87, 53-66.

Lee, J., E.-H. Lee, and K.-H. Seol. (2019). Validation of Integrated Multi-satellite Retrievals for
GPM (IMERG) by using gauge-based analysis products of daily precipitation over East
Asia. Theoretical and Applied Climatology, 137: 2497-2512. DOI: 10.1007/s00704-018-
2749-1

70

Page 85 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 86 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Li, N., Tang, G., Zhao, P., Hong, Y., Gou, Y., & Yang, K. (2017). Statistical assessment and
hydrological utility of the latest multi-satellite precipitation analysis IMERG in Ganjiang
River basin. Atmospheric research, 183, 212-223.

McCuen, R.H. (1989). Hydrologic Analysis and Design, 867 p. (Englewood Cliffs, NJ: Prentice

Müller, M. F., & Thompson, S. E. (2016). Comparing statistical and process-based flow duration
curve models in ungauged basins and changing rain regimes. Hydrology and Earth System
Sciences, 20(2), 669-683.

Nathan, R., Jordan, P., Scorah, M., Lang, S., Kuczera, G., Schaefer, M., & Weinmann, E. (2016).
Estimating the exceedance probability of extreme rainfalls up to the probable maximum
precipitation. Journal of hydrology, 543, 706-720.

Nazemi, A., M. Alam, and A. Elshorbagy. (2014). Uncertainties in future projections of extreme
rainfall at fine scales: The role of various sources.

NOAA (National Oceanic and Atmospheric Administration), 2006. Frequency Alas of the United
State, 1 (4)

Ossa-Moreno, J., Keir, G., McIntyre, N., Cameletti, M., & Rivera, D. (2019). Comparison of
approaches to interpolating climate observations in steep terrain with low-density gauging
networks. Hydrology and Earth System Sciences, 23(11), 4763-4781.

Paredes Trejo, F. J., Alves Barbosa, H., Peñaloza-Murillo, M. A., Moreno, M. A., & Farias, A.
(2016). Intercomparison of improved satellite rainfall estimation with CHIRPS gridded
product and rain gauge data over Venezuela. Atmósfera, 29(4), 323-342.

Peralta-Hernadez, A.R., R.C. Balling and L.R. Barba-Marinez, 2009. Comparative analysis of
indices of extreme rainfall event: variation and trend from southern Mexico. Atomosfera
22 (2):219-228.

Pipunic, R.C., et al. (2015). An evaluation and regional error modeling methodology for near-real-
time satellite rainfall data over Australia. J. Geophys. Res.: Atmos. 120 (20), 10767–10783.
https://doi.org/10.1002/2015jd023512.

Prakash, S. (2019). Performance assessment of CHIRPS, MSWEP, SM2RAIN-CCI, and TMPA

precipitation products across India. Journal of hydrology, 571, 50-59.

71

Page 86 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 87 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Quijano, J. A., Jaimes, M. A., Torres, M. A., Reinoso, E., Castellanos, L., Escamilla, J., & Ordaz,
M. (2015). Event-based approach for probabilistic agricultural drought risk assessment under
rainfed conditions. Natural Hazards, 76, 1297-1318.

Raiford, J.P.; Aziz, N.M.; Khan, A.A.; Powell, D.N. (2007). Rainfall depth-duration-frequency
relationships for SC related papers rain1fall depth-duration-frequency relationships for
South Carolina, North Carolina, and Georgia. Am. J. Environ. Sci. 3(2), 78–84

Ramos Filho, G. M., Coelho, V. H. R., da Silva Freitas, E., Xuan, Y., Brocca, L., & das Neves
Almeida, C. (2022). Regional-scale evaluation of 14 satellite-based precipitation products in
characterising extreme events and delineating rainfall thresholds for flood
hazards. Atmospheric Research, 276, 106259.

Reference

Rezacova, D., P. Pesice and Z. Sokol, 2005. Estimation of probable maximum precipitation for
river basins in Czech Republic. Atmos. Res. 77:407-421.

Saeidizand, R., Sabetghadam, S., Tarnavsky, E., & Pierleoni, A. (2018). Evaluation of CHIRPS
rainfall estimates over Iran. Quarterly Journal of the Royal Meteorological Society, 144,
282-291.

Sharannya, T. M., Al-Ansari, N., Deb Barma, S., & Mahesha, A. (2020). Evaluation of satellite
precipitation products in simulating streamflow in a humid tropical catchment of India using
a semi-distributed hydrological model. Water, 12(9), 2400.

Silva Lelis, L. C., R. W. Duarte Bosquilia, and S. N. Duarte. (2018). Assessment of Precipitation
Data Generated by GPM and TRMM Satellites. Revista Brasileira de Meteorologia, 33:
153-163.

Sorooshian, S., K. Hsu, X. Gao, H. V. Gupta, B. Imam, and D. Braithwaite. (2000). Evaluation of
PERSIANN system satellite-based estimates of tropical rainfall. Bull. Amer. Meteor. Soc.,
81 , 2035–2046.

Trenberth, K. E., Dai, A., Rasmussen, R. M., & Parsons, D. B. (2003). The changing character of
precipitation. Bulletin of the American Meteorological Society, 84(9), 1205-1218.

72

Page 87 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 88 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

WMO (World Meteorological Organization. (2009). Manual on estimation of probable maximum

precipitation (PMP). Geneva, Switzerland: World Meteorological Organization.

World Meteorological Organization, 1973. Manual for Estimation of Probable Maximum

Precipitation. Operational Hydrology Report No. 1, WMO-no.332, Geneva.

World Meteorological Organization, 1986. Manual for Estimation of Probable Maximum

Precipitation. Operational Hydrology Report No. 1, WMO-no.332, 269.

Yuan, J., Emura, K., Farnham, C., Alam, M.A. (2018). Frequency analysis of annual maximum
hourly precipitation and determination of best fit probability distribution for regions in
Japan. Urban Clim. 24, 276–286. https://doi.org/10.1016/j.uclim. 2017.07.008.

Zambrano F, Wardlow B, Tadesse T. (2016). Evaluating satellite-derived long-term historical

precipitation datasets for drought monitoring in Chile. Remote Sens Agric Ecosyst Hydrol
XVIII 9998:999823. https://doi.org/10.1117/12.2241032

73

Page 88 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 89 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

ANNEX

Annex 1. Reduced variate value for exceedance probability for the CHIRPS annual maximum
rainfall Computed using normal probability distribution for Gode station.

Year Rainfall (mm) Rainfall Order Rank P W Z XT

28.36 63.36
1990 1 0.03 2.62 1.85 59.85
20.66 58.71
1991 2 0.06 2.34 1.52 56.05
52.43 56.60
1992 3 0.10 2.16 1.30 53.54
40.31 53.44
1993 4 0.13 2.02 1.13 51.59
48.43 52.43
1994 5 0.16 1.91 0.99 49.96
46.80 48.43
1995 6 0.19 1.81 0.86 48.53
34.89 46.80
1996 7 0.23 1.73 0.75 47.24
32.06 45.09
1997 8 0.26 1.65 0.65 46.05
35.66 43.47
1998 9 0.29 1.57 0.55 44.93
34.96 43.06
1999 10 0.32 1.50 0.46 43.87
39.31 42.90
2000 11 0.35 1.44 0.37 42.86
37.90 42.63
2001 12 0.39 1.38 0.29 41.88
37.16 42.53
2002 13 0.42 1.32 0.20 40.92
35.66 40.31
2003 14 0.45 1.26 0.12 39.98
42.53 39.31
2004 15 0.48 1.20 0.04 39.05
21.03 37.90
2005 16 0.52 1.15 -0.04 38.12
45.09 37.16
2006 17 0.55 1.10 -0.12 37.19
43.47 35.66
2007 18 0.58 1.04 -0.20 36.25
43.06 35.66
2008 19 0.61 0.99 -0.29 35.30
42.90 34.96
2009 20 0.65 0.94 -0.37 34.33
17.16 34.89
2010 21 0.68 0.88 -0.46 33.32
63.36 33.69
2011 22 0.71 0.83 -0.55 32.28
33.69 32.95
2012 23 0.74 0.77 -0.64 31.19
58.71 32.06
2013 24 0.77 0.72 -0.74 30.04
42.63 28.36
2014 25 0.81 0.66 -0.85 28.79

74

Page 89 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 90 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

32.95 26.95
2015 26 0.84 0.59 -0.97 27.43
56.60 25.79
2016 27 0.87 0.53 -1.10 25.89
22.08 22.08
2017 28 0.90 0.45 -1.26 24.11
26.95 21.03
2018 29 0.94 0.37 -1.45 21.91
53.44 20.66
2019 30 0.97 0.26 -1.72 18.83
25.79 17.16
2020 31 1.00 0.00 -2.52 9.64
Annex 2. Reduced variate value for exceedance probability for the CHIRPS annual maximum
rainfall Computed using Log-Pearson-Type III probability distribution for Gode station

Year Rain Fall (RF) RF Order Log (RF) Kz YT XT

1990 28.3562 63.3568 1.80 1.58 1.79 61.14
1991 20.6588 58.7098 1.77 1.56 1.78 60.70
1992 52.4319 56.5979 1.75 1.55 1.78 60.49
1993 40.3148 53.4386 1.73 1.53 1.78 60.16
1994 48.4305 52.4319 1.72 1.52 1.78 60.05
1995 46.8046 48.4305 1.69 1.50 1.78 59.59
1996 34.8941 46.8046 1.67 1.49 1.77 59.40
1997 32.0618 45.0861 1.65 1.48 1.77 59.18
1998 35.6553 43.4713 1.64 1.47 1.77 58.97
1999 34.96 43.0642 1.63 1.46 1.77 58.91
2000 39.3113 42.9046 1.63 1.46 1.77 58.89
2001 37.903 42.6289 1.63 1.46 1.77 58.86
2002 37.163 42.5338 1.63 1.46 1.77 58.84
2003 35.6614 40.3148 1.61 1.44 1.77 58.53
2004 42.5338 39.3113 1.59 1.44 1.77 58.39
2005 21.0272 37.903 1.58 1.42 1.76 58.18
2006 45.0861 37.163 1.57 1.42 1.76 58.06
2007 43.4713 35.6614 1.55 1.41 1.76 57.82
2008 43.0642 35.6553 1.55 1.41 1.76 57.82
2009 42.9046 34.96 1.54 1.40 1.76 57.71
2010 17.1615 34.8941 1.54 1.40 1.76 57.70
2011 63.3568 33.6912 1.53 1.39 1.76 57.50
2012 33.6912 32.9479 1.52 1.38 1.76 57.37
2013 58.7098 32.0618 1.51 1.37 1.76 57.21
2014 42.6289 28.3562 1.45 1.33 1.75 56.50
2015 32.9479 26.9539 1.43 1.32 1.75 56.21
2016 56.5979 25.7903 1.41 1.30 1.75 55.95
2017 22.0822 22.0822 1.34 1.25 1.74 55.06

75

Page 90 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

 Page 91 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

2018 26.9539 21.0272 1.32 1.24 1.74 54.78

2019 53.4386 20.6588 1.32 1.23 1.74 54.68
2020 25.7903 17.1615 1.23 1.17 1.73 53.61
Annex 3. Reduced variate value for exceedance probability for the CHIRPS annual maximum
rainfall Computed using Log-normal probability distribution for Gode station

Year Rain Fall (RF) RF Order Log (RF) Rank P W Kz YT XT

1990 28.36 63.36 1.80 1 0.02 2.87 2.14 1.86 73.25
1991 20.66 58.71 1.77 2 0.05 2.46 1.66 1.80 62.77
1992 52.43 56.60 1.75 3 0.08 2.24 1.40 1.76 57.73
1993 40.31 53.44 1.73 4 0.11 2.09 1.21 1.73 54.32
1994 48.43 52.43 1.72 5 0.15 1.96 1.06 1.71 51.70
1995 46.80 48.43 1.69 6 0.18 1.86 0.92 1.70 49.55
1996 34.89 46.80 1.67 7 0.21 1.77 0.81 1.68 47.71
1997 32.06 45.09 1.65 8 0.24 1.68 0.70 1.66 46.09
1998 35.66 43.47 1.64 9 0.27 1.61 0.60 1.65 44.63
1999 34.96 43.06 1.63 10 0.31 1.54 0.51 1.64 43.30
2000 39.31 42.90 1.63 11 0.34 1.47 0.42 1.62 42.07
2001 37.90 42.63 1.63 12 0.37 1.41 0.33 1.61 40.91
2002 37.16 42.53 1.63 13 0.40 1.35 0.24 1.60 39.82
2003 35.66 40.31 1.61 14 0.44 1.29 0.16 1.59 38.78
2004 42.53 39.31 1.59 15 0.47 1.23 0.08 1.58 37.78
2005 21.03 37.90 1.58 16 0.50 1.18 0.00 1.57 36.81
2006 45.09 37.16 1.57 17 0.53 1.12 -0.08 1.55 35.87
2007 43.47 35.66 1.55 18 0.56 1.07 -0.16 1.54 34.94
2008 43.06 35.66 1.55 19 0.60 1.02 -0.24 1.53 34.03
2009 42.90 34.96 1.54 20 0.63 0.96 -0.33 1.52 33.13
2010 17.16 34.89 1.54 21 0.66 0.91 -0.41 1.51 32.23
2011 63.36 33.69 1.53 22 0.69 0.86 -0.50 1.50 31.33
2012 33.69 32.95 1.52 23 0.73 0.80 -0.59 1.48 30.41
2013 58.71 32.06 1.51 24 0.76 0.74 -0.69 1.47 29.47
2014 42.63 28.36 1.45 25 0.79 0.69 -0.80 1.45 28.50
2015 32.95 26.95 1.43 26 0.82 0.62 -0.91 1.44 27.49
2016 56.60 25.79 1.41 27 0.85 0.56 -1.03 1.42 26.40
2017 22.08 22.08 1.34 28 0.89 0.49 -1.18 1.40 25.22
2018 26.95 21.03 1.32 29 0.92 0.41 -1.35 1.38 23.87
2019 53.44 20.66 1.32 30 0.95 0.31 -1.57 1.35 22.23
2020 25.79 17.16 1.23 31 0.98 0.18 -1.92 1.30 19.84

76

Page 91 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

Page 92 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489

77

Page 92 of 92 - AI Writing Submission Submission ID trn:oid:::1:3081449489