FEDERAL DEMOCRATIC REPUBLIC OF ETHIOPIA
MINISTRY OF WATER, IRRIGATION AND ENERGY
DABUS HPPs
Consultancy Services for Pre-Feasibility and Feasibility Study 
of the Dabus Hydropower Project
Procurement Reference Number - DHPP/MoWIE/03/2014
DABUS HPPs
PRE-FEASIBILITY DESIGN
HYDROLOGICAL REPORT
July 2016
J
studio pietrangeli
consulting engineersstudio Ing. g. pietrangeli s.r.l.
via cicerone, 28 - rome - ITALY tel: 06.32.10.880 fax: 06.32.27.276
e-mail: dabus@pletrangell.lt
pag. 1 of 1
Ministry of Water Irrigation & Energy
Halle G/Sllase Road P.O. Box No.: 5744
Addis Ababa, Ethiopia
Rome, 01st July 2016
Att.ne:
Mr. Getu Tilahun
Email:
aetu tllahun2005@vahoo.com
Telephone:
+251 11 11 / 11 6637222
Telefax:
+251 11 6610710
DABUS
Consultancy Services for the PRE-FEASIBILITY and FEASIBILITY Studies of the DABUS Hydropower Project
PRE-FEASIBILITY DESIGN and EIA BASELINE rev. B- Transmission Letter
Dear Mr. Getu TILAHUN,
following:
-    your comments to the pre-feaslblllty design report and EIA baseline report (reference Is made to the attached matrix);
-     the workshop held In Addis Abeba on 16th June 2016;
-    our meeting held at your office on 17th June 2016,
We have updated the following reports:
-     100 WHO R SP 003 B vol. 1 / 2 - Main Report;
-     100 WHO R SP 003 B vol. 2 / 2 - Drawings;
-     120 HYD R SP 001 A - Hydrological Report;
130 GEO R SP 001 A - Geological and Geotechnical Report;
-     140 SEA R SP 001B - BASELINE Study.
Yours faithfully, Studio Ing. G. Pietrangeli S.r.i.
Antonio PIETRANGELI
Enclosures : n. 20 hard copies; 2 CD
5OC. € 98.130 l.v. - c.c.I.a.a. roma n 492427 - trlb. Roma n. 2626/8 2 - c.f. 05440170586 - p.I. 01401561004 capWORKSHOP 16 JUNE 2016
P.W)*’ >
cVWSW MATS' flf.WCWSfS TO CL IEN T ORSER VA TIONS
No.
Comments by Client
Answer by Consultant
------------
Q7
yvfv»? < 1               Sbnrr.                            Hearty, at which site the Drone survey oykAxThX cv» HP ch-105 (DAM area) or on HP ch- J55 (DAM area).
A7
The drone survey was carried out on the ch-155 dam area.
Tills typo will be corrected In Revision B of the Prcfeaslbillty Report.------------------------------ -------- t
QS
Sectwi  3.2:  Describe  briefly  the  problems  for  post  processing  of  the  PLEIADES 5JiYtvsuy>c                   withxi the last four months. In order to compare and
oX\>e refined one than only using SRTM 30x30m as usual.
*
A8
All the topographic activities regarding the Pleiades have been concluded
processing). In April 2016 a ground survey was carried out to acquire Ground Control ho (GCPs) to georeference and orthorectify the Images. Processing to obtain contour lines a_> rec r been completed so the Pleiades topographic database will be used, Instead of the SRTM, for tne
Feasibility Study since It has a much better accuracy.
—------------—-
Section   3.3:   It   needs   further   clarifications   why   the   drone   survey   did   not conduct on PH ch-IOS.
Q9
A9
There  were  problems  finding  suitable  landing  area  near  PH  ch-108.  The  Consultant  will  start  from the PLEIADES Images which are sufficient for design purposes at this stage.
Another  drone  survey  will  carried  out  In  a  future  stage  once  good  landing  areas  have  been Identified.
Q10
Figure   3.3.4:   Please   justify   name   whether  the   figure  name  'DAM  CH  150, surveyed areas 'is correct or not.
A10
The  name  of  the  figure  should  read  "ch  155".  This  will  be  corrected  In  Revision  B  of  the Prefeasibility Report.
Qu
Table 3.3.1: Droned surveyed areas  features, there are no ground control points shown at Dam Alt CH 155 and Alt CH 108. If not necessary for analysis, please give your justification dearly.
AU
The Consultant explained that another type of drone (RTK) was used that does not require GCPs. This drone is currently under testing.
Q12
Section  7.3.2:  Discuss  in  detail  the  tvoe  model  or  software  and  method  used for carrying out the Reservoir operation and energy production simulations
A12
A mathematical model created by SP was used for the reservoir routing. The model is described at length in the report.
Q13
The Reservoir volume and area curves as a function of elevation should be determined based on detailed topographic survey.
A13
Agreed.
In the Feasibility Study the accuracy will be improved (Plealdes Database).
Q14
As  per the TOR section 5.10: The Consultant should estimate the reservoir dead storage and dead levels  based on sediment measurement, but the report was done based on secondary data from literature against the TOR.
A14
More detailed modelling will be done during the Feasibility Study.
Dabus Comment (1): Dam safety expert
Q15
Most of the activities  are executed as  per the given TOR for Pre-feasibility study. However, most of the important site investigations  was  not done and completely  left for feasibility  stage investigation. Similarly, during the current geotechnical field investigation the emphasis  given to dam site and reservoir area is  quite low. This  means, dam foundation, reservoir water tightness, presence of weak  zones  not considered as  required for next phase wort planning.
A15
The  Consultant explained  that  investigation  planning  will  be  done  in  the  near future  in  collaboration with the Client.
______
Q16
Construction   matena!   available   quantity   and   quality   should   be   addressed   tc some detail at least for prefeasibility stage.
A16
For  the  site  at  ch  -  130  the  Consultant  estimated  the  quality  and  the  quantity  of  the  quarry material (ref. page 86/223).
For the site at ch - 108, detailed indications were not considered necessary given the abundant evidence of easily available construction material outcropping everywhere.
Q17
Instead of detailed petrography  analysis  of rock  the project area, th consultant should give more emphasis  to direct or indirect field investigate or structural geological investigation is  more important If necessary  at th stage such analysis should be done specifically a certain construction matena
e
n Ar
s
7   The  Consultant  pointed  out  that  this  information  was  included  in  the  report  130  GEO  R  SP  001 (Geology Report) of the Prefeasibility Report.
160627, DABUS, Reply to OBS at PRE-FEAS Report (Workshop 160615)
studio pietrangeli, romeDABUS hpp
WORKSHOP 16 JUNE 2016
CONSUL TANTS RESPONSES TO CLIENT OBSER VA T1ONS
Pago 4 of 9
No.
Comments by Client
Answer by Consultant
Q27
Though J  awment gnen on the inception report this  study  still focuses  on the S3me few sites  along the river course which are identified by  previous studies
A27
The entire river course was studied.
As  can  be  seen  clearly  from  the  longitudinal  profile  of  the  river  (ref.  Fig.  7.2.2),  the  areas  studies were the only ones feasible from a technical-economic viewpoint.
Q28
Selection of dams  made by  considering different factors  which had impact on the cost of investment and maintenance. This  mainly  depends  on the distance of the location of construction material. However, without consideration of this and other factors  simply  roller compacted concrete (RCC) and concrete surface rock fill dams are proposed.
A28
The main factors  (geology, quarry, maintenance etc.) have been considered In this  preliminary evaluation, more than adequate for a pre-feslblllty  design. More detailed elements  will be considered In the next stage (Feasibility Design).
Q29
Some  quarry  sites  are  identified  in  this  study  but  no  account  related  to  quality and volume of the material to be quarried has been given.
A29
Reference  Is  made  to  quality  and  volume  of  the  Identified  quarry  sites  at  Page  86/223  of  the Prefeasibility Report.
DabusHPPreFeasibilityDesign_HydrologicalComponent: Hydrology &
Water Quality Directorate (Surafel M)
Q30
The site map shall incorporate the location of the proposed dam sites and the nver networks for ease for visualization even though shown in (Vol. 2
drawing). Page 30.
A30
Noted and agreed.
Revision B of the Prefeasibility Report will address this.
In calibration processes  using concurrent data (precipitation and flow discharge data), you used a 3 years  data only  (1965-1967), why? And similarly for validation (1969 and 1972-1974). Page 54 and 55.
Q31
A31
In statistical tests (hypothesis testing), randomness, independence, homogeneity and stationarity methods were adopted but not shown summarized results of each test, except trend and stabonarity tests. Moreover, these tests are inferred as neglected (trend), and conservabve to use in practice (the stabonarity test), why and how? Page 66 and 67.
Periods  spanning Jan 1965, Dec  1967 and Jan 1972, Dec  1974 provide the more complete set of significant daily  discharge measurements  at the station of Dabus  near Assosa and were therefore selected for calibration and validation purposes. Different periods  were selected to perform the calibration and the validation of the model In order to guarantee that the calibration resulting from one period was  actually  efficient for the application of the model to another period, in other words in order to guarantee the generalization to other period of the calibration performed on a specific period.
Q32
A32
A table presenting the results of the performed tests in terms of the specific statistics used by each test is included in the revised version of the report.
The assumption of neglecting the observed trend stems  from the observation that the trend is negative, i.e. the observed maxima tend to decrease in the spanned period (1961-2008). The analysis  of the extremes  considering the sample as  a not-stationary  sequence, i.e. considering statistical models  whose parameters  are dependent on time (e.g. Reiss  and Thomas, 2007, Coles, 2000, Fawcett, 2013), would have produced a not conservabve decreasing in time of the discharge quintiles.
Reiss,  R.D.  and  Thomas,  M.  (2007).  Statistical  Analysis  of  Extreme  Values.  With  application  to Insurance, Finance, Hydrology and Other Heids. Birkhauser. Third Edition.
Coles, S. (2000). An introduction to statistical modelling of extreme values. Springer.
Fawcett,   L.   (2013).   Topics   In   Statistics.   Environmental   Extremes.   Lecture   Notes.   University   of Newcastle.
Q33
Please correct nrooeriv the
uie nyuromemc stabon name of D/dessa near Ario
r.
instead of Ario throughout the documents.
A33
Q34
On page 47, whxfi plot on left of Fig. 4.19indtetes? Similarly Fig 2 on page
------------------------i.uj  win uc uuiirvieu in Revision b or tne Report.
A34
Noted and agreed.
i mb win De corrected in Revision B of the Report Noted and agreed.
DAbUb, Reply to OBS at PRE-FEAS Report (Workshop 160615)------------------------------------------------- ---------------------------------------- ------------------------------------------ ---------------------------------------------------
Studio Dietranapli- rnmo----■■■■VftlllllJ
DABUS hpp
WORKSHOP 16 JUNE 2016
________
____________________________________________ CONSULTANTS RESPONSES TO CLIENT OBSERVATIONS
No.
-
Comments by Client
Answer by Consultant
Q35 |
Genera!  comment,  in  Vol.3  Support  Data,  it  is  not  a  good  idea  to  put  all  the daily data as printed and/ or avail publidy due to security reasons.
A35
Noted.
This will be removed from Revision B of the Prefeasibility Report
-------------------------------
Hydropower Development and Dam Administration Directorate
Although CRGE (Climate Resilient Green Economy) strategy  is  one of the most important issues  for Ethiopia to build up environmental friendly  economic growth in particular due to the fact that hydropower is  one of the pillars  for the implementation of the strategy, it is  ignored in the report. Thus, the issue should have been indicated and well addressed m the report
Agreed to discuss the CRGE In the final report
Legal provision like Ethiopian water resource management proclamation 197/2000 and Regulation (115/2005), which are likely  to have an association with construction of dam for power generation are not incorporated in the report.
Agreed to incorporate the proclamation 197/2000 and Regulation (115/2005) in the final report
Q36
Q37
Due to the fact that there will be huge amount of solid waste generated during the construction time, it is  important to address  the issue with regard to Solid Waste Management Proclamation (513/2007)
Agreed to address Proclamation 513/2007 In the final report
Q38
Annex  1.1 of page 132, It is  stated that Hippopotamus  amphibious  is  one of the wild animals  living in the study  area. Even though, this  wild animal is recorded in the IUCN red list, the consultant report does  indicate this  fact. On the other hand, there are threatened and rare wild animals  indicated m the report Like pantherapardus  (Leopard), Panthera Leo (Lion), Geocchelonepardalts  (Leopard tortoise), Nilobcus, Corocodyiusnilobcus  (Nite Crocodile) and it does  not indicate any  mitigation mechanism to protect the wild animals
As  per the TOR, this  is  an EIA Baseline Study  phase. Therefore, the resource and the potential impacts  are discussed. However, detailed mitigation measures  will be presented in the Full ESIA report.
Q39
Annex2 (table) of page 132, There are list of plants  which are found m the direct impact area. But, the list does  not seem complete or exhaustive. For instance, there is  no a single climber life form speoe in the list Does  that mean there is no dimber m the area?
Yes,  the  list  is  not  complete  or  exhaustive.  Detailed  sampling  and  identification  of  plants  will  be earned out during full EIA
Plant status or vulnerability is not well described.
Q40
Once detailed sampling and identification of plants is done during full EIA, we will give information on spedes vulnerability using IUCN RED LIST OF ETHIOPIA
Biophysical environment
Q41
Regarding water quality  assessment, it is  not dearly  described the techniques used to collect the samples. For instance, how the number of samples  is statically  determined and other technical issues  like sampling penod (Dry/ramy season) and sample distnbution are not well addressed m the report The report should have been presented m such way  to illustrate the statistical representatives  of the samples, the rationale behind the distnbution of the samples and the method of analysis which insures the reliability of the output.
Agreed to discuss how the sampling sites are selected and the technique used to collect the sample in the final report
160627, DABUS, Reply to OBS at PRE-FEAS Report (Workshop 160615)
studio pietrangdi, romeDABUS hpp
WORKS! IOP 16 JUNE 2016
CONSUL TANTS RESPONSES TO CLIENT OBSERVATIONS
Page 6 of 9
__________________________________________ _________
___________
No.                              Comments by Client                                1
Answer by Consultant
U not cfew and                     described what standards are used to evaluate the parameters   of   water   quality   (Turbidity,   Dissolved   solid,   hardness
Q42      etc.).   Moreover.   chemical   parameters   like   BOD,   COD   are   not   addressed.   On the    other    hand,    biological    issues    related    to    water    quality    are    totally
overlooked.
I
Agreed to elaborate In the final report
Watershed Management
Q43
The watershed of Dabus  area is  studied as  unit. Beyond this, the consultant expected to approach tn depth from sub catchment point of view which seems forgotten.
According to CBPWDG (Community  Based Participatory  Watershed Development Guideline), the watersheds size ranges 6,000-20,000 ha Is used to Identify communities and ranges of community where watershed planning can start at this  level. However, before Implementation starts  an In depth study will be expected to do based on the Guideline.
Q44
As indicated in the report, estimation of soil loss is made by USLE. This does not consider losses from gullies and river banks which suppose to under estimate the prediction of the loss. Thus, it is recommended to incorporate other technique which helps to consider the lose caused by gullies and river
banks
RUSLE Is used for soil loss estimation. The main soil loss part and cause for Gully formation is the untreated or unmanaged part of big watershed. So that focusing on assessing soil loss from the first phase of erosion is essential at this stage. Therefore, that is why this assessment does not consider losses from gullies and river bank.
Q45
Instead  of  using  USLE,  it  would  have  been  better  to  use  the  Revised  universal sot/ lose Equation (RUSLE) which is the predecessor USLE.
RUSLE is used for soil loss estimation. Please see the collected data
Q46
Projecting future land use land cover changes  of a water shade area based on the previous  trends  has  an massive importance for a proper management of a watershed are which is  highly  associated to protect a dam from soil erosion induced sediment accumulation. One way  or the other, understanding of the rate of patterns  or change of the water shade area critically  important to determine life time of the dam. So that, the consultant should have made the land use land cover change of Dabus  watershed area in a manner to show the water shade area dynamics.
The current land use and land cover has already been done and shown In the document. However, assessing the dynamics or change of land use is not covered In the TOR and beyond the scope of this assignment.
Preliminary Impact assessment
_________________________
Q47
The consultant did not strictly  follow the Hydropower Production
Transportation And Distribution guideline (2004 EPA).
As per the TOR this is an E1A Baseline Study phase. Therefore, during the next phase the Full ES1A
will strictly address all the issues as per the EPA Guideline.
Q48      Air pollution (Degradation of air quality by dust and vehicles emissions)
The same as above
Q49
Sound pollution (Machines  like crashers, concrete mixers, dozers  etc
Increase in ambient noise near the substations)
/
The same as above
Q50                       Water pollution (Contamination of surface and underground water, hazardocIS                        The same as above
studio pietrangeli, rome
160627, DABUS, Reply to OBS at PRE-FEAS Report (Workshop 160615)DABUS hpp
WORKSHOP 16 JUNE 2016
CONSUL TANTS RESPONSES TO CLIENT OBSERVATIONS
Page 7 of 9
No
.                             Comments by Client
Answer by Consultant
materials.
Q51
Soil pollution (Erosion of riverbed downstream of the dam, Soil compaction. Reduction in soil fertility. Soil destabilisation as a result of excavation, Risk of soil contamination from substations) are not well assessed
Yes  there  will  be  soil  erosion  during  construction  and  this  impact  will  be  elaborated  and  mitigation measures identified and recommended in the Full ESIA.
Regardless  of the uncovered environmental issues, the potential negative impact that the consultant has  identified for the project will cause a tremendous  amount of destruction on the environment However, the consultant recommendation and conclusion which implies  no environmental grounds  not to precede the proposed DABUS HPP seems  a paradox  with regard to the loss  identified by  the consultant. Does  that mean all the environmental loss  is  nothing? Or did the consultant conduct the study  for sake of procedure7
Based on this  EIA Baseline Study, there will be both positive and unavoidable negative impacts associated with the implementation of this  Project. These potential beneficial as  well as  significant adverse impacts on the physical, biological and socio-economic environment are identified.
However, based on this baseline survey result, it Is concluded that there are no environmental grounds for not proceeding the Proposed Dabus Hydropower project to the next phase and therefore, it Is recommended to proceed with the Feasibility Phase of the Project. This 
| recommendation is consistent with EPA's EIA Guideline.
Q52
Attention should be given for this issue and the consultant is expected to propose the proper mitigation measure identified on prefeasibihty level to
I which the negative impacts suppose to be reduced to insignificant level according to Environmental policy of Ethiopia (1997), Proclamation on Environmental impact assessment (299/2002), Guideline for dams and 1
reservoirs (2004).
As  per  EPA's  EIA  Guideline  requirements,  a  full  Environmental  and  Social  Impact  Assessment  will 1   be   prepared   during   the   Feasibility   Study   and   all   relevant   matters   related   to   the   potential I environmental Impacts Identified In this Study will be fully addressed.
i The significant environmental impacts require avoidance, mitigation and/or compensation 1 measures. These measures will be identified and elaborated further in the next phase.
Other comment
Q53
Most of the tables, graphs, maps etc m the EIA report are not visible This document will be used for furthers studies and reference for the future.
Will be corrected in the final version of the EIA Baseline report.
Q54
The preliminary environmental impacts including forest resources in the project area should be studied and incorporated m the report. These may include impact on aquatic ecosystems, terrestrial ecosystems (i.e. flora and fauna), forest biodiversity, floods and droughts m downstream areas due to the reduction of forest coverage, land use and land cover patterns (such as agncu/ture, settlement, forest, vegetation, rocks etc).
1  Preliminary  environmental  impacts  including  forest  resources  in  the  project  area  was  described. 1   The  existences  of  unique  riparian  and  wetland  vegetation  were  discussed  and  the  recommendation 1 is to avoid any damage on these resources.
However, as a compromise, it was also recommended to asses areas in the Dabus drainage system as a substitute that is going to be lost by the dam.
1
25(5      !
/J
//
Hydropower projects are usually located in areas with forests. Large forest ireas have been cleared for use as dam construction and water reservoir. Therefore, the preliminary environmental mitigation measures and costs for litigating forest loss and other environmental issues are not bnefly reported
the pre-feasibility study of the Dabus Hydropower Project
1 During the next phase, the Full ESIA will identify and recommend mitigation and/or compensation 1 measures for forest losses and estimate all costs to implement these measures.
Q
56     4 cc
5-  per the  TOR of section  5.9, the  preliminary social mitigation measures and vnpensation costs (e.g. resettlement, support land and matena! losses, and |
The TOR requires an EIA Baseline Study for this phase and a detailed and Full ESIA in the next | phase. Therefore, during the next phase the Full ESIA will strictly address all the issues as per the
160627, DABUS, Reply to OBS at PRE-FEAS Report (Wortshop 160615)
studio pietrangeli, romeDABUS hop
WORKSHOP 16 JUNE 2016
A
P<i()O B of B
CONSULT
NTS R
ESPONSES to client observations
No.
Comments by Client
---------------------------
Answer by Consultant
are                    dfeOTfce*/
report.
EPA  Guideline.  Tills  will  Include  oil  environmental  mitigation  and  management,  buffer  zones  and watershed development, environmental monitoring, capacity building and training costs.
Q57
The patrnnory total project cost inducting environmental and soda/ costs is not nduded m the feasibility report. 77>e term environmental and soda/ costs is not included tn the pre-fessibikty report. The term environmental and scoal cost refers to fi 03003/ expenses that are being pad and/or should be pad and accounted for the total financial investments and operation of
hydropower plants.
The same as above
Sodo-Economic Environment
To address  regional soda/ csues  (page 66-71) like literacy  & education, ethnic  & religious  affiliation, sodo economic  profile CS4 2007 report is referred. This  holds  true for physical disability  rate and health infrastructure (page 90). iVas  it not possible to refer latest references  or to take projected figure instead of taking out-dated figures7
Q58
None out of the Ove Tables and Charts below that are based on CSA can be updated because of lack  of up to date and disaggregated data - disaggregated such as  urban/rural, male/female, reglon/zone/woreda, etc. Such comprehensive and disaggregated data can only  be obtained from the CSA and the next national population census  Is  yet to be conducted. In the absence of such data, another source is  administrative data complied by  the woredas. In addition to Its  limited reliability, very  often, administrative data is  either not comprehensive, or not disaggregated by gender and the rest of variables, or unavailable altogether.
Despite these limitations, efforts  will be made during the upcoming ESIA fieldwork  to gather, where possible, data and Information required updating those Tables  and Charts  (1 to 5 below) that are based on CSA's outdated data.
1.     Table 2.4: Literacy Rates of Population of the Project Area by Sex and Urban/Rural Residence (CSA 2007)
2.      Chart  2.1:  Proportions  of  Major  Ethnic  Groups  Living  In  the  Proposed  Dabus  Hydropower Project Areas (West Wollega Zone and Asossa Zone) (CSA: 2007
3.      Chart  2.4:  Proportions  of  Populations  by  Religious  Affiliations  in  Dabus  Hydropower  Project Areas (Based on CSA:2007)
4.     Chart 2.5: Economic Activity Rate of Population (Aged 10 Years and Above) in the Proposed Dabus Hydropower Project Area by Urban-Rural Residence and Sex (CSA: 2007)
5.     Chart  2.9:  Physical  Disability  Rates  (Per  10,000  People)  among  Populations  in  the  Proposed Project Area (Based on CSA: 2007)
On the other hand, Table 5.20 is  correct except that the year (2007) is  mistakenly  written in Ethiopian calendar instead of Gregorian. Therefore, the year 2007 will be converted Gregorian 2014/15.
Q59
As  per the TOR, the sooo-econom/c  aspectynsks  of the Dabus  Hydropower project are not dearly  stated in the pre-feasibility  report. Positive as  well as negative socio-economic  impacts  likely  to be accrued due to the project are to be listed. For instance, demographic profile, water use pattern like agricultural
Yes, due to field constraints the socio-economic aspects/risks of the Dabus Hydropower project are not dearly stated in this EIA Baseline Survey report.
160627, DABUS, Reply to OBS at PRE-FEAS Report (Worttshop 160615)
studio pietrangeli, romeDABUS hpp
WORKSHOP 16 JUNE 2016
CONSULTANTS RESPONSES TO CLIENT OBSERVATIONS
Page 9 of 9
No.
Comments by Client
Answer by Consultant
practices, water quality analysis, forest dependent economic activities, impact
on human health and economic status, incidence of water related diseases,
resettlement, rehabilitation and social response programme of the project, etc
_________
1
160627, DABUS, Reply to OBS at PRE-FEAS Report (Workshop 160615)
studio pietrangeli, romedabus hpps PRE-FEAS
HYDROLOGICAL Report
page i of I
INDEX
Paragraph Subject
Pa 9e
1      CONTENT1
2      INTRODUCTION3
3      AVAILABLE HYDROMETRIC and METEOROLOGICAL DATA4
4      ANALYSIS of AVAILABLE DATA5
4.1          HYDROMETRIC DATA5
4.2          METEOROLOGICAL DATA12
5      EXTENSION of the DAILY DISCHARGE TIME-SERIES at DABUS near ASSOSA17
5.1          SYNTHETIC STREAMFLOW GENERATION17
5.2          TRANSFER FUNCTION19
5.3          RAINFALL-RUONFF MODELLING28
5.4          CONCLUDING REMARKS34
6      DESIGN FLOWS35
6.1          NATURAL MONTHLY AVERAGE DISCHARGES at DABUS ALT 130 at DABUS ALT 10835
6.2          FLOW DURATION CURVES38
7      FLOOD FREQUENCY ANALYSIS
7.1          BASIC ASSUMPTIONS
7.2          MODEL INFERENCE........................................................................................................... ......"’.'*44
7.3          CONCLUSIVE REMARKSZZZL".49
8      REFERENCES
Ref. Drawings 
- 120 GEN D SP 001 Hydro and Meteo Stations
-     120 GEN D SP 002 Catchment Areas
120 HYDRSP001A (HYDROLOGICAL Report).docx 120 HYD R SP 001A
’
ftudlo pletrangell,'^’DABUS HPPs PRE-FEAS
HYDROLOGICAL Report
ABBREVIATIONS ACRONYMS
page 1 of 1
AFDC            Annual Flow Duration Curves ANN               Artificial Neural Network CRU              Climatic Research Unit
FDC               Flow Duration Curve
GEVD
Generalized Extreme Value Distribution
GIS
Geographic Information System
GPD
Generalized Pareto Distribution
MAE
Mean Absolute Error
MBE
Mass Balance Error
MSE
Mean Square Error
NSE
Nash-Sutcliffe Efficiency
PDF
Probability Density Function
UEA
University of East Anglia
USACE
U. S. Army Corps of Engineers
WMO             World Meteorological Organization
120 HYD R SP001A (HYDROLOGICAL Report).docx 120 HYD R SP 001A
studio pietrangell, romeDABUS HPPs PRE-FEAS
HYDROLOGICAL Report
1  CONTENT
page 1 of 51
The object of this report is to study the hydrological behaviour of the Dabus river basin subtended by the dam sites. In particular the hydrological elaborations which are reported hereafter deal with the following aspects:
o mean monthly runoff;
o flood runoff.
Determination of the mean runoff, which is a key issue in estimating available water resources and reservoir regulation analysis, was carried out following four different approaches:
-    synthetic streamflow generation on the basis of the observations directly collected at Dabus near Assosa;
-     transposition to Dabus near Assosa of the observations collected at Didessa near Arjo by means of a power-law transfer function;
-     transposition to Dabus near Didessa of the observations collected at Didessa near Arjo by means of
an artificial neural network transfer function, and
-     rainfall-runoff modelling on the basis of the data by the Climatic Research Unit (CRU) of the University of East Anglia (UEA).
Among  these  approaches,  the  one  that  was  preferred  for  defining  a  time  series  to  be  used  in  energy production evaluation is the ANN transfer function approach.
To evaluate the flood runoff a standard frequency  analysis  was  then carried out. For the sake of conservativeness, the time series recorded at the station of Didessa near Arjo was preferred. The analysis was carried out adopting the standard 'block maxima approach.
The  following  three  different  probability  distributions  were  fitted  to  the  single  sample  of  dimensionless maximum annual discharge:
•     Extreme Value Type I, or Gumbel, or EV-I distribution;
•     Generalised Extreme Value or GEV distribution;
•     Log-Normal distribution.
The parameters  of probability  distributions  were estimated by  means  of the Maximum Likelihood (ML) method, and the goodness-of-fit was  checked by  means  of both diagnostic-plots  and the Kolmogorov- Smirnov Test.
The report Is structured as follows:
_ CONTENT
This chapter shows the scope of the present study.
- INTRODUCTION
“^qHYD R 5P001A (HYDROLOGICAL Reportj.dooc 120 HYD R SP 001A
studio pietrangeh, romeCARD’S
HYDROLOGICAL Report
>?sc-z*?s the main characteristics of the Dabus River Basin.
M ALLASlE HYDROMETRIC and METEOROLOGICAL DATA
S-r“e-^es the hydrometric and meteorological data available for the catchment area. ANALYSIS Of AVAILABLE DATA
In ths chapter, a deep analysis of the hydrometric and meteorological data Is illustrated. EXTENSION of the DAILY TIME SERIES at DABUS near ASSOSA
Deas with the approaches applied to determine the monthly runoff. DESIGN FLOODS
Summarizes the mam results of the design floods.
page 2 of 51
- FLOOD FREQUENCY ANALYSIS
ElusTatss the approaches utilized to carried out the frequency flood analysis and the main results. REFERENCES
I/O HfD P SKiOlA (KfDPZcOGICAL P/^xt).(kxz
120 HYDRSP 001A
studio pletrangell, romeDABUS HPPs PRE-FEAS
HYDROLOGICAL Report
page 3 of 51
2  INTRODUCTION
The Dabus HPP project are located on Dabus river, a north-flowing tributary of the Abbay River In south western Ethiopia.
The Dabus  river drains  an area of about 14,400 square kilometres  and Its  course defines  part of the boundary between Benishangul-Gumuz and Oromla Regions. It originates In the high volcanic mountains that border the drainage area to the south, and flows generally northwards In a large and flat basin known as the Dabus swamp, then continuing northwards up to its conjunction with Blue Nile River.
According to the modified Kopen system of climate classification, the basin falls within Tropical Climate II, being characterized by a mean annual rainfall around 2000 mm with a mono-modal distribution along the year.
The  proposed  layout  envisages  a  cascade  with  two  power  plants,  being  the  two  dams  located  at  the following coordinates:
Table 2.1:                 Drained area
LOCATION                        AREA
SUE
WGS 84
UTM Projection 36 N
East         South
DABUS Alt. Ch.0+130               705605               1087644
2
(km )
10'009
DABUS Alt. Ch.0+108               708957               1103040              10'400
The catchment area and the location of the hydrometrical/meteorological stations are shown In the following drawings:
o 120GENDSP001 - Hydro and Meteo Stations o 120GENDSP002 - CATCHMENTS AREA.
120 HYD R SP001A (HYDROLOGICAL Report).doo< 120 HYD R SP 001A
studio pietrangell, romeHYDROLOGICAL Report
3         AVAILABLE HYDROMETRIC and METEOROLOGICAL DATA
page 4 of 51
4 soecSc                 to Addis Ababa (February 15*19, 2016) was carried out In order to collect the data for the %’crostudy.   In   particular,   the   official   available   hydrometric   (daily   discharges)   and   pluviometric   (daily "c“*c  heatt)  data  were  requested  and  collected  at  the  Hydrology  and  Water  Quality  Directorate  of  the E*rOpen M mstry of Water, Irrigation and Electricity, and at the Ethiopian National Meteorological Agency.
T^>e 3 1 and 3.2 report the list of the hydrometric  and meteorological stations  within and close to the catd".>~ent of interest for which data were made available. Figure 3.1 reports the location of these stations also reported in drawing 130GENDSP002 - CATCHMENTS AREA.
Hydrometric stations
Table 3.2:                  Meteorological stations
LAT
x(£«J
V (Nortf)
A
m
n*i
m
m
Km’
Meteorological
Station
LON
(‘El
LAT
(•N)
X(East)
m
Y(Nord)
m
M 90
9-87
70E3S2
109 ue
10067
A Ira
35.33
906
756125
1002306
?o
AsSOSa
3452
1007
666568
1113S38
MG
10.15
678946
U22447
161
Begl
34.32
920
645018
1017231
!S_D
9 63
756072
1065746
94
53
•4 <8
Dembidolo
34.80
8 52
698130
942247
M£3
9_97
730978
1102284
194
■a
Kamashe
35.45
10.02
768570
1108623
3522
970
743214
1073C37
562
Nedlo
35.50
950
774-130
1051120
35®
719560
1050764
1(4
Ba ba si
34.44
9.45
658092
1044930
950
8
Dengoro
35.73
920
8000C7
1018107
IS 38
957
761613
105M07
112
T> 5
Gul Iso
m*
s
35.52
9.20
776915
1017938
3642
8 GB
215720
960811
9961
s
Jarso
35 32
9.45
7S4744
1045451
3652
9 42
M625J
104M11
4674
■o
Mendl
35.10
9.79
730346
1082912
3605
903
■15421
1000000
3281
Seko Umbl
3498
8.72
717836
964467
Amba 10
34.60
9.75
675507
1078187
E
Amba 16
34.65
9.92
680900
1097016
_e
o
Flgakobra
35.40
9.05
763831
1001250
5€
Kebe
3497
898
716583
993222
2
Klltukara
35.22
9.72
74356$
1075252
■U
Mukeleml
35.53
933
777912
1032332
Werejlru
35 37
9 62
760108
1061298
figure 3 l
Location of the Hydrometric and Meteorological stations
1/0 Mb R SK61A (HYDROLOGICAL f'rpoayticc/ 120 HYD R SP 001A
studio pletrangell, romeDABUS HPPs PRE-FEAS
HYDROLOGICAL Report
4 ANALYSIS of AVAILABLE DATA
page 5 of 51
4.1 HYDROMETRIC DATA
Table 4.1.1 reports the annual efficiency of the considered hydrometric stations (Table 4.1.1, Figure 3.1). The efficiency  was  evaluated as  the ratio between the number of the valid daily  observations  and the number of days within each considered year.
Table 4.1.1:             Hydrometric stations efficiency
ANNUALEFFICIENCY (%)
DOOM
PARK)
4001
ANGIR#
NOTMNTE
4002
DARANA0
ABASIA
4005
DABUMJ
AiSOSA
5002
HOHA^
ISSOM
5003
WUJUR# ASOSA
5004
HAFA>
AJJOM
5005
MOM
MIN DI
5006
AXELTU*
NIDK)
SOO4
KOMS0
oon
5010
ARIA Km’
9981
4674
2881
10067
161
94
194
562
168
112
I960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
198S
1986
1997
1993
1999
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
20W
2005
2006
2007
2008
MEAN
39.6
100.0
79.7
100.0
84 2
100.0
71.8
70.4
65.9
100.0
63 0
100.0
100.0
91.5
915
41.4
25.7
54.3
356
77 0
100 0
76.7
100 0
100 0
84 7
100.0
100.0
84.1
79.5
88.2
73.2
42 2
100.0
100.0
100.0
940
69 7
99 7
100.0
100.0
100.0
1000
100.0
98.4
100 0
100.0
100.0
1000
100.0
85.3
-
-
91.8
99.5
926
100.0
956
100.0
100.0
100.0
90.4
937
100.0
80.8
92.9
90.1
99.5
910
79.5
98.6
93 7
77.5
89 6
99 2
82 2
-
-
-
562
•
88.8
100.0
91.5
92 4
94.0
100.0
91.8
95 9
72.1
849
13.2
29.0
44.7
430
49.6
329
501
97.0
76.8
93.2
76.4
96.2
98 6
98 4
1000
984
99.7
992
99.2
65.2
98 9
863
93 0
71.7
-
-
-
-
-
-
100.0
100.0
100.0
86.6
100.0
100.0
86.3
86J
100.0
1000
71.6
92.3
100 0
100.0
100.0
100.0
100.0
100 0
100.0
100.0
1000
100.0
100.0
100.0
67.1
40.3
18.3
100.0
26.9
53.7
100.0
100.0
74 0
100 0
100.0
86.6
89.3
816
100.0
100.0
1000
100 0
100 0
89B
-
-
-
•
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
•
-
-
-
-
-
-
-
-
-
-
-
-
-
-
76.2
100.0
100.0
1000
100.0
98.9
1000
57.0
78.1
100.0
66.0
•
-
1000
1000
1000
1000
1000
96.4
1000
100 0
92.1
100.0
937
82.5
•
-
93 1
-
48.4
89 0
80.3
67.1
76.5
1000
1000
100.0
1000
49 0
100.0
100.0
79 5
53.7
100.0
1000
100 0
100.0
1000
100.0
1000
100 0
100 0
100 0
99.7
1000
100.0
992
98 9
91.1
57.3
839
100 0
100.0
100 0
100 0
100.0
1000
74 3
1000
100 0
400
712
100 0
100 0
100 0
100 0
964
100.0
1000
99 5
100.0
97 8
981
24.9
•
89.7
-
95.1
100 0
997
100 0
997
1000
100 0
1000
1000
73 4
100 0
80.0
-
-
100 0
100 0
100 0
100.0
1000
100 0
100.0
1000
100.0
100 0
100.0
100 0
100 0
94.8
65.3
96 2
100 0
100 0
100 0
1000
100 0
100.0
100.0
880
1000
1000
100.0
1000
718
100 0
1000
100.0
100.0
100.0
1000
100.0
100 0
100.0
100 0
968
Among the hydrometric stations located within the Catchment of Dabus river, the discharge observations collected at ST 5002 (Dabus near Assosa) are clearly the most significant for the project. Unfortunately, the
120 HYD R SP001A (HYDROLOGICAL Report).docx             120 HYD R SP 001A
studio pietrangeli, rome
cn 
in 
O'
£HYDROLOGICAL Report
page 6 of 51
t~»e senes observed at this station extends over 17 years, from 1963 to 1979, with an efficiency ecus tr *1%  urtherrnore, only the time-series' track extending from 1965 to 1974 is characterised by an
c
dxexy c"eater than 90% (Figure 4.1.1).
~abe *1.2 -eoorts the monthly average discharge observed at station Dabus near Assosa (ID 5002) within
the
recorded penod, while Figure 4.1.2 shows the multi-annual monthly average discharge. It Is to
be -ccced that the gaps in the time series within the period 1965-1974 occur mainly in the dry season. The ave-aoe mu-ti-ennual discharge (not ngorously calculable due to gaps) is close to 160.0 m3/s.
Figure F. 1.1:
Efficiency of station Dabus near Assosa (ID 5002)
“jde Z2.*
Monthly average discharge observ
ed at station Dabus near Assosa (ID 5002) within the available recorded period
-AN
MONTHLY AVERAGE DISCHARGE (m’/s)
FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
f 1M3
LW
ZMS
.^Jk
iwi
1-
-
-
-
-
-               227.66           426 25           498.99           37683            18630            98.13
•
•
-
£4 4X             3€ £4              2424              2129
-
19026           314.35           476.22           466.21           261.00           112.91
£7 38
-
ia£i 2436 1531 39.64 116.28 238 13 362.05 541.77 384 39 24635 135.69
4001 27.71 1774 21.68 6002 132.95 18822 571.22 574 24 267.13 122.63
£4 73              4122              2338              1445                                    75 71           153.43           415.67            49730           406.76           180.41           91.66
£536              3457              2743              19.10             2156             80.22
-               445.64           545.87            37669            180.68           96.72
1
IfX               52.17              3136              21.75             14.42              1527             58.09           140.23           261.20           480 62           428 76           22235           104.02 6057              3342              1S22                                     1801             47.70            143.40           322.43            46044            376.59           203.13            10507
1972 57 2£ 32 70 1714 1942 6936 16676 280.12 425.05 3S7.17 166.11 73.32
1        ’.571      1      <: n               2123
-
-
•
-                274.15           469.31           408.39           202.79            84.25
:>7<              <623              2555              1646
-                100.76          178 34           314.79           47802            415.63           193.17           80.74
<152
-
-
-
-
:vh
-
13.65 2258 61.28
-
-
-
-
-
;»7t        1     4CM              2121              14.18              1325              2534
7642 126.66 206 31 344.63 343.07
-
:>7>                                    30 73             1771              12.13              2909              74.69           174.46
-
-
-
VMM 5247 32 31 2133 1569 2361 74.64 170.20 317.60 482.45 409 56 21006 100.56
V*J 44 71 C 22 27.71 21 19 39 64 116.28 23823 445.64 571J2 574 24 267 13 135.69
v*a 40 96 2121 14.18 12.13 15 17 47.70 126.66 188.22 344.63 343.07 166.11 73.32
cut
I 11               6 76                459                300               7.24              19.66             3637             82.81             59.49             6163              34 65             18.48
120 HTD P. SPC01A (ffifMCCOGICAL P/-p<xt).d«/
120 HYD R SP 001A
studio pletrangell, romedabus hpps PRE-FEAS
HYDROLOGICAL Report
page 7 of 51
Figure 4.1.2:           Multi-annual monthly average discharge observed at station Dabus near Assosa (ID 5002) within the available recorded period
As far as the other stations located within the Dabus catchment area are concerned (ID 50XX), the
respective catchment areas are largely smaller with respect to that of Station Dabus at Assosa (ID 5002),
being of the order of magnitude of the hundreds of square kilometres. This circumstance does not support
the use of the data recorded at these station for a regionalization.
It is furthermore to be noticed that, although these stations present an efficiency greater or equal to 90% (Table 4.1.1), the quality of the data Is questionable. By way of example, Figure 4.1.3 provides the whole time series of dally discharge recorded at the station of Hoha near Assosa (ID 5003), Figure 4.1.4 provides that recorded at the station of Hafa near Assosa (ID 5005) and Figure 4,1.5 provides that recorded at the station of Komis near Gori (ID 5010). The single mass curve of the collected time series Is also given In the figures.
ume (yoars)
Figure 4.1.3:            Daily discharge at Hoha near assosa (ID 5003) and annual single mass curve.
120 HYD R SP001A (HYDROLOGICAL Report),docx             120 HYDRSP 001A
studio pletrangeli, romeHYDROLOGICAL Report
page 8 of 51
OK
i
I
iI
-gire 4. l. 4.
Daily discharge at Ha fa near assosa (ID 5005) and annual single mass curve.
Figure 4.1.5:
Daily discharge at Komis near Gori (ID 5010) and annua! single mass curve.
In tne first case, all the data recorded between 1985 and 1995 show a clear inconsistency with that recorded b^ore and afterwards. Evidence of such inconsistency  Is  also provided by  the single mass  curve which career/ deviates  from a linear relationship. In the second case, evident Inconsistencies  are present In the data observed within 2007 and 2008, while a marked difference is evident between the data observed in the perods 1980-1990 and 1990-2005 as also evidenced by the deviation in the single mass curve. A marked
between the data observed in the periods 1985-1993 and 1994-2008 is also evident in the third case (tme senes of daily discharges collected at the station of Komis near Gori), being the difference the opposite of that shown by the observations at the station of Hafa near Assosa.
O' parocular interest appear the observations  recorded along Didessa river and Its  tributaries. It is  to be ntxced that Didessa is  the closest to Dabus  of the tributaries  to the Blue (Abbay) Nile. The fact that the catchment area of the Didessa river is hydrologically similar to that of the Dabus river in terms of both land 'ea*ures ard meteorological disturbances (storms) makes the discharge observations collected at Its stations qute interesting for the scope of this study.
Three stations in particular were considered within the Didessa river catchment:
•     Odessa near Arjo (ID 4001);
•    Anger near Nekemnte (ID 4002); and
•     Dabana near Abasina (ID 4003).
The catchment area drained at the station of Didessa near Arjo (ID4001)
Is equal to 9981 km ,
2
Le. quite
similar to th»at drained by the station of Dabus near Assosa (10067 km ).
Th<- co1  erred time series extend from 1960 up to 2008 with an average annual efficiency of 85.3% (Figure
2
4 1 6), while a slightly decreasing trend In the average discharge Is apparent from the time series (Figure 4.17).
1/0 HfD R SP001A OffDPOtrZJGAl
drzz              120 HYDRSP001A
studio pletrangell, romeDABUS HPPs PRE-FEAS
HYDROLOGICAL Report
page 9 of 51
Figure 4.1.7:             Daily discharge at Didessa near Arjo (ID 4001) and annual single mass curve.
Fig. 4.1.8 compares the concurrent dally discharges collected at the stations of Didessa near Arjo (ID 4001) and of Dabus near Assosa (ID 5002) in the years 1965-1968. The assumed hydrological similarity appears to be confirmed by the observations. Two important features are evident by this comparison:
(a) The overall wet season hydrograph observed at Dabus near Assosa appears generally shifted In time with  respect  to  that  observed  at  Didessa  near  Arjo,  therefore  showing  that  the  meteorological disturbances  that  cross  the  region  tend  to  affect  the  Didessa  catchment  more  than  the  Dabus catchment.
T^HYD R SP001A (HYDROLOGICAL Report).dooc
120 HYDRSP 001A
studio pletrangeli, romeHYDROLOGICAL Report
page 10 of 51
time (month)
1200----------------------------- --------- T-------- T------- 1------- 1-------- 1------- 1-------- 1         T 45 1000
time (month)
F^jre 4 ! 8 - Comparison between the concurrent discharges collected at the station of Didessa near Arjo
(ID 4001 - black line) and Dabus near Assosa (ID 5002 - red line) in the years 1965-1968.
(b) The occurrence of a succession of floods during the wet season Is evident In the Didessa record while the record observed at the considered station along Dabus appears much smoother, therefore showing that the presence of flooding-land areas tends to moderate the flood flows upstream the stzhen of Dabus near Assosa.
As far as the stations Anger near Nekemnte (ID 4002) and Dabana near Abasina (ID 4003) are concerned, Figure 4.1.9 present the efficiency and the time series of collected dally discharges. It Is to be noticed that tte tendency of the recorded data Is consistent with that of the data collected at Didessa station near Anger. Fuftherrriofe, the data extend to different periods of time (1995-2008 in the case of station 4002 while 1963-
1/0 kyc, p SP001A (KfDPOlOGICAl P^xt).doa
120 HYD R SP 001A
studio pietrangell, romeDABUS HPPs PRE-FEAS
HYDROLOGICAL Report
page 11 of 51
1994 In that of station 4005). Ttierefore, as will be shown later on In this report, the data collected at these last two stations were used to Infill the gaps In the time series collected at station ID 4001.
Anger near Nekemnte (ID 4002)
Dabana near Abasina (ID 4005)
Figure 4.1.9:            Efficiency of stations Anger near Nekemnte (ID 4002) and Dabana near Abasina (ID 4005), and corresponding time series of collected daily discharges and single mass curves.
120 HYD R SP001A (HYDROLOGICAL Report).docx                    120 HYD R SP 001A
studio pletrangell, romeHYDROLOGICAL Report
page 12 of 51
4,2 METEOROLOGICAL DATA
Tatfe 4 2 1 "eoorts the annual efficiency of the considered pluviometric stations, being the efficiency 
as the raho between the number of the valid daily observations to the number of days within the corsjc^ec veer Among the pluviometric stations, the ones with the longer available period of observations e-e *me 5rst dass station of Assosa (1961-2015, efficiency 89.8%) and the third class station of Mendi
1955-2015. efficiency 79.3%). Plots of the efficiency variation at these stations are given in Figure 4.2.1.
l/ti Kfb k JXA\K (HYDROLOGICAL PAprxt) docx
r
120 HYDRSP001A
studio pletrangell, romeDABUS HPPs
PRE-FEAS
HYDROLOGICAL Report
13 of 51
page
lime (years)
time (years)
Figure 4.2.1:           Efficiency of stations Assosa (upper plot) and Mendt (lower plot).
ANNUAL CUMULATE RAINFALL HE1CTH (mm)
Table 4 2 2 reports the annual cumulative precipitation at available pluvlometric stations characterised by an annual efficiency > 80%.
T^SlsPOOlA (HYDROLOGICAL Report).doa
120 HYDRSP 001A
studio pietrangell, romeHYDROLOGICAL Report
-'x <«wn kxzMcns, a 'ready shown in Figure 3.1, are also reported in Figure 4.2.2. In particular, the
page 14 of 51
were crowed depending on their location: a group for North-locations; a group for Centre-East a croup South-East locations and a group for South-West locations. Figure 4.2.2 also reports 
of interest of the cumulative-monthly precipitation dataset reconstructed by the Climatic =J25e—umt (CRU) of the University of East Anglia (UEA) on a grid characterised by a resolution equal to C s' Ths reconstruction was carried out using complex spatialization techniques which take into account
dmaatoocal vanability ([12], —r AWA.cnj.uea.acuk/).
[13],     [14]) and calibradon with ground-based obsen/ations
ETHIOPIAN NATIONAL METtOROLOGlCAl AGENCY
■nd
QJMAT1C RESEARCH UNIT
EAST ANGUA UNIVERSITY
□ 1st & 2 nd dais stations
□ 3rd dsss stations
04th dass stations
OCRUdata
Figure 4.2.2:            Location of grouped p/uviometric stations and of grid point ofCRU-UEA dataset.
figure 4.2-3 shows  the non-dimensional annual cumulative precipitation observed at the different piuvorr-etric stations within and dose to the catchment of interest in the period 1980-2010, Ze. In the period dur.ng z/tuch a wider range of observations is available (Table 4.2.1 and 4.2.2). Each annual cumulative precp/^tiori was  made non-dimensional by  dividing it by  the corresponding multi-annual average value in (Zder to rriake the comparison easier.
The vanabilrty  in the chosen period (1980-2010) of the non-dimensional annual cumulative precipitation srr/z/s that the stations located in the northern portion of area of Interest experienced higher precipitation arcxjrxJ 1999 (1998-2001). The inspection of the daily cumulative time series revealed that particularly high tot/ rainfall was observed at the stations of Assosa, Mendl and Klltukara during the wet season of 1998,
1999 and 2900,
T ’’"' feature is not observed either In other portions of the basin or In other period of time reasonably pjggeshrq that the meteorological disturbances that cause the wet season precipitations are usually more unformiy spread over the area of Interest and that only m the period around 1999 these disturbances crrxted the area with a more Intense centre located over Its northern portion.
1/0 HfD P SK6IA (KrtfcOlOGfCAL P/^xt).drzy  120 HYDRSP001A
studio pletrangell, romeDABUS HPPs
PRE-FEAS
HYDROLOGICAL Report
page
15 orsi
Figure 4.2,4 provides a plot of the non-dimensional annual cumulative precipitation resulting from CRU-UEA dataset In the grid points of interest for the period 1980-2010. Also in this case, each annual cumulative precipitation was made non-dimensional by dividing It by the corresponding multi-annual average value.
Figure 4.2.3:            Non-dimensional annual cumulative precipitation observed at the different pluviometric stations
120 HYD R SP001A (HYDROLOGICAL Report).docx 120 HYD R SP 001A
studio pietrangell, rome1- the case of the CRU data, the variability in the chosen period (1980-2010) of the non-dimensional annual CLmuUCve precipitation appears  less  evident, being the general pattern more uniform all over the basin. Hcwe/er, an increment in the non-dimensional annual cumulative at the NE CRU point (LAT 9.75-LON35.25) is jresert during wet season 1985,1991, 1999-2000 and 2010.
Rgure 4.2.5 shov/s the average annual cumulatives of precipitation resulting from CRU-UEA dataset. These values car. be located at the centre of considered CRU grid points. The whole period (1961-2008) covered by the oczer/aton of daily discharge at the station of Didessa near Arjo was considered In this case. It is to be r<Crtxed that - coherently  with the daily  discharge observations  (Figure 4.1.7) - the average annual curr.ulatz/e precipitation shows a slightly decreasing tendency (trend).
120 MYD R SK61A (HYDROLOGICAL f'xyxX) 6</x
120 HYDRSP001A
studio pletrangell, romeDABUS HPPs PRE-FEAS
HYDROLOGICAL Report
page  Y1 of 51
5 EXTENSION of the DAILY DISCHARGE TIME-SERIES at DABUS near ASSOSA
It was shown that the time series of daily discharge actually observed at the station of Dabus near Assosa (ID5002) extend over 17 years with an efficiency of 71%, and that only in the decade between 1965 and 1974 did the efficiency achieve a value greater than 90%.
Three different approaches were adopted In order to extend this time series:
-     synthetic streamflow generation;
-     use of Dldessa near Arjo observation by means of a transfer function; and
-     rainfall-runoff modelling.
5.1   SYNTHETIC STREAMFLOW GENERATION
The time series of daily discharge observed at Dabus near Assosa (ID 5002) in the period between Jan 1, 1965 and Dec 31, 1974 was previously filled for gaps by means of linear Interpolation. As already shown, the gaps present In this track of the whole time series are concentrated in the dry season making such a simple procedure of infilling feasible.
The resulting ten years of data were used for the application of the Quimpo and Yevjevlch model ([1], [2], [3]), l.e. a 2nd order autoregressive scheme, and for the synthetic generation of a 48-year dally discharge (which can be considered the 48 years between 1961 and 2008).
Figure 5.L1: Synthetic streamflow generation approach: Comparison between observed and synthetic data at Dabus near Assosa.
^ThYOR SP001A (HYDROLOGICAL Report).doo< 120 HYD R SP 001A
studio pletrangeli, romeHYDROLOGICAL Report
page 18 of 51
XXwo    the    observed    and    synthetic    daily    discharge    resuftmg    from    the    application    of    the
X’ «
-ei.
5.1.1 provides 43 years of month,, average discharges respiting
X  Finally  figure  5.1.2  and  S.1.3  she  the  annual  average  discharges  and  the  mulh-annual -X m-ede »«harges result.™, from the appiicaben of the autoregressive model.
» s a « noticed that the assumptions «1). (21. [3)> »"
’ s“>chKK mOde'15 teSed
X ** 'ev^ch (Oahcaii, me respect of some spastic ^pefties of the original time series,dea y 
00 « atd" the synthetic tepmOuchon of features (eg a bend) that may have appeared after 1979 (the « o( the actual penod of observation at the station of Dabus near Assosa. Therefore such an
Xcn L « considered valid on., if stabonadty of me actua, (not direct!, cbse™d) time sehes can be assumed.
Synthebc stream flow generation approach: 48 years of monthly average discharges at Dabus near Assosa.
T&eS-Ll:
MONTHLY AVERA -------- m                MM APR mai____________________ jyi
NOV
168.07
DEC
85 IB
53 56
59 02
4845
5169
3U2
1196
)■£»             27 75
33 42
3047
2532
19 92
9 44
23.15
1455
20 36
58-70            3597             19 41            1355
1145
2611
30.17
2148
19.84
GDEISCMAnut
JUL
193 15
AUG
287.58
SEP
501 64
OCT
35895
45 KJ            2794             2147             2131             1951
5511
59 86
69 54
55 01
54 19
50 44
4920
5689
44 15
3253
20 39
16 28
3461
37J3             2532
21-84            3452
4352
30 09
29 74
17 79
20 96
1521
2636
27 M
32-13            2175             1843             3321
2632
3394
3703
16 08
20 74
2421
10 45
13 70
17.M
738
1493
26 64
12 C4            16 45            1603             23 28
5*71              40 04            25 3 9           2241             28 63
55 79
44 89
63 16
55 04
6734
60 11
62 00
6224
47 19
6131
51 15
72 49
54 21
4441
72 97
73 92
35 !■
3037
37 48
36 66
19 15
1331
24 95
27.02
32 4}            19 85
3457
33 97
36 70
2746
3119
21 AS
44 04
2955
3540
44 50
39 15
2283
2238
26 13
21.09
20 87
17.20
25.21
20 72
25 24
31 79
26 78
10 90
581
2343
1638
10.11
17.77
1948
2146
2144
1516
9 62
22.12
1423
1980
1885
2548
<1 J) 3175 22 74 1809
<0*4 3447 24 34 1607
54M 15 74 26 00 22 70
23 74
87 81
65 88
81 89
11292
74 68
33.87
10015
80.19
7137
79.90
89.12
43 58
48.79
77.11
7256
92.62
52-51
4740
111.86
80.21
32.87
59.29
7532
7882
74.67
111.09
42 10
5634
9793
55.60
7973
89.16
7623
71.91
79 06
15 81
76 21
37 74
1846
1657
1242
34.67
2123
5.79
12-73
3036
28 63
22 11
2634
10 24
2L34
20.26
18.20
28 93
4550
2801
20.92
33.16
3243
*/>92            3160             25 33            19 64           3244             8963
53 13
•2 U
25 99
40 54
17 19
18 24
1333
9 09
12.65
10 80
70 21
46 09
C4 12            15 19           2544             1949            33 92            116 17
UU
*-i n
3164
23 60
14 95
y»4i              1931             14 96
Cl M 39 14
54 U 2907
27 40
20 8)
1793
3293
36 69
22 to             17 96            2559
44 jt
44 n
V. M
Uto
1144
16 54
781
24 >2            14 70            7 20
35 92
2132
25 99
9 07
12.07
48.47
6637
11101
11041
92.54
26.75
72.27
10455
4 J 44           2122             1961            32 77           70 20
190.25 353 28 472.99 42607 18233 7199 196 06 314.11 441.64 356.40 210.03 90.91 244 42 362.84 51136 402 69 20336 9700 11052 8339 345.94 31404 215 17 97.25 125 00 267.13 45238 43745 197 06 92.79 219 IB 381.92 598.42 523 96 216.17 106 55 18224 33521 532 94 539.00 24339 101-86
183 40          31533          499.76          435 7S           21871           102.28
193 76          37733          492 26          431.73           20864           100.75
209 23         373.79          533 26          438 74           21469            8738
146 90         241.62          37739          257.05          166.01           94 58 180 11         353.84          542.14          502.80          237.74           97.44 15637          308 84          46312          319 94          158 78            7795 16239          296 38         477.67          396 89          200 06           93 89 188 56         290 02          451.94           37937           180.05           93 83 164.85         299 19          507 02          32931           17630            74.05 108.78         202.74          40838          382 49           21431           12231 24)92          39930           55033          456 98          189 81           93 71 18654         379.60         576.47          493.49          228.09          106.97 10249          247 96          440.63           34139           195.92           95.67 148-25         21797          409 35          40139          216 88           10881 ieo8o          32530          449 97          368 76          21054           106 37
18166         365 68          569 60          414 50          205.06           95.21
166 8$         304.61          459 95          4C2.89          208 06           99 34
22742          343 68          523.24          394 05          20735           100 73
123 06         23739          431.55          367 IB          184 44           11379
13338         262.18          435 18          370 86          168 78           92.08
239 74         386 30          57135           40735          198.93           7823
140 70         27236           46209          424 08          251.06          140 79
192 80         370.05           53289          404 36          171 78          106.22
21734         454 20          546 18          47733          233 49          115 97
16080          328 90          521 21         440.01          208 70           11733
250 60         366 22          484 70         421 96          206 00           89 98
151.18         290 23          483.99         368 88          228 21          126.57
173 97         304 78          459 59         344 16          172 23           94.63
20880         360.80          42737          338 76          252 26           10559
208 80         439 89          64331          58152          305.59          138 IS
8)39           144 16         389 65         352.77          186 21          105 05
231.24 399 68 585 85 497.98 207.29 101.81
14649          26130          45841          375 26          14737           85.47
202 69         322.67          51431          41436          223 60          112.86
208 75          39934          500.42         409 42          222 53           9465
15131          24931          460 35         426 75          255 41          127.37 129 95         264 92         457 15         396 73          219 62           11076
16199          301 89         510.29         427 06          193 53           8856
166 35         250 54          45941          440 86          198 24           10250
144 94         293.21         477 28         424 04          228 14           12808
15064
161.40
15358
173 78
115.55
145 15
192 11
182 84
168 06
169.20
175.49
119.85
17461
143.73
151-82
154 3 2
145.09
132.88
185 80
184 0 2
132.72
14250
157 14
173.81
154 64
171 92
134 71
142 09
176 77
157.80
171.41
195 17
167.93
170 68
158 78
15198
162.89
209.15
12234
193 35
140 80
171.04
177.77
15938
144 80
155 17
159 73
16257
►
UH
MX
UM             MOI              72 57           17 12            74 63           KM i. K             utA_______ >1 79           75 99            4847            116.17 « JI /2C4                          1151             5 81              5 79            26.75
175.09         312 80         487 26         409 78          207.85          101.84
160.40
250 60          45430          64331          58132          305 59          140 79
8)09            8339           345 94         257.05          147.37           71 99
'Ji 1           44.
« J4             4 14____ 4 90                  0 71             22.70           40 17          72 02            6039             6133             }g,42                    15 14
120 HYD ft SPQ01A (TDftOLOGICAL P/?port).dOCX
120 HYDRSP 001A
studio pletrangell, romeDABUS HPPs
PRE-FEAS
HYDROLOGICAL Report
19 of 51
page
Figure 5.1.2:          Synthetic streamflow generation approach: average annual discharges at station at Dabus near Assosa.
Figure 5.1.3:          Synthetic stream flow generation approach: multi-annual monthly average discharges at at Dabus near Assosa.
5.2 TRANSFER FUNCTION
As already observed (par. 4.1)z the time series collected at the Dldessa station near Arjo (ID 4001) along the river Dldessa extends from 1960 up to 2008 with an average annual efficiency of 85.3% (Figure 4.1.7),
It's been furthermore already noticed that the catchment area drained by station ID 4001 Is of 9981 km2 Ze. almost equal to the area drained by the station Dabus near Assosa ID 5002 (10139 km2), being the two catchments hydrologlcally similar In terms of both land features and meteorological disturbances (storms).
Being the two catchments hydrologlcally similar makes it possible to transfer the data recorded at the Didessa near Arjo (ID 4001) to the station Dabus near Assosa (ID 5002), provided that a suitable transfer function could be obtained on the basis of the analysis of a set of concurrent observations collected at the two stations.
Two kinds of transfer function were applied In the present case:
- a Power-law transfer function; and
120 HVD R SP001A (HYDROLOGICAL Report).docx                  120 HYDRSP001A
studio pietrangell, romeHYDROLOGICAL Report a-               Neural Network (ANN) transfer function.
page 20 of 51
oe^tcr of  a fc^sfer function providing the best fitting of  the concurrent  daily  discharges observed at staxns of Ddessa near Arjo and Dabus near Assosa underlies both the approaches.
of caps in the original time senes of daily discharges collected at Didessa near Arjo (ID 4001) between 1960 and 2008 (efficiency equal to 85.3%) has made it necessary to carry out the Infilling of the ^e-senes.  For  short  gaps  a  simple  linear  infilling  was  adopted,  while  for  longer  gaps  an  ANN  transfer Vrtior from stations Anger near Nekemnte (ID 4002) and Dabana near Abasina (ID 4005) was preferred. Z-^acti as already observed (par. 4.1), the tendency of the data recorded at stations ID 4002 and ID 4005 Is czrsstert with that of the data collected at Didessa near Anger. Furthermore, the data extend to different
□e“ocs of time (1995-2008 in the case of station 4002 while 1963-1994 in that of station 4005). Therefore, tne  collected  at  these  two  stations  have  made  it  possible  to  infill  the  whole  of  the  gaps  present  in  the time senes collected at station ID 4001.
F-jrmemore, it has already been observed that the daily discharges collected at the station of Didessa near Ano (ID 4001) appear invariably shifted in time with respect to those observed at the station of Dabus near Assosa (ID 5002). In particular (Figure 4.1.8), the overall wet season hydrograph at ID 4001 tends to be anoapated with respect to that at ID 5002. Furthermore, while a succession of floods  is  evident in the Odessa record during the wet season, the record observed at Dabus  appears  much more smooth. As Greedy  observed, this  feature provide evidence of flood-flows  moderation upstream the station of Dabus rear Assosa.
A shifting and a filtering of the infilled time series observed at the station of Didessa near Arjo (ID 4001) was rerefore performed In order to obtain the more convenient time-series to be transferred to the location of the station of Dabus near Assosa. In particular, a 15-day shift was applied to the original series, together wtn a high-pass digital filter charactensed by a cutting frequency equal to 1.16‘iO'6 Hz (/.e. 10 days).
Smilarty to Figure 4.1.8, Figure 5.2.1 compares the daily discharges resulting from the shifted and filtered time-senes  collected at the station of Didessa near Arjo (ID 4001) with the concurrent daily  discharges observed at the station of Dabus near Assosa (ID 5002) in the years 1965-1968. As evident from the figure, Te shafted arid filtered time-series (blue line) collected at the station 4001 superimposes more consistently than the original one (black line) to that collected at station 5002 (red line).
120 HYD P. SRZdA (HTDROtOGICAL Report) 6vy
120 HYDRSP001A
studio pletrangeli, romedaily discharge (m 3/s)
daily discharge (m 3/s)
daily discharge (m 3/s)
daily discharge (m 3/s)
DABUS HPPs PRE-FEAS
HYDROLOGICAL Report
page  21 of 51
12001-------- r--------t-------- 1------- r 1000
time (month)
Figure 5.2.1:             Comparison between the concurrent discharges collected at the station of Dtdessa near Arjo (ID 4001- original: black line; filtered and shifted: blue line) and Dabus near Assosa (ID 5002 - red line) In the years 1965-1963.
SP001A (HYDROLOGICAL Report).docx
120 HYDR5P 001A
studio pietrangeli, romeHYDROLOGICAL Report
page 22 of 51
FUNCTION
Oh r*e ^esrs of the anaJyss of the available set of concurrent observations resulting from the shifted and r-^e se~es co ected at the station of Didessa near Arjo and the original time series collected at the
scato- of Detvs near Assosa the following power-law transfer function was obtained
Q7»s_st' - 4 6O6'Qrz€SS*(t)0 732
“he cwt on the left of Fig. 5.2.2 provides the comparison between concurrent observations collected at the hare stations that have made possible to obtained this power law. The plot on the right shows the results 
car te achieved by transferring the data recorded at Didessa using this power law. As shown by the C cjre tne concurrent observations tend to organize along the X=Y line, although an important degree of
around this line is still present
3 IX ICE IOC *00 SCO 600 700 830 900 1000 1100 1200
Didetu Q Arlo
0 100 200 300 400 500 600 700 800 900 100011001200
Dabus (® Assosa from Power Law Iran func (m*/s)
fog-j? 5-2-2. Actual concurrent observations collected at the stabons of Didessa near Arjo and Dabus near Assosa (left).
Concurrent observabons after the applicabon of the obtained Power Law transfer function to the
data recorded at Didessa near Arjo (right).
Table 5 2.1 provides 48 years of monthly average discharges resulting from transferring by means of the power-law function the observations (filtered and shifted) collected at the station of Didessa near Arjo to the
•ocaticr. of tte station of Dabus near Assosa. Figure 5.2.3 and 5.2.4 show the annual average discharges ar«d tte mufti-annual monthly average discharges resulting from the application of the power law transfer function.
Tr^ a/erage multiannual discharge results  in 148.6 m3/s, therefore lower than that obtained with the ^ezousfy  dustrated stochastic  approach. The main reason Is  the slightly  negative trend present in the cxigir^ai oezzr/ations at the station of Didessa (Figure 5.1.7). As already observed this trend could not be present m the results of the stochastic approach which Is based only on the data recorded at station Dabus near Assosa between 1965 and 1974,
120 HYD P SKZHA (KrOPfATjCfOL PApcxt) Oct,
120 HYDRSP001A
studio pletrangell, romeDABUS HPPs PRE-FEAS
HYDROLOGICAL Report
page  23 of si
Table 5.2.1:            Power-taw transfer function: 48 years of monthly average discharges at Dabus near Assosa.
MONTHLY AVERAGE DISCHARGE
JAN             FEB            MAR           APR            MAY            JUN             JUL            AUG            SEP            OCT            NOV            DCC
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1991
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
200S
2006
2007
2008
6736          64.92           5655           94 88           75 20          167.04         360.80         42622          42105         360 64         169 49         129 60
71-91           52 96          48 70          47.92           6634           15154          25430          J1936         368.72         277.63         10132          70X6
57.14           4194           37.69           60.63          11448         181.24         265.42         39259          34405          19185         135 95         12S66
71.60           5942           44 67           55 84           68.28           15330         335.98         38057         368.72         37930          14169           8176
6946           4851           36 15           65 48           46.75          124.45         299 85         341 13         27338          31069         183 04         12186
80 21           7952           71 28          82 12           7283          212.97         303.28         346.41         366.03         174 66         131.17          89 04
67.63           52.62          41.26           4L10           61.27          147 86         27536         378 OS         373J2          35] 19         215 72         142 65
94 05           7053            4176           5028           59.26          167.10         29832          39931          310.76         271 46         11532          74 97
67.07           6738           7683           58 19           7938          191.03         32124         401.22         326.49         168 21          9139            6352
54.11           57.00           5754           4461           58.14          167.45         317.79         411.75         409 08         10456         133 29          74 18
5630           38.19          28 63          31.63           62 28          155.28         294.20         36181          32956          28150          176 69         98.48
6795           52.11          39.85           53.02           6930           10439         290.73         372.22         28231          141.70         106 56         67.85
4630           3121           18.21           2151           97.05          159.81         27801          366 14         40438         243.15         120 92          76.07 5535           37.17          33 43           26.25           82.97          162 98         261.04         376.11         390 09         23807         107 19         65 25 45 10           4454          29 8 7          50.02           6332           13031         240 66         25332         353 61         29739          12235          90 11 6483           5B.72           4521           51.76           71.43          137 61         22094          31232          334 67         20032          14098          83 80 57.11          49.04           38 13           3058           44 48          159 82         23925          31739          39952          29L49          18159         110.11 69.18          43 00          58 33           3150           85.21          156.74         267 65         31057          34449         352.70         160 70         10534 84 04          64 14          50 23          42 99           6130           106 98         234.97         29894          259X1          16087          93.65           61.73 38 05          36.19          27.44          75 88          135.49         230.05         31662          381.48         380.14         189 56          9943            9237
85.38           6498           6152           55.81           69.21            7859           186.23         32234         36L61          22104          10755          68.17
73.01           5348           48.45           39 84           6L19           15557          27234          32175         276.03         280 15         132.67          92 85
61.40          54.48          52.97           4483           64 02          121.75         240.24         355.48         35934         357.67         16499          10685
57.40 37.75 3041 27.41 46.49 132.82 27256 266 79 235 30 10834 7595 59 04 3058 2088 12.84 30.15 72.74 126.23 21254 31895 306.47 15750 8637 6234 3748 28.72 35.13 33.27 3133 99.71 215 05 226.41 28297 14391 7606 49 82 32.02 2150 3031 37.28 5533 141.22 240.64 27237 271.42 196 40 12439 7732
$9 28          49.88          41.74           24.42           51.58          169 58         24051         364 01         325.05         26135         135.46         86 12
63.72          46.17           4051           5331           46.56          101.61         16S37         226.48         35343          21931         100.68         106 86
68 10           5299           53.01           73.88           5992           123.83          18559          34100         330 21         16081           7739            5331
32 10          24 60          30 09           2850           26.84           85 93          237.91         35235          25432         162.03          75 60           57.21 45.27          46 04           3337           40.23           70.70          126 28          19235          29832          24548         282 70         125.76          8134 60.82           513$           45.02          69.93           9752           176.94         24953          35937         262.70         20119          123.41          75 13 58.09           4232            3536           36.16           7259           145 70         23622          36844          30730         120 95          79 69           56 23 4091           32.93           3353           38.77           57.17           86.07          148.66         24433          22755          12144           7433           95.04 81.77           4237           47.03           51.29          112.03         197.16         27593          30L61         254 67         177.58         132.02          8952 5980           4133           34 70          65.81          112.08         204.42          24482          30512          25893         262.42         239.12         152 66
111.72 8647 88 72 7639 9797 154.08 248.20 363 71 31652 34553 17602 8132 38 97 1866 732 2.70 71.20 177.16 228.51 27737 257.71 33150 13199 71.77 4951 4288 60.02 67 73 103 65 162.02 216 48 270.00 271.18 245 06 156.78 11357 83.40 77.04 7633 73.65 107 29 202.61 28051 33945 331.07 267 24 146.13 10234
9588           75.42           7008           76.63           63.61          130.60         217 89         249 89         25133         141.74         10553          87 71
70.78          58.96           70.29          82.28           64.71           10453          221.95         26393          28032          16992         101.66          81.75
65.04          61.00          50 72          S0.08           7555           14L62          233.47         258 67         26496         230.43         117.65          9137
74.62          57.72          60 31          55.83           8657           149 48         20853         305.41         33502         198.99         127.29          9331
74 48           6332            5956           62 81           8059           15258          264 61         404 76         34135          21738         152 22         11446
9144           7995           66 95           7781           10553          172.42         300 04         32434         330 00         217.00         11034           9160
83 18          66.92           48 98           7541          128 97         218.71          23733         298.88         25195         149.19         163 87         108 99
199 47
15256
16239
178 43
160 06
157.46
179DO
16239
15932
174 1J
159 54
13733
155 23
153 03
14346
143 95
15992
165 44
12657
16691
14030
150 61
16534
112-51
119.78
' 104 99
135.11
150,75
127.04
13167
11400
13235
147 76
12992
100 10
14690
165 09
17833
134 61
146 57
17434
13053
13092
136 96
146 09
16$ 68
163.96
152 69
MEAN
64 09          50.73           4577           5050            7431           14979          25035         326.05         31409          22938          12734          87 41
143.59
MAX
111.72         86.47           88 72           82.28          135 49         230.05          33598         41175         409 08         379 30         239.12         152.66
MIN
3897           1866            7 82             2.70           57.17            8607           148 66        244 83         22755         120 95          7433            5633
ST. DEV
18 27          16.17           17 05           1876           24 04            3554            4152            5017            5035            7168            363C           22.71
~ 250
M
>
E 7 200 p
ro
a iso
2 ioo
c
<0
0>
« 50
Qj
1 1 1—1—
1  48.59 m3/  s
*               Illi
p
p
p
__
r■
S333
29£S
3
333
c* cn
1 eo> feM p
c
w i/>
2 2 3    O' a a   a
<o <ns
CA
K         3  •/*
M   DC
g
0*
2 § O i I i 1sm 3■»•8••.8rM§ r8. §
3£
Figure 5.2.3:             Power-law transfer function: average annual discharges at station at Dabus near Assosa.
120 HYD R SP001A (HYDROLOGICAL Report).doo<                    120 HYD R SP 001A
studio pietrangeli, rome
2001
2002HYDROLOGICAL Report
page 24 of 51
S.2.4:
Power-law transfer function: multi-annual monthly average discharges at at Dabus near Assosa.
-=~CIAL NEURAL NETWORK TRANSFER FUNCTION
•.eurai computing (Bishop, 1995; Lippmann, 1987) has emerged as a practical technology and is now widely applied in many engineering fields, including, In particular, the application area of hydrology (Zhang et al., 1996).
Figure 5.2.5:              Feed-forward Artificial Neural Network - I05H05001.
In order to define a function capable of transferring the data observed at Dldessa near Arjo to the location of the stzton of Dabus near Assosa the feed-forward network shown in Fig. 5.2.5 was adopted. The two adap-ve-weght layer network Is characterized by 5 input units plus bias, 5 hidden units plus bias and one output unit (I5H5O1). The network's inputs consist of the daily discharges observed at station ID 4001 at time t-2, t-1, t, t+l and t+2, while the network's output consists of the dally discharge observed at station ID 9Zi2 at time t The specific discharges (q=Q/A) were opportunely pre-processed so as to re-scale them I "early by in the range (0; 1],
fur-xon represented by the network diagram (Fig. 5.2.5) can be expressed as
120 HfD k SP001A (KfbPOJjGfCAL Report).docx
120 HYDRSP 001A
studio pietrangell, romeDABUS HPPs PRE-FEAS
HYDROLOGICAL Report
page
25 of si
where   =-2,-1,   0,+l,+2   (for   i=l,..,5),   wff   and   wty   are   the   adaptive   weights   connecting   input   units and bias to the hidden units, and h^2) and those connecting hidden units and bias to the output unit.
Moreover, gO and ^() represent the hidden and output unit activation functions. In particular, the logistic-sigmoid activation functions  characterize both hidden and output unit. It is  to be noticed that the structure of the algorithm network  outlined is  the result of Investigating the effect of different numbers  of both hidden layers and units, and of different activation functions.
Unlike the coefficients  of the power-law transfer function that are derived once and for all from the interpolation of available concurrent data, the adaptive weights  of the ANN function result from the network's supervised learning (Ze. phase of calibration).
In particular, being available a time series  of actual concurrent observations  collected at the station ID 4001 (Didessa near Arjo) and ID 5002 (Dabus near Assosa) makes it possible to present the network with an input array  [qo,40oi] of observations  collected at the station ID 4001 (Didessa near Arjo) and a corresponding vector {qo,5oo2} of concurrent observations collected at the station ID 5002 (Dabus near Assosa), Ze. with a training set. The concurrent data observed during 1963, 1967 and in the period 1970-1979 was  used In the learning phase, while the concurrent data observed in the period 1965-1966 and 1968-69 was  used for the testing phase (Ze. phase of validation).
The array  of adaptive weights  [w] result from minimizing the error function chosen to express  the difference between the calculated outputs  {qp,soo2} and the actual observations  {qo,soo2}. In particular, the network presented here uses:
a)   the mean square error (MSE) as the error function (Bishop, 1995);
b)   the standard back-propagation technique to calculate the MSE derivatives with respect to weights 
(Rumelhart et al., 1986); and
c)   the Levenberg-Marquardt optimization scheme for iteratively adjusting the network weights at the
end of
d)   each epoch by means of the calculated derivatives (Levenberg, 1944; Marquardt, 1963; Bishop,
1995).
Fig. 5.2.6 compares the actually observed hydrograph (dally discharges) at the station of Dabus near Assosa (ID 5002) with that predicted by  the ANN transfer function during the ANN testing phase (1965-1969). The predicted hydrographs  follows  with sufficient approximation the observed ones, therefore the corresponding array  of adaptive weights  [w] was  used for transferring the overall time series  (filtered and shifted) observed at station of Didessa near Arjo to the location of station Dabus near Assosa.
120 HYD R SP001A (HYDROLOGICAL Report).docx
120 HYD R SP 001A
studio pietrangeli, romeHYDROLOGICAL Report
c jrrx*~xrc the on the left of Fig. 5.2.7 provides the comparison between concurrent observations 
page 26 of 51
at the two stations, being the time series of observations collected on Dldessa filtered and shifted as oesr'tec n par. 5.2.
-s= pct on the ncht shows the result achieved by transferring the data recorded at Didessa using the ANN hjncbon. As shown by the figure the concurrent observations tend to organize along the X=Y line
bees’ than n the case of transfer by power-law, although a degree of spread around this line Is still present.
52 6:
Comparison between the actually observed hydrograph at the station of Dabus near Assosa (ID 5002)
with that predicted by the ANN transfer function during the ANN testing phase (1965-1969).
fwe 522
Actual concurrent observations collected at the stations of Didessa near Arjo and Dabus near Assosa (left).
Ccmrrtr.t obcer/abons after (he application of the ANN Law transfer function to the data recorded at Didessa near Arjo (right).
k>                              Table 5.2.2 provides 48 years of monthly average discharges resulting from transferring by means  of  the  considered  ANN  function  the  observations  (filtered  and  shifted)  collected  at  the  station  of Didessa near Arjo to the location of the station of Dabus near Assosa. Figure 5.2.8 and 5.2.9 show the annual average discharges and the multi-annual monthly average discharges resulting from the application of the power lav/ transfer function.
120 HYD R SR001A (HYDROLOGICAL Report).dOCZ
120 HYDRSP001A
studio pletrangell, romeDABUS HPPs PRE-FEAS
HYDROLOGICAL Report
page
27 of si
Table 5.2.2:               ANN transfer function: 48 years of monthly average discharges at Dabus near Assosa.
MONTHLY AVERAGE DISCHARGE (m'/l)
JAN              FEB             MAR             APR              MAY              JUN               JUL             AUG              SEP              OCT             NOV              DEC
1961
1962
196)
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
197$
1976
1977
1978
1979
1980
1981
1982
1983
1984
198$
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
200)
2004
200S
2006
2007
2008
47.12            36 11            3L27             4735              51.91 ’         83.28            266.58          491.02          646 17          564.05           24037           167 59 63.75            34.44             2737            27.27             3236             9031            186 04           35131           40) 82          474 24           13354            5159
3362             2635             2535             25 99            60 17            88 47            227 66          426.25          498.99          3763 3           18630           98 13
S8.83            34.85            28.50            27.29            34.14             9526            217.46           48838           467.17          530 75          294 00            7294
58 44            36.84            24.14            21.19             1464             47.47            19016           31435           47632           466 21          261.00           11291
5738            38 63           24 3 6            1531             3964            116.28          238.13          362.05          541.77           38439           24695           135 69
43 19            40 01            27.71            17 74            21.68             6002            132.95           18822           57132           574.24          267.13          122.63
6473            41 28             2338             14 45            28.70            75.71            153.43          415.67           49730           406.76          180 41           91.66
55.96            3437            27.63            19.10            21 56            80 22            250.06          445.64           54597           376 69          180 68           96 72
52 17            3136            21.75            14 42            15.17            58.09            140.23           26120           480.62          428.76           22235           104.02
6037            33 62            19 22            21.61            18.66            47 70            143.40          322.43          460 44           37659           20J 13          106 0 7
57.26            32 70            17.14            17.14            19.42             6936            166.76          280.12           42555           357 17          166 11            7332
41.13            2123             18 68            19 65            35.95            114.97           185.05          274.15           46931           408 39          202.79           84 25
4623             2535             1686            20.97            27.35            100.76           17834           31479           478.02          415.63          193 3 7          60 .74
4332             3030            24.49            24 62            29.27            60.16            159.64          237 70           32053           455.12          205.67            8338
4222            33 41            2631             13.65             2258             61.28            158.47          27235          40935           329.98          157 76           87 12
41.14            27.90            26 09            23.96            20.02            73.04            126.66           20631           344 63          343 07          242.83           13254
4096            21.21            14.18            13.25             2534             8188            195.90          326.72          398.07          474 34           29557           11831 5939            30 78            17.71            1233             29 09            74.69            174.46           28226           33102           220 24          118.29           41.12 2736            23.98            21.97            2739             93.97            16363           30332           434.45          515.95          387 72           87.86             5935 6032            41 79            33.20            33.74             3332             42.45             99.58            26959           429.12          383.83           153 38           4939 3837             3539            29.81            2558             28.68             6537            213.97           35837           37137           30231           213 80           86 90 4239             3139            27.21            30.05            27.82            63 30            159.21          367.80           44741           49236           26664           136.71 4933            26.86            2236            21.19             2234             56 74            19439           31337           261.03          173.23           48.69            37 60 26.20            2L12             1947            20.6S             3008             63.01            17034           274.14           40431           26834            8559             41.47
2729            21.96            2236            23 68            22.60            31.97            170.14           21937           25232           237.06           6985             33 44
24 41            2134            21.21            22.23            24.65            71 29            191.16           25732           305 94           22856           17132            6236
3529            30.04            27 28            2135             2139             10134           207.23           35977           42633           347.82          215.68           7146
4308            29 99            2454             2850            26 11            40.16            135.96           20137          306.73          387.09          142.42           61 08
6474             3034            29.53            39 81            33.91             6034            160.01          249.73           48035           290.88           76 00             3433
2484            21.09            2139             2232             21.55             2732             15830           32952           411.19           22520           94.19            3564
2731             2735            23.72            23.14            32.17             7733            147.03          245.16           32939           28231           20733           64 91
3831            29.63            2734             29.26            51.27           109 31           212.97          350.18          405.19          230.27          162.24           5832
35 14            28.11            23.92            23.28            30.36            64 93            200.76          334.64          470.74           21591           61.17            37.61
26.84            23 31            22 04            2351            30.08            33 78            11063           185 73           24032           183.98           62.70             4437
8285            28 03            23 82            31.25            42.88           150.05           234.70           36633           25030           254.75           14593            7734
43 44            2831            22 89            26.92            55.80            16167           221.18          312.96          333.17          225 06          303 68          210.64 12224           6433            52.98            50.85            5331            101 66          191.41           36733           43990           358 78          323.75           13225 26 68            22.08            1930            18.74            22.80            140.30           194 40          271.84          288.04          320.64          24659           62.43 32.82            24.59            31 62            31.83            52.21            131.06           19758           220.72          308.82          262.07           21550          124 10 7034            47.77            4055             4809             46.17           151.07          228.38          391.16          43230          282.44          229.93           99 79 67.60            5182            40 06            43 42            36 68            64 00            174.87           22157           25756           20550           10099           59.6J 4837             3538            33.99            40.76            50.16            42.25            166.86          228 47          292.65          257.40           10035           58 56 4002             3757            29.20            28 09            33.17             8138            176.47          225.77          301.19          25237           159.85           7634 47.49            3766             28.16            3603            45 83            77.46            187.12          267.08          360.00          33757           156 64           86 41
49.20            38 22            3108             3562            37.84             8733            194 93          40836          524 42          323 80          202.08          118 08
6932            51.20            4255            42.06            54.19            115.42           24186           364.44          402.63          34457           137 41           69 86
5130            45 65            30.44            3408            82.07            158.79          220 64          29135          338 72          208 79          169.70           11161
222.74
156 45
17289
195 80
168 63
183 43
172 23
J6637
177 39
15255
15132
14017
156 30
15820
13959
D451
1)402
167 71
115 93
17895
13563
14767
174 46
10233
11881
94 35
116 79
15537
118 92
129.16
116 09
124.13
142 07
127.21
82.26
14071
16239
138 25
13635
136 08
17237
11031
11235
120.10
138 95
17036
16138
14525
MEAN
48 22            32.62            2631             2625            34 89            83 84            185 06          30764           39809          333.43          178 08           82 86
146.40
MAX
12234           6453             52 98            50 65             9337            163 63           30352           48838           57122          574.24          323.75          210 64
MIN
26.68            22 08             1930            18.74            22.80            33 78            110.63          185.73          24032           183.98           61.17            37 61
ST. DIV
1737             9.17              7 19              917             16.21            36 04            39.09            7331            89 3 7            97.41             7037             35 79
POHYD R SP001A (HYDROLOGICAL Report), docx
120 HYD R SP 001A
studio pietrangell, romeHYDROLOGICAL Report
page 28 of 51
Figure 5-2.9:
ANN transfer function: mufo-annual monthly average discharges at at Dabus near Assosa.
T-e average multiannual discharge results in 146.4 m3/s, therefore almost equal to that obtained using the
□ewer law transfer function. Being based on the actual data recorded at the station of Didessa near Arjo, a so me ANN transferred time series show a slightly negative trend.
5.3  RAINFALL-RUONFF MODELLING
A complete rainfall-runoff modelling was carried out using the available reconstructed rainfall data by CRU- UEA (par. 3.2) and the available infilled data observed at the station of Dabus near Assosa (par. 4.1 and 5 1). Tne well-known HEC-HMS software ([15], [16]) developed by  the U. S. Army  Corps  of Engineers <USACE) was used in order to carry out this modelling activity, being the geomorphological parameters of me catchment at the station of Dabus near Assosa derived from a specific digital terrain model obtained through CIS tool. As usual the model definition has resulted from a calibration and a validation phase.
MODS-CALIBRATION
Smilarty to tte network's learning phase, the objective of the model calibration is to match simulated runoff vdixnes, runoff peaks and timing of hydrographs with the observed ones, by optimizing the values of the rvzei parameters.
HEC-HMS foresees  a calibration tool which looks  for the optimal parameter values. In other words, the ca’utxaoon phase foresees the running of the model using a selected number of historical events for which concurrent stream flows  and precipitations  are available, since the alm of the calibration runs  is  that of dining - by means of an iterative adjustment - the optimal values of the model parameters (Table 5.3.1) capable of guaranteeing the stated best match between the characteristics of the simulated and observed hydrographs.
120 HfD R SWjIA (KfDQOLOGIC/4 P/^),docz
120 HYDRSP 001A
studio pletrangeli, romeDABUS HPPs PRE-FEAS
HYDROLOGICAL Report
Table 5.3.1: List of Calibration parameters
page 29 of 51
parameter 1 - LOSS
constant rate (mm/h)
parameter 2 - LOSS
initial deficit (mm)
parameter 3 - LOSS
maximum defidt (-)
parameter 4 - TRANSFORM
lag time (min)
parameter 5 - CANOPY
Initial storage (%)
parameter 6 - CANOPY
Max storage (mm)
parameter 7 - BASEFLOW
Initial discharge (m3/s)
parameter 8 - BASEFLOW
recession constant (-)
parameter 9 - BASEFLOW
Threshold discharge (m3/s)
In the present case, the calibration phase was carried out using the concurrent data of daily precipitation (obtained disaggregating the original monthly CRU data) and flow discharge within the three years between 1965-1967 and evaluating the calibration process performance by:
(a) visual inspection and comparison of calculated and observed hydrograph and total run-off volume;
(b)  Nash-Sutcliffe efficiency coefficient NSE           'TSQqbs   "  QsimY
.Sfetw ’QobsI ,
(c) Mass balance error coefficient MBE = 1 - 0^-—• and
Qobs
(d) Mean absolute error coefficient mae = - f z - Qan ,1.
n /«/
Fig. 5.3.1 shows the results of the calibration phase in terms of comparison between the observed and simulated hydrographs  and between concurrent observed and simulated discharges. The simulated hydrographs appear to follow the observed one quite closely. Therefore, although concurrent observed and simulated discharges still show a tendency to spread around the X=Y line, the calibration results were considered satisfactory and the obtained model parametrization assumed for the validation phase.
Figure S3.1:            Rainfall-Runoff Model: Comparison between observed and simulated discharge - calibration period (1965-1967).
■^HlYDR SP001A (HYDROLOGICAL Report) docx
120 HYD R SP 001A
studio pietrangeli, rome□AfiUSMPTK
HYDROLOGICAL Report
page 30 of 51
~2b 3 3 2 3“>c 5-3.3 summarize trie peak flow and trie flow volume In the calibration phase as well as the ues obte~>ed for foe coefficients NSE, MBE and MAE.
Table 53.2:               Simulated vs observed Flow volume and peak Flows
Simulated
Observed
Difference
Percent Different
Volume (mm)                     2049.5
1660.1
389
23.45%
Peak flow (m3's)                    731.1
764.2
-33
-4.3%
Table 533:               Values of the NSE, MBE and MAE coefficients for calibration period 1965-1967
NSE
MBE
MAE
Penod 1965-1967                   0.71
0.85
61 m3
validation
The set of parameters  obtained during the calibration phase were used in the validation phase. Two G^erent time windows  characterised by  the availability  of concurrent data of rainfall and runoff were considered:
•    a time window spanning the year 1969; and
- a time window spanning the three years 1972 and 1974.
Figure 5.3.2 provides trie results of the validation phase in terms of comparison between the observed and simulated hydrographs  and between concurrent observed and simulated discharges. The simulated
h mcraphs appear
r
to provide a good approximation of the observed ones. Therefore, the validation
results 
were considered satisfactory and the parametrization obtained during the calibration phase assumed also for foe application.
Tab. 5.3.4 and 5.3.5 summarize the peak flow and the flow volume in the calibration phase as well as the /a.jes obtained for the coefficients NSE, MBE and MAE.
Table 5.3.4:              Simulated vs observed Flow volume and peak Flows for the year 1969
Simulated
Observed
Difference
Percent Different
Penod 1969
Volume (mm)
521
556
-36
-6%
Peak flow (m3/s)
429
578
-149
-26%
Year 19721974
Volume (mm)
1578
1432
146
10%
Peak Bow (m3/s)
536
557
-21
-4%
Table 5.3.5:               Values of the NSE and MBE coefficients for the year 1969
NSE
MBE
MAE
Year 1969
0.926
0.963
35 rtf
Year 1972-1974
0.81
0.90
4707/5
Similarly, as in the case of the application of the preceding approaches, Table 5.3.6 provides 48 years of monthly average discharges resulting from the application of the calibrated and validated rainfall-runoff
120 HYD R SP001A (KfDROCOGICAL P/wt) docx
120 HYDRSP 001A
studio pletrangell, romeDABUS HPPs PRE-FEAS
HYDROLOGICAL Report
page 31 of 51
model. In particular, the model was applied using the reconstructed rainfall data obtained by the Climate Research Unit (CRU) of the University of East Anglia (ref. par. 4.2).
0 10 20 E
a.
50 60
o
o
o
o
1200
1100
1000
1200
1100
1000
800
E
700
<      800
E
7 700
1
&
3
I 600
- S0C J
o     400
o 400
o 100 200 300 400 SOO 600 700 800 900 100011001200
Dibui ff ajjom tlmulaied by HEC-HMS model
0      100 200 300 400 500 600 700 800 900 100011001200
Debut ff Astou simulated by HEC-HMS model (m‘/i)
Figure 5.3.2:           Rainfall-Runoff Model: Comparison between observed and simulated discharge - validation periods (1969 and 1972-
■^YdTsPOOIA (HYDROLOGICAL Report).docx 120 HYD R SP 001A
studio pietrangell, romeAvti*£* annual
CASUS
HYDROLOGICAL Report
page 32 of 51
5_' £
Rj^^Runoff Modelling:
nearAssosa.
MONTHLY AVtMCC CISCMAWGl n-'/tl
*»
"T»             M*«
*r«
JUL
491-32
sir              OCT            NOV             OK
15 M             • «*              469             9 45
:•> rs            *C«’            44 59           2790                                                   224 55
Mac            22 79           24 99 813?            18*2           10.79           1927
•X“l             M41            22 87           1889 -w             • IX             4*3 »           285J           28 SO
159 76
82.78
170 25
298.03
_-w-         , 13*1            SO 4 J          35 69           2*40                                                   389 60
AUG
75960
417.48
24956
364 17
272*8
429.30
47334
.384            5*24            K 62            1989 44 1«           MOO           3174
1 M51 JI 42 20 61 16 72
39722
268 63
38871
1 l’”l             UM             1X78           19 01           1114                                                   15531 rr-T            <854            28*7            1456            1536
W35           2 7 63           15*5            1438
374            4-455           7934            ISA!            1532
4329            2530            1724
1 ’V5’
CT*            6*28             35 «            22*3            2109
7*73            *3*3            2X07           1241
172 02
382.70
34625
45989
12793
106 SO
28459
m
33
:wb
3C
1*0
3*4
385
4X30            223*             3*2             *5.19
M 47 >684 20*2 14 63 17051
1134 17 47 1457 1991
r*>4 1453 958 926
4933 1236 26 96 19 91
136.67
162.18
19421
33452
132*1
MU 3627 2025 1201 15429
13 M uxi 1022 624 54242
1 IM*
:w
30
30
:rr
391
1892
:m
17*4
:y?5
1996
ITT
:m
1999
:=
sc:
sci
XC4
scs
-Z4
SC7
xci
*724 2538 1438 1357
ISfcO 45 49 2639 1*74
1257 1749 1021 1033
3«              15.77           1125             929
147.43
37655
119.93
19123
22234
25636
5131            2*21            1920            1627                                                    14139         284 67
41.71 28 73 1534 1158
41X8 2336 15*5 20*6
11189 59*2 32.71 20 83
14173
-’ll!
IS 18            10*3             980
121.06
403.02
6X53             3*30            20 37            1564
121.78
4321 2132 17*1 18 42 8325 17128
28*1 1722 1720 2397 15132 161 27
23*5             1221              8*1               979
62.70
15 45            9.73              7*5              961                                  1837
74*7 4084 23*8 19*0
84 3 J 4452 2420 use
GC21 J170 1167 1*51
5131 3031 17.49 1055
8737
103 13
30227
J 7 27            22*5            13 99            1076                                                    20105 50 71            2837            15 95           10.48
2027            1131              917             1081 42.73            2323            1335              730
4IM             24 48           15 97            1451
24552
30337
17984
9X22
160 70
160.61
24108
25935
227.21 175.04
200 01
132*1
146.67
143*9
216.93
146*7
208*4
127.18
14069
94*5
9022
13300
103*8
107*2
199*4
156 41
148*8
8131
113.11
12352
120*7
216*2
71*0
163 40
146.77
11.26
7917
4034
12166
136*2
192.15
16668
74*4
126.77
5129
122*9
110*0
11808
129 69
43 53            2237            1215            13-89
7702 3         684.30          46434          306*5
461.13 305*5 18458 107.95
19250         148.20          95.23           59.81
57130          38860          228 15          12856
409 03         434.20          267*4          150.93
61924          611 40         378 71         213 97
61957          60830          368.07         206 94
587*8          404 38         234 72         130.74
42490          347.73         205 13         111.97
489*2          34S.7S          217*5           11909
349 61         255.82         155.92          89.94
350.05         258.20         157.13          8977
406*1          314,90          18863          103.66
534.25          43239          255 72         142.70
316.92         257.20          U1.00          109 72
44930          365.15          218*8          129.19
34997          230.40         137.05          77.23
403.88 304.79 199.03 116.17
17536          15030           9831            57.88
239 65         153.02           91*1            50.77
385*3          296.59          171*5           93*3
91.15          151.13         193.99          12030
28256          183.28          112*2           63.76
44458          256 88          14125          80 81
36230          480.71         290 34          16137
298.90          187*0           11156          61 35
221*7 144.81 8778 4B BO
260.83 22221 167 74 96.04
329.20 24834 14950 8S54
35958           245*4           142*8            79*5
594 OS 637.67 379*6 211*0
157.27 107*2 70*9 44.21
627 09         385.02          218*2           118*3
35754          22125          13127           75.76
209.43 1393 4 84 64 49.02
130 26          87.29           53.73            3309
6153            49.11            3837            2535
396 14         383 89          251*1          140.7S
304 20          45855           282*6          157.20
451.06 328 14 195.12 11155
426.28         284.02          17039           96.69
18855          12835           87*2             6016
32332          275 14          16601           9234
10384           80 05            54*5            34 41
39032           247*6           14592           81.16 249.62         205.97          131*8           76.23 30033          207.74          12753           74.20 241.15          IS 131           8959            51.24
143*4
IZLA3            UM             3126             19*2            16 78
MAS •.£**•. 90 97 4959 45.19
6211
105*1
338.5 7
542.42
351-04         278.96          17258           9831
15*5
973
785
730
13 70
23 05
26*7
13B54
crars
2161
15 20
■ 33
7.16
11387
13487
5,33
Rainfa
ll-Runoff
Modelling
: average annual disc
harges at station a
t Dabus near Assosa.
PO KfD R SR001A (HYDROLOGICAL P^xtj.docx             120 HYDRSP 001A
studio pietrangell, romeDABUS HPPs PRE-FEAS
HYDROLOGICAL Report
33 of 51 page
Time (months)
Figure 5.3.4: Rainfall-Runoff Modelling: multi-annual monthly average discharges at at Dabus near Assosa.
3
Figure  5.3.3  and  5.3.4  show  the  annual  average  discharges  and  the  multi-annual  monthly  average discharges resulting from the application of the power law transfer function.
The average multiannual discharge results In 143.4 m /s, therefore slightly lower than that obtained using the power law and the ANN transfer function. Being based on the precipitation data reconstructed by the Climate Research Unit (CRU) of the University of East Anglia (UEA), the time series resulting from the application of the rainfall-runoff model also shows a slightly negative trend.
T^HYOR SP001A (HYDROLOGICAL Report).docx 120 HYD R SP 001A
studio pietrangell, romeHYDROLOGICAL Report
page 34 of 51
5.4 CONCLUDING REMARKS
o «- . «
wtous wragraphs, d(fere t approate were adopted |n M
„
oteewd m senes of daily discharge al the station of Oahus near Assosa (10S002) » a period of
t'ne scanning the years from 1961 to 2008, Ze.:
Synthetic streamflow generation;
Use of Didessa near Arjo observation by means of a Transfer factions; and
Rainfall-runoff modelling.
Tab. 5.4.1 recalls the multi-annual average discharge at the station of Dabus near Assosa resulting from the application of each applied approach.
Table 5.4. J: Values of the NSE and MBE coefficients for the year 1969
Q^(m3/s)
Synthetic stream flow generation
160.5
Power law transfer function
applied to Didessa near Arjo observations
148.6
ANN transfer function
applied to Didessa near Arjo observations
146.4
Rainfall-runoff modelling based on
CRU-UEA reconstructed rainfall data
143.4
As  already  observed, the extended time series  resulting from the first approach should not be used for ericray  production evaluation as  it is  based on data recorded in the decade 1965-1974 and could not account for possible trends in the following years.
The multi-annual average discharge resulting from the other approaches varies from a minimum of 143.4 m3/s (rainfall-runoff modelling) and a maximum of 148.6 m3/s (power law transfer function).
eeweei these approaches, the one that was preferred for defining an extended time series to be used in energy production evaluation is the ANN transfer function approach (Tab. 5.2.2 Fig. 5.2.8 and 5.2.9). In fact, this approach is based on actual observed discharges (observation at Didessa near Arjo) and provides the best results in terms of comparison between predicted (or simulated) and actually observed discharges.
studio pietrangeli, rome
120 HYdTswIa (HW-OtDGICAt Pz^xtJxJocz
120 HYD R SP 001ADABUS HPPs PRE-FEAS
HYDROLOGICAL Report
page 35 of 51
6 DESIGN FLOWS
As  Illustrated In the previous  section, the chosen dally  discharge time-series  (1961-2008) at the station of Dabus near Assosa (Tab. 5.2,2, Fig. 5.2.8 and 5.2.9) Is the result of the extension - by means of an Artificial Neural Network  transfer function applied to the filtered and shifted time series  observed at the station of Dldessa near Arjo - of the actually observed data (1963-1979).
As far as the time-series at the section of the proposed dams (Tab. 2.1: Dabus Alt. 130, A= 10009 km2 and Dabus Alt. 108, A= 10400 km2) are concerned, these were obtained simply multiplying the values of the time series obtained at the station of Dabus near Assosa for a coefficient corresponding to the ration between the drained areas, the two coefficients being 0.994 for Dabus Alt. 130 and 1.039 for Dabus Alt. 108.
6.1 NATURAL MONTHLY AVERAGE DISCHARGES at DABUS ALT 130 at DABUS
ALT 108
Tab. 6.1.1 provides 48 years of monthly average discharges at Dabus Alt.130, while Figure 6.1.1 and 6.1.2 show the annual average discharges and the multi-annual monthly average discharges.
Tab. 6.1.2 provides 48 years of monthly average discharges at Dabus Alt.130, while Figure 6.1.3 and 6.1.4 show the annual average discharges and the multi-annual monthly average discharges.
120 hyd R
5P001A (HYDROLOGICAL Report).docx
120 HYD R SP 001A
studio pietrangeli, romeAverage annual discharge (m’/s)
DASUSHPPs
HYDROLOGICAL Report
page 36 of 51
Table 6.1.1:              4S years of monthly average discharges at Dabus Alt. 130.
MONTHLY AVERAGE DISCHARGE
LAN             EEB            MAR            APR             MAY            JUN             JUL            AUG             SEP            OCT            NOV            DEC
1*61
i%:
1943
1944
19*5
19*6
1967
1961
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
7007
2008
46 JIS           35 90            3109            47.08           51.61            8240           265 04         488.19         642 45          56030          23899         166.62
6338            3424           27 41           27.11           32.17            9049           18497          349*8          401.49          47151          132.77          51 29
33 *3            26.79           25.20            2544            5942            8796           22635         423.79         496.12         374.66          18523           9756
5849            34 65            2134            27.13           33 94           94 71           21621          48557          464.48         S27.69         29231          72 52
S8JD            3643               2400              21.07            1456           47.20          189.06         31254          473.48         46352          25950          11226
5765            38 41           24 22            1522            39.41           11561         236.76         359.96         538,65          38218          24553          13491
4294            3978            2755            17.64            2156           59.67          132.18         187.14         56733          57093          26559         121.92
6436               4144               23*4               1437            2853            7527           15255          41328          494.43         404.42          17937          91.13
55 64            3437            27 47           18 99           21.44            7976           24862         443.07         542.73          37452          179 64          96.16
5187            3108            2162             1434              15.06           57.76          139.42         259.70         4773 5          42629          22167          103.42
60 22           33.43           19.11           21 49           18 55           4743           14257          32057          457.79         374 42         20196          105.46
5693            3251            1704            17.04            1931           69.46          165.80         27851          422.60          35511          165.15          72.90
40 89            2141             1*57             1954           3S.74           11431          181.98         27257          46661          406 04         201.62          83.76
4596            25.40           16.76           2045            27.19          100.IB          17731         312-98         47527          41324          19206          80.27
4347            3042            2435            24 48             29 10           S941          158.72         23633          31868          45250          204 49          82.70
4198            3342            2616            1357            22.45           6033           15756          27128          406.99         328.08          15635          86.62
4090            77.74            2594            2342            1990           72.62          125.93         205 12          34264          34109          241.43          13178
40.72            2109            1440            1337           25 19            8837           194.77         324*4          395.78          47141          29337          117.63
5905            3040            1761            1106            2892            7426           173.45         280.63         329.11          21897          11761          40.88
Z7 40            2344              2144            2723            93.43          162.69         301.77         431.95          51298          385.49          8735            59 01
60 17            4155            3301            3355            33.13           4221            9901           268.04         426.65         381.62          15250          49.11
3845            3539            29 64           25 43            2851           64 99          212.74         35680          36933          301.07          21257          86.40
42-15            3141            27.05           29.88           27.66           63.43           15829         365.68          44433          49002          265.10          13532
4944            2641            22.43           21.07           22-21           56.41           19327          31156          25953          172.23          48.41            3738
26.05            2100            1936            2053            29.91            6265           16986          27256          402.48         266.79          85.10            4123
27 13            2143            22-23            2354           22.47           31.79           16916         21B.01         251.07         235.69          69.45           33.25
24.27            2122            21.09            2110            2451            7048           190.06         25534          304 18          22724          17023          61.80
3509            2947            27 12            2142            21.76          100.95         206 04         357 70          42437          345.82          21444           71.05
4243             2942            2440            2834            2556            3993           13518          20021          30456          384 86          14160          60.73
64 3 7            3017            2936            3958            33.71           59.99          159 09         24829          47751          289.20           7556            34.13
24.70            2097            21-27            2119            21.43           27.66           15739         327.62          40832          223.90           9365            35.43
2735            2739            2358            23.01            3198            77.48          146 18         243.75          32739          28128          20663           64 54
38.09           29.46           27 38           29.09           50 97          108.68         211.74         348 16          40236          228 94          16131           57.98
34 94            2795            23.78           23.15            3049            6456           199.60         332.71          46803          214.67          603 2            3739
2649            23 18           21.91            2337            2991            3359           109.99         184 66         238.94         182.92          62.34           43.92
82-37 2747 23 68 3107 42 63 14919 23335 36442 24936 253.28 14459 7639
43.19 28 74 2176 26 76 55.48 160 74 219.91 311 16 33125 223.76 301.93 209.43
12154 64 16 52 67 5056 53.00 10L07 190.31 36521 43737 356.71 32138 131.49
2653 2195 19.19 18.63 22.67 139 49 19328 270.27 28638 318.79 245 17 62.07
3263 24.45 31-44 3165 51.91 13030 196.44 219.45 30734 260 56 21426 12339
6993 47.49 40.32 4741 4590 15020 227.06 38B.91 43031 28031 22861 99.22
6721 5152 3943 43.17 36.47 63.63 173.86 22029 256 08 20432 100.41 59 29
4829            35.18            33 79            4055            4937            42.01          165.90         227.15         290.96         255 92          99 77            58.22
39 79            3755            29.03            2793             3258            80.91          175.45         224.47         299 45         250 72          15893           75.90
4722             3744            28.00            35 82            4557            77 01          186 04         26554          357.93         335.63         155.74          85 91
4892             38.00            30.90            35 41            37.62            8732          193.81         406.01          521.40         321.93          200.92         117.40
6892             5091             4230             4182            S3 48           11476         247.43         36234          40031           34258          136.62          69.46
51.00            4539            30 26            33.88            81.60           157.88         21937         289 57         336.77          20759          168.72         110.97
221.45
15554
171.90
194 67
167 66
18237
171.24
165.22
17637
15167
1502S
13936
155.40
15729
138.79
13331
13324
166.74
11526
17731
135.04
14632
173.45
10L74
118.13
9380
116.12
15467
11823
12842
11542
123.41
14125
12648
81.78
139 90
16126
187.16
13537
13529
17138
10967
11230
119.41
131.15
169.97
160.94
144 42
MEAN
4754             3244             2615            26.10            34.69            8336           184.00         30537          395 79          33151          177.06           1238
145.55
MAX
12154           64 16            5267             5056            93.43           162 69         30L77          48557          56733          570.93          32138          209 43
MIN
2653             2195             19.19            18.63            22.67            33 59          109 99         184.66          23854          182.92          60.82            3739
STXCV
1747             9.12              7.15              912             16 12             354)            3187           73 48           88.85            96 35            6977            3559
Figure 6.1.1:              A verage annual discharges at Dabus Alt. 130.
120 HYD R SP001A (HYDROLOGICAL Report).docx
120 HYD R SP 001A
studio pletrangell, romeHYDROLOGICAL Report
page  37 of 51
DABUS HPPs PRE-FEAS
Table 6.1.2
48 years of monthly average discharges at Dabus Alt. 108.
M 0 NTH IV AVDtAGE DlSCMAWCt I'"*/*)
OCT           XOV
66754
417.11
512 71
419 93
24432
440J5 51550 319 29
541 JI
246.01
1373$
15151
429 42
51375
563 93
49652
23032
429 JI
245 56
356 03
354 42
202 ’I
3375J
411-24
34157
41942
22753
25016
30535
12220
533 02
140 12
22151
44331
22105
370 74
462.21
26017
316 06
4409$
22211
JJ61I
1X'
167 56
22150
29752
14054
212 JO
30233
116 69
311 15
124 07
37191
14355
17661
20131
42147
541 77
4J5 95
167 23
rs i.
31356
S043J
24127
120 HYDRSP 001A
studio pietrangell, romeDABUS HPPs PRE-FEAS
HYDROLOGICAL Report
6.2 FLOW DURATION CURVES
Both the total period Flow Duration Curve (FDC) and the Annual Flow Duration Curves  (ADFCs) were evaluated according to Vogel and Fenessay  ([10], [11]) on the basis  of the dally  discharge time series (Figure 6.2.1 and 6.2.2 - upper plots) and estimating the discharge quantiles  as  suggested by  Parzen ([28]). It is here recalled that this time series span the period between 1961-2008
The 95% confidence Intervals  for the median of the Annual Flow Duration Curves  (Figure 6.2.1 and 6.2.2) lower left) and AFDCs  characterised by  a return period of 2 and 10 years  (Figure 6.2.1 and 6.2.2 lower right) were evaluated according to Vogel and Fenessay ([10]).
The  total  period  FDC,  as  well  as  the  mean  and  median  AFDCs,  show  the  expected  perennial-character  of  the Dabus river flow regime.
120 HYD R SP001A (HYDROLOGICAL Report).docx
studio plelrangell, romeDABUS HPPs PRE-FEAS
HYDROLOGICAL Report
page 39 of 51
Annual Flow Duration Curves (AFDCs)
Frequency of exceedance F[Q>q)x100
Frequency of exceedance F(Q>q]x100
Median and Mean AFDC
—--------- — Madian
Frequency of exceedance F(Q>q)x100
Frequency of exceedance F[Q>qlx100
Figure 6.2.1:             Dabus Alt. 130: Totalpenod 1961-2008 Flow Duration Curve (upper left), Annual Flow Duration Curves (upper right), Median and Mean of Annual Duration Curves (lower left).
Total period Flow Duration Curve (FDC)
Annual Flow Duration Curves (AFDCs)
Frequency of exceedance F[Q>q]x100
Dabus Alt. 108: Total period 1961-2008 Flow Duration Curve (upper left), Annual Flow Duration Curves (upper right), Median and Mean of Annual Durabon Curves (lower left).
120 HYD
R SP001A (HYDROLOGICAL ReportJ.docx
120 HYDRSP 001A
studio pletrangeh, romeDASUSHPPs PRE-FEAS
HYDROLOGICAL Report
page 38 of 51
Figure 6.1.4:             Multi-annual monthly average discharges at Dabus Alt. 108.
6.2 FLOW DURATION CURVES
Both the total period Flow Duration Curve (FDC) and the Annual Flow Duration Curves (ADFCs) were evaluated according to Vogel and Fenessay ([10], [11]) on the basis of the dally discharge time series (Figure 6.2.1 and 6.2.2 - upper plots) and estimating the discharge quantiles as suggested by Parzen ([28]). It Is here recalled that this time series span the period between 1961-2008
The 95% confidence Intervals for the median of the Annual Flow Duration Curves (Figure 6.2.1 and 6.2.2) lower left) and AFDCs characterised by a return period of 2 and 10 years (Figure 6.2.1 and 6.2.2 lower right) were evaluated according to Vogel and Fenessay ([10]).
The total period FDC, as well as the mean and median AFDCs, show the expected perennial-character of the Dabus river flow regime.
120 HYD R SP001A (HYDROLOGICAL Report) docx
120 HYDRSP 001A
studio pietrangeli, romepage   39
DABUS HPPs PRE-FEAS
HYDROLOGICAL Report
Total period Flow Duration Curve (FDC)
Annual Flow Duration Curves (AFDCs)
Frequency of exceedance F(Q>qJx100 Median and Mean AFDC
Frequency of exceedance F(Q>q)x100 Generalized AFDC
Frequency of exceedance F(Q>q]x100
Frequency of exceedance F[Q>q]x100
Figure 6.2.1: Dabus Alt. ISO: Total penod 1961-2008 Flow Duration Curve (upper left), Annual Flow Duration Curves (upper
right), Median and Mean of Annual Duration Curves (lower left).
Total period Flow Duration Curve
(FDC)
Annual Flow Duration Curves (AFDCs)
E
or
s
E
|
Frequency of exceedance F(Q>q)x100
Generalized AFDC
o
8
Figure 6.2.2:
Dabus Alt. 108: Total period 19
61 -2008 Flo
w Duration r, ,
right), Median an
d Mean
Annua! Duration
Dur3ti°" Curves (upper
120 HYD R SP001A (HYD
ROLOGICAL Report) docx
120 H
YD R SP 001A
studio PietrangeliTTomeDABUS HPPs PRE-FEAS
HYDROLOGICAL Report
page 38 of 51
Figure 6.1.4:                   Multi-annual monthly average discharges at Dabus Alt. 108.
6.2 FLOW DURATION CURVES
Both the total period Flow Duration Curve (FDC) and the Annual Flow Duration Curves (ADFCs) were
evaluated according to Vogel and Fenessay ([10], [11]) on the basis of the daily discharge time series 
(Figure 6.2.1 and 6.2.2 - upper plots) and estimating the discharge quantiles as suggested by Parzen ([28]).
It is here recalled that this time series span the period between 1961-2008
The 95% confidence intervals  for the median of the Annual Flow Duration Curves  (Figure 6.2.1 and 6.2.2) lower left) and AFDCs  characterised by  a return period of 2 and 10 years  (Figure 6.2.1 and 6.2.2 lower right) were evaluated according to Vogel and Fenessay ([10]).
The total period FDC, as well as the mean and median AFDCs, show the expected perennial character of the Dabus river flow regime.
^HYD^rTsPOoTa (HYDROLOGICAL Report).docx
120 HYDRSP 001A
studio pietrangell, romeHYDROLOGICAL Report
Total period Flow Duration Curve (FDC)
Annual Flow Duration Curves (AFDCs)
Frequency of exceedance F[Q>q]x100
Median and Mean AFDC
Frequency of exceedance F[Q>q]x100 Generalized AFDC
Frequency of exceedance F[Q>q]x100
Frequency of exceedance F[Q>q]x100
Figure 6.2.1:
Dabus Alt. 130: Total period 1961-2008 Flow Duration Curve (upper left), Annual Flow Duration Curves (upper right), Median and Mean of Annual Duration Curves (lower left).
Total period Flow Duration Curve (FDC)
Annual Flow Duration Curves (AFDCs)
1000
Frequency of exceedance F[Q>q]x100 Median and Mean AFDC
— Median
“ ■ Mean
95 % Conf, Interval
40             60             80 Frequency of exceedance F(Q>q]x100
^gure 6.2.2:
Dabus Alt. 108: Total period 1961-2008 Flow Duration Curve (upper left), Annual Flow Duration Curves  uu
(
right), Median and Mean of Annual Duration Curves (lower left).
P00lA (hydrological
120 hydITsp ooi a
studio P»etr Jrxjell, romi*Figure 6.3.3:              Dabus AH 130 and Alt 108: Comparison between total period Flow Duration Curve at (FDC) and
the mean of Annual Flow Duration Curves (AFDCs).
Figure 6.3.3 shows the comparison between the total period Flow Duration Curve (FDC) and the mean of the Annual Flow Duration Curves  (AFDCs), together with the values  of the corresponding main discharge quantiles, evaluated at Dabus Alt 130 (left) and at Dabus Alt. 108 (right)
120 HYD R SP001A (HYDROLOGICAL Report).doo<
120 HYDRSP001A
studio pietrangeli, rome41 orsi
DABUS HPPs PRE-FEAS
HYDROLOGICAL Report
7 FLOOD FREQUENCY ANALYSIS
page
7.1 BASIC ASSUMPTIONS
Two general approach are possible to carry out a frequency flood analysis.
- a local approach; and
a regional approach.
As  a general rule, the availability  of several hydrographic  stations  along the river of Interest or within Its catchment area makes  it possible to perform a regional analysis. This  type of analysis  should be preferred to the local one as  it makes  possible to extend the length of the extreme-value sample on which the analysis  Is performed and therefore to enhance the reliability  of the analysis  Itself. Unfortunately, In the present case, not only  the catchments  of the other stations  located within the overall Dabus  river basin are largely  smaller with respect to that of the station of Dabus  at Assosa (ID 5002) but, as  already  observed (par. 4.1), the quality of the data is questionable.
Given the above, in order to perform a local frequency  flood analysis, an extended time series  of daily discharges  at the station of Dabus  near Assosa should be used. For sake of conservativeness, the time series recorded at the station of Didessa near Arjo was  preferred In the present case. In-fact, not only  does  this last time series  result from actually  observed data but, differently  from the one actually  recorded at the station of Dabus  near Assosa, It presents  several peaks  on the underlying strictly  seasonal (wet-season) large-flood occurrence (e.p. Figure 7.1.1).
As  already  observed (par. 4.1), the much more smoothed seasonal hydrographs  observed at the station of Dabus  near Assosa can be explained with the presence of floodlng-land areas  (e.g. swamps) that tend to moderate the flood flows  upstream this  station. Furthermore notice that, for the same reason, using the dally  discharge series  actually  recorded at the station of Didessa near Arjo makes  It possible to avoid any conversion from dally to peak discharge.
time (month)
Figure 7.1.1:                Comparison between the concurrent discharges collected at the station of Didessa near Ario (ID 4001 - black line) and Dabus near Assosa (ID 5002 - red line) during 1967.
}
120 HYD R SP001A (HYDROLOGICAL Report).docx
120 HYD R SP 001A
studio pietrangell, romeDABUS HPPs PRE-FEAS
HYDROLOGICAL Report
page 42 of 51
The availability  at the station of Didessa near Arjo of a time-series  of daily  discharges  spanning the periods 1961-200S, i.e. a senes  of 4S years, makes  it possible to perform a reliable local flood-frequency  analysis. Two main approaches are possible (e.g. [19]-[25]):
- modelling of block maxima (i.e. maxima of blocks of equal duration, e.g. one year); and
modelling of threshold excesses.
As it ts known (eg. [19], [20]), the block maxima approach is closely associated with the use of GEVD family (Genera'ced Extreme Value Distribution), while the one of the threshold excesses with the use of GPD family (Genera', zed Pareto Distribution). Furthermore, it was  shown that (eg. [19], [20]), if block  maxima of a random process have an approximate distribution within the GEVD family, then the threshold excesses of the same process will have an approximate distribution within the GPD family.
Ln the present case, being large-flood occurrence mainly  a seasonal phenomenon suggests  the adoption of the standard 'block  maxima' approach, being the blocks  (as  usual in hydrology) extended over a period of one year. Tab. 7.1.1 presents the chosen sample of yearly maxima between 1961 and 2008.
Taiwf 7. LI:                  Chosen sample of yearly maxima at the station of Didessa near Arjo (ID 4001) between 1961 and2008.
YEAR
rrf/s
YEAR Qyr.mx
m3/s
YEAR
Qyr,mx
m3/s
YEAR Qyr.mx
m3/s
YEAR Qyr.mx
m3/s
1361             1172-9
1362              893.3
1363              788.5
1364             849.6
1365             950.2 1566             814.0
1367            1164.9
1368             796.4
1369             721.4 1970             940.3
1971              673.7
1972              595.8
1973              765.9
1974              720.6
1975              680.3
1976              469.1
1977              829.9
1978              891.3
1979              489.6 1920              645.0
1981 624.8
1982 554.3
1983 1083.0
1984 368.9
1985             480.0
1986             502.3
1987             511.4
1988             689.1
1989             574.4
1990             689.1
1991 771.4
1992 749.7
1993 645.6
1994 804.3
1995 353.3
1996             525.1
1997             586.4
1998             655.7
1999             660.8
2000             443.0
2001 744.4
2002 311.7
2003             395.0
2004             601.0
2005             613.3
2006             749.7
2007 598.6
2008 582.8
a general rule, the series  of 'block  maxima should meet certain statistical criteria (hypothesis) such as randomness, independence, homogeneity and stationarity in order for the results of a frequency analysis to be theoretically  valid ([19]-[24], [17], [27], [23]). Standard statistical testing of the available series  was therefore performed according to WMO ([17]) and Helsel and Hirsch ([27]). In particular, the tests by Grubs- Becks (outliers), Wald-Wolfowitz (independence and stationarity), Spearmann and Mann-Kendall (trend) and Mann-M/hitney (homogeneity and stationarity) were performed.
120 HYD R SP001A (HYDROLOGICAL Report).doo<
120 HYD R SP 001A
studio pletrangeli, romeDABUS HPPs PRE-FEAS
HYDROLOGICAL Report
page 43 of 51
As  expected, Mann-Kendall, Spearman and Mann-Whitney  null-hypothesis  are rejected when these tests  are performed on the selected time series  of yearly  maxima. In-fact, the sample of yearly  maxima shows  a clear decreasing trend, making the sequence not-stationary  (Fig. 7.1.1). Tab.7.1.2 presents  the results  of each test comprehensive of the values of the tests' statistics as defined by WMO ([17]).
Figure 7.1.1: Chosen sample of yearly maxima at the station of Didessa near Arjo (ID 4001) between 1961 and2008.
Table 7.1.2:              Results of Hypothesis Tests comprehensive of the values of tests'statistics as defined by WMO ([17]).
HO
Test Hypothesis
Xh
Xi
Grubbs and Beck
0
Not Rejected
1470
292
HO             Test Hypothesis                S
var[S)
P
Mann and Kendall                          1
Rejected
-432                12700
K
-3.83 0.00013
HO
Test Hypothesis
R
meanfRj
var[R]
u
Ulf
u
UD
Wald and Wolfowitz
0
Not Rejected
2.3EW7
2.2EW7
6.3E+10
1.68
-1.96
1.96
HO
Test Hypothesis
RSP
t
*LB
tuf
Spearman
1
Rejected
-0.56
-4.61
-2.01
2.01
HO
Test Hypothesis
U
mean[U]
var[U]
u
Ulf
Uu«
Mann and Whitney
1
Rejected
112
288
2350
-3.62
-1.96
1.96
-------------- ---------------
Despite the observed trend, a standard flood frequency analysis of yearly maxima of the dally discharges can be performed if this  trend is  not considered (neglected), Ze if the selected time-series  Is  assumed as  a stationary sequence. Notice that, this assumption, widely If not always considered In practice, Is In the
present case also conservative as modelling the not-stationary series, Ze. the trend, would Imply taking into account the decreasing magnitude of floods with time.
-^TsroOlMHYDROLbGICALReporq.docx 
120 HYD R SP 001A
studio pietrangeli, romeDABUS HPPs PRE-FEAS
HYDROLOGICAL Report
page 44 of 51
7.2 MODEL INFERENCE
Three different models  (probability  distributions) were fitted to the resulting single sample of dimensionless maximum annua1 oeak  discharge (Extreme Value Type I, or Gumbel, or EV-I distribution; Generalised Extreme Value cr GEV distribution; Log-Normal distribution), being the parameters  of each model estimated by  means  o* me Maximum Likelihood (ML) method, and the goodness-of-fit checked by  means  of both d*agnosbc-oiots  {eg. [20]) and the Kolmogorov-Smirnov  Test ([14]). Confidence Intervals  of parameters' estimate we-e calculated fry means of the Delta Method ([20], [21], [22], [24]).
Rgj*es ‘m-- 7.2 1 to 7.2.3 show the results of the inference of the chosen models In terms of diagnostic 
^j-mermore providing the parameters' estimate, the variance-covariance matrix, the main quantiles 
expresses n terms of return penod Tr, and the results of the Kolmogorov-Smirnov goodness-of-fit test at
95 m prubatxlitv.
A< it £5 oovr tne Kdmogorov-Smimov  goodness-of-fit test Is  applied on the assumption that the treoreoca: cxrnnuo'js  CDF is  completely  specified [(23), (26)], /.e. its  parameters  are a priori known. In other words, tne parameters  of the distribution should not be estimated from the same sample that is  used for the test The test was  therefore performed by  splitting the available sample In two part: one used to estimate the parameters  and one to carry  out the test In particular, by  sorting the single sample of the d men&oniess  maximum annual peak  discharge so as  to obtain an ordered statistics, the two samples  were obtained b/ seectng ether the 1st, 3rd, 5th,... or the 2nd, 4th, 6th,... order statistic.
As  far as  the aacnosbc  plots  are concerned, sample frequencies  (empirical probabilities) and probabilities resulting from the chosen model (model probabilities) were compared in the Probability  Plot (upper left panels), sample quantiles  (empirical) and quantiles  resulting from the chosen model (model) were compared m the Quantile Plot (upper right panels). The lower left panels  show the return level plot, Ze. a different form of comparison between empirical and model cumulative (probability) distribution. Because of the simplicity of interpretation, and because the choice of the scale compresses the tall of the distribution so that the effect o< extrapolation is  highlighted, return level plots  are particularly  convenient for both model presentation and validation, Finally, the lower right panels  show the comparison between the empirical and the model probability denoty function (PDF).
Figure 7.2.4 shows the comparison between the fitted models both In terms of return level plot, and In terms of return period matrix (quantiles).
120 HYD R SPOO1A (HYDROLOGICAL     Report),<Rz.x               120 H YD R SR 00 J A
r.tudlo plctrangell, rome------- .. ----------------------------------------------------------------------------------------------------------------------
DABUS HPPs PRE-FEAS
HYDROLOGICAL Report
page 45 of 51
0 0.5 1
Empirical
Return Level Plot
Model
|n. of data:                                                                                                                       Complete Variance-Covariance Matrix
342.29
156.95
156.95
677.35
Parameter
value
confidence Interval
location
590.670
539.660
641.680
scale
170.466
134.204
206.727
Return Period Matrix
Tr
yr
Qmax
m’/s
Cl
±mJ/s
2
653
57
5
846
86
KolmoRorv-Smlmov goodnes-of-flt test
10
974
109
Null hypothesis
Not rejected
20
1097
133
□n,0.05
0.196
50
1256
165
d„
0.062
100
1375
190
200
1493
214
500
1650
247
1000
1768
271
2000
1886
296
5000
2043
329
10000
2161
354
Figure 7.2.1:             E VI (Gumbe!) mode! fitting on the selected sample of maximum annual discharges.
RcPorl)'docx             120 IIYD R SP 001A
studio pIctraiKjell, routetWLjtA MIX*
MZXraCVZGU Report
page 46 of 51
^robabilitv Pk>1
Quantile Plot
Empirical
Return Level Plot
Model
**
Return period Tr [year]
I «8|
Complete Variance-Covariance Matrix
1 v»lue                            confidence interval
-tX -z-
SX35                 549.35               659.76
175.52
137.01
214.04
Return Period Matrix
-C 149
-0.336
0.027
Tr
E
Qmax
m3/s
Cl
±m3/s
2
667
59
5
840
72
- '-1**
Rz
S
C-
test
reir-Od
0.196
ooce
10
940
86
Zsz.
20
1026
110
50
1124
154
100
1189
194
200
1247
239
500
1315
301
1000
1361
351
2000
1402
401
5000
1450
467
10300
1483
516
'*>>•* ///
C^ v rrrtH ftvng on the selected sample of maximum annual discharges.
1K» V ->X1A 'KO//.
I/O HfD P SP f/j1A
studio pletrangell, romeDABUS HPPs
PRE-FEAS
HYDROLOGICAL Report
Probability Plot
1000 800 600 400
Empirical
page 47 of 51
Quantile Plot
Model
Matrix
0.00
0.00
Parameter
value
0.00
0.00
location scale
6.486
n 7qi
6n
U.Zod
0.349
Tr
yr
Qmax
m3/s
Cl
±m3/s
2
656
54
Kolmogorv-Smlmo
v goodnes-of-fittest
5
838
77
Null hypothesis
Not rejected
10
952
101
On.0.05
0.196
20
1058
132
dn
0.054
50
1192
183
100
1291
232
200
1388
290
500
1515
384
1000
1612
470
2000
1709
572
5000
1838
735
10000
1936
Figure 72.3:
883
LOGNORMAL mode! fitting on the selected sample
fe of maximum
120 HYD R SP001A (HYDROLOGICAL Report).d(Kx ™HYDR^T6oiA
studio pit
I’/ romeMnwliniim (Inily ili«< h»iHP ° I'" '
D*R.!$, MX-
h^DROLOGICAL Report
page 48 of 51
120 HYD R sp001 A (HYDROLOGICAL Report).docx
120 HYD f< SP 001A
studio pletrangell, rome-------
— -
dabus hpps
PRE-FEAS
HYDROLOGICAL Report
page 49 of 51
7.3 CONCLUSIVE REMARKS
Although none of the fitted models  rejected the null-hypothesis  and therefore EVI, GEV and Log Normal model can all be assumed as  underlying distributions  of the population of the yearly  maxima, visual inspection of either Figure. 7.2.3 or Fig. 7.2.4, shows  that the best fit of the overall selected sample is provided by  the GEV model. In fact, this  model is  capable of following the concavity shown by the selected sample of extreme values. The two major events that appear not fitted by the model may be considered as exceptional events  that have a probability  of exceedance lower (Ze. a return period greater) than that resulting from the sample frequencies (plotting positions).
Nevertheless,  for  sake  of  safety  the discharge quantiles provided by the EVI  (Gumbel) model were selected for design purposes.
T^HYcTr SP001A (HYDROLOGICAL Report).docx                 120 HYD R SP 001A
studio pietrangell, romeMix-h.
Report
S REFERENCES
page 50 of 51
”’<x KG :96’
V" IS Cohns. Colorado, 1967.
-\ve Of daily over flow sequences, Colorado State Univ. Hydrology
Ojtxk     K   G   av   >evev>dx   V   (1967).   Stochastic   descnption   of   daily   river   flows,   Proceedings -r»—Symposium, l. A.S.H., Fort Collins, Colorado, 1967.
mkv. r '
1^'
Hydrology, laxmi Publications
5-hvnc  c  *•’  199S  Vara  %  eh  vol  for   Pattern  Recognition.  Oxford  University  Press,  Oxford.
^•OTr^g-t- 196’ An introduction to computing with neural nets. IEEE ASSP Magazine 4 (2), 4-
Kjrwer%a^w     0   E   Hrton,   GE.,   Williams,   RJ.   (1986).   Learning   internal   representations   by   error zroGaarxr- 1- ^a-a'ei Distributed Processing: Explorations in the Microstructure of Cognition, vol. 1,
X 316*362
*     19M A method for the solution of certain problems in least squares. Quarterly of
*42*me*.ebcs 2, 164-168.
r              ^a*e. D 1963 ', An aiaonthm for least-squares estimation of nonlinear parameters. Journal of the y Inc and Applied Mathematics 11 (2), 431-441.
-
.-
. *     1969) mow duration curves. Manual of Hydrology. Part 2. Low-Flow Techniques. USGS. M  and  Fennessey,  N.M.  (1995).  Flow-duration  curves  I:  New  interpretation  and  confidence jo-rv of Hydraulic Engineering, 120 (4).
vw  P  ”  &rc  Fer.nessey,  N.M.  (1995)  Flow  duration  curves  II:  A  review  of  applications  In  water nesoijns pumnc Water Resource Bulletin, 31 (6).
. Jones P.O , Osborn TJ. And Lister D.H. (2013). Updated high-resolution grids of monthly g oc<zr.a^cr< - the cru ts3.10 dataset Int Jour. Climatology.
♦> ►fjrre, M And Jones, P. (1999). Representing twentieth-century space-time climate
pa- I Development of a 1961-90 mean monthly terrestrial climatology. Journal of Climate.
Arerrar
Sooety.
w                   M. And Jones, P. (2000). Representing twentieth-century space-time climate ><- n Development of 1901-96 monthly grids of terrestrial surface climate. Journal of
O                           WZf&rcAcqxzl Society.
/X£ ZJX;
Modeling System HEC-HMS. Technical Reference Manual. US Army Corps of Er^r^firig Center.
'.’A.rxi (2015) *tyGrcAogc Modeling System HEC-HMS. User's manual v. 4.1. US Army Corps of Engr^ertf’ig Center.
Czgzr./z*xz> (2009).Guide to Hydrological Practices. Vol. II. WMO-N°168.
1 lr
-
( (1974) Protores for deeding outlying observations In samples. NTIS-BRL N°1713.
M p , G Urxjyw, p/xxzen, H. (1982), Extremes and related properties of random
procjr^^
Jr-rUg
(20)
Coles, S (2W/>) Ar, introdudio
n to stotMcal modelling of extreme values. Springer.
1/ HVDRSiqOU J'/:/j
\/b H<b P X/>ia
studio pletrangell, romeDABUS HPPs PRE-FEAS
HYDROLOGICAL Report
page 51 of 51
[21] Belrlant, J., Goegebeur, Y., Segeres, J., Teugels, J. (2004). Statistics of extremes. Theory and applications. John Wiley and sons.
[22] Castillo, E., Hadi, A.S., Balakrishnan, N., Sarabia, J.M. (2005). Extreme value and related models with application in engineering and science. John Wiley and sons.
[23] Kottegoda, N.T., Rosso, R. (1998). Statistics, Probability and reliability for dvil and environmental engineers. MacGraw Hill.
[24] Smith, R.L. (1989). Extreme value analysis of environmental time series. An example based on ozone data (with discussion). Statistical Science, 4, 367-393.
[25] Smith, R.L. (2001). Environmental statistics. Department of statistics. University of North Carolina. Chapel Hill.
[26] Lilliefors, H. W. (1967). On the Kolgomorov-Smirnov test for normality with mean and variance unknown. J. Am. Stat. Assoc., Vol. 62.
[27] Helsel, D.R., and R.M. Hirsch (1992). Statistical Methods in Water Resources, Elsevier, NY. [28] Parzen, E. (1979). Non-parametric statistical data modeling. Journal of the American Statistical
Association, 74 (365).
■^HYDR SP001A (HYDROLOGICAL Report).docx 120 HYD R SP 001A
studio pietrangeli, romeI
I
I
I
ffl
n
IDABUS HPPs
PRE-FEAS
HYDROLOGICAL Report
Ref. Drawings
120 GEN D SP 001 B - Hydro and Meteo Stations ± 120 GEN D SP 002 B - Catchment Areas
120 HYD R SP 001A
studio pietrangeli, romeI
I I I I I I I I I I I I I I
1
I
t
II
I
1
DABUS ch. -108 CATCHMENT
A = 10,400 km2
KOMIS-REA^GOF
7 LTUK'
DABUS ch. -130
CATCHMENT
A =10,013 km2
■■ ■
'' ’ V NEDJO
.
We—- FI<?*KOUI<iA) 'v’T*"'1
/^LEGENDn oGlCAl5TATlON
.
1 ME,E^X STATION
f h^^tchment
MINISIKY ol WATFA IRRIGATION * ENERGY ftDEAAL DEMOCRATIC REPUBLIC »T ETHIOPIA I
DABUS Hydropower Project |
1 r ‘ gp a‘“'°-n 0 p«r»V*» '.
H PRE'- FEASIBILITYI
"l20 = HYDROLOGY
GEN = GENERAL
CATCHMENT AREAS
r. 1                030/
MJO
KUUCI
DABUS ch. -1 10,400 km2
1 'vocfrioo
—i
i
i
/ -5’
lekemte
DABU Sth. -130 CATCHMENT AREA
I'
NOTES
. k. l.lN. IMMinM„l WAtl l<
I
HYDROMETRIC STATIONS METEOROLOGICAL STATIONS
DABUS I lydropowor Profoct
BP
PRE - FEASIBILITY
120 - HYDROLOGY
QEN • GENERAL
HYDRO und METEO STATIONS
50Km
modification ■

Summery

Dabus Hydropower Project Summary

Dabus Hydropower Project Pre-Feasibility Study

Project Overview

The document presents the pre-feasibility study for the Dabus Hydropower Project in Ethiopia, conducted by Studio Ing. G. Pietrangeli S.r.l. for the Ethiopian Ministry of Water, Irrigation and Energy.

Key Project Details:

Location: Dabus River, a tributary of the Abbay (Blue Nile) River in southwestern Ethiopia
Catchment Area: Approximately 14,400 km²
Proposed Layout: Cascade with two power plants
Dam Sites:
- Dabus Alt. Ch.0+130 (Area: 10,009 km²)
- Dabus Alt. Ch.0+108 (Area: 10,400 km²)
Climate: Tropical Climate II with mean annual rainfall around 2000 mm

Hydrological Study

The hydrological report focuses on analyzing the river's behavior to determine mean monthly runoff and flood runoff.

Data Collection

Hydrometric and meteorological data were collected from various stations:

Station Type	Key Stations	Data Period
Hydrometric	Dabus near Assosa, Didessa near Arjo	1960-2008
Meteorological	Assosa, Mendi, Kiltukara	1961-2015

Methodology

Three approaches were used to extend the daily discharge time series:

Synthetic Streamflow Generation: Using Quimpo and Yevjevich model
Transfer Function:
- Power-law transfer function
- Artificial Neural Network (ANN) transfer function (preferred method)
Rainfall-Runoff Modeling: Using HEC-HMS software with CRU-UEA rainfall data

Key Findings

Parameter	Dabus Alt.130	Dabus Alt.108
Average Annual Discharge	145.55 m³/s	151.24 m³/s
Flow Duration Curve Q50%	229 m³/s	238 m³/s
Flow Duration Curve Q90%	74 m³/s	77 m³/s

Workshop and Client Responses

The document includes responses to client observations from a June 2016 workshop, addressing:

Survey methodology and site selection
Geotechnical investigations
Hydrological modeling approaches
Environmental impact assessment baseline
Watershed management
Socio-economic considerations

Key Consultant Responses:

Drone survey was carried out on the ch-155 dam area
Pleiades topographic database will be used for Feasibility Study
Detailed mitigation measures will be presented in Full ESIA report
RUSLE is used for soil loss estimation
ANN transfer function approach was preferred for hydrological analysis

Next Steps

The study recommends proceeding with the Feasibility Phase of the project, noting that:

More detailed site investigations are needed
Full Environmental and Social Impact Assessment will be prepared
Detailed mitigation measures will be identified
Construction material availability needs further assessment