Академический Документы
Профессиональный Документы
Культура Документы
SUPERVISOR
DR. M. O. NYADAWA
A Project submitted in partial fulfillment of the award of BSc. Civil Engineering of the Jomo
Kenyatta University of Agriculture and Technology
i
DECLARATION
I, Kevin Ochola Obambo do declare that this report is my original work and to the best of my knowledge,
it has not been submitted for any degree award in any University or Institution.
CERTIFICATION
I have read this report and approve it for examination.
Signed_______________________________________________Date_____________
Dr. M. O. Nyadawa
ii
DEDICATION
This Design work is dedicated to my dear parents and entire family. Its through you that God has
provided for me in all ways up to this point in my life. Thank you for being there
iii
ACKNOWLEDGEMENTS
I wish to thank my supervisor Dr. M.O.Nyadawa for his guidance and assistance at every stage of this
project, Eng. Wangai for his help on the watershed model and Mr.J.K. Mwangi on his invaluable support
and advice during data collection and analysis.
I also wish to thank the Mr. Muriuki GEGIS dept. for providing some Landsat images, Mr. Lawrence
Okello of the RCMRD (regional centre of mapping for resource development) for advice on the GIS and
remote sensing methods and providing recent Landsat images.
I am appreciative of the assistance offered to me at the Meteorological department, on meteorological
data of the study area, the Chief Electrical Engineer Ministry of Energy, Eng. Muiru and the Director of
Operations KENGEN, Eng Richard Nderitu for assistance in obtaining generation data on the Sondu
miriu power project.
I am indebted to Ken Kemucie for his assistance in GIS related procedures during preparation of data.
I am grateful to Mr. Thooko and Mr. Njiru of WRMA (Water Resources Management Authority), for
providing streamflow data and the operation of the rating curve.
Finally, I would like to thank my friends and colleagues for assistance and encouragement.
iv
Impacts of land use changes on hydrological processes and implication for hydropower potential:
case study of Mau catchment and Sondu-miriu project
ABSTRACT
To demonstrate the relation in land use change to streamflow variability, a hydrological modelling at a
catchment scale has been used to investigate the impact of land use change on streamflow in Mau
catchment using a distributed deterministic modelling system. The model divides a catchment into
equally sized grid cells, providing the ability to impose changes in land use spatially. Model parameters
were established by calibration against measured discharge data available over the year 1986. The model
simulation-period was divided into calibration 1986 and validation 1995 periods. The model provided
relatively good fits between measured and simulated discharge both at a daily and monthly scales.
To illustrate the effect of land use change on streamflow, the challenge is to determine whether the
hydrological variability is as a result of change in land use/cover. This was done by keeping all other
datasets in the model constant and varying only land use / land cover data. This study provides an
approach to assess the actual changes in hydrologic response of a watershed to land use transformations
made in the past 30 years. The approach combines land sat image data from satellites from 1973- 2008 of
the SW. Mau catchment, climatic data and hydrometric data. In 1973, 262 km2 (42%) of the watershed
area was covered by natural forest. Due to continued encroachment into forest land, the cover decreased
to percent 212 km2 (34%) of the area in 2008. Most of the cover change occurred between (2002-2008)
i.e. from 237km2 (38%) in 2002 to 212km2 (34%) in 2008 a decrease of about 10%. Three main classes
of land use have replaced the closed canopy. These are: poor natural cover, agricultural areas and
residential areas. The hydrology of the Sondu River also has changed, with sharp increases in peak flows
and an earlier occurrence of these peaks between 1973 and 2008. It can be concluded that land use change
is a factor affecting the (changes in the) hydrologic regime of the Sondu river drainage.
An active management strategy aimed at the conservation and regeneration of the natural vegetation is
recommended, in order to improve the consistency of streamflow, during both dry and wet periods.
In order to use the results of this kind of model for decision making and water resources management, the
model should be tested under different environment and different scenario conditions. Rainfall
measurement and stage-discharge rating curves should be given priority to improve model performance.
Table of contents
DECLARATION .......................................................................................................................................... ii
CERTIFICATION ........................................................................................................................................ ii
DEDICATION ............................................................................................................................................. iii
ACKNOWLEDGEMENTS ......................................................................................................................... iv
ABSTRACT .................................................................................................................................................. v
LIST OF FIGURES ................................................................................................................................... viii
LIST OF TABLES ....................................................................................................................................... ix
LIST OF IMAGES ........................................................................................................................................ x
LIST OF ABBREVIATIONS ...................................................................................................................... xi
CHAPTER ONE ........................................................................................................................................... 1
1.0 INTRODUCTION .................................................................................................................................. 1
1.1 Background ......................................................................................................................................... 1
1.2 Study justification ............................................................................................................................... 3
1.3 Problem statement ............................................................................................................................... 3
1.4 Objectives ........................................................................................................................................... 4
1.5 Research hypothesis ............................................................................................................................ 4
1.6 Scope of study ..................................................................................................................................... 4
CHAPTER TWO .......................................................................................................................................... 5
2.0 LITERATURE REVIEW ....................................................................................................................... 5
2.1 Introduction ......................................................................................................................................... 5
2.2 Model overview ................................................................................................................................ 15
CHAPTER THREE .................................................................................................................................... 17
3.0 RESEARCH METHODOLOGY. ......................................................................................................... 17
3.1 Study Area ........................................................................................................................................ 17
3.1.1 Location ..................................................................................................................................... 17
3.2 Input data .......................................................................................................................................... 19
3.2.1 Remote sensing and GIS geo-data. ............................................................................................ 19
3.2.2 Meteorological data ................................................................................................................... 22
3.2.3 Hydrometric data ....................................................................................................................... 23
3.3 Data Analyses ................................................................................................................................... 24
3.3.1 Remote Sensing Data Processing ............................................................................................... 24
3.3.2 GIS Processing and Analysis ..................................................................................................... 24
vi
4.1 Comparisons of Streamflow discharge and generation of the power project ................................... 35
CHAPTER FIVE ........................................................................................................................................ 37
5.0 CONCLUSIONS AND RECOMMENDATIONS ............................................................................... 37
5.1 Conclusions ....................................................................................................................................... 37
5.2 Recommendations ............................................................................................................................. 37
REFERENCES ........................................................................................................................................... 38
APPENDICES ............................................................................................................................................ 40
Appendices 1: STEAMFLOW DATA 1986 (sample) ............................................................................ 40
Appendices 2: sondu miriu power generation ......................................................................................... 42
Appendices 3:METEOROLOGICAL DATA ......................................................................................... 43
Appendices 4:The parameters file lists of vectors of the Sondu river .................................................... 44
Appendices 5: Changes of river flow response with change land use/ land cover data .......................... 45
vii
LIST OF FIGURES
Figure 1: Sondu Miriu power project layout ................................................................................................. 2
Figure 2: Schematization of the principal components of an integrated flow model ................................... 6
Figure 3: Schematization of the relations between the model and the system's environment ...................... 7
Figure 4: General layout of model .............................................................................................................. 16
Figure 5: Geographical location of the Sondu river basin .......................................................................... 17
Figure 8: Upper catchment of Sondu river within the Southwest Mau area ............................................... 18
Figure 6: General layout of the entire Sondu river basin ............................................................................ 18
Figure 7: General layout of the entire Sondu river basin ............................................................................ 18
Figure 9: General structure of the CEQUEAU parameters optimization module ....................................... 26
Figure 10: Model user interface .................................................................................................................. 28
viii
LIST OF TABLES
Table 1: Characteristics of the images used in the study ............................................................................ 19
Table 2: Meteorological data ...................................................................................................................... 22
Table 3: Values of rating curve ................................................................................................................... 24
Table 4: List of parameters that can be optimized ...................................................................................... 27
Table 5: Lists of vectors of the parameters file of Sondu river. .................................................................. 31
Table 6: Change of river flow response with change land use/ land cover data ......................................... 33
ix
LIST OF IMAGES
Image 1: Large scale destruction of forests and bamboo forests in South West Mau forest reserve .......... 14
Image 2: Land sat image P182R60 year 1973, showing watershed box ..................................................... 20
Image 3: Land sat image P169R60 year 1986, showing watershed box ..................................................... 20
Image 4: Land sat image P169R60 year 1995, showing watershed box ..................................................... 21
Image 5: Land sat image P169R60 year 2002, showing watershed box ..................................................... 21
Image 6: Land sat image P169R60 year 2008, showing watershed box ..................................................... 22
LIST OF ABBREVIATIONS
ArcGIS---------------------------Geographic information system software
DHSVM--------------------------Distributed Hydrology Soil Vegetation Model
ERDAS Imagine-----------------Remote sensing software
GIS--------------------------------Geographic Information System
GoK-------------------------------Government of Kenya
HRU-------------------------------Hydrologic response units
HSPF-----------------------------Hydrologic Simulation Program - Fortran
KenGen---------------------------Kenya Electricity Generating Company Ltd
KPLC------------------------------Kenya Power and Lighting Company
MIKE-SHE-----------------------Model Systeme Hydrologique European
n----------------------------------Mannings roughness
PRMS (MMS) -------------------Precipitation-Runoff Modeling System
SWAT----------------------------Soil-Water Assessment Tool
UBCWM--------------------------UBC Watershed Model
UNEP------------------------------United Nations Environmental Programme
xi
CHAPTER ONE
1.0 INTRODUCTION
1.1 Background
Forests cover only about three per cent of Kenyas land area they provide the following ecosystem
services: they trap and store rain water; regulate river flows and prevent flooding; help recharge groundwater tables; improve soil fertility; reduce soil erosion and sediment loads in river water; help regulate
local climate conditions; and act as carbon reservoirs and sinks.
There has been progressive reduction in forest cover over the years as economic pressures lead to
encroachment into protected forest areas; this has been experienced in the Mau catchment which is the
largest closed canopy in Kenya about 400,000ha, it is also the catchment to most drainage systems
including, Sondu, Mara, Nzoia,Yala, Ewaso Nyiro rivers.
The Sondu river is one of the major drainage systems whose catchment lies within the Southwest Mau
region. It is one of the large river basins across the western Kenya region. The basin is considered one of
the important sub catchments of the Lake Victoria basin as the Sondu River is one of the six major rivers
that drain into the lake; it has a catchment area of 3470sq.km and a length of 200km. With the recent
commissioning of the Sondu Miriu power project the hydrology of the river has attracted attention due to
the low flows that have impeded the achievement of the power plant capacity. Hence focus has been on
the Sondu river drainage basin and there has been growing concerns regarding degradation in the rivers
watershed that lies in the Southwest Mau forest reserve, suspected to be affecting the flow regime and
discharge. The river originates from the western slopes of the Mau escarpment and flows through a
narrow gorge penetrating the Nyakach escarpment, it then meanders into Odino falls before entering the
flood plains of Nyakwere where it drains into the Winam gulf of Lake Victoria. The Sondu Miriu power
project is based on a run-of-river diversion from a weir structure on the Sondu River above Nyakach
escarpment. Accurate scientific data is required to forecast and avert any more degradation including
planning, and implementing appropriate mitigation measures on the river.
This study utilizes the CEQUEAU hydrological model, a distributed deterministic model for stream flow
simulation and geographical information system (GIS) to determine the extent of land use/cover changes
in the Sondu river watershed and analyze the effects of these changes on the river flows over the last three
decades.
In 2001, 61,023 ha of forest in the Mau Complex were excised including one quarter of Southwest Mau
forest reserve was excised, the Southwest Mau forest is the primary source of the Sondu River. Between
1973 and 2005, Maasai Mau Forest lost over 8,214 ha of forest within its official boundaries, which were
established to protect the forest. Almost 43 per cent of that loss occurred in just two years from 2003 to
2005.Forest loss in critical catchment areas for the Sondu, Mara, Molo, Naishi, Makalia Nderit, and Njoro
Rivers will result in ecological and hydrological changes, which threaten the sustainable future of areas
downstream. In addition, people have encroached into some 43,700 ha in the Mau Complexs remaining
protected forests. The hydrology of the Sondu River has also increased with sharp increases in the flood
peak flows. There is evidence of increased soil erosion in the upper catchments with silt built-up in the
downstream especially in the flood plains of Nyakwere where it drains into the Winam gulf of Lake
Victoria. (UNEP, 2009), Kenya: Atlas of Our Changing Environment.Division of Early Warning and
Assessment (DEWA), Kenya
The Sondu River is the lifeline of the power project at the Sondu town and the basin is considered one of
the more serene sub catchments of the Lake Victoria Basin and ultimately the Nile Basin. The basin also
contains forests, large-scale farms, smallholder farms, pastoral grazing lands and fishers. Hence there is
need for urgent action to stem land degradation of the Sondu river basin including planning and
implementing appropriate measures.
1.4 Objectives
The basic objective of the study is to investigate the spatio-temporal information on the status of land use/
cover changes on the watershed and subsequent variability on the hydrological regime hence the changes
on the streamflow discharge.
The Specific objectives of the study include:
To process and display remotely sensed land use / land cover datasets at chosen time intervals
over 30 years in the Sondu river upper catchment in the Southwest Mau forest reserve.
To calibrate and validate the CEQUEAU model hence simulate streamflow and flow
characteristics of the river.
To ascertain whether or not changes in the flow regimes were attributable to land cover changes
in the basin.
CHAPTER TWO
and has proved very useful to answer questions raised by regional water resources problems associated
with surface water and groundwater (Hansen and Dyhr-Nielsen 1983). For example, such a tool can
enable the study of the effects of aquifer abstraction on surface water regimen, particularly during lowwater periods (Miles and Rushton 1983; Refsgaard and Hansen 1982b; Besbes et al. 1981; Emsellem
1971).
Characteristics of Integrated Flow Models
The model cannot represent all the mechanisms involved in the hydrological cycle and, to avoid undue
complexity, must simplify reality by integrating only its principal components. The aim of modeling is to
estimate the total flow at the outlet of the watershed. It is recognized that total flow is a combination of
three processes: surface flow, interflow and base flow (Fig. 2). We are also interested in the eventual
interactions between these hydrological processes and the systems' environment (Fig. 3): exchanges with
the atmosphere, surface outflows, lateral contribution by surface water, exchanges between aquifers and
exchanges between the non-modelled surface areas and aquifers.
However, in addition to the integration of the main processes and relations, the model should, for
economic and operational considerations, satisfy particular characteristics (Freeze and Harlan 1969;
Jacquet 1971). First, the model must take into account the system's physical properties. The model's
representation of the system must be based on the maximum amount of geological, geomorphological,
topographical and physiographical information. In addition, the functions used in describing the
hydrological processes must be derived from known physical law or include conceptual functions in
which parameters represent a physical reality. Because the model is indeed an image of the system's
physical structure, simulation results can be interpreted with an increased level of confidence.
Second, the model must represent the spatial and temporal variations in the system's structure and
behavior. This condition is relevant to the variability of hydrological phenomena and to the model's
usefulness as a tool for planning, development and management purposes. In most cases, modeling work
is not concerned with the conditions that prevail at the monitoring stations themselves:
Figure 3: Schematization of the relations between the model and the system's environment
(after Ledoux 1980).
The tool must therefore be helpful in extrapolating the watershed's response to points where there may be
particular needs.
Third, the model must be suitable for a vast range of hydrological and geographical conditions. In practice
this is desirable because it imparts a great deal of versatility with respect to the model's eventual
applications to different hydrological system configurations. The model must use a minimum number of
parameters for calibration and must incorporate simple relations to represent the hydrological processes
so that its usefulness is unhindered when going from one watershed to another.
7
Furthermore, it is preferable to use a model that can be easily adapted to data obtained at different levels
of discretisation at any time of the modeling process. This permits 1) to use the model as soon as the
minimum amount of data required to run the program is collected; 2) to update the model as new
information and data become available. This gives, for example, the modeler the possibility of
progressively defining areas where particular studies have to be undertaken to respond more accurately to
the objectives of the modeling study.
Since the model is to be applied by users who have not worked on its design, it is necessary as Bachmat et
al. (1980) point out, that the model be supplied with a set of relevant documents, including a complete
description of the model, a user's manual, print out copies of the program and modeling examples.
Existing Integrated Flow Models
Until now, much attention has been focused on management models that integrate the use of surface
water and groundwater with respect to economic, technological and political objectives (Bachmat et al.
1980). In watershed hydrology, the first attempts at modeling the interaction between surface and
groundwater flows were aimed at the study of the hydrologic cycle's mechanisms and to the development
of new tools to improve hydrologic methods and research. Pinder and Sauer (1971) were the first to
present a model that treats alluvial-aquifer-stream interactions to calculate the modification of a flood
wave due to bank storage effects. Freeze (1972) introduced a deterministic mathematical model that
couples three-dimensional transient saturated-unsaturated subsurface flow and one-dimensional gradually
varied unsteady channel flow to study the role of subsurface flow in generating storm runoff.
In practice, usable models have been built for various objectives according to two approaches. The
general approach includes models conceived and designed to be applied to watershed of variable sizes
and characteristics. The Knapp et al. (1975) model is a very good example of an integrated watershed
flow model designed to be general in practice. However, it should be noted that it requires the estimation
of a relatively large number of parameters (- 40). Another example is the Cunningham and Sinclair (1979)
coupled flow model involving the simultaneous numerical solution of the surface and subsurface flow
equations using the Garlekin finite-element method. It includes specific methodologies worked out by
Pinder and Sauer (1971).
The second approach is more specific to a particular watershed's physical characteristics. With this
approach, model application is restricted to watershed where the general configuration and the physical
characteristics are similar. The main objective of the model reported by Refsgaard and Hansen (1982a)
was to predict the hydrological consequences of groundwater abstraction on the river discharges and on
the hydraulic heads of the aquifers for the SusA watershed in Denmark.
Hydrological model- is a conceptual or physically based procedure for numerically simulating a process
or processes which occur in a watershed.
Modelling is the process of implementation of the model in simulating the actual phenomena
Most of the knowledge on the effects of land use change on catchment runoff comes from experimental
catchment studies, statistical methods and hydrological modelling. Some of models used include:
DHSVM
The Distributed Hydrology Soil Vegetation Model (DHSVM) is a physically based model developed at
the University of Washington to represent the effects of topography and vegetation on water fluxes
through the landscape (Wigmosta et al.1994; Wigmosta et al. 2002).
Model Description
The DHSVM is a fully distributed hydrologic model that subdivides the model domain (typically the
watershed or group of watersheds under investigation) into small computational grid elements using the
spatial resolution of an underlying digital elevation model (DEM). It is typically applied at high spatial
resolutions on the order of 10 to 100 m (Beckers and Alila 2004) for watersheds of up to 10 000 km2 and
at sub-daily time scales (best results are obtained at hourly time steps) for multi-year simulations. The
model includes two vegetation layers (forest canopy overstorey and understorey) and multiple soil layers.
Model Applications
The DHSVM has been widely applied in the PNW in a forest hydrology and management context, for
assessing the effects of forest roads on flood flows, for hydrologic forecasting, and for assessing the
potential effects of climate change, including but not limited to the following studies:
HBV-EC
The HBV (Hydrologiska Byrns Vattenbalansavdelning) model is a conceptual hydrological model
designed for use in mountainous environments. It was originally developed in the early 1970s at the
Swedish Meteorological and Hydrological Institute (SMHI), and has been used extensively for
hydrological forecasting, engineering design, and climate change studies (Lindstrm et al.1997). The
HBV-EC model, reviewed here, was adapted by Environment Canada and UBC (Moore 1993) to better
represent glacier processes. Several additional improvements have been made since, as described below.
9
Model Description
The HBV-EC model uses the Grouped Response Unit (GRU) concept to group DEM/GIS grid cells into
bins having similar land cover, elevation, slope, and aspect (Stahl et al. 2008). The model incorporates
two soil layers, no explicit vegetation layer, and operates at a daily time step
HSPF
The Hydrologic Simulation Program-Fortran (HSPF13) was designed by the US Envionmental Protection
Agency (EPA) to simulate a broad range of surface and sub-surface hydrologic and water quality
processes in watersheds (Bicknell et al. 2001). The HSPS model is based on the Stanford Watershed
Model and can simulate one or many pervious or impervious unit areas discharging to one or many river
reaches or reservoirs.
Model Description
HSPF is a lumped parameter hydrologic model that can simulate all primary natural hydrological
processes (Bicknell et al. 2001). The unsaturated zone is approximated using a single storage reservoir,
while explicit representation of vegetation in the model is limited. The model runs at any time step, from
1 minute to 1 day, which can be divided equally into 1 day.
Model Applications
Many of the published applications of the HSPF model have been in the central and east portions of
theUSA (Pike 1995). In British Columbia, data from Carnation Creek, on the eastern coast of Vancouver
Island, have been used in a test calibration for using HSPF in forested watersheds (Hetherington et al.
1995). An important finding of this study was that forest cover in the HSPF is represented as a single
index, making the representation of differing forest cover types very limited
MIKE-SHE
The integrated hydrological modelling system, MIKE-SHE17, simulates the major processes in the
hydrologic cycle and includes process models for evapotranspiration, overland flow, unsaturated flow,
groundwater flow, channel flow, as well as modelling how all these interact. The MIKE-SHE model was
developed, starting in 1977, as consortium of three European organizations (Systme Hydrologique
10
Europen, [SHE]) (Abbott et al. 1986a, 1986b) and the full model is currently distributed by the Danish
Hydrological Institute (DHI).
Model Description
MIKE-SHE is a fully distributed, grid-based hydrologic modelling system that incorporates multiple soil
layers and a single vegetation layer (i.e., overstorey and understorey are not distinguished) and flexible
model time step.
Model Applications
MIKE-SHE has proven valuable in hundreds of research and consultancy projects covering a wide range
of climatic and hydrologic regimes, many of which are referenced in Graham and Butts (2005).
PRMS (MMS)
The Precipitation-Runoff Modeling System (PRMS) is a modular-design, deterministic, distributed
parameter modeling system developed by the USGS23 to evaluate the impacts of various combinations of
precipitation, climate, and land use on streamflow, sediment yields, and general watershed hydrology
(Leavesley et al. 2005). To better allow for model updates and adding process components, the
architecture and modular structure of PRMS were redesigned and formed the basis for the Modular
Modeling System (MMS) (Leavesley et al. 2005).
Model Description
Watersheds are divided into HRUs based on characteristics including slope, aspect, elevation, vegetation
type, soil type, land use, and precipitation distribution (Leavesley et al. 2005). The model incorporates
a single vegetation layer, two conceptual soil zones (recharge zone and a lower zone), subsurface and
groundwater reservoirs, and operates at a daily time step.
RHESSys
RHESSys24, the Regional Hydro-Ecological Simulation System, is a GIS-based, modelling framework
that integrates water and biogeochemical cycling and transport over spatially variable terrain at small
(first order streams) to medium (fourth- and fifth-order streams) scales (Tague and Band 2004). RHESSys
combines both physically based process models and a method for partitioning and parameterizing the
landscape (Tague and Band 2004).
Model Description
11
RHESSys uses a hierarchical approach to partition a landscape into hydrologically distinct units,
including patches (essentially HRUs, areas with similar soil and land use characteristics), hillslopes
(portions of a watershed that drain to a single point or stream reach), climate zones and watersheds (to
organize stream routing). Multiple vegetation strata can be included in the model, while soil water storage
is conceptually divided into an unsaturated and a saturated zone (Tague and Band 2004). The model
operates at a daily time step
SSARR
The Streamflow Synthesis and Reservoir Regulation (SSARR26) model was developed by the US Army
Corps of Engineers, North Pacific Region to provide mathematical hydrologic simulations for systems
analysis as required for the planning, design, and operation of water control works (US-Army Corps of
Engineers 1991). The SSARR model has been further developed for operational river forecasting and
river management activities (US-Army Corps of Engineers 1991).
Model Description
The model is semi-distributed and works with elevation bands for snowmelt calculations. It includes a
single soil layer and a single vegetation layer (i.e., overstorey and understorey are not distinguished). The
model can operate at time intervals ranging from 0.1 to 24 hours
SWAT
The Soil Water Assessment Tool (SWAT) is a watershed-scale model developed to predict the impact of
management on water, sediment, and agricultural chemical yields (Gassman et al. 2007).
Model Description
In SWAT, watersheds are divided into multiple sub-watersheds, which are then subdivided into
hydrologic response units (HRUs) (Gassman et al. 2007). The HRUs are areas of homogeneous land use
and soil characteristics, and represent percentages of the sub-watershed area that are not considred in a
spatially explicitly manner in the model simulation (Gassman et al. 2007). The model considered two soil
layers (root zone and unsaturated zone), together with conceptual shallow and deeper aquifer stores, a
single vegetation layer, and operates on a daily time step.
Model Applications
Gassman et al. (2007) reviewed historical SWAT development as well as a total of 115 model
applications in the US, Europe, and worldwide. These studies illustrate the versatility of SWAT in
12
simulating the hydrology of diverse watersheds. Generally, this review illustrated the ability of SWAT to
replicate hydrologic processes at a variety of spatial scales on an annual or monthly basis. However,
model performance has been inadequate in some studies, especially when comparisons of predicted
output were made with time series of measured daily flow. These shortcomings in simulating watershed
hydrology at relatively short time scales (days) is attributed to the simplified representation of runoff
processes with the SCS curve number method. One drawback of the HRU approach is that explicit spatial
representation of riparian buffer zones, wetlands, and other features is not possible.
UBCWM
The UBC Watershed Model (UBCWM) was originally developed for daily streamflow forecasting on the
Fraser River system in British Columbia (Quick et al. 1995). The UBCWM is being used operationally
for forecasting daily streamflows in sub-watersheds of the Fraser River system that are subject to
snowmelt floods from the mountain snowpacks of the Coast, Columbia, and Rocky Mountains in British
Columbia.
Model Description
The model was originally designed for forecasting runoff from mountain watersheds and, for this reason,
the model is divided into area-elevation bands (Quick et al. 1995). The model incorporates a conceptual
soil moisture storage reservoir and a single canopy layer
WaSiM-ETH
The Wasserhaushalts-Simulations-Modell (WaSiM-ETH30) is a fully distributed, physically based model
for estimating climate change impacts for subalpine and alpine regions. Compared to PREVAH,
WaSiMETH is a physically based runoff model and has more flexibility in separating surface runoff from
interflow, allowing, in general, a better reproduction of flood events (Gurtz et al. 1999). The application
of WaSiM-ETH to climate change studies in cold mountainous terrain has been reviewed by Werner and
Bennett (2009).
Model Description
The spatial and temporal discretizations of the model can accommodate grid cell sizes of virtually any
dimensions (from centimeters to several kilometers), and time steps between 1 minute to several days
(Werner and Bennett 2009). The model includes a layered soil model and layered vegetation
13
The Mau Complex forms the largest closedcanopy forest ecosystem of Kenya, as large as the forests of
Mt. Kenya and the Aberdares combined. It is the single most important water catchment in Rift Valley
and western Kenya. Through the ecological services provided by its forests, the Mau Complex is a natural
asset of national importance that supports key economic sectors in Rift Valley and western Kenya,
including energy, tourism, agriculture and water supply. The Mau Complex is particularly important for
two of the three largest foreign currency earners: tea and tourism. UNEP (2009), Kenya: Atlas of Our
Changing Environment. Division of Early Warning and Assessment (DEWA), Kenya
Image 1: Large scale destruction of forests and bamboo forests in South West Mau forest reserve
14
Detailed model operation is schematically illustrated in Fig. Fundamental hydrologic equations are Water
balance, Runoff Curve Number method and Muskingum channel routing method.
16
CHAPTER THREE
3.0 RESEARCH METHODOLOGY
3.1 Study Area
3.1.1 Location
The study area shown in the figure below is the Southwest Mau forest reserve a part of the expansive Mau
forest, located southeast of Kericho town. It forms the upper catchment of the Sondu River that drains to
Lake Victoria. The western flanks of the Mau forest complex are influenced by Lake Victoria
macroclimatic region and are generally wetter with rainfall greater than 2000mm and more evenly
distributed with a mean annual temperature minimum of 12 C maximum of 16 C. The soil of the Mau
forest is rich volcanic loam.
The watershed outlet is taken at the Itare gauging station 1JA04 the gauging station used for calibration
of the model is Kiptiget 1JA02 whose position is 35 15E
upper catchment of the river in the Southwest Mau forest reserve.
17
Figure 7:
6: General
general layout
layoutof
ofthe
theentire
entireSondu
Sonduriver
riverbasin
basin
Figure 8: Upper catchment of Sondu river within the Southwest Mau area
18
Type of sensor
period
Bands available
Scene
Spatial resolution
Landsat TM
1973
1-5 and 7
P182r60
28.5m
Landsat ETM
1986
P169r60
1995
P169r60
2002
1-5 and 7
P169r60
2008
28.5m
P169r60
Landsat 1986, 1995, 2002 and 2008 image was in path 169 and row 60
These were in Datum WGS 1984 but different projected coordinate system zones.
2. GPS points of:
3. Scanned Topographic-sheets number 131-2, 131-1, 117-4 and 117-3. These were in the datum
Arc 1960.
4. Vector Data Set-Sondu digitized rivers. This was in the datum Arc 1960.
19
Image 2: Land sat image P182R60 year 1973, showing watershed box
Image 3: Land sat image P169R60 year 1986, showing watershed box
20
Image 4: Land sat image P169R60 year 1995, showing watershed box
Image 5: Land sat image P169R60 year 2002, showing watershed box
21
Image 6: Land sat image P169R60 year 2008, showing watershed box
STATION NO.
GRIDS
9035324
29S
35
38E
9035323
37S
35
12E
9035298
21S
35
31E
9035292
27S
35
29E
9035261
33S
35
3E
KERICHO TIMBILIL
9035244
21S
35
21E
9035241
25S
35
44E
9035079
45S
35
22E
BALLOBRACH,CHEMELIL
9035053
4S
35
9E
9035044
26S
35
8E
9035279
22S
35
16E
22
Meteorological data were obtained from the Kenya Meteorological Department. The data included mean
daily temperature and rainfall data of the years 1973, 1986, 1995, 2002 and 2008.
The mean daily temperatures were obtained from the Kericho meteorological station, station number
9035244 while precipitation data were obtained from volunteer stations within the watershed their station
numbers are 9035044, 9035053, 9035079, 9035241, 9035244, 9035261, 9035292, 9035298, 9035323,
9035324 and 9035330.
Apart from the meteorological data the grid positions of the stations were obtained to map out their
positions in the watershed area. The mapping out is carried out using GIS methods hence the geographic
coordinate system of the stations is very important as alignment and accuracy problems may arise unless
there is correct transformation between geographic coordinate systems.
The digitized topographical sheets used for watershed delineation used the Arc 1960 coordinate system
while the stations used the WGS1984 coordinate system, this information assisted in the accurate
positioning of the meteorological stations.
The meteorological data provide was very consisted for the study periods hence minimized errors of
inconsistent data. The stations are;
23
Constant
Value provided
20.6718
1.2786
Ho
0.3
Hmin
0.37
Above a Hmin, Each segment applies from Hmin until Hmin of the next segment or until Hmax of the rating
curve it belongs to. In our case we had only one segment. The constants are all given. All that is required
is to input the water level (H) for each day and get the discharge (Q) from the equation.
MS excel spreadsheet was utilized in the conversion from gauge heights to discharges.
The gauging stations grid positions were also noted for positioning in the watershed area using GIS their
geographic coordinate system is WGS1984 hence required conversion to Arc 1960 for accuracy in
mapping them out.
The observed streamflow data had inconsistencies but was sufficient enough for calibration purposes.
24
All the shape files used were created and attribute set up in Arc Catalogue
1. Shape files of the area of study i.e. Sondu river watershed was created and digitized from the
scanned topographic sheets (ArcMap).
2. The rivers in the watershed were clipped from the provided data set of rivers (ArcMap).
3. Watershed box was also delineated to cover the entire area of study (ArcMap).
4. Using the Landsat images as the template and the watershed box, the land use classification of the
area was done. Forest area was digitized in all the years to give forest cover of 1973, 1986, 1995,
2002 and 2008. This was also done in ArcMap.
5. The GPS points of meteorological and Kiptiget gauging stations were displayed and projected to
WGS 1984 UTM Zone 36N.
6. In ArcMap a grid dividing the entire watershed box into equal regions was created.
7. Map composites consisting of the grid, clipped rivers, watershed box, delineated watershed,
meteorological stations, Kiptiget gauging station and each of the years forest cover were
developed in ArcMap.
The constants are determined using the hydrological and physiographical characteristics of the
watershed under study (e.g.: average latitude, concentration time, etc...)
25
The parameters relating to the physical phenomena are fixed by studies external to the model
(e.g.: snowmelt parameter, coefficient of correction of precipitations as a function of the altitude,
etc...)
As for the other parameters, no fixed rule is used to determine their value; only experiment and
knowledge of their interactions provide the better adjustment. To determine the values of these
parameters for the first trial, one can take the values used for the simulation of neighboring
watersheds.
The verification of the adjustment of the model is done by analysis of numerical criteria and the analysis
of the results presented on graphs. Then if necessary, some parameters are modified and the simulation is
redone. This process is repeated until satisfactory results are obtained.
To facilitate the adjustment of the parameters and to allow the analysis of the accuracy of the simulations,
the CEQUEAU model computes some numerical criteria and can produce several types of graphs
showing the observed and calculated streamflows. These numerical criteria and the possible types of
graphs will be presented in the section on analytical results
3.3.3.1.2 Adjustment of the parameters by optimization
In the Tools menu of Project window, the CEQUEAU model makes it possible to launch the optimization
program. This program allows finding the value of some parameters of the model, while maximizing or
minimizing an objective function in order to reproduce the observed streamflows with minimal error.
The optimization algorithm (BOTM) based on M.J.D. Powell's method (1964) is used. The main program
and an account of the method are given in Optimization Techniques with FORTRAN (1973).
The program makes it possible to optimize simultaneously twenty parameters chosen amongst a total of
twenty-eight
26
Name
CIN
CVMAR
CVNB
CVNH
CVSB
XINFMA
HINF
8
9
10
11
HINT
HMAR
HNAP
HPOT
12
13
HSOL
HRIMP
14
15
COEP
EVNAP
16
17
18
19
20
TRI
XAA
XIT
COET
EXXKT
Description
Infiltration coefficient in the LOWER ZONE reservoir
Drainage coefficient of the LAKES and MARSHES reservoir
Coefficient of low draining of the LOWER ZONE reservoir
High draining coefficient of the LOWER ZONE reservoir.
Low draining coefficient of the UPPER ZONE reservoir.
Maximum infiltration (mm/time step).
Infiltration threshold of the UPPER ZONE reservoir towards the
LOWER ZONE reservoir (mm).
Intermediate draining threshold of the UPPER ZONE reservoir (mm).
Draining threshold of the LAKES and MARSHES reservoir (mm).
Threshold of high draining of the LOWER ZONE reservoir (mm).
Threshold of water taken away at a potential rate by evapo-transpiration
(mm).
Height of the UPPER ZONE reservoir (mm).
Depth of water required to begin the runoff on impervious surfaces
(mm).
Correction of the precipitations according to altitude (mm/meter/year).
Fraction of the evaporation-transpiration taken in the LOWER ZONE reservoir (from 0.0 to
1.0).
Fraction of impervious area on whole squares (from 0.0 to 1.0).
Exponent of Thornthwaite formula.
Value of Thornthwaite thermal index.
Correction of the temperatures according to altitude (oC/1 000 m).
Adjustment parameter of the transfer coefficients from a partial square
to the next, for a time step of one day.
=( )
( )
eq. (ii)
Where:
NTD: Nash Coefficient;
qoi, qoi and observed flows of day i;
qoi : averages of observed streamflows for the n days used for the calculation of the coefficient .
The Nash coefficient represents the ratio of the residual variance to the variance of the observed
streamflows. Its value is 1 when the simulated streamflows are identical to the observed streamflows. As
the difference between the calculated and observed streamflows increases, the coefficient decrease and
can even become negative.
27
The calibration was carried out at the Kiptiget 1JA02 outlet as the gauging station had sufficient observed
streamflow data. The parameters were used to simulate the streamflow at the Itare gauging station that is
the outlet of the catchment.
The following simulations were conducted with CEQUEAU to evaluate the impact of the change in land
use on streamflow:
1. 1986 was considered as model calibration period and the same parameters used in 1995 as the
validation period.
2. To demonstrate the effect of change in land use/ land cover data on river flow time distribution; was
done by keeping all datasets in the model constant.
The CEQUEAU model was calibrated for 1986 using the model parameters such as curve number
(abstraction coefficient), available soil water capacity, and soil evaporation compensation factor were
adjusted interactively to get reasonable match between observed and simulated mean monthly stream
flow. Streamflow was simulated using the calibrated parameters for the validation period and results were
compared.
28
CHAPTER FOUR
4.0 RESULTS AND DISCUSSIONS
29
After classification of the land sat images the forest area was digitized to give the forest cover in
the watershed area vis--vis the non-forested area for every time period selected. The area is
subdivided into grids to enable spatial analysis.
30
19860101
-1.
80
80
0.15
65.
.75
70.
0.001
37
47
19860330
19860724
19860220
19860323
19860630
19860718
19860629
19861225
0
0
100. 125. 100. 125.
100.
100. 125. 100. 125.
100.
622.89
20.71
191.58
151.97
31
0803
-1.0
0
0.
50.
1.
5.663
11
15
14
15
0
10
28
1
1
-3.0
0.7
3
3
0
0.
0.35
60.
75.
30.
4530.
+14.
0.0
16
11
15
13
0
0
125.
100.
125.
88.02
1138.68
1188.72
1149.86
1273.56
0
125
85.43
1
1.0
24
10.
0.
0.
237.15
Mean monthly simulated streamflow and associated standard deviations at the basin output (1JA04) and
the Kiptiget (1JA02) sub basin was compared well with the observed data. However, simulated values
were lower to observed values at the sub basin stream gauge (1JA02). Monthly flow patterns compared
reasonably well with the pattern of rainfall data used in the simulation, since volunteer weather stations
were used for the entire basin, the spatial variability of rainfall was adequately represented
The purpose of calibration is to adjust the model parameters so that the model closely matches the real
system. The real system is represented by observed discharge data. Adjustment of parameters is essential
to overcome errors due to unknowns and, unmeasured quantities in the basin.
From available observed discharge data at 1JA02 complete datasets were only found in the years 1986
and 1995, hence the year 1986 was used for calibration. Fig.14 shows a comparison of observed
streamflow and simulated. There were larger disparities at high flows but the simulated hydrograph
matched the trend of the observed during low flows. Parameter adjustments resulted in a coefficient of
determination (R2) of 0.63. Coefficient of determination (R2) equals one for a perfect case where
simulated values equal. These results were used to simulate streamflow at the 1JA04 gauging station. The
same parameters were used to demonstrate the effect of change in land use on basin runoff response. The
reason for poor predictability by the model at high flows could be attributed to inaccuracies in
interpolated rainfall due to poorly distributed observation network. Because the correlation is good at low
flows it is reasonable to say that rain gauge positions are not dense enough to capture extents of spatial
rainfall accurately.
Figure 16: Hydrograph of 1973 rainfall on land use of 1973 and 2008
32
Table 6: Change of river flow response with change land use/ land cover data
33
This is expected as the rainfall pattern and magnitude are similar in the two cases leaving any cause of
shift to change in area of different classes land use / land cover, the model parameters are maintained
from calibration stage.
There was a general declining trend in the streamflow with reducing land use cover from 1973 to 2008
and this was coupled with a larger variability in streamflow values between minimum and maximum
discharges.
Figure 17: Comparison of streamflow trends of the Itare and Sondu gauging stations
The discharge at the Itare gauging station 1JA02 was compared against discharge at Sondu gauging
station 1JG03.
34
There is a general trend of reduction in power production against with reducing runoff at the Sondu
gauging station, though this wasnt very consistent as power production is influenced by several other
factors apart from discharge that were not considered in the scope of this research.
36
CHAPTER FIVE
5.0 CONCLUSIONS AND RECOMMENDATIONS
5.1 Conclusions
The objectives which were set for this study were generally achieved to a satisfactory extent taking into
account the constraints of data and available models. The results are summarized as follows:
1. Land sat images were processed and classified the cover change assessed; most of the decline
occurring between (2002-2008) about 11% then (1973-1986) 1.5 % , (1986 and 1995) 2.9%,
(1995-2002) 3%.
2. There has been a declining trend in streamflow discharge over the duration characterized by erratic
flows with high peaks and very low flow discharges, hence a high hydrological variability,
attributable to land use change.
3. Calibration of CEQUEAU model in Sondu basin is accomplished and results has shown a coefficient
of determination of 0.63.
5.2 Recommendations
1. Replanting of trees and agro forestry should be undertaken in the catchment. To improve the
reliability of the streamflow hence of power project.
2. A large watershed model should be used to assess the entire Sondu river catchment as in this study
we are only able to assess the sub catchment within the Southwest Mau area.
3. Gauging stations data should be regularly recorded as this was a major problem in the study since
most gauging stations had data inconsistencies.
4. Analysis of effect of land use / land cover on river regime should be attempted by high spatial
resolution images and close time intervals.
37
REFERENCES
Akotsi, E., Gachanja, M., Ndirangu, J. (2006). Changes in the Forest Cover in Kenyas Five Water
Towers 20032005. Department of Resource Surveys and Remote Sensing, Kenya Forests Working
Group. Nairobi. Kenya Forests Working Group.
Bachmat, Y., Bredehoeft, J., Andrews, B., Holtz, D., and Sebastian, S. (1980). Groundwater management:
the use of numerical models, Water resources monograph, series 5, American Geophysical Union,
Washington, D. C., 127 p
Besbes, M., Duermael, B., Ledoux, E., De Marsily, G., and Talbot, A. (1981). Soutien des Ctiages d'une
riviere par pompage dans la nappe qu'elle draine: mythe ou rCalitt? Bulletin du B.R.G.M. 2 (111).
Crawford, N. G., and Linsley, R. K. (1966). Digital simulation in hydrology: Stanford watershed model
IV. Dept. of Civil Engineering, Stanford University, tech. rep. no. 39, 220 p.
Fleming, G. (1979). Deterministic models in hydrology, FAO irrigation and drainage paper, Rome 80 p.
Freeze, R. A., and Harlan, R. L. (1969). Blueprint for a physically - based digitally - simulated hydrologic
response model, Journal of Hydrology, Vol. 9, 237-25
Landsat satellite images for the years 1973, 1986, 1995, 2002 and 2008.
Knapp, R. M., Green, D. W., Pogge, C., and Stanford, C. (1975). Development and field testing of a basin
hydrology simulator, Water Resources Research Vol. 11(6), 879-888.
Miles, J. C., and Rushton, K. R. (1983). A coupled surface water and groundwater catchment model,
Journal of Hydrology, Vol. 62, 159-177
MORIN, G. (2002). Chapter 13 CEQUEAU hydrological model. p. 507- 576. Singh, V. P. and D. K.
Frevert (Ed.), 2002. Mathematical Models of Large Watershed Hydrology, Water Resources Publications,
Highlands Ranch, CO, 80163-0026 USA, (2002) 914 p.
Mulvaney, T. J. (1851). On the use of self-registering rain and flood gauges in making observations of the
relations of rainfall and of flood discharges in a given catchment, Trans. Inst. Civil Engrs. Ir. (Dublin)
Vol. 4(2): 18.
Nash, J.E. and Sutcliffe, J.V. (1970). Riverflow forecasting through conceptual model. Journal of
Hydrology, 10: 282-290
Ododo, E., Olago, D., Ochola, W. Eds. (2006). Environment for Development: An Ecosystems
Assessment of Lake Victoria Basin. Nairobi, United Nations Environment Programme, Pan African
START Secretariat.
38
Refsgaard, C.J. ,1996. Parameterization, calibration and validation of distributed hydrological models.
Journal of hydrology.198:69-67
Siriwardena, L., Finlayson, B.L., and McMahon, T.A, 2006. The impact of land use change on catchment
hydrology in large catchments: The Comet River, Central Queensland, Australia. Journal of Hydrology
326, 199-214.
Trescott, P. C., Pinder, G. F., and Larson, S. P. (1976). Finite difference model for aquifer simulation in
two dimensions with results of numerical experiments. Techniques of water resources investigations of
the U. S. Geol. SUN. book, 7, chap. cl., 116 p.
UNEP (2009), Kenya: Atlas of Our Changing Environment. Division of Early Warning and Assessment
(DEWA), Kenya
39
APPENDICES
Appendices 1: STEAMFLOW DATA 1986 (sample)
1JG03 MIRIU SONDU- 3447'40.157"E 020'46.869"S
1JA02 KIPTIGET- 3515'24.655"E 033'2.832"S
C -20.6718 20.6718
n -1.2786 1.2786
HO - 0.3
Hmin-0.37
1-Jan-86
0.3
0.3
0.21451
1-Jan-86
0.3
0.3
0.21451
2-Jan-86
0.29
0.29
0.20541
2-Jan-86
0.29
0.29
0.20541
3-Jan-86
0.28
0.28
0.196398
3-Jan-86
0.28
0.28
0.196398
4-Jan-86
0.27
0.27
0.187474
4-Jan-86
0.27
0.27
0.187474
5-Jan-86
0.27
0.27
0.187474
5-Jan-86
0.26
0.26
0.178643
6-Jan-86
0.26
0.26
0.178643
6-Jan-86
0.26
0.26
0.178643
7-Jan-86
0.26
0.26
0.178643
7-Jan-86
0.26
0.26
0.178643
8-Jan-86
0.27
0.27
0.187474
8-Jan-86
0.27
0.27
0.187474
9-Jan-86
9-Jan-86
0.26
0.26
0.26
0.26
0.178643
0.178643
40
4.4343
05
4.4343
05
4.2461
99
4.2461
99
4.0598
93
4.0598
93
3.8754
31
3.8754
31
3.8754
31
3.6928
63
3.6928
63
3.6928
63
3.6928
63
3.6928
63
3.8754
31
3.8754
31
3.6928
63
3.6928
10-Jan-86
0.26
0.26
0.178643
10-Jan-86
0.26
0.26
0.178643
11-Jan-86
0.26
0.26
0.178643
11-Jan-86
0.26
0.26
0.178643
12-Jan-86
0.26
0.26
0.178643
12-Jan-86
0.26
0.26
0.178643
13-Jan-86
0.26
0.26
0.178643
13-Jan-86
0.26
0.26
0.178643
14-Jan-86
0.25
0.25
0.169905
14-Jan-86
0.25
0.25
0.169905
15-Jan-86
0.25
0.25
0.169905
41
63
3.6928
63
3.6928
63
3.6928
63
3.6928
63
3.6928
63
3.6928
63
3.6928
63
3.6928
63
3.5122
42
3.5122
42
3.5122
42
Appendices 2:
sondu miriu
power
generation
Date
(Month
)
Oct-07
Nov-07
Dec-07
Jan-08
Feb-08
Mar-08
Apr-08
May-08
Jun-08
Jul-08
Aug-08
Sep-08
Oct-08
Nov-08
Dec-08
Jan-09
Feb-09
Mar-09
Apr-09
May-09
TGU
(kWh)
5,563,987
7,201,000
11,395,000
5,344,000
3,403,000
13,404,000
29,572,000
35,549,000
38,741,000
30,016,000
40,777,000
40,432,000
41,948,000
40,251,000
30,507,000
14,295,000
8,600,000
5,135,000
17,333,000
36,816,000
(MWh)
5563.987
7201
11395
5344
3403
13404
29572
35549
38741
30016
40777
40432
41948
40251
30507
14295
8600
5135
17333
36816
Jun-09
27,039,000
27039
11464,000
16,404,000
21,113,000
22,755,000
16,582,000
18,442,000
39,092,000
21,862,000
14164
16404
21113
22755
16582
18442
39092
21862
Jul-09
Aug-09
Sep-09
Oct-09
Nov-09
Dec-09
Jan-10
Feb-10
42
KAMINJEIWET
SECONDARY SCHOOL
KAMINJEIWET
SECONDARY SCHOOL
LITEIN MISSION KERICHO
KOIWA ESTATE - KERICHO
KOIWA ESTATE - KERICHO
KOIWA ESTATE - KERICHO
KAPLONG GIRLS
SECONDARY SCHOOL
KAPLONG GIRLS
SECONDARY SCHOOL
KERICHO
METEOROLOGICAL
STATION
KERICHO
METEOROLOGICAL
STATION
KERICHO
METEOROLOGICAL
STATION
KERICHO
METEOROLOGICAL
STATION
KERICHO
METEOROLOGICAL
STATION
AROKET TEA ESTATE
SOTIK
AROKET TEA ESTATE
SOTIK
Precipitation;
daily total
Precipitation;
daily total
Precipitation;
daily total
Precipitation;
daily total
Precipitation;
daily total
Precipitation;
daily total
Precipitation;
daily total
Precipitation;
daily total
20
03
20
08
19
73
19
73
19
83
19
93
19
73
19
83
Precipitation;
daily total
19
73
Precipitation;
daily total
19
83
60.7
45
64.1
24
6.7
21
2.9
38
1.7
22
7.3
Precipitation;
daily total
19
93
145
126.
6
38.8
18
3.6
28
2.2
21
3
15
0.9
Precipitation;
daily total
20
03
16.2
25
191
33
5.6
34
2.4
15
1.6
12
9.1
Precipitation;
daily total
Precipitation;
daily total
Precipitation;
daily total
20
08
19
83
19
93
42.3
67.5
294
19
5
20
8.9
13
7.8
83
160.
7
91.3
79.5
20
3.3
57
82.
4
16
6.1
13
3
68.3
64.9
47.
6
18
1.5
26
3.2
44
43
90.6
49.6
151.2
30
6
27
6.4
13
3.7
70.2
147.
2
140.
4
109.
3
195.
2
48.4
416.8
45
2.7
17
2.3
19
0.7
70.4
225.
9
7.8
15
6
16
9.2
59.
3
8.1
19
4.5
35
0.1
14
0.1
66.6
154.
6
89.4
26
1.8
24
6.5
15
4.4
74.5
21
4.1
24
5.1
16
7.2
22.2
23.1
40.3
44
6
32
0
27
5
36.1
81.9
55.7
24
2.9
88.
3
67.
7
12
7.7
24
6.6
31.
8
88.
7
10
9.4
96.
7
19.
1
12
6.5
95.
2
19860101
-1.
80
80
0.15
65.
.75
70.
0.001
37
47
19860330
19860724
19860220
19860323
19860630
19860718
19860629
19861225
0
0
100. 125. 100. 125.
100.
100. 125. 100. 125.
100.
622.89
20.71
191.58
151.97
44
0803
-1.0
0
0.
50.
1.
5.663
11
15
14
15
0
10
28
1
1
-3.0
0.7
3
3
0
0.
0.35
60.
75.
30.
4530.
+14.
0.0
16
11
15
13
0
0
125.
100.
125.
88.02
1138.68
1188.72
1149.86
1273.56
0
125
85.43
1
1.0
24
10.
0.
0.
237.15
Appendices 5: Changes of river flow response with change land use/ land
cover data
DEC. 1973 (day
334 to day 365)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Totals
45