Gumara Watershed and Stream link - Data Availability And Analysis

3.2. Data Availability And Analysis

3.2.3. Gumara Watershed and Stream link

The process of delineating watersheds by using DEM isreferred to as terrain pre-processing(Merwade 2012). In this study, the watershed was delineatedusing the

"Hydrology" tool within the Spatial Analyst tools in Arc Toolbox. Stream network can be defined from DEM using the output fromFlow Accumulation function.In this study 1% of the maximum flow accumulation was used to define the stream, which is 0.01*1343697 (13436) used as a threshold value to define the stream in the watershed. When threshold value becomes smaller in number perennial as well as non perennial streams was defined through the watershed. Selecting the hydropower site on non perennial streams require the construction of dam for long term storage. But in this study micro to small hydropower potential was assed. This much potential more of the time addressed with the help of runoff river type of hydropower plant, which is constructed on Perennial River to

get power throughout the year. So the threshold value must amplify the streams which flow throughout the year.

Figure 3.8: Gumara sub- Watershed and Stream link 3.2.4. METEOROLOGICAL DATA

3.2.4.1. Rainfall

The relevant data in the research area is precipitation in the form of rainfall. Monthly precipitation data of 13 stations namely Bahirdar,Wereta, Addiszemen, Amedber, Arbgebaye, Amebesami, Lewaye, Hamusit, Zenzelima, Mekaneyesus, Wanzaye and Yifagehave been collected from the National Meteorological services Agency of Ethiopia from Baherdar branch. The rainfall data obtained from such nearest station to the Gumera catchment should be taken in to consideration to develop Rainfall grid map of the catchment. The respective distance of these rainfalls gauging station with respect to the respective base station was obtained by using GIS from Rain gauge stations map and authenticated by their respective coordinate system. There is an intermittent or break of data for a short period of time or even for a particular month. Missing rainfall data with in such gauging stations have been calculated and even also their consistency and

homogeneity was also checked. The selected meteorological stations are also presented in Table 3.5and Figure 3.9.

Table3.5: Selected Meteorological Stations

Station Name

Station Class

Period of

Record Degree Altitude

Years Easting Northing m

yifage 3^rd 2003-2016 37.60 12.10 2020

Amebesami 3^rd 2008-2016 37.625 11.700 2076

AmedBer 3^rd 1984-2015 37.886 11.914 2051

Wanzaye 3^rd 2003-2015 37.631 11.758 1928

Mekaneyasus 3^rd 2003-2016 38.054 11.608 2374

Bahirdar 1^st 1961-2016 37.360 11.680 1800

Debratabor 1^st 1951-2016 37.995 11.867 2612

Lewaye 4^th 2003-2016 38.072 11.720 2709

Hamusit 3^rd 2003-2016 37.562 11.783 1930

Werata 3^rd 1978-2015 37.696 11.922 1819

Addiszemen 3^rd 1997-2016 37.773 12.117 1940

Arbegebaye 4^th 2006-2015 37.750 11.636 2228

Zenzalim 3^rd 2008-2016 37.462 11.625 1910

Source: - Information collected from Metrological Agency Bahirdar Branch.

Figure 3.9: Location of selected Rain gauge station

From the above stations Lewaye, Arebegebaye and Wnzaye were located within the watershed, the other stations located around the watershed. The estimator stations for the base stations ware selected based on their relative distance from the Base stations.

3.2.5. ESTIMATING MISSING PRECIPITATION DATA

Measured precipitation data are important to many problems in hydrologic analysis and design. The record at many rain gauge stations may consist of short breaks due to several reasons such as absence of the observer, instrumental failures etc. it is better to estimate these missing records and fill the gaps rather than to leave them. This applies especially when data is processed with automatic equipment like an electronic computer (Reddy 2005)

A number of methods have been proposed for estimating missing rainfall data. The Arithmetic Mean Method is the simplest method. The normal-ratio method is based on the weights on the mean annual rainfall at each gage and Distance power method is based on the weights on distance of each estimator station from base station.

In this study the mean monthly rainfall values have been determined and the missing monthly rainfall data have been filled using Simple Arithmetic mean, Normal Ratio method and Distance power method.

3.2.5.1. Arithmetic Mean Method

If the normal annual precipitations at the adjacent stations are within 10% of the normal rainfall of the station under consideration, then the missing rainfall data may be estimated as a simple arithmetic average of the rainfalls at the adjacent gauges. Thus, if the missing precipitation at station X is and are the rainfalls at the n surrounding stations.(Reddy 2005)

( )

Usually the data from three surrounding gauges will give good results.

3.2.5.2. Normal - Ratio Method

If the normal annual rainfalls at the surrounding gauging stations differ from the normal annual rainfall of the station in question by more than 10%, the normal ratio method is preferred. In this method, the rainfall values at the surrounding stations are weighted by the ratio of the normal annual rainfalls. The general formula for computing is:

[ ]

Where: is the normal annual rainfall at station X and , are the normal annual rainfalls at the m surrounding stations respectively. A minimum of three surrounding stations are generally used in the normal ratio method(Reddy 2005)

3.2.5.3. Distance power method

The rainfall at a station is estimated as a weighted average of the observed rainfall at theneighboring stations. The weights are equal to the reciprocal of the distance or some power of the reciprocal of the distance of the estimator stations from the estimated stations. Let Di be the distance of the estimator station from the estimated station. If the weights are an inverse square of distance, the estimated rainfall at station A is:

P_A =^∑

∑ 3.3

Note that the weights go on reducing with distance and approach zero at large distances.

A major shortcoming of this method is that the orographic features and spatial distribution of the variables are not considered. The extra information, if stations are close to each other, is not properly used. The distance of each estimator station from the estimated station whose data is to be estimated is computed with the help of the coordinates using the formula:

Di² =,( ) ( ) -3.4

Both the mean monthly rainfall values and the summarized annual rainfall values in mm are given in Appendix – IIB. The graphs showing the monthly and yearly variability of rainfall at Bahirdar, which have longest series of data and Lewaye, which is the closer station to the watershed are given below. For other stations, the graphs are attached in Appendix – IIB.

Figure 3.10: Mean monthly Rainfall at Bahirdar station

Figure 3.11: Annual Rainfall at Bahirdar station

Figure 3.12: Mean monthly rainfall for lewaye station.

0.0

JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC

Rainfall in mm

months

mean monthly variability of Rainfall at Bahedar

Monthly Variability of Rainfall for Bahirdar Station

0.0 500.0 1000.0 1500.0 2000.0

1997 1999 2001 2003 2005 2007 2009 2011 2013 2015

Rainfall in mm

Years

Yearly Variability of Rainfall for Bahirdar Station

yearly variability of

JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC

Rainfall in mm

months

Mean monthly Rainfall variability for Lewaye

Mean monthly Rainfall of Lewaye

Figure 3.13: Annual average Rainfall for Lewaye station

The figures given above indicate that the rainfall characteristics are a bimodalrainfall pattern. The main rainy season among the above given stations are from July to measured variables plotted as a double-mass curve may give indefinite resultsbecause we may be unable to say which of the variables caused a break in slope. The pattern, which is composed of the average of many records, is less affected by an inconsistency in the record of any one station. After constructing the double mass curve, it was found that there is no inconsistency observed for all stations. The Double Mass Curve for Lewaye (M/eyasus and D/tabor as Base station), D/tabor (Lewaye and Amedber as a Base station) and Wnzaye (Hamusit, Werata and A/gebaye are base station), which are the nearest Raingauge station for Gumera Watershed, were shown below. The double Mass Curves for the rest of stations were plotted in Appendix III.

0.0

2003 2005 2007 2009 2011 2013 2015

Annual average variability of Rainfall for Lewaye Station

Annual average variability of Rainfall at Lewaye

Year

Rainfall mm

Figure 3.14: Double Mass Curve plot for Lewaye Station

Figure 3.15: Double Mass curve for D/tabor station

Figure 3.16: Double mass Curve for Wanzaye Station

R² = 0.9974

0 5000 10000 15000 20000 25000

Cummulative Annual Rainfall for lewaye in( mm)

Cummulative annual Rainfall for pattern(Base stations in mm)

Double Mass curve for Lewaye

Double mass curve for Lewaye station

R² = 0.9993

0 5000 10000 15000 20000 25000

Cummulative Annual Rainfall D/tabor station in mm

Cummulative annual Rainfall for pattern(Base station)in mm

Double Mass Curve for D/tabor station

Double mass Curve for Debratabor Station

3.2.7. HOMOGENEITY CHECKING FOR RAINFALL STATIONS

Checking homogeneity of group stations is essential. The homogeneity of the selected base gauging stations average monthly rainfall records were made to be non-dimensional using equation.

Where: - - Non - dimensional value of rainfall for month i

𝑖 - Over years - averaged monthly rainfall at the station i - The over year - averaged yearly rainfall of the station

Figure 3.17: Homogeneity Test for Lewaye’s Base Station

Figure 3.18: Homogeneity Test for M/eyasus‟s Base Station

Figure 3.19: Homogeneity Test for Wanzaye‟s Base Station

As shown in the above figure there is extreme homogeneous nature between the Estimator station of Lewaye (M/eyasus and D/tabor) and Lewaye, Estimator station of M/eyasus(Lewaye and A/gebaye) and M/eyasus as well as Estimator station of Wanzaye (Hamusit,Werata,Ambesami) and Wanzaye. The Homogeneity natures of the other stations are attached in Appendix IV.

3.2.8 STREAM FLOW DATA

Stream flow data is one of the most important and predominant data for hydropower potential analysis of the Gumera river basin. Stream flow records are among the most valuable of all hydrologic data because they represent an integration of all hydrologic factors.Furthermore, the flow of streams is a sensitive indicator of climatic variations, because runoff is the residual of precipitation after the demands for evapotranspiration are satisfied. With evapotranspiration losses fairly constant from year to year in a given area, variations in annual runoff are much greater in percentage than variations in annual precipitation. In this study monthly stream flow data of Gumera river basin for 40 years was used from the Gumara gauging station. From the total order of data, 33 in number or 7.05% of data were missed. Therefore, to fillmissed data a simple Arithmetic average method from its data set is applied. The filled stream flow data of Gumara wasattached in appendix v.

3.2.8.1. Checking consistence and homogeneity of stream flow data

Before the data was used as an input for any types of process, its consistency and homogeneity must be checked. To detect the presence of inconsistencies or non – homogeneities in Gumera river flow data , split record tests on Variances and Means are applied (from the total order of 40 yearly observational data, in one group 20 number of data set and in the next group 20 number of data set were prepared for checking stability of variance and mean). Among different test types, simple but efficient procedures for screening the data in annual time series are selected. These are F-test for stability of Variances, t-test for stability of Mean and Spearman`s rank correlation test for indicating absence of trends.

3.2.8.2. Test for absence of trend by using plotting the data and spearman`s rank correlation method

To authenticate the absence of trend or discontinuity of the given time series stream flow data of Gumera river basin, plotting average annual stream flow data with respect to duration at which the flow occurred.

A).plotting the time series data

Figure 3.20: Gumera river stream flow trend analysis graph

The figure shows, Gumera river Annual average stream flow data for 40 year time series, has some discontinuities or trends. This is due to the value of mean monthly rainfall between (1997-2001) was out layer in some extent. But this out layer is due to the fact that, there is a vagary of environment to some extent (from the office of Abay Basin Authority Bahirdar Branch). Despite the fact that there is an out layer to some extent, the recorded data taken as it is without adjustment, because the change is due to vagary of environment.

b) Spearman`s Rank-correlation coefficient Method

The Spearman`s Rank-correlation coefficient, Rsp is expressed as:

RSP = 1-^∑

( ) 3.6

Where n is the total number of data, D is difference, and i is the chronological order number.The difference between rankings is computed as:

D_i =K_x – K_y3.7

Where Kx, is the rank of the variable, x, which is the chronological order number of the observations. The series of observations, y, is transformed to its rank equivalent, Ky, by assigning the chronological order number of an observation in the original series to the corresponding order number in the ranked series, y.

t

= R

SP*

3.8

Where has Student‟s distribution with v = n-2 degrees of freedom. Student‟s t-distribution is symmetrical around t = O. AppendixIXcontains a table of the percentile points of the t-distribution for a significance level of 5 per cent (two-tailed). At a significance level of 5 percent (two-tailed), the time series has no trend if:

t* +< tt < t* +3.9

The table which show the computation procedure of Spearman`s Rank-correlation coefficient Method for checking the trend of long term Stream flow data was attached in appendixIB

The table of percentile points for the t-distribution (Appendix IX) gives the critical Values at the 5-percent level of significance for 40 - 2 = 38 degrees of freedom as:

Table 3.6: percentile points for t-distribution t {(v,2.5%)} tt t {(v,97.5%)}

-2.02 0.314 2.02

When checking this result against the permissible Range in Equation 3.9, the condition is satisfied. Thus, the time series doesn't have trend. Therefore, it is possible to use the data for further analysis.

3.2.8.3. Checking the stability of Mean and variance

a

) Checking the stability of Variance using F-test

The Fisher Distribution, F-test is the statistical test which is analyzed the ratio of the Variance of two split, non-over lapping, and sub-sets of the time series. In order to carry over the F-test, the River flow time series are grouped (split) in to two non-over lapping sub-sets (i.e. 1 - 20 and 21- 40).

The Fisher Distribution test statistically expressed as:

Ft =

3.10

Where S² is variance and its convenient formula for computing the sample Standard deviation, S is expressed as follow:-

S =^∑ ( ) variance; the alternate hypothesis is H1: S12

<> S22

The variance of the time series is stable, and one can use the sample standard deviation, s, as an estimate of the population standard deviation if;

F V1, V2, 2.5% < Ft< V1, V2, 97.5

Where V1 = n1-1, is the number of degree of freedom for the nominator.

V₂= n_2-1,is the number of degree of freedom for the denominator. n₁ and n₂ are the number of data in each sub-set. The F-distribution is not symmetrical for v1 and v2. One should therefore enterTables properly, usually by taking v₁ horizontally and v₂ vertically.

(See AppendixXIfor a table of the F-distribution F{V_1,V_2, P}for the 5-per-cent level of significance (two- tailed).

b) Checking the Stability of Mean by using t-test

In order to compute the t-test for the Mean stability, the same sub-sets from F-test (for stability of variance test) are used. A suitable statistic for testing the null hypothesis, HO; X1 = X2.,Against the alternative hypothesis, H1; X1<> X2.

t

=

( ) ( )

(

)

3.12

Where, n is the number of observation in subset, x is the mean of the subset and s is the variance.

The Mean of the time series is considered stable if:

t {v,2.5%} <tt< t{v,97.5%}

The computation of stability of Variance, Ft and stability of Mean, tt, by using the same two sub-sets of Gumera River Mean monthly stream flow data is attached in appendix V.

Fisher testComputation with referring tables of appendixIX.

Table 3.7: percentile points for F-test

F {v₁,V₂, 2.5%} 0.406

F(t) 0.39982477

F {v₁,V₂,97.5%} 2.46

The result of variance stability and Mean stability analysis shows that,Ftand ttvalues are being within the permissible stable range. So that the variance and mean of the time series stream flow data of Gumera river basin have been stable at 5% significance level.

Based on the consistency and homogeneity analysis were preformed in the above, the

data has been Consistence and Homogeneous. So that, it is possible to utilize stream flow data for further studies of hydropower potential assessment on Gumara River basin.

3.2.9. FLOW DATA ANALYSIS

3.2.9.1. Transfer data to Runoff Factor estimator sites

Even though hydrometric stations are available in a river basin, usually it is not common for these gauges to be located precisely at rivers confluence and site of interest. The most recommended guideline to transfer stream flow data to the point of interest is to use area ratio methods described by equation 3.13. This method uses the drainage areas to interpolate flow values between or near gauged sites on the same stream. Flow values are transferred from a gauged site, either upstream or downstream to the ungauged site (Admasu 2000)

(

)

Where: Discharge at the Site of Interest

Discharge at the gauge site

Drainage area at the site of interest

Drainage area at the gauging site n Varies between 0.6 – 1.2

If the ungauged Area is within 20% of the gauge Area (0.8 ≤ (Agauge/A_ungauge) ≤ 1.2) then (n=1) to be used. The estimated discharge at the site will be within 10% of actual discharge(Awulachew 2007). When ungauged A is within 50% of the gauge A, two station data are considered for data transferring. Relation can be developed to estimate a weighted average flow at a site lying between upstream and downstream gauges.

( ) ( ₎

( )

Where: - Gauge1 upstream gauging site and gauge2 downstream gauging site.

When the site of interest is more than 50% of the area of gauge, estimate the value of n from annual flow data in the basin. The ratio of average annual discharge at the site

(estimated) and at gauge (recorded) can be used to transform the flow duration curve from gauge to the site of interest. The stream flow data of Gumara river basin at Gauged station transferred to the un Gauge (Runoff Factor and Parametric duration curve estimator sites) was done by area ratio method for those sites suitable for area ratio method and by Regionalization of monthly flow characteristics using GIS and SpatialInterpolation Algorithmfor those sites deviate from the requirements of area ratio method. Runoff factor estimator Site number 2, 4, 7 & gauge station are also used as parametric duration curve estimator sites. Estimator sites for development of representative Runoff factor and parametric duration curve equation was selected based on the criteria of area ratio method and have relatively the same land use, soil type, main stream slope, topography, morphology and Drainage area and even also projected along the same stream line. The drainage area ratio between the gauged site and the estimated sites for Runoff Factor are 1.015, 1.017, 1.026, 1.042, 1.049, 1.15, and 1.18 respectively from the downstream to upstream in the Gumara watershed. Most of these sites found at the downstream side of Gumara watershed in order to detect the effect of whole watershed Runoff Formation factors upstream of estimator points. This estimator points lay on relatively lower elevation of Gumara watershed. Due to this reason it is batter to analysis the runoff formation factors, which relate rainfall and runoff of among part of the watershed in the upstream side was detected through these points. The distance between the gauged station and the last estimated site (estimator station -7) is 25.6km following the natural River structure. The distance between the estimated stations arranged according to the topography of the watershed. More distance was given for those sites, which have drainage area similar to the topography of drainage area of gauging station. The topography of Gumara is extremely flat at the downstream side of the watershed. So that the drainage area ratio of gauged site to estimated sites, which are lay on 25 km River stretch starting from the outlet, are in between 0.8 and 1.2 or within 20% of the gauge station. This implies that the drainage area difference of extremely flat surface is not that much huge within a considerable distance. Figure 3.21and table 3.8

In document GIS BASED ASSESSMENT OF HYDROPOWER POTENTIAL (A case study on Gumara River Basin, Ethiopia) (Page 48-0)