Evaluating Discharge Capacity of Major Chara’s of Sylhet City Using GIS

Heavy rainfall is one of the most frequent and widespread severe weather hazards that affect Sylhet city. In Sylhet, nine natural channels locally called “chara” are mainly responsible for the drainage of heavy rainfall to the Surma River. The present condition of this natural drainage is not feasible due to the unplanned land development and manipulation by people. So water logging is a common scenario of Sylhet city. Many parts of Sylhet city are experiencing the severe inundation problem due to heavy rainfall. Time series analysis of rainfall data (1974-2015) is important for knowing and predicting of rainfall variation. Mann-Kendall analysis showed no trend for monthly and yearly rainfall data. The rainfall intensity of Sylhet is higher than other districts of Bangladesh. In this study, IDF (Intensity Duration Frequency) curve has been developed that is commonly used in engineering planning and design. By using IDF curve, ArcGIS software, DEM map, and normal discharge, maximum discharge for a different point of Mongoli chara and Bolramer chara has been calculated (for 25 years return period). Discharge through Mongoli chara and Bolramer chara has been represented in ArcMap using ordinary kriging. To keep the catchment of those chara free from inundation and water logging problem, this calculated discharge needed to be managed.


Introduction
Bangladesh is affected by various types of natural disasters almost every year because of the global warming as well as climate change impacts. Bangladesh is a low-lying country with more than 230 waterways, is one of the most disas- series. The purpose of the Mann-Kendall (MK) test [6] is to statistically assess if there is a monotonic increase or decrease trend of the rainfall of 1974 to 2015 [7].

Development of Short Duration Rainfall Data
Short duration data are rare in development countries like Bangladesh. The data of 24-hour rainfall records are available in rain gauge stations.
[ ] 1 3 24 24 t P P t = (1) Journal of Water Resource and Protection where, P t is the required rainfall depth in mm at t hr. duration, P 24 is the daily rainfall in mm and t is the duration of rainfall for which the rainfall depth is required in hr.
For estimation of various duration like 1-hr rainfall values from annual maximum daily rainfall data, Chowdhury et al. [8], used Indian Meteorological Department (IMD) empirical reduction formula to estimate the short duration rainfall from daily rainfall data in Sylhet city and found that this formula gives the best estimation of short duration rainfall [9]. In this study, this empirical formula was used to estimate the short duration rainfall in SCC.

Probability Distribution and Chi-Square Test
To identify a specific theoretical distribution for the available data it is important to find an acceptable method. The aim of the test is to find how good a fit is the observed and the predicted data. Chi-square is one of the most widely used tests to find the best fit theoretical distribution of any specific dataset which is represented by this formula [10].
( ) where, i O and i E represent the observed and expected frequencies respectively. If the observed frequencies are close to the corresponding expected frequencies, the 2 X value will be small, indicating a good fit; otherwise it will be a poor fit. Chi-Square test was performed by sigma magic software.

Short Duration for Gumbel Distribution
Short duration rainfall data such as 5 min, 10 min, 15 where, T t P is the depth of t-minute, T-year return period rainfall, 10 t P is the depth of a t-minute, 10-year return period rainfall and 60 T P is the depth of a 60-minute, T-year return period rainfall.

Development of Rainfall Intensity Duration Frequency (IDF) Curve
IDF curve is used for estimating the maximum rainfall intensity for the different duration and return period. A specified return period T and duration t is calculated for the rainfall depth. Its mean intensity m I (mm/hour) has been obtained dividing it by the duration t (hour). Then the IDF curve is obtain by plotting, on a graph, the main intensity m I (mm/hour) against the duration D (min). Journal of Water Resource and Protection Mathematically, this curve can be represented in different forms as follows [9]: where, i is rainfall intensity in mm/hr, D is duration of rainfall in minutes, the parameters a, b, n, x, y define the shape and appropriate units for curve fitting to IDF data.

Catchment Area of Chara
The catchment area of Mongoli Chara and Bolramer Chara has been identified through field survey and contour map. The direction of runoff flow has been identified during field survey and consulting local people. And then the catchment area was divided into some regions.

Runoff Coefficient
The runoff coefficient C is the variable of the rational method. The runoff coefficient (C) is a dimensionless coefficient relating the amount of runoff to the amount of precipitation received. It is a larger value for areas with low infiltration and high runoff, and lower for permeable, well-vegetated areas [11]. Where watershed is not homogeneous, a weighted runoff coefficient should be determined. A weighted runoff coefficient is computed using the following equation where, j A is the area for land cover j, j C is the runoff coefficient for area j, N is the number of distinct land covers within the watershed, and w C is the weighted runoff coefficient.

Time of Concentration
Time of concentration ( c T ) is the time required for runoff to travel from the hy- shorter the c T , the larger the peak discharge [14].
( ) where, C is the cover factor, L is the hydraulic length in ft, and S is the slope in Percent.

Discharge Due to Rainfall
The rational formula has developed the relation between rainfall and discharge.
Discharge due to rainfall calculated by this formula [15].
where, r Q : Peak discharge (m 3 /s) C: Runoff Coefficient. i: Rainfall intensity during time of concentration (mm/hr). A: Drainage Area (ha).

Discharge Due to Population
There are many formulas for population prediction. The geometric mean increase is used to find out the future increment in population. Assuming that, the future population is increasing at a constant rate. Geometric increase formula [16] is: Here, n p = population after n year, o p = Present population, r = rate of growth and n = number of year.
Discharge Due to Population [14]: where, P Q = Discharge due to population, X = Per capita discharge and P = Total population.

Monthly Rainfall Trend Analysis
There are several statistical tests available for testing stationeries of time series.
In the present study, the Man-Kendell test (Table 1) for linear trend has been carried out for the monthly rainfall series from 1974 to 2015 of Sylhet. There is no trend observed in monthly rainfall data series to find the best fit

Chi-Square Test
Chi-Square test has been performed by 1-hr rainfall data that shown in Table 2.

Development of Rainfall Intensity Duration
Within same return period rainfall depth depends on rainfall duration. Rainfall depth is increased with the increase of rainfall duration. In 2 year return period for 5 min and 120 min rainfall depth is 20.55474 mm and 86.05012 mm respectively. Average intensities of rainfall for different return periods have been calculated by dividing t minute rainfall depths by the corresponding duration of t (hour) ( Table 4).

Intensity-Duration-Frequency (IDF) Curve
Short duration rainfall data such as 5 min, 10

Derivation of IDF Equation
To derive an equation for calculating rainfall intensity for the regions of interest  where it is possible to convert the equation into a power equation (Table 5), and thus to calculate all the parameters related to the equation. Table 5. Rainfall IDF curve equation and regression value for Sylhet city.

Catchment Area Selection & Dividing into Sub-Areas
Catchment Area of was selected by using "Polygon" tools and then converted it to "Shape file" by using "Feature to Polygon" tools of ArcGIS. Separation into sub-area was done by the same process ( Figure 3).

Development of Digital Elevation Model (DEM)
From Elevation map of Sylhet city DEM was created by using "geo-statistical analysis" tools of ArcGIS (Figure 4). Journal of Water Resource and Protection

Time of Concentration
Time of concentration ( c T ) of six zone was estimated by Federal Aviation Agency (FAA) formula ( Table 6). Length of flow and average slope was found from ArcMap. Length of flow and average slope is calculated from Shape file and DEM map. Time of concentration is highest at zone 3 and lowest at zone 6.

Discharge Due to Rainfall
Discharge was calculated by the rational formula. Rainfall intensity has been calculated from IDF curve for the duration equal to the time of concentration. 25 year return period has been selected for calculating rainfall calculated discharge has been given in below. Drainage area is calculated from Shape file. At zone 1 discharge is maximum as it has largest area, at zone 5 discharge is minimum as its catchment area is lowest. Peak discharge is the sum of discharge of this zone and peak discharge of previous zone. Peak discharge at zone 4 is the sum of discharge of zone 4, zone 6 and peak discharge of zone 3.

Discharge Due to Population
Averagely per capita use of water 100 -120 liter/day (PHE) has been considered. Meantime, discharge due to population was measured. Table 7 shows at zone 1 discharge is maximum, at zone 5 discharges is minimum. Peak discharge is the sum of discharge of this zone and peak discharge of previous zone. Peak discharge at zone 4 is the sum of discharge of zone 4, zone 6 and peak discharge of zone 3 (Table 8).

Total Discharge
Total discharge is the sum of discharge due to rainfall and discharge due to the population. Table 9 shows the peak discharge for different zones, which is maximum (60.87 m 3 /s) at zone 5 and minimum (8.58 m 3 /s) at zone 6.

Discharge in Chara
From the equation the corresponding discharge at different point of Chara in m 3 /s was obtained and plot them in the ArcMap software by using "kriging" tools of ArcGIS. After that, Geostatic analysis was carried out with Kriging which gave the following map ( Figure 5).

Existing and Required X-Sectional Area
The required cross-sectional area has been calculated which is shown in Table   10. Before meeting with Bolramer chara the existing x-section of Mongoli chara satisfy the required cross section. But after meeting with Bolramer chara the existing cross section is not sufficient.

Conclusions
Based on the result of the study, following conclusions were made: • There is no trend in rainfall that could be detected at the 5% significance level in Sylhet.