2.1 Study Area
Melamchi-Indrawati watershed consists of three main tributaries: Melamchi, Yangri and Larke in Koshi River Basin. Melamchi catchment has an area of 324 km2 and very steep river gradient and the shortest (41km) from Himalaya origin i.e. from 5800m at its highest points to 773masl at the confluence of Melamchi and Indrawati River. The topographic variation of the Melamchi-Indrawati catchment is diverse which ranges from 629 to 6075masl and the Melamchi river stretches very steep gradient and high flow velocity, river section is narrow and deep with a slope variation of between 17–8% (Pandey et al. 2021). The mean annual flow of the Melamchi River is 9.7m3/s and received more than daily 12mm of rainfall during monsoon season (DHM 2021). The forest is the dominant land use type in the basin followed by agricultural land and grassland. The snow/glacier land occupied 35.91km2 which is 2.92% of total land coverage in the Melamchi-Indrawati basin (Uddin et al, 2015). Mainly, eight numbers of hydro-meteorological stations and locations from Dumredovan to Bhemathang (damming area) were covered for this study (Fig. 1; Table 1).
Table.1 Hydro-meteorological stations detail in Melamchi-Indrawati Watershed
SN | Name of Station | Index No | Latitude | Longitude | Altitude (m) | Remarks |
1 | Nawalpur | s1008 | 27.8130 | 85.6241 | 1653 | Precipitation |
2 | Sermathang | s1016 | 27.9445 | 85.5951 | 2574 | Precipitation |
3 | Duwachaur | s1017 | 27.8571 | 85.5663 | 1481 | Precipitation |
4 | Tarkeghang | s1058 | 27.9993 | 85.5544 | 2596 | Precipitation/Tem |
5 | Dhap | s1025 | 27.9124 | 85.6333 | 1284 | Precipitation |
6 | Nakote Station | s 627.5 | 28.0108 | 85.5353 | 1750 | Discharge |
7 | Bhaunepati | s1018 | 27.7924 | 85.5726 | 774 | Precipitation |
8 | Ganjala, ICIMOD | SnowAMP | 28.1545 | 85.5625 | 4962 | Ppt/Tem/Snow depth |
The best available daily and hourly precipitation, temperature, snow depth (Ganjala, SnowAMP), discharge (Nakote) data from the Melamchi-Indrawati basin were collected from the hydro-meteorological stations (Table 1). The Ganzala pass AWS station (Index: SnowAMP Ganja La) is located at the upper part of the Melamchi watershed at an elevation of 4962masl which was installed as part of the snow accumulation and melt process in the Himalayan catchment (Saloranta et al. 2019) and Bhaunepati (s1018) is located at the lower altitudinal region below Melamchi-Indrawati confluence at an altitude of 774masl.
2.2 Data Used
The daily and hourly precipitation data of hydro-meteorological stations of the basin during 1992–2021; gauge height and 10m interval discharge data of Nakote during flood time were collected from the Department of Hydrology and Meteorology, Government of Nepal. The hourly temperature, precipitation and best available snow depth data of Ganzala pass AWS station (Index: SnowAMP Ganja La) were collected from ICIMOD (https://rds.icimod.org). The collected data was carefully examined, cleaned, and organized to ensure accuracy and consistency. The drone survey GIS data such as flooded areas, inundated households, drone images i.e. geo-tiff, ortho-mosaic data for both flood periods were collected from National Disaster Risk Reduction and Management Authority (NDRRMA) Government of Nepal. The drone survey was undertaken by Trimax IT Infrastructure with DJI drones between July 6th and August 14th under NDRRMA. The auxiliary data such as SRTM (DEM) from NASA, USGS site and Topo sheet layers were collected from the Department of Survey, Government of Nepal. The field visit was conducted on June 15–22, 2021 on flooded sites and observed post flood scenario (Damaged households, infrastructures, river profile and flood benchmarks) and cross-sectional features of the river.
2.3 Methods
The spatial and temporal pattern of precipitation in the basin was analyzed using daily rainfall data from stations (Nawalpur, Sermathang, Duwachaur, Tarkeghang, Dhap, Baunepati and Ganjala). Hydrologic Engineering Center's River Analysis System (HEC-RAS), 1D model was used to estimate sub-basin wise (regional) flow contribution in 8 different segments of the river during both MF21st and MF21nd event. The relationship between rainfall and runoff; Temperature and SWE to discharge at upper catchment region were examined. The correlation between rainfall (Sermathang nearest station) and runoff (Nakote station) for both events was analyzed. For the MF21st, historical data on discharge and the water level was collected from the Nakote hydrological stations but for the MF21nd, the discharge and water level data were collected from Dolalghat to predict the discharge of Nakote station using the G2G correlation method (Beven K.J. and Binley A 1992; Gupta et al. 1999). Regression method was applied to the historical statistical data from both stations to develop an equation that could be used to predict the discharge and water level of the Nakote hydrological station (Wang et al. 2016; Xiong et al. 2017).
The peak discharge of both events was estimated which involved comparing real-time gauge readings to estimate peak discharge. Real time discharge of the Melamchi River at 10 min intervals during both events at the Nakote hydrological station was estimated using the gauge-to-gauge correlation. The rating equation was also applied to estimate extreme flood discharge during the flood events. Furthermore, the SCS curve number of the HEC HMS Model was applied to verify the peak discharge. This study analyzed the relationship between temperature and snow water equivalent (SWE) of the Melamchi catchment using regional equation (DHM, 2006) and then SWE was converted in to stream flow (Modi et al, 2022). The terrain processing and georeferencing technique in ArcGIS were employed to develop a flooded map and river cross sections using drone survey geo-tiff data tools. The obtained data were carefully examined, analyzed and validated as mentioned in following research flow (Fig. 2).
During post flood field observation, sample discharge in some river segments were measured using a current meter (area and velocity method). Beside it, river channel geometry, flood benchmarks at 8 different segments of the river from Dumbredovan to Bhemathang (as shown in Fig. 1) were observed to estimate the discharge in HEC-HMS model and ground truthing the results. The detail methods of regional flow estimation using HECRAS, and relationship of flood discharge with temperature induced melting (temperature-SWE- stream flow) are presented in below as follows.
2.3.1 HEC RAS modeling and flood discharge
A one-dimensional (1D) steady HEC-RAS model was used for flood analysis purpose (Basnet & Acharya 2019, Hicks & Peacock 2005, Shrestha et al. 2010). In this study, data for 8 cross sections was fetched from the field observation. The 1D is applicable for flow scenarios that vary gradually with time and distance. In the flow analysis study for different sub-basin, the continuity and momentum equations were applied (Eqs. 1 and 2).
$$\text{Q}=\text{A}\text{*}\text{V}\dots \dots \dots \dots \dots \dots \dots \dots \dots \dots \dots \dots \dots \dots \dots \dots \dots \dots \left(1\right)$$
$$\frac{\partial \left(\frac{{\text{Q}}^{2}}{\text{A}}\right)}{\partial \text{x}}+\text{g}\text{A}\left(\frac{\partial \text{h}}{\partial \text{x}}+{\text{S}}_{\text{f}}-{\text{S}}_{0}\right)=0$$
2
………………………
Where, A represents a cross-section area; Q is the water flow, x is a measured distance in the direction of the channel, g is the acceleration due to gravity, h is the height of the water level above the datum, S0 is a slope of the river bed, and Sf is an energy slope.
The real time discharge of both events in Nakote hydrological stations was estimated using stage discharge relations of the following equations (Adhikari et al., 2000)
Q = C*(h - a) n ………………………(3)
Where, Q is a discharge (m3/s), h is a water level height (m), a is a constant value representing the stage at zero discharge and C and n are coefficients. The theoretical value of probability is p < 0.05.
Peak discharge at Nakote station was estimated during flood events using the Soil Conservation Service (SCS) curve number (CN) method. The following mathematical equation (USACE, 2005) was used to estimate the peak discharge in Melamchi River
Qp = (P − 0.2S) ² / (P + 0.8S) * (1000 / CN − 10) ………..(4)
Where: Qp is a peak discharge (m³/s); P is a precipitation depth (mm); S is a potential maximum retention after runoff begins (mm) and CN is the curve number. The peak discharge (Qp) takes into account the precipitation depth (P), potential maximum retention after runoff begins (S), and the curve number (CN) for the Melamchi River watershed.
2.3.2 Temperature, snow water equivalent and discharge
Pearson correlation was used to measure the correlation between daily mean temperature and snow water equivalents. The simple linear model was fitted with SWE and stream flow (Modi et al, 2022) to convert SWE to stream discharge of the Melamchi river using following equation (Eq. 5)
Q = ai SWEi +bi…………………...(5)
Where, Q is the stream flow volume during summer months (June, July, August, Sept), i represents the SWE at a given date, a and b are the model coefficients which values are − 0.0021 and 9.693 in the region respectively. SWE is estimated using the snow depth of the region based on the following regional equation (DHM, 2006).
SWE = 80.85×D + 658.47…………….(6)
Where, SWE is snow water equivalent (mm), D is snow depth measured (m) and the values 80.85 and 658.47 is a coefficient values developed in the region.