Computation of runoff in the Jamuneswari river basin of north west-Bangladesh

Abstract

A study has been conducted for computation of runoff in the Jamuneswari basin located in the North West Region of Bangladesh. The basin is a moderate size of approx. 1055.77 km2 drainage area. The main river systems of the basin are the J amuneswary channel of about 123 km length and its two tributaries, the Burikhora Chikly in the west of about 41 km length and the Bullai in the east of about 41 km channel length. In conducting the study data on landuse, soil, climate, digital elevation model, infrastructures, river systems, hydro-climatic gage locations, long term flow, water level, river cross sections and other model results were collected from relevant organizations within the country. To fulfill the objectives of the study hydrologic modeling tools was used like HEC-HMS (version 2.0) that have functionality of simulating hydro graphs at any components of a hydrologic basin. The Geographic Information System (GIS) tools e.g Arc/Info 7.2.1 and Spatial Analyst of ArcView 3.1 having grid based hydrologic analysis functionality of any watershed was also used in analyzing spatial data and to provide inputs to the HEC-HMS model. The hydrologic processes and complexity was investigated, analyzing flow data, climatic data, soils permeability, landuse and land cover data. The spatial variability of runoff potentials of the area was also studied and thus hydrological behavior of the basin was clearly understood. It shows that the basin has moderate to high runoff potentials with very quick runoff peaks and therefore the basin is highly responsive to the heavy shower. In the study GIS techniques for stream delineation, sub-catchments/watershed delineation and computation of hydrologic modeling parameters (e.g. SCS curve number-CN, Muskingum K and X etc.) were used. A best method among the options in ArcView (spatial analyst) to create an accurate OEM (digital elevation model) surface from the raw point observed data was also investigated. The computed hydrologic modeling parameters were written to HEC-HMS compatible file format. The hydrologic model developed under HECHMS to compute runoff uses a basin model for SCS unit hydro graph transformation, SCS Loss method, Muskingum channel routing method and theissen polygon weight method to account rainfall distribution. The minimum slope method of baseflow separation was implemented in the modeling. Model results were compared with the observed runoff and also with the NAM (a part of MIKE 11 software) result. Some of the model parameters were further investigated and re-build the model. The study proved the usefulness of OEM data to delineate streams, ephemerals rivers, channels, creeks etc. for hydrological modeling. The OEM should be accurate and at representable scale for the area. The study showed the usefulness of LANDSAT TM image to obtain landuse classification for SCS-CN value calculation for hydrologic modeling specially for the SCS loss rate method.

An interface between GIS and HEC-HMS developed and implemented in the study to create a comprehensive HEC-HMS model for the study basin. The interface was quite successful with the SCS-unit hydrograph, SCS- Transformation, thiessen polygon meteorological model and Muskingum River routing options. The model result shows reasonable fit with the observed data in the early monsoon period and a relatively poor fit in the monsoon period. This reflects re-investigation of some of the model parameters that were assumed due to nonavailability of data e.g. overland flow velocity, velocity in the streams, evaporation loss and antecedent moisture condition to compute CN value. Finally recommendations were put forward for further study.