STUDY OF EXTREME HYDROLOGICAL EVENTS AND IMPACT OF CLIMATE CHANGE ON RAINFALLS - A CASE STUDY OF NORTHWESTERN BANGLADESH

Abstract

In general, Bangladesh is prone to different types of floods with different magnitudes because of its unique geographical location and topography. Teesta sub-basin in the Northwestern part of Bangladesh is more vulnerable to flood if compared to other parts. In this context, a study of hydrological extreme events and impact of climate change on rainfall was undertaken, aiming at (1) developing the isopluvial maps of extreme rainfall quantiles in the northwestern part of Bangladesh, (2) estimating the flood quantiles of different return periods and their corresponding flood stages at Kaunia gauging station; and, (3) downscaling of daily rainfall for impact assessments of climate change. The maximum annual daily rainfalls of 10 rainfall stations and annual flood series from Kaunia gauging station in the Northwest of Bangladesh were used in the study. The data were checked through the tests of outlier, short-term dependence and long-term dependence and trend identification for checking homogeneity and consistency. The fitness of probability distributions used was tested using Chi-square (x2), Kolmogorov-Smirnov (K-S), D-index, and L- moment ratio diagram tests. The fitness of good tests shows that GEV distribution best fit the maximum annual daily rainfalls and PT-3 fits the annual flood series. Design rainfalls and flood quantiles of different return periods were then computed. The isopluvial maps of 24 hours rainfall for 2-, 5-, 20-, 25-, 50-, 100- and 1000-year return periods were developed using interpolation techniques in GIS. Three rainfall stations from Jamuneswari catchment were considered for assessing the impact of climate change using LARS-WG model. Three variables namely, length of wet and dry series distributions, daily rainfall distributions for each month, and monthly mean rainfalls along with their variances, were used for the validation of the model. The test of Kolmogorov-Smirnov (K-S), Student (t-test), and Fisher (f-test) tests have shown average to good performance. The MPEH5 GCM's predictions, for A2 emission scenario over the period of 2011- 2030 centered in 2020, were used to perturb site parameters and generate daily site-specific climate change scenarios. The predicted climate data at site have shown an increase in annual mean rainfalls by 7 %, and 12 % at Ghoraghat and Nawabganj respectively whereas a slight decrease of annual mean rainfall is expect at Phulbari rainfall station. The seasonal mean rainfalls will increase in winter (i.e. Dec, Jan and Feb), pre-monsoon (i.e. Mar, Apr and May), and post-monsoon (i.e. Sep, Oct and Nov) at all three rainfall stations; the monsoon mean rainfall (i.e in Jun, Jul and Aug) will decrease by about 6 percent. The monthly mean rainfall is expected to increase mostly during December, March, April, August and October. June and July will face slight decrease in monthly mean rainfalls. The extreme rainfall quartiles, estimated through GEV probability distribution, are expected to increase at all three sites up to 100 year return periods. These results on extreme hydrological events may be useful to decision makers for design, operation, and maintenance of the hydraulic structures and for reducing vulnerability to flooding and improving community preparedness. The downscaled rainfall data can be used from the planning stage up to the operation of different water systems. Such a policy constitutes one of the adaptation measures to climate change impacts.