Downscaling of general circulation model precipitation for assessment of impact on water resurces at basin level | |
| Author | Sharma, Devesh |
| Call Number | AIT Diss. no.WM-06-04 |
| Subject(s) | Climatic changes--Water-supply--Thailand Hydrological models--Thailand |
| Note | A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Engineering in Water Engineering and Management, School of Engineering and Technology |
| Publisher | Asian Institute of Technology |
| Series Statement | Dissertation ; no. WM-06-04 |
| Abstract | Climate change and its potential hydrological effects are progressively contributing to uncertainties in availability of water, particularly at basin scale, as well as in water demand. The scientific study at the global level indicates that Southeast Asia is more vulnerable to climate change and its variability, including rise in sea level, shifts of climatic zones and occurrence of extreme events (IPCC 2001). Many general circulation models (GCMs) have been developed to understand the impact of increased concentrations of greenhouse gases on the climatic parameters. But, GCM precipitation is generally characterized by the biases and low spatial resolution. These two are the major limiting factors for direct application of GCMs scenarios in studies dealing with the assessment of impact of climate change. However, climate model output can be utilized more efficiently in impact studies with due consideration of its coarse resolution and biasness. The objective of the present study is to improve the selected GCM scenarios by applying the bias-correction and downscaling techniques and use these scenarios to assess the impact of climate change on availability of water resources. The study is conducted for two river basins of Thailand with different climatic characteristics: the Ping River Basin and the Mae Klong River Basin. Trend analysis in extreme indices is done based on observed daily temperature and rainfall data using 'RclimDex' software to detect change in climatic situation within the study domain. A total of 21 precipitation and temperature indices are calculated for each observation station. The statistical significances are evaluated using the Kendall-tau test. The results of the analysis depict significant increase in the annual number of warm days and warm nights, with corresponding significant decrease in the annual number of cool days and cold nights. The trends for temperature indices are more consistent in the study basins compared to precipitation indices. There is a decrease in annual total precipitation for almost all stations. The maximum number of consecutive dry days is increasing. It is also observed that there is less spatial coherence in other extreme indicators like maximum 1-day rainfall (RXIday), 5-days rainfall (RX5day) and simple daily intensity index (SDII). A suitable GCM is selected by determining statistical characteristics of precipitation and temperature scenarios at observation station locations and grid nodes. Results reveal a considerable variability in different GCMs simulations for the observed climate. HadCM3, ECHAM4, GFDL-R30 and US NCAR models are good in simulating the magnitude and spatial variability of mean temperature. Compared to mean temperature, high variation and poor statistical agreement is found in precipitation scenarios of all GCMs. By considering the statistical characteristics, both the Australian model (CSIRO Mk2) followed by the German model (ECHAM4) represent the observed magnitude better than other models. However, the ECHAM4 scenarios are selected for improvement and application due to their high spatial resolution and daily data availability. Three bias-correction techniques namely, scaling, gamma-gamma transformation and empirical-gamma, are applied on a daily time scale for nine years (1991-1999) to improve the quality of raw ECI-IAM4/OPYC SRES A2 and B2 precipitation scenarios at the grid nodes. Scaling method is the bias correction method to correct the biases from the mean rainfall amount and two other techniques are considered to improve the rainfall frequency and intensity of GCM precipitation. The output of bias correction methods has been compared with observed data based on statistical parameters. Out of three techniques, gamma-gamma transformation method is found to be more effective in correcting the rainfall frequency and intensity simultaneously. The bias corrected daily precipitation scenarios are then used for spatial downscaling. Spatial disaggregation method, based on multiplicative random cascade theory, is used in dealing with coarse resolution problems in ECHAM4IOPYC precipitation. Disaggregation model parameters (b, ơ²) are estimated using Mandelbrot-Kahane-Peyriere (MKP) function at q=1 for each month to preserve the spatial heterogeneity of rainfall. Values of b (intermittency) decrease with increase in rainfall while ơ² (variability) values remain almost insensitive to rainfall amount. Values of ơ² show a seasonal trend but the trend is opposite to that of β values. On an average, rainfall with high amounts is less intermittent and there is less effect of rainfall variability. The spatial disaggregation model satisfactorily reproduces the observed trend and variation of average rainfall amount except during heavy rainfall events with a certain degree of spatial and temporal variations. The United States Army Corps of Engineers (USACE) model HEC-HMS is used to assess the impact of climate change on future water resource availability. MODCLARK transformation is used with Standard Hydrologic Grid (SHG) size of 2x2 km² for incorporating the distributed precipitation data for the model. The model is calibrated for the year 1999 and is verified for the year 2000 with observed daily inflow. The coefficient of determination (R²), Nash-Sutcliffe efficiency (EI) and absolute percentage volume error (APVE) are calculated to evaluate the model performance. This calibrated model is then used to simulate flow for a two-year period (1999-2000) with modified precipitation scenarios and the raw GCM precipitation as input. The results of the hydrologic simulation using bias-corrected downscaled precipitation scenarios provide the most reasonable representation of the physical situation when compared with the observed flow data, indicating the proficiency of the improved scenarios in reproducing basin-scale observation. Finally, the impacts of future climate change scenarios on streamflow in the upper part of the two river basins are also examined for three future periods (2025, 2050 and 2095). Results obtained with application of the improved ECHAM4/OPYC precipitation scenarios (A2 and B2) indicate that there is a decrease in the streamflow in the upper Ping River Basin and an increase in the streamflow in the upper Mae Klong River Basin |
| Year | 2007 |
| Corresponding Series Added Entry | Asian Institute of Technology. Dissertation ; no. WM-06-04 |
| Type | Dissertation |
| School | School of Engineering and Technology |
| Department | Department of Civil and Infrastucture Engineering (DCIE) |
| Academic Program/FoS | Water Engineering and Management (WM) |
| Chairperson(s) | Babel, Mukand Singh; |
| Examination Committee(s) | Clemente, Roberto S. ;Honda, Kiyoshi ;Gupta, Ashim Das; |
| Scholarship Donor(s) | DANIDA; |
| Degree | Thesis (Ph.D.) - Asian Institute of Technology, 2006 |