Drought assessment in the Pampanga River basin, the Philippines – Part 2: A comparative SPI approach for quantifying climate change hazards

Abstract

The change of climatic patterns is a major concern in the Pampanga River basin, the Philippines, and may intensify the magnitude of drought and flood hazard events under climate change uncertainty. Most of municipal water demand requirements for the Metro Manila relies on the Angat dam, which drains only 546 km 2 at the headwaters of the Pampanga River basin, as mentioned in Part 1 of this study. In this study, the climate change impacts were investigated for the RCP8.5 greenhouse gas emission scenario using 13 of higher-resolution general circulation models (GCMs) in CMIP5 and 4 ensemble members of the 20-km super high resolution atmospheric GCM, MRI-AGCM3.2S, with different sea surface temperature (SST) and sea ice distributions. The simulated daily precipitation was bias-corrected with the daily gridded rain gauges dataset as a reference observation, after the bilinear interpolation from the coarse grids of GCMs to the fine grid of the observation dataset. To estimate the meteorological hazards of droughts and floods due to climate change, we compared the precipitation characteristics between present (1979-2003) and future (2075-2099) climate projections using the standardized precipitation index (SPI) and the comparative SPI (cSPI). The cSPI is a new concept designed to intuitively evaluate the meteorological drought in target datasets on the basis of the reference dataset. Computing the SPI by substituting future climate precipitation in the cumulative distribution function (CDF) with parameters derived from the corresponding present climate, we can compute the cSPI in future climate projections on the basis of the present climate projection. The cSPI could estimate not only the probability change of extremes but also the change of mean precipitation due to climate change simultaneously. From cSPI results, the interannual variability of the cSPI in future climates increased significantly, but the average of the cSPI was almost the same as in present climate projections in the SST ensemble average of MRI-AGCM3.2S. As the result, both the meteorological drought and flood increased in MRI-AGCM3.2S simulations. In the higher-resolution model ensemble average CMIP5, the interannual variability of the cSPI in future climate was almost the same as in present climate projections, but the average of the cSPI signif- icantly increased. As the result, the meteorological drought might be decreased in future but the heavy wet conditions could be increased in CMIP5 multi-model ensembles. It would be required to adapt to the heavier wet condition in the Angat dam in the end of the 21st century under the RCP8.5 emission scenario than the present climate condition. The comparative standardized index concept of the cSPI approach, computing standardized index by substi- tuting target datasets in the CDF with parameters derived from the reference dataset, is applicable to the other standardized indices, such as the standardized streamflow index (SSI), the standardized reservoir storage in- dex (SRSI), and the standardized precipitation evapotranspiration index (SPEI), which is obtained with various probability distributions. For example, the comparative SRSI is utilized to evaluate socio-economic droughts under climate change in Part 3 of this study.