High-resolution dynamic downscaling over the southeast Asia domain of the Coordinated Regional Climate Downscaling Experiment (CoRDEX) was performed using the regional Weather Research and Forecasting (WRF) model forced by the European Centre for Medium-Range Weather Forecasts reanalysis (ERA)-Interim dataset as well as the current climate and future projections of the hadley Centre Global Environmental Model version 2–Atmosphere–ocean (hadGEM2–Ao). Changes in the near and far future in downscaled low-level circulation and precipitation for the hindcast experiment and historical run were analyzed and compared with observations and/or corresponding large-scale forcing.

overall compared with the hindcast experiment in 50 km resolution, not only the biases in the low-level circulation field were reduced, but also the skills of precipitation in terms of RMSE and spatial correlation coefficient were improved in the hindcast experiment with 25 km resolution in this study. The decrease of the bias is largely caused by the increase of the horizontal resolution. The WRF model reasonably reproduced the quantity and location of precipitation better than the ERA-Interim forcing. The added value over the ocean was smaller than that over land (figure not shown) owing to the biases in the lateral boundary forcing (LBF) from the ERA-Interim, lack of strong underlying forcing that may reduce errors in the WRF simulation, as well as shortage of observations over the ocean.

In the control experiment, the low-level circulation (precipitation) bias from the hadGEM2-Ao was enhanced (largely reduced). The reduction occurred mainly because of the strong forcing of the underlying surface, which was better simulated in the control experiment by the increased grid resolution than in the global hadGEM2-Ao.

For the future projections, the increased precipitation occurred near the climatological rain belt and the decreased precipitation occurred in the climatological dry region. The wet region became wetter and the dry region became drier in the future projection of both models. Besides the increase in the spatial variability of precipitation, the strength of both wet and dry conditions will be increased in the future under the global warming scenarios. Thus, extreme flooding or droughts will become more harmful. Generally, the two models were more consistent in projecting the mean rainfall than rainfall variability. however, both models suggested diversity among the land areas in wet and dry anomalies. In March-April-May (MAM) under scenario of RCP4.5, the Indo–China Peninsula (IC) would become slightly drier while wetter conditions would happen over sumatra, Borneo, and New Guinea. The change over Java was uncertain. This diversity become even larger in case of the scenario RCP8.5, under which most land areas in the Maritime Continent would likely have severe drought in summer and autumn. In winter, both models suggested increase of precipitation only under the RCP4.5 scenario.

1. INTRODUCTION

2. Model and Data

3. Experimental Design

4. Results

4.1 hindcast Experiment

4.1.1 Low-Level Circulation

4.1.2 Precipitation

4.2 historical Experiment

4.2.1 Low-Level Circulation

4.2.2 Precipitation

4.3 Future Projection

4.3.1 Low-Level Circulation

4.3.2 Precipitation

5. Conclusions

REFERENCES

TABLES

FIGURE CAPTIONS

FIGURES