Abstract:
This study presents an assessment of the Conformal Cubic Atmospheric Model
downscaled climate projections at 8 km spatial resolution which are forced by sea surface
temperature and sea ice concentration of three global climate models centred on Sri
Lanka. Model simulations were compared with observed daily gridded precipitation and
temperature data for the period from 1981 to 2005 over Sri Lanka. Statistical measures
such as pattern correlation, mean biases and root mean square errors are calculated using
observed and simulated data. The results show that the temperature biases range between
-1.4 and -2.4 °C and pattern correlations vary among the seasons between 0.92 and 0.96.
The model underestimates the temperature over the selected domain. The pattern
correlation of rainfall varies among the seasons and it is relatively low for the southwest
monsoon season which is 0.67 with RMSE about 3.8 mm/day. Simulations show small
bias values during the northeast season with a strong pattern correlation of 0.76 and
RMSE of 1.7 mm/day.
