This suggests that, in the future climate, the less frequent storms will be more intense. The implications are that the present climate extremes will become more frequent in the future. For example, 100 RP year depths increase from 342 to 545 mm for NMIA and 260 to 330 mm for SIA, by 2100. The associated

frequency will also change such that the 25 year RP will become the 17 and 19 year RP in the future and the 100 year RP in the present climate will become the 42 and 56 year RP, see Table 5. There is an increasing trend of the scale parameter that range from −0.007 to 0.22 mm/year, and averaged 0.072 for both stations. This implies that the scale at the end of 90 years in 2100 will increase by 6.5 mm. Concurrent with this, the mean for both stations are estimated to be decreasing at −0.123 and −0.104 mm per annum for NMIA and SIA intensities respectively. In check details other words, the variability is projected to increase, and the mean is projected to decrease for both stations into the future. Against this background of increased projected intensities, the variability increases are, responsible for the projected increases in intensities. Predicted future depth-RP curves for the four models indicate an increase over the stationary model. Drainage and flood control planning should contemplate the increasing trends particularly for the longer 25–100 years RP. Existing IDF curves

were confirmed with frequency re-analysis for both stations with high CC and SRC of 0.98 and 0.96 respectively and confirm the suitability of the PDF, PPF and PEM used in the former study by UWA. PDF buy NU7441 is the most important factor in the frequency analysis configuration, with Weibull performing better than the Gumbel and Logistic PDF. Variations in PPF do not add substantial improvement to the statistical analysis. This is similar

to observations by Seckin et al. (2010). L-Moments and standard statistics experiments ifoxetine did not show considerable improvements relative to PWM results. IDF curves derived from the Weibull experiment differ from the control experiment up to 41% and highlight the implication of extrapolation to the 100 year RP with different PDF models. Historical observations of extreme events support the increases predicted by the Weibull model that better mapped the extreme tail of the distribution. Extreme precipitation should, therefore, consider an assessment of the performance of a number of PDF and report GOF. Empirical equations and statistical downscaling methods were able to predict AMS. Chowdhury’s equations predicted the 5 min to 12 h duration better than Nhat’s and thus provide a means for converting 24-h durations to shorter durations. ANN downscaling model predicted the 2–10 day durations accurately. Extension and infilling for the period 1885–2010 reduced the gaps significantly to between 0 (for the 12-h and shorter duration) and 47% (for the 2–10 day durations).