Artificial neural networks in downscaling and projections of long term future precipitation and development of future intensity-duration-frequency curves

This research presented a holistic approach for downscaling of precipitation in both space and time dimensions. Two novel neural network models were developed for spatial downscaling of climate variables and projection of future precipitation. A temporal downscaling method was then used to estimate intraday precipitation from the projected daily precipitation. This information has enabled estimation of future annual maxima precipitation quantiles to be made. As a result, the Intensity-Duration-Frequency curves of future precipitation were plotted. Outcomes of this research are beneficial for policy-makers, in terms of providing valuable input for mitigation and adaption strategies, along with better water resources management.