Climate Change

* “Climate Change or Periodicity”, the Dichotomy in Hydro-Climatic Analysis, Prediction and Design.

Hydrologic design information, prediction of hydro-climatic and other environmental time series frequently require an understanding of interactions between random processes, short-term, long-term natural and anthropologically induced climatic signals inherent in the data. These interactions can pose a challenge when evaluating climate change forcings and their potential impacts. According to common definition, climate change generally refers to variabilities over a long period of time with respect to the growing accumulation of greenhouse gases in the atmosphere.

Analysis of annual hydro-climatic time series typically require a minimum of 30 years to sufficiently define trends due to climate change. Due to data inadequacies, analyses are frequently conducted on much shorter time series. However, data misinterpretation can arise if the environmental process has teleconnections with atmosphere-ocean oscillations. While shorter hydro-climatic quasi-periodic fluctuations such as those modulated by Quasi-Biennial Oscillations (approx. 2.4 year cycle) may not pose analytical problems, determination of long-term annual statistics of processes associated with long periodic fluctuations, e.g. Pacific Decadal Oscillations (PDOs ~30 year or greater periodicity), can lead to erroneous conclusions. For instance, if the data available were only observed during rising or falling part of a periodic cycle, this could be misinterpreted as climate change induced trend.

Faced with data inadequacies and complex interactions between hydro-climatic processes, data augmentation and signal decomposition techniques can be beneficial in understanding and studying climate change impacts. A variety of standard methodologies to augment short data time series such as spatial cross-correlation modelling and prediction, tree rings analysis can be found in published literature. Novel techniques, such as analysis in time and frequency domains, fractal analysis, etc., can provide insight into complex time series data interaction and help isolate signals caused by climate change.