Low-frequency variability in the hydrological cycle: implications for detection and attribution of climate change

Prediction of the response, of the global water cycle to the increasing greenhouse gases, is a pending problem. Scarce, short and noisy observational records lead often to the ambiguous results. This also contributes to the fact that climate models often struggle to reproduce correctly the statistical features of precipitation. Its noisy spectrum, apparent especially on the regional scale, poses a challenge for detection and attribution of a robust signal.

This study attempts to identify the low-frequency variability and quantify its contribution to the total variability of precipitation on a global and regional scale. Consequently, the effect of natural variability on detection and attribution of anthropogenic change is analysed. To identify decadal/multi-decadal time-scale variability, we analyse thermal and hydrological variables provided by the long control run of Geophysical Fluid Dynamics Laboratory CM3 model. Derived fingerprints are used to verify whether long-term natural variability can significantly contribute to the simulated and observed precipitation records and whether they are separable from the response to the anthropogenic forcing.