Australia: The Land Where Time Began

A biography of the Australian continent 

Interdecadal Pacific Oscillation and its Modulation of the ENSO-Precipitation Teleconnection

In this study Westra et al. evaluate the role of the Interdecadal Pacific Oscillation (IPO) in the modulation of the El Niño-Southern Oscillation (ENSO)-precipitation relationship. Westra et al. described the standard IPO index together with several alternatives that were derived by the use of a low-frequency ENSO filter, and their results indicated that an equivalent IPO index can be obtained as a low-frequency version of ENSO. Several statistical artefacts arising from the use of a combination of raw and smoothed ENSO indices in the modelling of the ENSO-precipitation teleconnection were then discarded. Included among these artefacts is the identification of a modulation of a predictor-response relationship by the low frequency version of the predictor under model misspecification that is potentially spurious. Westra et al. evaluated the role of the IPO index in the modulation of the ENSO-precipitation relationship by the use of a global gridded precipitation dataset which was based on 3 alternative statistical models: stratified, linear, and piecewise linear. The information brought by the IPO index is in general beyond that which is already contained in the Niño-3.4 index, is limited and statistically not significant. In northeastern Australia there is an exception using annual precipitation data, and only for the linear model. A non–linear ENSO-precipitation relationship is induced by stratification by the IPO index, which suggests that the apparent modulation by the IPO is likely to be spurious and attributable to a combination of sample stratification and model misspecification. Westra et al. say caution is therefore required when using climate indices that have been smoothed to model or explain variability in precipitation that is of low frequency.

Conclusions

The outcomes of this study are:

1.      The IPO can be derived as a smoothed version of standard representations of ENSO;

2.      The apparent modulation is due to a combination of data stratification and model misspecification, though it is theoretically possible to spuriously attribute modulation to ENSO-precipitation teleconnection modulation to the IPO;

3.      It was found, based on observational data of land surface precipitation, that the information brought by the IPO index, which was beyond the information that had been already contained in ENSO is quite limited, and could be explained by the role of stratification and model misspecification, i.e., assuming a linear model to simulate a nonlinear predictor response association).

In this paper Westra et al. focused entirely on the possible modulation of the teleconnection between ENSO and precipitation by the IPO index; therefore the analysis doesn’t require assumptions as to whether the ENSO phenomenon itself exhibits variability of low frequency, and whether this low-frequency variability is internally forced or is external to the ENSO system.

They also described a number of potential issues that were associated with the use of the IPO index to model statistically the ENSO-precipitation relationship, which included the difficulty in explaining physically the mechanism that is the cause of modulation, and they also caution against using the IPO index for the prediction of future precipitation. Westra et al. do not recommend the IPO index as the basis of statistical modelling of seasonal or annual precipitation because of the potential of statistical artefacts when stratifying and using smoothed series to simulate the ENSO-precipitation relationship, in the absence of the identification of a physical mechanism for modulation by the IPO of the ENSO-precipitation relationship.

Sources & Further reading

1. Westra, S., B. Renard and M. Thyer (2015). "The ENSO–Precipitation Teleconnection and Its Modulation by the Interdecadal Pacific Oscillation." Journal of Climate 28(12): 4753-4773.