The hydroclimate-ENSO relationship becomes
weaker for longer lags. However, because the IPO
value changes very slowly compared to the SOI
value, the strength of the hydroclimate-IPO
relationship is similar even for lags of more than one
year. For example, the Jan-Feb-Mar streamflow
versus Oct-Nov-Dec SOI correlation is statistically
significant at α = 0.05 in about 65% of the
catchments (40% for rainfall) but the Jan-Feb-Mar
streamflow versus Jul-Aug-Sep SOI is significant in
only about 40% of the catchments (25% for rainfall).
In contrast, the number of catchments showing
statistically significant Jan-Feb-Mar hydroclimate
versus Oct-Nov-Dec IPO and versus Jul-Aug-Sep
IPO and longer lags are similar (about 60% for both
streamflow and rainfall). Therefore, when the
hydroclimate-IPO relationship is relatively strong,
like in Jan-Feb-Mar, the relationship can be used to
provide long lead-time forecasts.