We use contingency tables to investigate the impact of the positive and negative phases of the PEO and SST (Niño 3.4) on the occurrence of the deficit and non-deficit runoff years, respectively. Table 3 shows the six contingency tables (for three rivers and two indices), where the contingency values are shown in percent from the lengths of the phase of the respective index. All the contingencies proved to be significant at the 95% confidence level (Chi-square test), except for the contingency between the occurrence of deficit/non-deficit years and the SST (Niño 3.4) phase in the Beijiang. It can be seen that deficit runoff years are contingent with the negative PEO phase and vice versa. In contrast, deficit runoff years are contingent with the positive phase of SST (Niño 3.4) and vice versa.
Contingencies (%) between the runoff years (deficit/non-deficit) and the phases (positive/negative) of PEO and SST (Niño 3.4) with Cramér's V statistics (non-significant value shown in italics)
| . | PEO . | SST (Niño 3.4) . | ||
|---|---|---|---|---|
| Year/Phase . | Positive . | Negative . | Positive . | Negative . |
| Weihe | ||||
| Deficit | 25 | 75 | 57 | 43 |
| Non-deficit | 50 | 50 | 40 | 60 |
| Cramér's V | 0.52 | 0.34 | ||
| Beijiang | ||||
| Deficit | 37 | 63 | 37 | 63 |
| Non-deficit | 81 | 19 | 33 | 67 |
| Cramér's V | 0.94 | 0.09 | ||
| Qingjiang | ||||
| Deficit | 40 | 60 | 50 | 50 |
| Non-deficit | 75 | 25 | 29 | 71 |
| Cramér's V | 0.66 | 0.40 | ||
| . | PEO . | SST (Niño 3.4) . | ||
|---|---|---|---|---|
| Year/Phase . | Positive . | Negative . | Positive . | Negative . |
| Weihe | ||||
| Deficit | 25 | 75 | 57 | 43 |
| Non-deficit | 50 | 50 | 40 | 60 |
| Cramér's V | 0.52 | 0.34 | ||
| Beijiang | ||||
| Deficit | 37 | 63 | 37 | 63 |
| Non-deficit | 81 | 19 | 33 | 67 |
| Cramér's V | 0.94 | 0.09 | ||
| Qingjiang | ||||
| Deficit | 40 | 60 | 50 | 50 |
| Non-deficit | 75 | 25 | 29 | 71 |
| Cramér's V | 0.66 | 0.40 | ||