liJadeRelationship intelligence

TSV-002 · ENSO/SOI Relationship Drift

JADE ANALYSIS

Has the predictive relationship between ENSO/SOI and Townsville rainfall changed through time?

VALIDATED RELATIONSHIP

SOI

Moderate

r = +0.39 · p = <0.001 · n = 75

NO SIGNIFICANT CHANGE DETECTED

STABLE RELATIONSHIP

ONI / ENSO

Moderate

r = -0.27 · p = 0.018 · n = 76

NO SIGNIFICANT CHANGE DETECTED

STABLE RELATIONSHIP

IOD / DMI

Weak

r = -0.04 · p = 0.699 · n = 85

NO SIGNIFICANT CHANGE DETECTED

STABLE RELATIONSHIP

SOI · 20-year rolling r

19712026

ONI · 20-year rolling r

19702026

Baseline windows — SOI

WindownrpStrength
historical to 1999
1952–1999
48+0.370.010Moderate
from 2000
2000–2026
27+0.400.038Moderate
recent 20y
2007–2026
20+0.330.158Moderate
first 30y
1952–1981
30+0.420.021Moderately strong
last 30y
1997–2026
30+0.280.140Moderate

Interpretation

  • SOI (Sep–Feb mean) versus Townsville wet-season rainfall: full-period Pearson r = 0.3898 (n=75, 1952–2026). Positive SOI (La Niña-like) is expected to associate with wetter seasons.
  • ONI (SON–DJF mean) versus the same rainfall: r = -0.269. Negative ONI (La Niña) is expected to associate with wetter seasons.
  • Direction call on the SOI relationship: NO SIGNIFICANT CHANGE DETECTED. Stability: STABLE RELATIONSHIP.
  • Direction call on the ONI relationship: NO SIGNIFICANT CHANGE DETECTED. Stability: STABLE RELATIONSHIP.
  • These are associations, not causal claims. IPO/PDO phase, IOD interaction, monsoon bursts and cyclone years can all dominate a single season.

Competing explanations

  • Internal climate variability / IPO-PDO phaseopen

    Not tested as a covariate in v0.1.

  • IOD interactionpartial

    IOD association is estimated separately; no formal interaction model.

  • Background warming / changing SST gradientsopen
  • Measurement or station-siting changeopen

    Airport record; homogenisation not applied.

  • Limited recent-window sampleopen
  • Statistical artefact of window choicetested

    Multiple baselines are reported; conclusion is flagged if baseline-sensitive.

Counter-evidence and limits

  • · A full-period association can remain while rolling-window r fluctuates inside bootstrap uncertainty.
  • · ENSO is not the only driver of Townsville wet-season rainfall; a weaker correlation is not proof that ENSO ceased to matter.
  • · Linear, seasonal-mean association only.
  • · No partialling of IOD, MJO or monsoon onset.
  • · ONI and SOI are overlapping representations of ENSO; they are not independent tests.