Rainfall Analogs and Planetary Configurations in Australian Forecasting

Analog forecasting finds past years with similar conditions to the current moment and uses their subsequent weather as a forecast template. The definition of "similar conditions" is what separates methods.

Standard climatological analogs use ENSO state, sea surface temperatures, and atmospheric circulation patterns. Planetary analogs use a different similarity criterion: years when the same planets occupied similar positions. This is the method used in long-range Australian rainfall forecasting from Inigo Jones onwards.

What a planetary analog is

The planetary configuration at any date is a specific set of positions, Jupiter at 45°, Saturn at 180°, their mutual aspect 135° and closing, and so on. To find analogs, you search historical records for dates with similar configurations and look at what rainfall did in those years.

The hypothesis is that the same planetary geometry produces similar conditions in the solar system (solar radiation output, interplanetary magnetic field, cosmic ray flux) which influence terrestrial atmospheric dynamics in consistent ways.

Whether that causal chain is real is still debated. The empirical observation, that years with similar planetary configurations sometimes show similar rainfall patterns, is in the historical data. The mechanism is the open question.

The method in Australian context

Inigo Jones (see our post on the Crohamhurst Observatory) used analog matching as a core forecasting technique. His procedure, as reconstructed from his published work:

1. Identify the current planetary configuration, primarily the positions of Jupiter, Saturn, and their mutual aspect
2. Search the historical Queensland rainfall record (available from 1870s onwards) for years with similar configurations
3. Average the rainfall outcomes across the best matching analog years
4. Use that average as the long-range forecast

This is conceptually identical to modern ENSO-based analog forecasting, same structure, different similarity criterion.

The advantage over ENSO analogs is the forecasting horizon. ENSO-based analogs are useful one to two seasons ahead; planetary configurations are calculable arbitrarily far into the future, enabling genuine long-range forecasting at 12-18 months.

Evidence quality

Australian agricultural research groups tested Jones-style planetary analogs in the 1970s and 1980s. Results were mixed but not uniformly negative. Drought prediction, the most practically important application, showed moderate skill in some assessments.

The methodological challenges are real. Analog forecasting requires a long historical record to find well-matched analog years. Australian rainfall records from 1870 give about 150 years of data. With 150 years and planetary cycles of 20-30 years (the Jupiter-Saturn synodic period is approximately 20 years), you get roughly 7 independent Jupiter-Saturn cycles. That's a thin sample for statistical testing.

Better-matched analogs are more likely to be predictive but rarer. If you require planetary configuration similarity within ±5° for all major planets, you may find only three or four analogs in the historical record, not enough for reliable averaging.

The practical compromise: use a wider matching criterion, accept more noise, and interpret the analog forecast as a probabilistic tendency rather than a precise prediction.

The ENSO interaction

Australian rainfall has a strong ENSO signal. La Niña years are systematically wetter across much of eastern Australia; El Niño years are drier. This is the dominant mode of Australian rainfall variability on interannual timescales.

Any planetary analog method that doesn't control for ENSO will pick up ENSO variance as part of the planetary signal. If years with a particular Jupiter-Saturn configuration happened to cluster in La Niña years, the "planetary" forecast is really an ENSO forecast in disguise.

This confounding is a genuine problem with older planetary analog research. Modern approaches should control for ENSO state explicitly, either by stratifying the analog search by ENSO phase or by removing the ENSO signal from the rainfall anomaly before doing the analog matching.

Application to agricultural planning

The practical value of long-range rainfall forecasting for Australian agriculture is substantial. Planting decisions, irrigation scheduling, livestock management, and input purchasing all benefit from extended outlooks, if the outlooks have any skill.

For vegetable gardening and small-scale farming, even a probabilistic tendency ("above-normal rainfall probability for southeast Queensland this autumn") is useful for planning. See gardenbuddy.au for practical gardening applications of seasonal forecasting in Australian conditions.

The Dayboro region sits in the Sunshine Coast hinterland catchment, the same zone where the original Crohamhurst Observatory was located. Long-range rainfall performance in this region can be tracked against historical forecasts using data available at dayboro.au.

Modern computational approach

The computational challenge that constrained Jones and his contemporaries, manually computing planetary positions and searching through printed rainfall tables, is no longer binding. Modern planetary ephemeris libraries (Swiss Ephemeris, JPL Horizons) give positions to arc-second accuracy for any date from ancient times to far future. See our planetary aspects computation guide for the technical details.

Crohamhurst.app implements planetary analog matching against the Australian historical rainfall record. Users can specify analog matching criteria, request analog years, and view the composite rainfall anomaly for matched periods.

The Bartels test as validation

Before constructing analog forecasts from a claimed planetary cycle, it's worth testing whether the cycle exists in the historical rainfall record at all. The Bartels significance test applied to annual Queensland rainfall with the candidate planetary period as the test frequency tells you whether the signal is above noise.

Cycles that fail the Bartels test at p > 0.10 should not be used as the basis for analog forecasting, the historical signal is too weak to support the method.

Related: Inigo Jones and the Crohamhurst Observatory | Bartels Cycle Significance Testing | Computing Planetary Aspects: Ephemeris and Precision