Verification & Statistical Parity Testing
[!TIP] Statistical Parity checks whether the generated synthetic series reproduces the statistical moments of the historical input record over long simulation horizons.
This guide explains how to compare synthetic output statistics against the historical record.
What is Statistical Parity?
Over a long synthetic realization (e.g., 50 to 100 years), the generated series should approximate the statistical moments of the historical analysis period:
- Mean Precipitation: Annual and monthly precipitation totals compared against historical averages (typically within $\pm 3\% - 5\%$ over long horizons).
- Wet Day Frequency ($P_{\text{wet}}$): The proportion of rainy days per month compared against historical observations.
- Transition Probabilities: Historical $P(W|W)$ and $P(W|D)$ rates governing wet-spell and dry-spell durations.
- Extreme Storm Depths: Upper-percentile daily rainfall depths (95th and 99th percentiles).
- Temperature Distribution: Monthly mean $T_{\max}$ and $T_{\min}$ values and diurnal temperature ranges.
Using the Statistical Parity Tab
To view the comparison table:
- After running a simulation of at least 30 years, click on the Statistical Parity tab in the Simulation Analysis panel.
- Review the side-by-side comparison table:
| Climate Metric | Historical Record | Synthetic Simulation | Difference (%) | Status |
|---|---|---|---|---|
| Annual Precipitation | 412.5 mm | 415.8 mm | +0.8% | Pass |
| Wet Days / Year | 86.4 days | 85.9 days | -0.6% | Pass |
| January Mean Depth | 4.8 mm | 4.9 mm | +2.1% | Pass |
| July Mean Depth | 8.9 mm | 8.7 mm | -2.2% | Pass |
| Mean Summer Tmax | 32.1 °C | 32.3 °C | +0.2 °C | Pass |
| Mean Winter Tmin | -4.6 °C | -4.5 °C | +0.1 °C | Pass |
- Status Indicators:
- Green Checkmark (Pass): The synthetic metric falls within the expected random sampling confidence interval ($95\%$ confidence level).
- Yellow Warning: Moderate deviation, often due to a short simulation horizon (e.g., 5 years instead of 30+ years).
- Red Alert: Larger deviation, which may occur on very short runs or when manual parameter edits were applied.
Sampling Variability in Short Simulations
A short simulation (such as 3 years) can produce an annual rainfall average that differs noticeably from the 30-year historical mean due to random sampling variability.
To evaluate convergence toward the input parameters, run at least 30 to 50 years (or multiple realizations). Over longer horizons, sampling variance decreases and the synthetic statistics converge toward the input parameters.