...
The default autocorrelation coefficient is set to zero to allow the same model run by different users to produce the same magnitude of loads, however you can specify the auto correlation coefficient if required (say if needing to calibrate against measured concentration data) and should use the values as set out in the table below:
| Time-step | Autocorrelation coefficient | |
|---|---|---|
| Baseflow | Stormflow | |
| 6-min | 0.94 | 0.95 |
| 12-min | 0.82 | 0.93 |
| 30-min | 0.51 | 0.84 |
| 1-hour | 0.41 | 0.77 |
| 3-hour | 0.37 | 0.62 |
| 6-hour | 0.35 | 0.50 |
| Day | 0.31 | 0.27 |
It is important to note that the autocorrelation coefficient will not significantly affect the treatment train effectiveness produced by music, but simply ensures that the variation over time in concentrations during storm events and baseflow conditions is more ‘realistic’. Depending on the time-step and coefficient used, there can be variations in mean annual loads for the same model run on different computers, however the maximum difference is usually within 10% of the previous run.
...