Mismatch in the hazard curve from Disaggregation and mean_rate_by_src

[Chetana KR]

Dear Openquake community,

I am comparing the hazard curve outputs obtained from the following:

1) Mean hazard curves from Classical PSHA run

2) Mean hazard curves from Disaggregation run

3) Mean hazard curves reconstructed from the mean_rate_by_src output file obtained from the disaggregation run.

The results from 1 and 2 match perfectly as expected. However, the hazard curves reconstructed from the 3 don't match the other two.

for the reconstruction from mean_rate_by_src, I have:

1) summed the annual frequency of exceedance (AFOE) over all sources at each IMT-IML level, and 

2) converted the resulting rates to probability of exceedance using the Poisson relation before comparison

 

Despite this, a discrepancy remains. 

The difference is illustrated in the attached figure

May I know if such discrepancy is expected due to disaggregation outputs correspond to projections of the multidimenaional probability matrix?

Thanks in advance,

Chetana