Source flux statistics reporting errors in MPI

[dkachuma]

Describe the bug
There are two issues identified with the source flux reporting in GEOS:

1. CSV Output Issue: The source flux CSV file contains only a single time step (appearing to be the final one), even though the time column is labeled as 0.

2. MPI Scaling/Calculation Issue: When running the case with multiple MPI ranks (e.g., 4), the reported production mass and rates are physically strange (orders of magnitude such as $10^{56}$ and $10^{47}$) and appear to increase over time. The same case produces reasonable results when run on a single MPI rank.

To Reproduce
Steps to reproduce the behaviour:

1. Attached case with source fluxes (e.g., INJECTOR.LEFT and INJECTOR.RIGHT).

2. A source flux report (fluxStatistics) to the configuration.

3. Run the case with mpirun -n 1. Observe reasonable rates ($\approx 11.57$ kg/s).

4. Run the same case with mpirun -n 4.

5. Compare the flux statistics table in the logs and the generated CSV file.

Expected behaviour

1. The CSV file should contain data for all logged time steps.

2. The production rates and mass values should remain consistent and physically accurate regardless of the number of MPI ranks used.

Screenshots
1 MPI Rank (Correct):

-------------------------------------------------------------------------------------------------
|              fluxStatistics, flux statistics for: INJECTOR.LEFT, INJECTOR.RIGHT               |
|-----------------------------------------------------------------------------------------------|
|     Flux(es)     |    Region     |  Element Count  |  Prod. mass [kg]   |  Prod. rate [kg/s]  |
|------------------|---------------|-----------------|--------------------|---------------------|
|   INJECTOR.LEFT  |  all_regions  |              1  |   [0, 9999999936]  |     [0, 11.574074]  |
|  INJECTOR.RIGHT  |  all_regions  |              1  |  [0, -9999999936]  |    [0, -11.574074]  |
|        flux_set  |  all_regions  |              2  |            [0, 0]  |             [0, 0]  |
-------------------------------------------------------------------------------------------------

4 MPI Ranks (Incorrect):

---------------------------------------------------------------------------------------------------------------------
|                        fluxStatistics, flux statistics for: INJECTOR.LEFT, INJECTOR.RIGHT                         |
|-------------------------------------------------------------------------------------------------------------------|
|     Flux(es)     |    Region     |  Element Count  |       Prod. mass [kg]        |       Prod. rate [kg/s]       |
|------------------|---------------|-----------------|------------------------------|-------------------------------|
|   INJECTOR.LEFT  |  all_regions  |              1  |  [0, 8.589625290559929e+56]  |   [0, 9.941695938148062e+47]  |
|  INJECTOR.RIGHT  |  all_regions  |              1  |  [0, -8.58962529055992e+56]  |  [0, -9.941695938148053e+47]  |
|        flux_set  |  all_regions  |              2  |  [0, 8.711228593176025e+41]  |   [0, 9.735556609752802e+32]  |
---------------------------------------------------------------------------------------------------------------------

Platform (please complete the following information):

• Machine: Pine
• Compiler: gcc 12.2.0
• GEOS Version: 1.1.0 (sha1: 8c90dae)
• MPI version: Open MPI v5.0.5

Additional context

• Files and logs attached in zip.

fluxStats.zip