The problem
When generating precipitation timeseries over global mean (and presumably also other sub-regions), the internal variability does not match what one gets from averaging over the field generated on the full grid. We expect the fully gridded output to be the true result,
Most likely there is some discrepancy in the two generation methods, and perhaps the problem is to do with ordering of operation, generating the noise on top of the original average result perhaps.
The plot shows the issue (erroneous plot annotation notwithstanding) the plots show two left panels monthly and annual timeseries of precipitation from the global mean timeseries generation, whereas the rightmost plot is the global mean of annual gridded data.
This is related to #82, however #84 addressed this in a way that solved the problem for the gridded data, but the spatially averaged timeseries remain erroneous
The problem
When generating precipitation timeseries over global mean (and presumably also other sub-regions), the internal variability does not match what one gets from averaging over the field generated on the full grid. We expect the fully gridded output to be the true result,
Most likely there is some discrepancy in the two generation methods, and perhaps the problem is to do with ordering of operation, generating the noise on top of the original average result perhaps.
The plot shows the issue (erroneous plot annotation notwithstanding) the plots show two left panels monthly and annual timeseries of precipitation from the global mean timeseries generation, whereas the rightmost plot is the global mean of annual gridded data.
This is related to #82, however #84 addressed this in a way that solved the problem for the gridded data, but the spatially averaged timeseries remain erroneous