Related to #279, for functions that perform reductions, there should be a keyword argument that keep the dimension (can be at the first element of the dimension).
This is needed for calibration (see CliMA/ClimaCalibrate.jl#267) as there is an implicit assumption that a time dimension is required. Alternatively, this solution can also be tackled in ClimaCalibrate by resolving the issue that is linked.
Related to #279, for functions that perform reductions, there should be a keyword argument that keep the dimension (can be at the first element of the dimension).
This is needed for calibration (see CliMA/ClimaCalibrate.jl#267) as there is an implicit assumption that a time dimension is required. Alternatively, this solution can also be tackled in ClimaCalibrate by resolving the issue that is linked.