saco.Dataset.infer_mean_abstraction

Dataset.infer_mean_abstraction(scenario: str = 'FL', percentile: int = 95, exclude_swabs_with_hofs: bool = True, exclude_gwabs: List[str] = None, exclude_swabs: List[str] = None)

Infer mean abstraction from impacts under a given scenario and percentile.

Parameters:

  • scenario – Abbreviation of artificial influences scenario used as basis for inferring long-term average abstraction.

  • percentile – Flow percentile (natural) used as basis for inferring long-term average abstraction.

  • exclude_swabs_with_hofs – Whether to exclude SWABS with HOFs from long-term average calculations.

  • exclude_gwabs – Groundwater abstractions whose long-term average should not be inferred. List should contain entries from UNIQUEID in GWABs_NBB.

  • exclude_swabs – Surface water abstractions whose long-term average should not be inferred. List should contain entries from UNIQUEID in SWABS_NBB.

Notes

WRGIS models the relationships between long-term average abstraction and impacts at different flow percentiles. Here we use these relationships to estimate long-term average abstraction given impacts at specific (single) flow percentile. This is done under an assumption that the relative seasonal/FDC profile of impacts remains constant.

The impact under a given scenario/percentile combination includes the effect of local consumptiveness. The long-term average numbers calculated using this method exclude local consumptiveness. This is reflected by the “WR” vs “NR” suffixes in the impact numbers for a specific percentile (WR = “water returned”) vs the long-term average abstraction numbers (NR = “no water returned”).

Lists of abstractions to be excluded from long-term average calculations can be supplied via the exclude_gwabs and exclude_swabs arguments. Abstractions to be excluded should be specified using their UNIQUEID. If a long-term average column is already present before this method is called, the existing value will be retained. If not, a NaN will be inserted for these abstractions.

This method operates only for the SWABS_NBB and GWABs_NBB tables. Complex abstractions/impacts in the SupResGW_NBB table are not handled. By default, SWABS with HOFs are excluded from long-term average calculations (see exclude_swabs_with_hofs argument). This is because the method may not yield a reasonable long-term average abstraction if the impact in SWABS_NBB at the reference percentile is being constrained by the HOF condition.

For example, a SWAB might be “off” at Q95 due to a HOF condition. However, it might well be unreasonable for its long-term average to become zero. Hence it is not appropriate to apply the seasonal disaggregation factors to obtain a revised long-term average in this case.