generate_errors¶
Constant error¶
- de.generate_errors.constant(ts, offset=1.5)[source]¶
Generate constant errors.
Constant errors are generated by multiplying with either constant positive offset or constant negative offset.
- Parameters
ts ((N,)array_like) – Observed time series
offset (float, optional) – Offset multiplied to time series. If greater than 1 positive constant offset and if less than 1 negative constant offset. The default is 1.5.
- Returns
ts_const – Time series with constant error
- Return type
array_like
Dynamic error¶
- de.generate_errors.positive_dynamic(ts, prop=0.5)[source]¶
Generate positive dynamic errors (i.e. Overestimate high flows - Underestimate low flows)
High to medium flows are increased by linear decreasing factors. Medium to low flows are decreased by linear decreasing factors.
- Parameters
ts (dataframe) – Dataframe with time series
prop (float, optional) – Factor by which time series is tilted.
- Returns
ts_dyn – Time series with positive dynamic error
- Return type
dataframe
- de.generate_errors.negative_dynamic(ts, prop=0.5)[source]¶
Generate negative dynamic error (i.e Underestimate high flows - Overestimate low flows)
High to medium flows are decreased by linear increasing factors. Medium to low flows are increased by linear increasing factors.
- Parameters
ts (dataframe) – Observed time series
prop (float, optional) – Factor by which time series is tilted.
- Returns
ts_dyn – Time series with negative dynamic error
- Return type
dataframe
Timing error¶
- de.generate_errors.timing(ts, tshift=3, shuffle=True)[source]¶
Generate timing errors.
Timing errors are generated by either shifting or shuffling.
- Parameters
ts (dataframe) – dataframe with time series
tshift (int, optional) – days by which time series is shifted. Both positive and negative time shift are possible. The default is 3 days.
shuffle (boolean, optional) – If True, time series is shuffled. The default is shuffling.
- Returns
ts_tim – Time series with timing error
- Return type
dataframe
