nwm_metrics.metric_functions#

Module containing functions to calculate various metrics for evaluating hydrological model performance.

Functions:

treat_values: Preprocess time series data by removing negative values, NaN values, and replacing zero values.
pearson_corr: Calculate Pearson correlation coefficient and p-value between observed and simulated values.
mean_abs_error: Calculate mean absolute error between observed and simulated values.
root_mean_squared_error: Calculate root mean squared error or mean squared error between observed and simulated values.
rmse_std_ratio: Calculate the ratio of RMSE to the standard deviation of observed values.
percent_bias: Calculate percent bias between observed and simulated values.
nse: Calculate Nash-Sutcliffe efficiency, with options for logarithmic transformation and normalization.
weighted_nse: Calculate a weighted average of NSE and logarithmic NSE.
kge: Calculate Kling-Gupta efficiency between observed and simulated values.
pbias_fdc: Calculate percent bias of flow duration curve segments (high, medium, low flow) based on exceedance probabilities.
categorical_score: Calculate categorical scores (POD, FAR, CSI, FBIAS) based on a specified threshold.
calculate_metrics: A wrapper function to calculate multiple metrics at once based on user-specified metric names.

Functions

`calculate_metrics`	Compute all statistical metrics between simulation and observation.
`categorical_score`	Compute probability of detection (POD), probability of false_alarm (FAR), cirtical success index (CSI) and frequency bias (FBIAS).
`get_threshold_value`	Get threshold value based on the specified threshold type and value.
`kge`	Compute Kling-Gupta efficiency between simulation and observation.
`mean_abs_error`	Compute mean absolute error between simulation and observation.
`nse`	Compute Nash-Sutcliffe efficiency.
`pbias_fdc`	Compute percent bias of flow duration curve (FDC) high-segment volume, midsegment slope and low-segment volume according to Yilmaz et al (2008).
`pearson_corr`	Compute mean squared error, or optionally root mean squared error.
`percent_bias`	Compute mean squared error, or optionally root mean squared error.
`rmse_std_ratio`	Compute ratio of RMSE between simulation and observation to standard deviation of observation.
`root_mean_squared_error`	Compute root mean squared error, or optionally mean squared error.
`treat_values`	Remove NaN, inf and negative values, and replace zero values of time series.
`weighted_nse`	Compute weighted average of NSE of raw time series and NSE of logrithmic time series.