Diagnostic output skill¶

Measurement diagnostics (type 1) support different type of skill assessments by comparison between the model result and the measurements stored in the file:

skill() - table with skill statistics
scatter() - scatter plot

The same skill assessment methods and properties are available for their main data properties .result, .forecast, .forecast_at_update, and .analysis (if the file contains no updates only .result and .forecast will be available and they will be identical).

In [1]:

import fmdap

In [2]:

fn = '../tests/testdata/Diagnostics_F16_EnKF.dfs0'
diag = fmdap.read_diagnostic(fn, name="F16")

skill()¶

In [3]:

diag.skill()

Out[3]:

	n	bias	rmse	urmse	mae	cc	si	r2
observation
F16 analysis	58	0.207470	0.419241	0.364307	0.311775	0.970885	0.087019	0.920583
F16 forecast	67	0.169513	0.393557	0.355179	0.281344	0.973028	0.088905	0.930080
F16 forecast@update	58	0.204999	0.418657	0.365033	0.307975	0.970494	0.087193	0.920804
F16 result	67	0.171653	0.394095	0.354748	0.284634	0.973410	0.088797	0.929889

In [4]:

diag.forecast_at_update.skill().round(3)

Out[4]:

	n	bias	rmse	urmse	mae	cc	si	r2
observation
F16 forecast@update	58	0.205	0.419	0.365	0.308	0.97	0.087	0.921

In [5]:

diag.analysis.skill(metrics=["mape","mef","nse","r2","lin_slope"])

Out[5]:

	n	mape	mef	nse	r2	lin_slope
observation
F16 analysis	58	7.902526	0.28181	0.923925	0.920583	0.991982

scatter()¶

diag.scatter() = diag.result.scatter()

In [6]:

diag.scatter(figsize=(6,6))

In [7]:

diag.analysis.scatter(figsize=(6,6), show_points=False)

In [8]:

diag.forecast.scatter(figsize=(6,6), show_hist=False)

In [ ]: