Possible EMOS amendments by gavinevans · Pull Request #11 · gavinevans/improver

gavinevans · 2021-06-14T14:26:45Z

Description
A branch containing some possible EMOS amendments. Tests have not been updated. or added to support the new features.

This PR is only to help facilitate an initial look and will not be merged.

Testing:

Ran tests and they passed OK
Added new tests for the new feature(s)

…d coordinate.

… for use in testing.

…nd longitude, so that the truth cubelist is mergeable.

…l standardisation.

…se of this distribution including to support the use of additional predictors for this distribution.

…tinue.

…ogy.

fionaRust · 2021-06-14T15:06:35Z

-                    predictor,
-                ).x.astype(np.float32)
-            )
+        for index, (truth_slice, fv_slice) in enumerate(zip(truth.slices_over(sindex), forecast_var.slices_over(sindex))):


Preprocessing of arrays for additional predictors before passing to minimising,

fionaRust · 2021-06-14T15:07:18Z

-            forecast_predictor_data = flatten_ignoring_masked_data(
-                forecast_predictor.data
-            )
+            fp_data_list = []


More preprocessing for additional predictors

fionaRust · 2021-06-14T15:08:44Z

            )
-            a_b = np.array([a] + b.tolist(), dtype=np.float64)
+
+        if predictor.lower() == "mean":


Again additional predictors and more initiall guesses

fionaRust · 2021-06-14T15:12:22Z

+                import statsmodels.api as sm
                forecast_predictor_flattened = flatten_ignoring_masked_data(
-                    forecast_predictor
+                    forecast_predictor, preserve_leading_dimension=True


All additional predictor related

fionaRust · 2021-06-14T15:13:53Z

-            for (truths_slice, fp_slice) in zip(
-                truths.slices_over(index), forecast_predictor.slices_over(index)
-            ):
+            for truths_slice in truths.slices_over(index):


Similar to above, pre processing of additional predictors

fionaRust · 2021-06-15T14:30:44Z

+                   "beta coefficients in order to create a calibrated forecast.")
+            raise ValueError(msg)
+
+        if self.standardise_cubelist:


All the equations in this section to do with applying coefficients if the standardisation was used when calculating the coefficients are now correct to the best of my understanding.

fionaRust · 2021-06-15T14:36:20Z

+    return std_forecast, std_truth, forecast_mean, forecast_sd, truth_mean, truth_sd
+
+
+def standardise_forecasts(historic_forecasts, hf_coords=["realization", "time"]):


This function and the one below aren't currently used.

fionaRust · 2021-06-15T14:41:10Z

+
+    if truths.coords("time", dim_coords=True):
+        truth_mean = truths.collapsed(truth_coords, nanmean)
+    else:


Why are these else statements needed?

This is dealing with a spin up problem. It will not impact the testing we are currently doing as we should always have more than one truth. For now it is probably best to fail if this else statement is reached.

fionaRust · 2021-06-29T09:48:13Z

+                    coord_values={y_name: lambda cell: any(np.isclose(cell.point, truth_slice.coord(axis="y").points)),
+                                x_name: lambda cell: any(np.isclose(cell.point, truth_slice.coord(axis="x").points))})
+            new_fp_data = []
+            for fp_cube in forecast_predictors:


I think I need some help explaining this

This is for static predictors where they don't change with time. Give them a time coordinate so they match the shape of the other predictor.

I've added a comment to indicate that this is to account for static predictors.

fionaRust · 2021-06-29T10:34:11Z

@@ -215,13 +218,38 @@ def filter_non_matching_cubes(
        if truth_slice:


Suggested change

if truth_slice:

if truth_slice and af_slices:

matching_historic_forecasts.append(hf_slice)

matching_truths.append(truth_slice)

matching_additional_fields.extend(af_slices)

I've modified this to avoid the possible issue of truths and historic forecasts being added potentially without a required additional predictor.

fionaRust

I'm now happy that these changes are suitable for use in the EMOS comparison experiments for temperature

… not in realizations space.

…andardisation option.

Add acceptance tests for QRF

gavinevans added 15 commits May 13, 2021 16:23

Modifications to support local standardisation of EMOS coefficients.

0fb1e0b

Edit to temporarily ignore metadata differences in the forecast perio…

de464dc

…d coordinate.

Initial edits to explore ways of incorporating additional predictors.

ef19a53

Initial rough version that seems to work for global EMOS corrections.

e7a0ee2

Modifications to support point-by-point and using default initial guess.

c1fb1e6

Modifications to add global standardisation capability to this branch…

801cf8d

… for use in testing.

Adding filtering of truth to ensure a consistent altitude, latitude a…

36896bd

…nd longitude, so that the truth cubelist is mergeable.

Add computation of a skew-logistic distribution to this branch.

dcc5fc6

Add masking of the forecast when the truth becomes masked by the loca…

145cd36

…l standardisation.

Amendments related to the skew logistic distribution to correct the u…

f9357f0

…se of this distribution including to support the use of additional predictors for this distribution.

Fixes to errors in code.

d721233

Modifications to raise errors in the standardisation, rather than con…

99855ca

…tinue.

Alterations to support standardising using only the forecast climatol…

36d41ea

…ogy.

Minor updates to support local standardisation using forecasts.

cd819c7

Corrections following further testing.

e5f53d3

fionaRust reviewed Jun 29, 2021

View reviewed changes

gavinevans added 2 commits June 29, 2021 13:06

Minor edits following review comments.

de17321

Correction to code.

4aebecf

gavinevans force-pushed the impro_2102_additional_predictors branch from b72eb70 to 4aebecf Compare June 29, 2021 13:41

Further corrections.

6691653

fionaRust approved these changes Jun 29, 2021

View reviewed changes

gavinevans added 3 commits June 30, 2021 09:47

Update to only check the points of the additional predictors fields.

149a788

Minor edits including fix for when applying coefficients to forecasts…

8108f91

… not in realizations space.

Update to provide handling for a NaN scale parameter for the local st…

0ca37f0

…andardisation option.

gavinevans pushed a commit that referenced this pull request Aug 6, 2025

Merge pull request #11 from gavinevans/mobt_877_implement_qrf2

af6f7ed

Add acceptance tests for QRF

		return std_forecast, std_truth, forecast_mean, forecast_sd, truth_mean, truth_sd


		def standardise_forecasts(historic_forecasts, hf_coords=["realization", "time"]):

		@@ -215,13 +218,38 @@ def filter_non_matching_cubes(
		if truth_slice:

-        if truth_slice:
+    if truth_slice and af_slices:
+        matching_historic_forecasts.append(hf_slice)
+        matching_truths.append(truth_slice)
+        matching_additional_fields.extend(af_slices)

Conversation

gavinevans commented Jun 14, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

fionaRust left a comment

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

gavinevans commented Jun 14, 2021 •

edited

Loading