1.5.9 zbry clean by jkousal32 · Pull Request #138 · ecmwf-ifs/ecwam

jkousal32 · 2026-04-29T18:10:44Z

Description

BYDRZ alternative source term package (ST6-like in the WAVEWATCH III convention)

At this point, I believe all is up to the ecwam standard except for the noted BYDRZ TODO, which I will address soon (EDIT*** completed this TODO now)

I believe this is now a good point to get some more eyes on this code.

Very minor changes in ifs-source and ifs-scripts to support those in ecWAM.

These PRs (across ifs-source, ifs-scripts, ecWAM) should be merged together.

ecWAM: 1.5.9 zbry clean #138 (this)
ifs-source: https://github.com/ecmwf-ifs/ifs-source/pull/46
ifs-scripts: https://github.com/ecmwf-ifs/ifs-scripts/pull/73

Update: these will go into 51R1 together, once declared, with ecwam 1.6

Contributor Declaration

By opening this pull request, I affirm the following:

All authors agree to the Contributor License Agreement.
The code follows the project's coding standards.
I have performed self-review and added comments where needed.
I have added or updated tests to verify that my changes are effective and functional.
I have run all existing tests and confirmed they pass.

… on SINFLX for bydbr;optional LLFACT; comments!!!

… setwavephys

Copilot

Pull request overview

Note

Copilot was unable to run its full agentic suite in this review.

Adds the BYDRZ (ST6-like) alternative wave physics package to ECWAM, including new wind-input and dissipation terms, and exposes the configuration via run scripts/tests.

Changes:

Introduces a new physics branch (IPHYS=2) with BYDRZ wind input, dissipation, swell dissipation, stress diagnostics, and air–sea interaction handling.
Extends core interfaces (e.g., SINFLX/SDISSIP/SINPUT/AIRSEA) to pass additional state needed by the new physics (e.g., CGROUP, CHRNCK, gustiness outputs).
Adds/updates test configurations and run-script parameters to exercise the new physics options.

Reviewed changes

Copilot reviewed 33 out of 35 changed files in this pull request and generated 10 comments.

Show a summary per file

File	Description
tests/etopo1_oper_an_fc_O48_cy50r1_bydrz.yml	Adds a BYDRZ-specific regression test config and expected validation stats.
tests/etopo1_oper_an_fc_O48_cy50r1.yml	Extends baseline test outputs to include additional stress-related fields.
src/ecwam/yowstat.F90	Adds new runtime switches/parameters (e.g., IPHYS2_AIRSEA, LLLOWWINDS, ZCDFAC).
src/ecwam/yowphys.F90	Adds BYDRZ physics parameters (including water viscosity constant and BYDRZ tunables).
src/ecwam/yowpcons.F90	Centralizes Euler’s constant (GAMMA_E) in constants module.
src/ecwam/yowfred.F90	Adds allocatable frequency-grid helper arrays for BYDRZ extensions.
src/ecwam/wsigstar.F90	Extends gustiness routine to also return SIG_U10.
src/ecwam/wdfluxes.F90	Updates calls to SINFLX/SDISSIP to match new signatures.
src/ecwam/w_maxh.F90	Switches to module-provided GAMMA_E constant.
src/ecwam/userin.F90	Wires new namelist/state variables and prints BYDRZ-related settings.
src/ecwam/tauwindsxy.F90	New helper to integrate wind-input spectra into stress components.
src/ecwam/tau_wave_atmos.F90	New routine to compute stress from negative input (waves → atmosphere), incl. HF extension.
src/ecwam/swldissip_bydrz.F90	New BYDRZ swell dissipation source term.
src/ecwam/sinput_jan.F90	Updates WSIGSTAR call-site for new SIG_U10 output and comment cleanup.
src/ecwam/sinput_ard.F90	Updates WSIGSTAR call-site for new SIG_U10 output.
src/ecwam/sinput.F90	Extends SINPUT interface (CGROUP, CHRNCK; Z0M/UFRIC intent changes).
src/ecwam/sinflx_bydrz.F90	New BYDRZ wind input + stress balance + gustiness + PHIWA handling.
src/ecwam/sinflx_ard_jan.F90	Refactors existing SINFLX logic into a dedicated ARD/JAN implementation.
src/ecwam/sinflx.F90	Dispatches to SINFLX_ARD_JAN vs SINFLX_BYDRZ based on IPHYS.
src/ecwam/setwavphys.F90	Adds default parameterization for IPHYS=2 and IPHYS2_AIRSEA options.
src/ecwam/sdissip_bydrz.F90	New BYDRZ dissipation source term.
src/ecwam/sdissip.F90	Dispatches dissipation to BYDRZ implementations for IPHYS=2.
src/ecwam/outbeta.F90	Extends local variables for new flux/stress-related diagnostics.
src/ecwam/mpuserin.F90	Adds defaults and documentation for new IPHYS2_AIRSEA/LLLOWWINDS/ZCDFAC controls.
src/ecwam/lfactor.F90	New reduction-factor solver to enforce stress constraints via spectral reduction.
src/ecwam/initmdl.F90	Adds BYDRZ-specific frequency-space setup and HF extension grid allocation.
src/ecwam/implsch.F90	Adjusts SINFLX call count and updated interfaces for new physics.
src/ecwam/frcutindex.F90	Minor whitespace-only change.
src/ecwam/calcphiwa.F90	New energy-flux-from-input computation (PHIWA) helper.
src/ecwam/airsea_jan.F90	New JAN-specific AIRSEA implementation extracted from AIRSEA.
src/ecwam/airsea_bydrz.F90	New BYDRZ-specific AIRSEA implementation.
src/ecwam/airsea.F90	Refactors AIRSEA into a dispatcher with optional HALP/RNFAC for JAN branch.
src/ecwam/CMakeLists.txt	Registers new source files in the build.
share/ecwam/scripts/ecwam_run_model.sh	Exposes IPHYS2_AIRSEA and LLLOWWINDS config into the generated namelist.
.gitignore	Ignores `x*.sh` files.

Comments suppressed due to low confidence (1)

src/ecwam/yowpcons.F90:1

GAMMA_E is a physical/mathematical constant and should be declared as a PARAMETER to prevent accidental runtime modification and to enable compiler optimizations. Consider switching this to REAL(KIND=JWRB), PARAMETER :: GAMMA_E = ....

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

jkousal32 · 2026-05-05T15:58:28Z

@jrbidlot @awnawab heads up, have now completed the last TODO. Now I believe all is up to the ecwam standard and good to go!

awnawab · 2026-05-07T07:27:27Z

As discussed offline, I've retargeted this PR to develop-1.6, which we can use to collect 51R1 contributions. We can also ignore the failing tests, these are due to failing loki builds, which I'll mop up after this PR is merged.

awnawab

Thanks a lot for this @jkousal32 🙏 I can't comment on the science but technically this is implemented quite well, I've just left a few minor comments which would be nice to address before we merge this. Just for my own understanding, does this new physics package offer similar forecasting skill at a reduced computational cost? I seem to remember hearing something about that, although it's quite possible I hallucinated that 😅

Once the above points are addressed, I would be happy to merge this, and just to reiterate don't worry about the failing loki builds. The cpu hpc-ci build doesn't use loki preprocessing, and this test passes.

After this PR, it would be good to have a discussion offline about whether we do want to continue to maintain three physics packages in the long term or whether one of these could potentially eventually be retired. Secondly, I'd like to get your thoughts on possibly have the branching between the various packages at a much higher level outside the OpenMP loop. Although this may lead to a bit of code duplication, we would end up with simpler and potentially faster (on CPU and especially on GPU) compute kernels. Let's discuss further on our dedicated waves channel 😊

awnawab · 2026-05-12T15:42:50Z

+!       Wind height
+      ZNLEV    = 10._JWRB
+
+      !$loki inline


Could you please remove this new loki inline directive? This particular one is misplaced and should preceed a call, but in general, the performance impact of inlining is not easily determined a priori, so we shouldn't add these by default.

awnawab · 2026-05-12T15:43:51Z

+!*    3. DETERMINE ROUGHNESS LENGTH (if needed).
+!        ---------------------------
+
+      !$loki inline


Same comment here about inlining.

awnawab · 2026-05-12T15:49:47Z

+#include "taut_z0.intfb.h"
+#include "z0wave.intfb.h"
+
+INTEGER(KIND=JWIM), INTENT(IN) :: KIJS, KIJL, ICODE_WND, IUSFG


The indentation is a little off here. Could we please align the declarations and the subroutine body itself with the module imports?

awnawab · 2026-05-12T15:51:08Z

+
+      IMPLICIT NONE
+
+REAL(KIND=JWRB), DIMENSION(NANG,NFRE), INTENT(IN)  :: SPOS ! POS Sin(sigma) in [m2/Hz]


Same comment here about indentation.

awnawab · 2026-05-12T15:51:53Z

+!/ 0) --- split integral into low/high frequency contributions ------------- /
+!
+!
+!     Th=2pi,f=inf                 Th=2pi,f=FR(NFRE)       Th=2pi,f=inf  


Nice ASCII math 🤓👏

awnawab · 2026-05-12T16:11:04Z

+#include "calcphiwa.intfb.h"
+
+
+INTEGER(KIND=JWIM), INTENT(IN) :: ICALL  !! CALL NUMBER.


Same comment here about indentation.

awnawab · 2026-05-12T16:15:02Z

+
+      IMPLICIT NONE
+
+INTEGER(KIND=JWIM), INTENT(IN) :: KIJS, KIJL


Indentation please.

awnawab · 2026-05-12T16:24:55Z

+! Determine value of spectrum at NFRE (i.e. at highest frequency). 
+!  - Note, direction dimension must remain
+
+ZA_SX       = SX(:,NFRE)


Explicit bounds on the LHS please.

awnawab · 2026-05-12T16:25:06Z

+
+#include "tauwindsxy.intfb.h"
+
+REAL(KIND=JWRB), DIMENSION(NANG,NFRE), INTENT(IN)  :: S ! Sin(sigma) in [m2/rad-Hz]


Indentation please.

awnawab · 2026-05-12T16:26:12Z

+
+      IMPLICIT NONE
+
+REAL(KIND=JWRB), INTENT(IN)  :: SDENSX_IN(*), SDENSY_IN(*)


Why do we need assumed size declarations here, would assumed shape not suffice? Also indentation please.

jkousal32 · 2026-05-18T12:15:09Z

Hi @awnawab
No, unfortunately no speed up brought at this stage relative to 50r1.
I did get a big speed up relative to my first implementation though, which might be what you are thinking of:

This was a first (big) pass at trying to reduce the cost. Perhaps more speed up can be found. Timings are for the source term operations in the O48 test.

Thanks for your other comments. Will address soon!

Plans for further offline discussion sounds good.

jkousal32 added 30 commits April 29, 2026 16:55

first commit of bydbr physics

d7da32e

bugfix, ecWAM with new bydbr physics now returning reasonable results

7f58f0a

small code tidy

49940fd

freq cut off bydbr;formatting;CGROUP thru implsch;test w/o iterations…

8ee35e4

… on SINFLX for bydbr;optional LLFACT; comments!!!

add frcutindex_bydb

af9a343

bring SINPUT_BYDBR into SINFLX_BYDBR

dc75235

contents of sinput_bydbr.F90 moved within sinflx_bydbr.F90

f9a5893

bugfix on CHRNCK

5059abd

tidy of moving of sinput_bydbr into sinflx

ceef4ab

output extra ocean parameters

680711e

bug fix and temporary fix on FRIC=32

6c8657a

implement different flavors of airsea for ZBRY

04029f5

airsea_iter.F90 replaced by airsea_zbry.F90

54bc68e

remove unnecessary line

42d2b3f

renaming bydbr to zbry

ebccf9c

renaming bydbr to zbry

6cbecd7

correct zbry implementation for NGST=1

069e617

bugfix on SIG_U10

17291df

for IPHYS2_AIRSEA=0, diagnose CD directly from U10 & UFRIC

79e99e5

further bugfix on implementation NGST=1

df58475

diagnose CD; output UPROXY

bec8006

bydbr config file replaced with zbry

428dbf6

small tidy

06abc33

add low winds parameterization from Mohammed Yasrab

ab7349a

rename IPHYS2_LOWWINDS to LLLOWWINDS

baefb1c

small tidy

6ce3296

bring IJ loops within (SINFLX_ZBRY)

bb20eb0

bring IJ loops within (S*DISSIP_ZBRY); bring zbry wave phys vars into…

3e7e139

… setwavephys

iron out a bug in IJ looping; clean up

d51792b

forgot to index A (action) spectrum

f5f3a9e

jkousal32 added 10 commits April 29, 2026 17:00

major restructure for ZBRY to homogenize with ecWAM

1799cb3

handle allocation in initmdl

9cc9ec4

proper indexing and loops

421765b

refactoring for efficiency

cc3c78c

tidy up in/out & comment documentation thereof

7e76673

further optimizations and refactorizations

90ab549

tauwinds into tauwindsxy for efficiency

d73ed52

modify indentation

8d05114

minor aesthetic tidy

e11b026

renaming of zbry to bydrz

55ac7a6

jkousal32 requested review from awnawab, Copilot and jrbidlot April 29, 2026 18:10

jkousal32 self-assigned this Apr 29, 2026

github-actions Bot added the contributor label Apr 29, 2026

Copilot AI reviewed Apr 29, 2026

View reviewed changes

jkousal32 and others added 6 commits April 29, 2026 18:22

iterative scheme simplification for IPHYS2_AIRSEA=3; indentation

ec10162

Update src/ecwam/sinput_jan.F90

4f2ca92

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

Update src/ecwam/airsea_jan.F90

081aaf5

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

update comments

194abdf

remove dummy inits

080a571

setup initializations cleanly for different physics

05e8b50

awnawab added the approved-for-ci label May 5, 2026

github-actions Bot removed the approved-for-ci label May 5, 2026

address B1 issue

c202a3c

awnawab changed the base branch from main to develop-1.6 May 7, 2026 07:25

awnawab requested changes May 13, 2026

View reviewed changes


		IMPLICIT NONE

		REAL(KIND=JWRB), DIMENSION(NANG,NFRE), INTENT(IN) :: SPOS ! POS Sin(sigma) in [m2/Hz]

		#include "calcphiwa.intfb.h"


		INTEGER(KIND=JWIM), INTENT(IN) :: ICALL !! CALL NUMBER.


		#include "tauwindsxy.intfb.h"

		REAL(KIND=JWRB), DIMENSION(NANG,NFRE), INTENT(IN) :: S ! Sin(sigma) in [m2/rad-Hz]


		IMPLICIT NONE

		REAL(KIND=JWRB), INTENT(IN) :: SDENSX_IN(), SDENSY_IN()

Uh oh!

Conversation

jkousal32 commented Apr 29, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Contributor Declaration

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

jkousal32 commented May 5, 2026

Uh oh!

awnawab commented May 7, 2026

Uh oh!

awnawab left a comment

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!

jkousal32 commented May 18, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

jkousal32 commented Apr 29, 2026 •

edited

Loading

jkousal32 commented May 18, 2026 •

edited

Loading