Ten diffusive domains w schism boundary #573

kumdonoaa · 2022-07-07T21:13:57Z

The changes made here enabled for diffusive wave channel routing to run in serial mode on all ten available mainstem domains where channel cross section survey data are available. Also, schism's water depth output was integrated with t-route so that the water depth values along shorelines were used as downstream boundary conditions for the diffusive wave.

Additions

diffusive.f90: 1) Using schism water depth values as diffusive wave's downstream boundary condition is activated with minimum depth enforced for numerical stability; 2) Commented out maxCourant variable as it is solely used for bookkeeping with the initial value undefined
nhd_io.py: 1) In Lines 505~510, if statement was added just in case when flowveldepth.index doesn't have an index of link_gage_df, 2) in line 1582, a new equation to compute water depth from a channel bottom using Schism ouputs was implemented.
input.py: diffusive_parameters dictionary was simply replaced by hybrid_parameters that include diffusive_parameters.
compute.py: In lines 1158 to1159 and 1172 and 1173, initialization of related numpy arrays was introduced for running multiple diffusive domains
diffusive_utils.py: Lines 1055~1060 prevents dbcd_df (dataframe of schism water depth for diffusive boundary condition) from having any zero or negative water depth.
HurricaneLaura/domain/coastal_10diffusive.txt: mask files for Hurricane Laura (or Atl/Gulf covering all ten diffusive domains) domain
HurricaneLaura/domain/coastal_domain_subset.yaml: lists of links of diffusive domains on original hydrofabric
HurricaneLaura/domain/refactored_coastal_domain_subset.yaml: lists of links of diffusive domains on refactored hydrofabric (refactored so as to remove too short stream segments by converging to adjacent segments)
HurricaneLaura/domain/final_diffusive_natural_xs.nc: natural channel cross section data on original hydrofabric including (x, z) vertices data and manning's N values at each point
HurricaneLaura/domain/refac_final_diffusive_natural_xs.nc: natural channel cross section data on refactored hydrofabric including (x, z) vertices data and manning's N values at each point

Removals

Changes

Testing

In order to run diffusive/MC on Hurricane Laura domain, that is, diffusive on ten mainstems and MC on the rest tributary networks), run test_AnA.yaml in test/HurricaneLaura. The command is "python3 -m nwm_routing -f test_AnA.yaml" after compiling using ./complier.sh in t-route directory.

Screenshots

Notes

Todos

Checklist

Testing checklist

Target Environment support

Windows
Linux
Browser

Accessibility

Keyboard friendly
Screen reader friendly

Other

src/troute-network/troute/nhd_network.py

src/troute-network/troute/nhd_network_utilities_v02.py

src/troute-nwm/src/nwm_routing/output.py

test/HurricaneLaura/test_AnA.yaml

hellkite500

I don't think we need to include any of the test/HurricaneLaura in the repository itself. We can make an archive of the relevant domain data and (relative) configurations, but they shouldn't get commited to the repo.

src/kernel/diffusive/diffusive.f90

hellkite500 · 2022-07-12T19:26:24Z

src/troute-network/troute/nhd_io.py

@@ -485,12 +502,20 @@ def write_chanobs(
    -------------

    '''
+    # TODO: for and if statements are improvised just in case when link of gage location is not in flowveldepth, which should be fixed
+    link_na = []
+    for link in list(link_gage_df.index):


I don't think you need to convert this to a list to simply iterate it, same with the next .index...

list() removed. Thanks.

hellkite500 · 2022-07-12T19:27:35Z

src/troute-network/troute/nhd_io.py

+    for link in list(link_gage_df.index):
+        if link not in list(flowveldepth.index):
+            link_na.append(link)
+    link_gage_df_nona = link_gage_df.drop(link_na)


could probably just drop this inplace...link_gaga_df.drop(link_na, inplace=True)

doesn't look like this change got committed/pushed???

hellkite500 · 2022-07-12T19:33:48Z

src/troute-network/troute/nhd_io.py

+    start_date = ds.variables['time'].units
+
+    start_date   = dparser.parse(start_date,fuzzy=True)
+    dt_schism    = 3600 # [sec]


could this be dynamically inferred from the timesteps array:

if len(timesteps) > 1: dt_schism = timesteps[1]-timesteps[0] else: raise RuntimeError("schism provided less than 2 time steps")

changed. Thanks for your code.

hellkite500 · 2022-07-12T19:34:08Z

src/troute-network/troute/nhd_io.py

+    start_date   = dparser.parse(start_date,fuzzy=True)
+    dt_schism    = 3600 # [sec]
+    dt_timeslice = timedelta(minutes=dt_schism/60.0)
+    start_date   = start_date # + dt_timeslice


remove this line

hellkite500 · 2022-07-12T20:19:21Z

src/kernel/diffusive/diffusive.f90

@@ -1547,7 +1612,8 @@ function rtsafe(i, j, Q_cur, Q_ds, z_cur, z_ds, y_ds)
    double precision            :: x1, x2, df, dxx, dxold, f, fh, fl, temp, xh, xl
    double precision            :: y_norm, y_ulm_multi, y_llm_multi, elv_norm, y_old
    double precision            :: rtsafe    
-
+    open(unit=301, file="./output/rtsafe_check.txt")


remove open file unit

hellkite500 · 2022-07-12T20:19:59Z

src/kernel/diffusive/diffusive.f90

      else                         ! newton step acceptable. take it.
        dxold  = dxx
        dxx    = f / df
        temp   = rtsafe
        rtsafe = rtsafe - dxx
-        if (temp == rtsafe) return
+        if (temp == rtsafe) then


remove this whole block since it is a noop now anyways.

removed. thanks.

hellkite500 · 2022-07-12T20:20:11Z

src/kernel/diffusive/diffusive.f90

      end if

-      if (abs(dxx) < xacc) return  ! convergence criterion.
+      if (abs(dxx) < xacc) then 
+        !write(301, *) "fnal:", i, j, iter, rtsafe


another noop block here

done. Thanks.

hellkite500 · 2022-07-12T20:20:27Z

src/kernel/diffusive/diffusive.f90


    allocate(el1(nel), a1(nel), peri1(nel), redi1(nel), redi1All(nel))
    allocate(equiv_mann(nel), conv1(nel), tpW1(nel))
    allocate(newdKdA(nel))
    allocate(compoundSKK(nel), elev(nel))
    allocate(i_start(nel), i_end(nel))

-    !open(unit=201, file="./output/xsec_natural.txt")
+    open(unit=301, file="./output/xsec_natural_check.txt")


remove this file unit

hellkite500 · 2022-07-12T20:21:05Z

src/kernel/diffusive/diffusive.f90

+          !iv = 0
+          !do ii2 = iel, nel
+          !  iv = iv + 1
+          !  dmyarr(iv) = newdKdA(ii2)


remove any unused/unwanted code from this funnction

hellkite500 · 2022-07-13T16:24:00Z

src/troute-routing/troute/routing/diffusive_utils.py

-    para_ar_g[2]  = 50.0    # lower limit of diffusivity (default: 50)
-    para_ar_g[3]  = 5000.0  # upper limit of diffusivity (default: 1000)
+    para_ar_g[2]  = 10.0    # lower limit of diffusivity (default: 50)
+    para_ar_g[3]  = 40000.0  # upper limit of diffusivity (default: 1000)


If these are the default values, then the comments should be updated appropriately...

Changed. Thanks.

hellkite500 · 2022-07-15T13:33:02Z

what is this file for? coastal_10diffusive.txt and is it the same in both lower colorado and hurricane laura?

kumdonoaa · 2022-07-15T14:01:50Z

what is this file for? coastal_10diffusive.txt and is it the same in both lower colorado and hurricane laura?

It is a mask file for diffusive domains. Even it can be used for both cases, I can slim it down for Lower Colorado. I will adjust it shortly.

hellkite500 · 2022-07-15T15:51:12Z

Is the one in the hurricane laura directory still needed? test/HurricaneLaura/domain/coastal_10diffusive.txt

hellkite500

I think this can be merged for now. I think some additional refactoring and cleanup would be useful in places, but we can address that once the functionality is merged and can changes can be tested against a "stable" implementation/

hellkite500 · 2022-07-15T18:47:25Z

src/troute-network/troute/nhd_io.py

+    for link in list(link_gage_df.index):
+        if link not in list(flowveldepth.index):
+            link_na.append(link)
+    link_gage_df_nona = link_gage_df.drop(link_na)


doesn't look like this change got committed/pushed???

Dong Kim added 7 commits July 7, 2022 10:31

f90 files

9df60f8

step2

5a6d3f7

run all ten diffusive domains in serial with schism depth boundary data

6dc8c53

code update 6/29/2022

9b86efe

run all ten diffusive domains in serial with schism boundary

ae989d4

more clean up

620e29e

comment out unused maxCourant variable

a67d85e

kumdonoaa requested a review from hellkite500 July 7, 2022 22:39

BrianAvant requested changes Jul 8, 2022

View reviewed changes

src/troute-network/troute/nhd_network.py Show resolved Hide resolved

src/troute-network/troute/nhd_network_utilities_v02.py Outdated Show resolved Hide resolved

src/troute-nwm/src/nwm_routing/output.py Show resolved Hide resolved

test/HurricaneLaura/test_AnA.yaml Show resolved Hide resolved

Dong Kim added 2 commits July 8, 2022 11:31

removed private file paths

3ba2126

further cleanup

9b93cb2

hellkite500 requested changes Jul 12, 2022

View reviewed changes

Dong Kim added 2 commits July 13, 2022 09:36

change dt_schisma and dt_db_g and cleanup

43ac381

remove coastal_domain_testing.yaml

a36e5a6

hellkite500 reviewed Jul 13, 2022

View reviewed changes

Dong Kim added 8 commits July 13, 2022 11:16

remove old files in LowererColorado_TX folder

f845c93

remove 2 more old files in LowerColorado_TX folder

97c64a7

updated hydrofabric & natural xsec files

fbf2f6b

remove private path in yaml

2308ea3

change default value of diffusivity limits

14e3366

remove list()

52c0dea

list() removed

35d189b

remove two natural xsec nc files

80be5c3

updated mask file only for Lower Colorado River

18808a1

remove data of HurricaneLaura

ee19b68

hellkite500 approved these changes Jul 15, 2022

View reviewed changes

BrianAvant approved these changes Jul 15, 2022

View reviewed changes

kumdonoaa merged commit 0c9ec38 into NOAA-OWP:master Jul 18, 2022

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Ten diffusive domains w schism boundary #573

Ten diffusive domains w schism boundary #573

kumdonoaa commented Jul 7, 2022 •

edited

Loading

hellkite500 left a comment

hellkite500 Jul 12, 2022

kumdonoaa Jul 13, 2022

hellkite500 Jul 12, 2022

kumdonoaa Jul 13, 2022

hellkite500 Jul 15, 2022

hellkite500 Jul 12, 2022

kumdonoaa Jul 12, 2022

hellkite500 Jul 12, 2022

kumdonoaa Jul 12, 2022

hellkite500 Jul 12, 2022

kumdonoaa Jul 14, 2022

hellkite500 Jul 12, 2022

kumdonoaa Jul 14, 2022

hellkite500 Jul 12, 2022

kumdonoaa Jul 14, 2022

hellkite500 Jul 12, 2022

kumdonoaa Jul 14, 2022

hellkite500 Jul 12, 2022

kumdonoaa Jul 14, 2022

hellkite500 Jul 13, 2022

kumdonoaa Jul 13, 2022

hellkite500 commented Jul 15, 2022

kumdonoaa commented Jul 15, 2022

hellkite500 commented Jul 15, 2022

hellkite500 left a comment

hellkite500 Jul 15, 2022

Ten diffusive domains w schism boundary #573

Ten diffusive domains w schism boundary #573

Conversation

kumdonoaa commented Jul 7, 2022 • edited Loading

Additions

Removals

Changes

Testing

Screenshots

Notes

Todos

Checklist

Testing checklist

Target Environment support

Accessibility

Other

hellkite500 left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

hellkite500 commented Jul 15, 2022

kumdonoaa commented Jul 15, 2022

hellkite500 commented Jul 15, 2022

hellkite500 left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

kumdonoaa commented Jul 7, 2022 •

edited

Loading