feat(output): CF time with bounds; clearer names, labels and defaults by xiangtaoxu · Pull Request #58 · EDmodel/ED-MORE

xiangtaoxu · 2026-10-29T23:33:45Z

What and why

The last of four stacked output pull requests (fixes, new variables, diel_, clean-ups). Stacked on
#57 (on #56, on #55): until those merge, this PR's diff includes them, so review the last commit,
c1161e7.

1. CF time with bounds.

The problem. time was the decimal calendar year of a record's start, with units = "year".
No CF reader decodes that, and a record had no end. ex05's README told users to take a record's
width as t[i+1] - t[i] and drop the last, partial record.
The change.
- time is each record's start in days since <run start>, with calendar = "proleptic_gregorian" (ED-MORE has leap years, and ex05 starts in 1600) and standard_name = "time".
- time_bnds(time, nv) holds the start and the end of the record's last step, so a partial record
  (run start or end, a restart) carries its true span. A snapshot's instant is the upper bound.
- The year, month and day companions (and F's hour, minute, second) stay.
- diagnostics.md §4 says how to turn a rate into an amount with the bounds. ex05's README and
  plot_pft_size.py (the one shipped reader of time) use the new axis.
The time units (days since the run's start, proleptic Gregorian) are the owner's choice, confirmed
in review.

2. Names, labels and defaults (the plan's §5, decision D9):

mean_transp_leaf_py → mean_leaf_transp_py, moved to the ecophys group and off by default.
Its cohort source mean_transp_co → mean_leaf_transp_co. Both are per leaf area (mol m⁻² leaf
s⁻¹); mean_transp_py, the transpiration in kg m⁻² s⁻¹, keeps its name. Their diel_ twins
follow. ex05's Palo Verde io_config is updated.
mean_abs_par_co's long name says what it holds, the incident-equivalent PAR (the absorbed VIS
over the leaf absorptance), not "absorbed".
mean_cas_co2_pa moves to the energy group, beside mean_cas_co2_py.
mean_soil_temp_top_pa defaults to F D M, like its siblings (it was F M).
diagnostics.md §7 no longer says the radiation group holds an albedo.
From the review: mean_rh_py/_pa → mean_rhet_py/_pa, so "rh" no longer reads as relative
humidity (feat(output): the water and energy budgets close from the output; sub-daily values everywhere #56 names the new autotrophic term mean_raut_py).
From the review: mean_ca_co and mean_pressure_co are removed. They repeated, for every
cohort, the patch's canopy-air CO₂ and the forcing's pressure, and only edmore.canopy's
docstring named them; it now points the offline leaf replay at mean_cas_co2_pa (the previous
sub-step's, the CO₂ the solve saw) and mean_psurf_py. Two cohort rows fewer (29 fast, 39 in
all).

Left for two follow-up PRs, after the owner's review: first a refactor of the internal "film"
vocabulary to veg_surface, then these: mean_cas_depth_py/_pa read different sources (the state
and the forcing's echo); mean_sw_up_py duplicates the VIS + NIR pair; the ET sources and the
residuals have no _pa twins.

Design: the plan agreed with the owner (§5 and D9). Drafted with Claude Code (see Credit).

Effect on results (tick exactly one; CI checks)

Bit-for-bit: output is identical to the base commit's in every run listed under
Verification. A new option that is off by default counts when the model is unchanged with it off.
Round-off: differences at round-off level only, e.g. from reordered sums.
Answer-changing: results move; give the largest changes below.

The model and every value are bit for bit. What changes is how the files label and select them: the
encoding of time, the new time_bnds, two renames and one default (listed under Verification).

User-facing changes (tick all that apply; each needs an "Upgrading" note in the changelog)

None
Config keys added, renamed or removed, or a value in the shipped config files changed
Output variables or output files changed
Input files (forcing, maps, initial conditions) or their format changed
Building or running changed: command-line or CMake options, dependencies, the Python API
State files written by the last release no longer load
Output: time units and calendar, and time_bnds; mean_transp_leaf_py, mean_transp_co and
mean_rh_py/_pa renamed (the first off by default); mean_ca_co and mean_pressure_co removed;
mean_soil_temp_top_pa on D by default. Each has an Upgrading note.

Verification

Base commit: feat(output): diel_, the mean diel cycle of a daily, monthly or yearly record #57's head, a209fc4.
Runs compared with the base commit: dev builds, ifx 2026.1 and gfortran, 3 months each, F on.
Two runs: the reference config with the shipped census, and example 04's calibrated BCI run on the
tower forcing. On both compilers:
- the state is identical, 95/95;
- every variable on every tier is identical, except time_bnds (new on every tier),
  mean_soil_temp_top_pa (now on D) and mean_transp_leaf_py (renamed, off by default);
- time decodes as days since 2001-06-01 00:00:00, proleptic Gregorian, with the monthly bounds
  [[0, 30], [30, 61], [61, 92]].
Tests: pytest output_files now checks time's units, calendar and bounds on M (each month),
D (each day) and F (hourly, end to end).
ctest: 98/98 on ifx and on gfortran dev builds; lint clean; ruff check clean.
ed_io_config.toml regenerated.
Debug build: at the top of the stack: see the comment below once it has run.
The long tests (label long): not run.
Other builds: not run (production, nvfortran).
Conservation warnings in those runs: none.

Credit and AI use

Collaborators:
AI assistance: Claude Code (Claude Opus 5.5) wrote the code, tests and docs from the plan agreed
with the owner, and ran the verification above.
I answer for this pull request: I can explain every change in it, and the verification above was actually run

Not in this PR

The leftovers listed under §2.
Example 04's hand-built diurnal composite could read diel_ records.

🤖 Generated with Claude Code

…writes The cohort order of a month's records. The stand is re-sorted every slow step (daily growth and mortality), but the output folded cohort slabs by slot, so a monthly cohort record averaged two cohorts in one slot after a mid-month swap, and staged fast-loop rows were reduced with the post-sort weights. Every cohort value of a month is now folded into its cohort's slot of the month (keyed by global_cohort_id: the live state, the diagnostic rows and the staged rows, whose ids and weights the fast loop stages once per step), and every record of the month is written in the month's end order (write_order, before the I/O phase). The axes' bookkeeping is also written on the F tier, read from the stand. No overwrites. A run lists every diagnostic file it would write and stops before its first step if one exists; the writer creates files with NC_NOCLOBBER. RunConfig.remove_diagnostics() removes a configuration's files; the examples' scripts and calibrate_fast's trials call it before they run again. file_chunk honoured. A tier with a cohort or patch variable was cut to monthly files (F: daily) whatever its file_chunk said. The cohort and patch axes are now unlimited and grow within a file; each record names its cohorts and patches. "run" is renamed "all"; a monthly chunk of a day and an annual one shorter than a year are refused. Model state is bit for bit (ifx and gfortran dev builds, two runs). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

…ion, dew Four polygon outputs split ET by where the vapour comes from, as the canopy air's vapour ledger booked it on each step: - mean_transp_py transpiration - mean_evap_canopy_py evaporation of the intercepted water (leaf and wood films) - mean_evap_ground_py ground evaporation: bare soil and snow pack - mean_dew_py condensate (dew, fog) sent to soil layer 1 ET = transp + evap_canopy + evap_ground - dew - the canopy air's vapour gain. Default tiers D M Y, F on request (they are patch-block fast rows). The films' evaporation is b-weighted into the step's boundary-flux amounts (stage_bflux_t / boundary_flux_t, both integrators); close_ledgers exports the four amounts in step_exports_t, and patch_diag_row turns them into rates. PA_TRANSP held le_flux / L_v (ET) under a transpiration label and no output read it; it now holds transpiration. The patch rows after it move up by three. test_fast_integ_step: on both schemes, each step's sources less the dew less ET equal the canopy air's net vapour inflow (worst 1.5e-16 ESDIRK2, 2.1e-16 RK45, relative), the morning rain's films evaporate, and the day's canopy and ground evaporation agree between ESDIRK2 and RK45. Bit for bit: example 04's calibrated BCI run on the production ifx build matches the archived v0.4.0 run in all 24 variables both write (43,824 hourly and 1,826 daily records). ctest 95/95 on ifx and gfortran dev builds. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

…-daily values everywhere Every term the whole-column water and energy ledgers book is now on the patch row and an output variable, so both budgets close from the files on any tier (docs/science/diagnostics.md §4, The budgets): shed water, film excess, store gain and the water's paths and stores; bottom heat, advected enthalpy, the stores' energy gains and H by source. Adds Ra, the respiration split and Rh on the patch row, ecosystem carbon, absorbed PAR, the leaves' wet fraction and patch ET. ET and runoff are read from the fast loop's exports (the polygon accumulators are gone), and the formerly sub-daily-only variables get coarse sources, so every fast-loop quantity is on F and on D/M/Y. A per-cohort row reduced to the patch or polygon is reduced on F from the staged rows. New pytest output_budgets closes both budgets from a run's D and F files. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

…y record Every mean_<x> with a sub-daily value has a diel_<x>: its mean at each time of day over the record, in bins of the new [output].diel_interval ("1 hour" to "4 hours"), on a diel axis (the last) with the coordinate diel_hour and its bounds. Each staged sub-step is folded into its bin of the coarse tiers' open records as the F tier reads it, before the step closes them; a cohort's cycle goes to its slot of the month. D, M and Y (D and M on the cohort and patch axes), never F; off by default, so the output is bit for bit. Needs the fast loop and a dt_fast that divides the bin. MAX_OUTPUT_VARS rises to 1024 for the ~150 new names. New pytest output_diel: a daily record is, bin by bin, its day's sub-daily records, on the polygon, patch, soil and cohort axes, at 1- and 3-hour bins; a diel_ variable on F is refused. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

time is now each record's start in days since the run's start (proleptic Gregorian, standard_name time), with time_bnds: the start and the end of the record's last step, so a partial record carries its true span and xarray and CDO decode the axis. The year/month/day companions stay. mean_transp_leaf_py -> mean_leaf_transp_py (ecophys, off by default) and mean_transp_co -> mean_leaf_transp_co, so the per-leaf-area transpiration has one name at every level apart from mean_transp_py. mean_abs_par_co's long name says it is the incident-equivalent PAR; mean_cas_co2_pa joins its polygon twin in the energy group; mean_soil_temp_top_pa is on D by default like its siblings; diagnostics.md §7 no longer promises an albedo. ex05's plot script and README read the new time axis. Model state is bit for bit. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

xiangtaoxu · 2026-10-29T23:35:59Z

Debug build at the top of the stack (this branch, ifx, runtime checks on): builds clean, ctest 96/96 passed (the Debug build registers 96 of the dev builds' 98 tests).

🤖 Generated with Claude Code

A run still stops before its first step when a file it would write exists, unless the new [output].overwrite = true (default false) lets it replace them (NC_CLOBBER, no start-up check). The examples' configs and the Python scripts set it, so rerunning an example replaces its own files; RunConfig.remove_diagnostics() and the scripts' clean-up helpers are gone. The output_files test reruns with overwrite on and checks every file is written again. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

xiangtaoxu · 2026-10-30T00:56:07Z

+   ! Rn + bottom heat - H - LE + advected enthalpy - store gains = the energy residual.
+   !---~---
+   integer(ik), parameter, public :: PA_HEAT_BOTTOM       = 23_ik  ! [W/m2] heat into the soil column through its bottom
+   integer(ik), parameter, public :: PA_ENTH_ADVECT       = 24_ik  ! [W/m2] enthalpy the water carries in, beyond H and LE


What water flux does this term cover?

xiangtaoxu · 2026-10-30T00:57:16Z

+   integer(ik), parameter, public :: PA_LEAF_RESP         = 33_ik  ! [umol CO2/m2/s] leaf dark respiration
+   integer(ik), parameter, public :: PA_STEM_RESP         = 34_ik  ! [umol CO2/m2/s] stem maintenance respiration
+   integer(ik), parameter, public :: PA_ROOT_RESP         = 35_ik  ! [umol CO2/m2/s] fine-root maintenance respiration
+   integer(ik), parameter, public :: PA_RH                = 36_ik  ! [umol CO2/m2/s] heterotrophic respiration


This can be mistaken as relative humidity. I suggest use RHet and for RA use RAut

xiangtaoxu · 2026-10-30T00:58:46Z

+   integer(ik), parameter, public :: CO_LEAF_PAR     = 27_ik  ! [umol photon/m2 leaf/s] incident-equivalent PAR per leaf area
+   integer(ik), parameter, public :: CO_GB_MOL       = 28_ik  ! [mol H2O/m2 leaf/s] boundary-layer conductance
+   integer(ik), parameter, public :: CO_PSI_PREDAWN  = 29_ik  ! [MPa] the stomatal limb's predawn leaf potential
+   integer(ik), parameter, public :: CO_CA           = 30_ik  ! [umol/mol] canopy-air CO2 the leaf sees


Is this the leaf surface boundary layer CO2?

xiangtaoxu · 2026-10-30T00:59:25Z

+   integer(ik), parameter, public :: CO_GB_MOL       = 28_ik  ! [mol H2O/m2 leaf/s] boundary-layer conductance
+   integer(ik), parameter, public :: CO_PSI_PREDAWN  = 29_ik  ! [MPa] the stomatal limb's predawn leaf potential
+   integer(ik), parameter, public :: CO_CA           = 30_ik  ! [umol/mol] canopy-air CO2 the leaf sees
+   integer(ik), parameter, public :: CO_PRESSURE     = 31_ik  ! [Pa] air pressure the leaf sees


Why do we need to output this variable, isn't it the same as CAS pressure?

xiangtaoxu · 2026-10-30T01:00:47Z

+   integer(ik), parameter, public :: PA_SNOWMELT          = 48_ik  ! [kg/m2/s] meltwater the snow pack drains to the pond
+   integer(ik), parameter, public :: PA_SUBLIMATION       = 49_ik  ! [kg/m2/s] snow pack -> canopy air (deposition < 0)
+   integer(ik), parameter, public :: PA_SHED_WATER        = 50_ik  ! [kg/m2/s] tissue water the slow step shed to the pond
+   integer(ik), parameter, public :: PA_FILM_EXCESS       = 51_ik  ! [kg/m2/s] film water above capacity, removed


I recommend avoid film, use veg_surface_water or something similar instead

…s enthalpy split by direction Review of EDmodel#58: "film" and "canopy" read unclearly, and "rh"/"ra" read as relative humidity. - mean_canopy_water_py -> mean_veg_surface_water_py, mean_evap_canopy_py -> mean_veg_surface_evap_py, mean_film_excess_py -> mean_veg_surface_overflow_py, mean_evap_ground_py -> mean_ground_evap_py, mean_ra_py -> mean_raut_py; the patch rows and exports this PR added follow (PA_RHET, PA_RAUT, ...). - mean_enth_advect_py ("advected" suggests horizontal transport) is split by direction: mean_precip_enth_py (the rain and snow's in) and mean_outflow_enth_py (runoff, drainage, overflow and the exported vapour's beyond LE, out). The budget test closes on the pair. Model state is bit for bit with the previous head. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

…CO2 and pressure removed Review of EDmodel#58: - mean_rh_py/_pa -> mean_rhet_py/_pa (FLD_P_RH -> FLD_P_RHET): "rh" read as relative humidity. - mean_ca_co and mean_pressure_co are removed with their rows (CO_CA, CO_PRESSURE): they repeated, for every cohort, the patch's canopy-air CO2 at the sub-step's start and the forcing's air pressure. Only edmore.canopy's docstring named them; it now points the offline leaf replay at mean_cas_co2_pa (the previous sub-step's) and mean_psurf_py. 29 fast cohort rows, 39 in all. Model state is bit for bit; mean_rhet_* equals the old mean_rh_* record for record. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

…-daily values everywhere (#56) * fix(output): one cohort order per month, file_chunk honoured, no overwrites The cohort order of a month's records. The stand is re-sorted every slow step (daily growth and mortality), but the output folded cohort slabs by slot, so a monthly cohort record averaged two cohorts in one slot after a mid-month swap, and staged fast-loop rows were reduced with the post-sort weights. Every cohort value of a month is now folded into its cohort's slot of the month (keyed by global_cohort_id: the live state, the diagnostic rows and the staged rows, whose ids and weights the fast loop stages once per step), and every record of the month is written in the month's end order (write_order, before the I/O phase). The axes' bookkeeping is also written on the F tier, read from the stand. No overwrites. A run lists every diagnostic file it would write and stops before its first step if one exists; the writer creates files with NC_NOCLOBBER. RunConfig.remove_diagnostics() removes a configuration's files; the examples' scripts and calibrate_fast's trials call it before they run again. file_chunk honoured. A tier with a cohort or patch variable was cut to monthly files (F: daily) whatever its file_chunk said. The cohort and patch axes are now unlimited and grow within a file; each record names its cohorts and patches. "run" is renamed "all"; a monthly chunk of a day and an annual one shorter than a year are refused. Model state is bit for bit (ifx and gfortran dev builds, two runs). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(changelog): cite #55 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * feat(output): ET's sources: transpiration, canopy and ground evaporation, dew Four polygon outputs split ET by where the vapour comes from, as the canopy air's vapour ledger booked it on each step: - mean_transp_py transpiration - mean_evap_canopy_py evaporation of the intercepted water (leaf and wood films) - mean_evap_ground_py ground evaporation: bare soil and snow pack - mean_dew_py condensate (dew, fog) sent to soil layer 1 ET = transp + evap_canopy + evap_ground - dew - the canopy air's vapour gain. Default tiers D M Y, F on request (they are patch-block fast rows). The films' evaporation is b-weighted into the step's boundary-flux amounts (stage_bflux_t / boundary_flux_t, both integrators); close_ledgers exports the four amounts in step_exports_t, and patch_diag_row turns them into rates. PA_TRANSP held le_flux / L_v (ET) under a transpiration label and no output read it; it now holds transpiration. The patch rows after it move up by three. test_fast_integ_step: on both schemes, each step's sources less the dew less ET equal the canopy air's net vapour inflow (worst 1.5e-16 ESDIRK2, 2.1e-16 RK45, relative), the morning rain's films evaporate, and the day's canopy and ground evaporation agree between ESDIRK2 and RK45. Bit for bit: example 04's calibrated BCI run on the production ifx build matches the archived v0.4.0 run in all 24 variables both write (43,824 hourly and 1,826 daily records). ctest 95/95 on ifx and gfortran dev builds. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * feat(output): the water and energy budgets close from the output; sub-daily values everywhere Every term the whole-column water and energy ledgers book is now on the patch row and an output variable, so both budgets close from the files on any tier (docs/science/diagnostics.md §4, The budgets): shed water, film excess, store gain and the water's paths and stores; bottom heat, advected enthalpy, the stores' energy gains and H by source. Adds Ra, the respiration split and Rh on the patch row, ecosystem carbon, absorbed PAR, the leaves' wet fraction and patch ET. ET and runoff are read from the fast loop's exports (the polygon accumulators are gone), and the formerly sub-daily-only variables get coarse sources, so every fast-loop quantity is on F and on D/M/Y. A per-cohort row reduced to the patch or polygon is reduced on F from the staged rows. New pytest output_budgets closes both budgets from a run's D and F files. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(changelog): the default files' new variables, an upgrading note Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(changelog): cite #56 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * feat(output): [output].overwrite, an opt-in to replace existing output A run still stops before its first step when a file it would write exists, unless the new [output].overwrite = true (default false) lets it replace them (NC_CLOBBER, no start-up check). The examples' configs and the Python scripts set it, so rerunning an example replaces its own files; RunConfig.remove_diagnostics() and the scripts' clean-up helpers are gone. The output_files test reruns with overwrite on and checks every file is written again. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * refactor(output): veg_surface, ground_evap and raut names; the water's enthalpy split by direction Review of #58: "film" and "canopy" read unclearly, and "rh"/"ra" read as relative humidity. - mean_canopy_water_py -> mean_veg_surface_water_py, mean_evap_canopy_py -> mean_veg_surface_evap_py, mean_film_excess_py -> mean_veg_surface_overflow_py, mean_evap_ground_py -> mean_ground_evap_py, mean_ra_py -> mean_raut_py; the patch rows and exports this PR added follow (PA_RHET, PA_RAUT, ...). - mean_enth_advect_py ("advected" suggests horizontal transport) is split by direction: mean_precip_enth_py (the rain and snow's in) and mean_outflow_enth_py (runoff, drainage, overflow and the exported vapour's beyond LE, out). The budget test closes on the pair. Model state is bit for bit with the previous head. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

…y record (#57) * fix(output): one cohort order per month, file_chunk honoured, no overwrites The cohort order of a month's records. The stand is re-sorted every slow step (daily growth and mortality), but the output folded cohort slabs by slot, so a monthly cohort record averaged two cohorts in one slot after a mid-month swap, and staged fast-loop rows were reduced with the post-sort weights. Every cohort value of a month is now folded into its cohort's slot of the month (keyed by global_cohort_id: the live state, the diagnostic rows and the staged rows, whose ids and weights the fast loop stages once per step), and every record of the month is written in the month's end order (write_order, before the I/O phase). The axes' bookkeeping is also written on the F tier, read from the stand. No overwrites. A run lists every diagnostic file it would write and stops before its first step if one exists; the writer creates files with NC_NOCLOBBER. RunConfig.remove_diagnostics() removes a configuration's files; the examples' scripts and calibrate_fast's trials call it before they run again. file_chunk honoured. A tier with a cohort or patch variable was cut to monthly files (F: daily) whatever its file_chunk said. The cohort and patch axes are now unlimited and grow within a file; each record names its cohorts and patches. "run" is renamed "all"; a monthly chunk of a day and an annual one shorter than a year are refused. Model state is bit for bit (ifx and gfortran dev builds, two runs). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(changelog): cite #55 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * feat(output): ET's sources: transpiration, canopy and ground evaporation, dew Four polygon outputs split ET by where the vapour comes from, as the canopy air's vapour ledger booked it on each step: - mean_transp_py transpiration - mean_evap_canopy_py evaporation of the intercepted water (leaf and wood films) - mean_evap_ground_py ground evaporation: bare soil and snow pack - mean_dew_py condensate (dew, fog) sent to soil layer 1 ET = transp + evap_canopy + evap_ground - dew - the canopy air's vapour gain. Default tiers D M Y, F on request (they are patch-block fast rows). The films' evaporation is b-weighted into the step's boundary-flux amounts (stage_bflux_t / boundary_flux_t, both integrators); close_ledgers exports the four amounts in step_exports_t, and patch_diag_row turns them into rates. PA_TRANSP held le_flux / L_v (ET) under a transpiration label and no output read it; it now holds transpiration. The patch rows after it move up by three. test_fast_integ_step: on both schemes, each step's sources less the dew less ET equal the canopy air's net vapour inflow (worst 1.5e-16 ESDIRK2, 2.1e-16 RK45, relative), the morning rain's films evaporate, and the day's canopy and ground evaporation agree between ESDIRK2 and RK45. Bit for bit: example 04's calibrated BCI run on the production ifx build matches the archived v0.4.0 run in all 24 variables both write (43,824 hourly and 1,826 daily records). ctest 95/95 on ifx and gfortran dev builds. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * feat(output): the water and energy budgets close from the output; sub-daily values everywhere Every term the whole-column water and energy ledgers book is now on the patch row and an output variable, so both budgets close from the files on any tier (docs/science/diagnostics.md §4, The budgets): shed water, film excess, store gain and the water's paths and stores; bottom heat, advected enthalpy, the stores' energy gains and H by source. Adds Ra, the respiration split and Rh on the patch row, ecosystem carbon, absorbed PAR, the leaves' wet fraction and patch ET. ET and runoff are read from the fast loop's exports (the polygon accumulators are gone), and the formerly sub-daily-only variables get coarse sources, so every fast-loop quantity is on F and on D/M/Y. A per-cohort row reduced to the patch or polygon is reduced on F from the staged rows. New pytest output_budgets closes both budgets from a run's D and F files. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(changelog): the default files' new variables, an upgrading note Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(changelog): cite #56 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * feat(output): diel_, the mean diel cycle of a daily, monthly or yearly record Every mean_<x> with a sub-daily value has a diel_<x>: its mean at each time of day over the record, in bins of the new [output].diel_interval ("1 hour" to "4 hours"), on a diel axis (the last) with the coordinate diel_hour and its bounds. Each staged sub-step is folded into its bin of the coarse tiers' open records as the F tier reads it, before the step closes them; a cohort's cycle goes to its slot of the month. D, M and Y (D and M on the cohort and patch axes), never F; off by default, so the output is bit for bit. Needs the fast loop and a dt_fast that divides the bin. MAX_OUTPUT_VARS rises to 1024 for the ~150 new names. New pytest output_diel: a daily record is, bin by bin, its day's sub-daily records, on the polygon, patch, soil and cohort axes, at 1- and 3-hour bins; a diel_ variable on F is refused. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(changelog): cite #57 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * feat(output): [output].overwrite, an opt-in to replace existing output A run still stops before its first step when a file it would write exists, unless the new [output].overwrite = true (default false) lets it replace them (NC_CLOBBER, no start-up check). The examples' configs and the Python scripts set it, so rerunning an example replaces its own files; RunConfig.remove_diagnostics() and the scripts' clean-up helpers are gone. The output_files test reruns with overwrite on and checks every file is written again. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * refactor(output): veg_surface, ground_evap and raut names; the water's enthalpy split by direction Review of #58: "film" and "canopy" read unclearly, and "rh"/"ra" read as relative humidity. - mean_canopy_water_py -> mean_veg_surface_water_py, mean_evap_canopy_py -> mean_veg_surface_evap_py, mean_film_excess_py -> mean_veg_surface_overflow_py, mean_evap_ground_py -> mean_ground_evap_py, mean_ra_py -> mean_raut_py; the patch rows and exports this PR added follow (PA_RHET, PA_RAUT, ...). - mean_enth_advect_py ("advected" suggests horizontal transport) is split by direction: mean_precip_enth_py (the rain and snow's in) and mean_outflow_enth_py (runoff, drainage, overflow and the exported vapour's beyond LE, out). The budget test closes on the pair. Model state is bit for bit with the previous head. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * build(output): regenerate ed_io_config.toml after the renames Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

…EDmodel#59) * fix(output): one cohort order per month, file_chunk honoured, no overwrites The cohort order of a month's records. The stand is re-sorted every slow step (daily growth and mortality), but the output folded cohort slabs by slot, so a monthly cohort record averaged two cohorts in one slot after a mid-month swap, and staged fast-loop rows were reduced with the post-sort weights. Every cohort value of a month is now folded into its cohort's slot of the month (keyed by global_cohort_id: the live state, the diagnostic rows and the staged rows, whose ids and weights the fast loop stages once per step), and every record of the month is written in the month's end order (write_order, before the I/O phase). The axes' bookkeeping is also written on the F tier, read from the stand. No overwrites. A run lists every diagnostic file it would write and stops before its first step if one exists; the writer creates files with NC_NOCLOBBER. RunConfig.remove_diagnostics() removes a configuration's files; the examples' scripts and calibrate_fast's trials call it before they run again. file_chunk honoured. A tier with a cohort or patch variable was cut to monthly files (F: daily) whatever its file_chunk said. The cohort and patch axes are now unlimited and grow within a file; each record names its cohorts and patches. "run" is renamed "all"; a monthly chunk of a day and an annual one shorter than a year are refused. Model state is bit for bit (ifx and gfortran dev builds, two runs). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(changelog): cite EDmodel#55 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * feat(output): ET's sources: transpiration, canopy and ground evaporation, dew Four polygon outputs split ET by where the vapour comes from, as the canopy air's vapour ledger booked it on each step: - mean_transp_py transpiration - mean_evap_canopy_py evaporation of the intercepted water (leaf and wood films) - mean_evap_ground_py ground evaporation: bare soil and snow pack - mean_dew_py condensate (dew, fog) sent to soil layer 1 ET = transp + evap_canopy + evap_ground - dew - the canopy air's vapour gain. Default tiers D M Y, F on request (they are patch-block fast rows). The films' evaporation is b-weighted into the step's boundary-flux amounts (stage_bflux_t / boundary_flux_t, both integrators); close_ledgers exports the four amounts in step_exports_t, and patch_diag_row turns them into rates. PA_TRANSP held le_flux / L_v (ET) under a transpiration label and no output read it; it now holds transpiration. The patch rows after it move up by three. test_fast_integ_step: on both schemes, each step's sources less the dew less ET equal the canopy air's net vapour inflow (worst 1.5e-16 ESDIRK2, 2.1e-16 RK45, relative), the morning rain's films evaporate, and the day's canopy and ground evaporation agree between ESDIRK2 and RK45. Bit for bit: example 04's calibrated BCI run on the production ifx build matches the archived v0.4.0 run in all 24 variables both write (43,824 hourly and 1,826 daily records). ctest 95/95 on ifx and gfortran dev builds. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * feat(output): the water and energy budgets close from the output; sub-daily values everywhere Every term the whole-column water and energy ledgers book is now on the patch row and an output variable, so both budgets close from the files on any tier (docs/science/diagnostics.md §4, The budgets): shed water, film excess, store gain and the water's paths and stores; bottom heat, advected enthalpy, the stores' energy gains and H by source. Adds Ra, the respiration split and Rh on the patch row, ecosystem carbon, absorbed PAR, the leaves' wet fraction and patch ET. ET and runoff are read from the fast loop's exports (the polygon accumulators are gone), and the formerly sub-daily-only variables get coarse sources, so every fast-loop quantity is on F and on D/M/Y. A per-cohort row reduced to the patch or polygon is reduced on F from the staged rows. New pytest output_budgets closes both budgets from a run's D and F files. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(changelog): the default files' new variables, an upgrading note Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(changelog): cite EDmodel#56 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * feat(output): diel_, the mean diel cycle of a daily, monthly or yearly record Every mean_<x> with a sub-daily value has a diel_<x>: its mean at each time of day over the record, in bins of the new [output].diel_interval ("1 hour" to "4 hours"), on a diel axis (the last) with the coordinate diel_hour and its bounds. Each staged sub-step is folded into its bin of the coarse tiers' open records as the F tier reads it, before the step closes them; a cohort's cycle goes to its slot of the month. D, M and Y (D and M on the cohort and patch axes), never F; off by default, so the output is bit for bit. Needs the fast loop and a dt_fast that divides the bin. MAX_OUTPUT_VARS rises to 1024 for the ~150 new names. New pytest output_diel: a daily record is, bin by bin, its day's sub-daily records, on the polygon, patch, soil and cohort axes, at 1- and 3-hour bins; a diel_ variable on F is refused. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(changelog): cite EDmodel#57 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * feat(output): CF time with bounds; clearer names, labels and defaults time is now each record's start in days since the run's start (proleptic Gregorian, standard_name time), with time_bnds: the start and the end of the record's last step, so a partial record carries its true span and xarray and CDO decode the axis. The year/month/day companions stay. mean_transp_leaf_py -> mean_leaf_transp_py (ecophys, off by default) and mean_transp_co -> mean_leaf_transp_co, so the per-leaf-area transpiration has one name at every level apart from mean_transp_py. mean_abs_par_co's long name says it is the incident-equivalent PAR; mean_cas_co2_pa joins its polygon twin in the energy group; mean_soil_temp_top_pa is on D by default like its siblings; diagnostics.md §7 no longer promises an albedo. ex05's plot script and README read the new time axis. Model state is bit for bit. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(changelog): cite EDmodel#58 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * feat(output): [output].overwrite, an opt-in to replace existing output A run still stops before its first step when a file it would write exists, unless the new [output].overwrite = true (default false) lets it replace them (NC_CLOBBER, no start-up check). The examples' configs and the Python scripts set it, so rerunning an example replaces its own files; RunConfig.remove_diagnostics() and the scripts' clean-up helpers are gone. The output_files test reruns with overwrite on and checks every file is written again. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * refactor(output): veg_surface, ground_evap and raut names; the water's enthalpy split by direction Review of EDmodel#58: "film" and "canopy" read unclearly, and "rh"/"ra" read as relative humidity. - mean_canopy_water_py -> mean_veg_surface_water_py, mean_evap_canopy_py -> mean_veg_surface_evap_py, mean_film_excess_py -> mean_veg_surface_overflow_py, mean_evap_ground_py -> mean_ground_evap_py, mean_ra_py -> mean_raut_py; the patch rows and exports this PR added follow (PA_RHET, PA_RAUT, ...). - mean_enth_advect_py ("advected" suggests horizontal transport) is split by direction: mean_precip_enth_py (the rain and snow's in) and mean_outflow_enth_py (runoff, drainage, overflow and the exported vapour's beyond LE, out). The budget test closes on the pair. Model state is bit for bit with the previous head. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * build(output): regenerate ed_io_config.toml after the renames Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * refactor(output): mean_rhet_*, and the cohort repeats of the patch's CO2 and pressure removed Review of EDmodel#58: - mean_rh_py/_pa -> mean_rhet_py/_pa (FLD_P_RH -> FLD_P_RHET): "rh" read as relative humidity. - mean_ca_co and mean_pressure_co are removed with their rows (CO_CA, CO_PRESSURE): they repeated, for every cohort, the patch's canopy-air CO2 at the sub-step's start and the forcing's air pressure. Only edmore.canopy's docstring named them; it now points the offline leaf replay at mean_cas_co2_pa (the previous sub-step's) and mean_psurf_py. 29 fast cohort rows, 39 in all. Model state is bit for bit; mean_rhet_* equals the old mean_rh_* record for record. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * refactor(fast): "film" identifiers become veg_surface / surface water The water the leaves and wood hold on their surfaces was called a "film" in the code; its state, config keys and outputs already say surf_water / veg_surface. Identifiers now follow: frozen%film -> frozen%veg_surface, film_evap_leaf/_wood -> leaf_/wood_surface_evap, cohort_film_water -> cohort_surface_water, limit_film_evaporation -> limit_surface_evaporation, clamp_canopy_film -> clamp_surface_water, leaf_film_coeff -> leaf_surface_evap_coeff, g_film_leaf/_w -> g_surf_leaf/_wood, and the tests' locals and names. No config key, state-file variable or output changes; comments and docs follow in the next commit. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(fast): "film" in comments, messages and docs becomes "surface water" The prose follows the identifiers of the previous commit: the water the leaves and wood hold on their surfaces is "surface water" throughout the comments, the tests' messages, the TOML comments and the science docs (vegetation_energy_dynamics.md's section is now "The canopy's surface water"). Kept: soil_biophysics.md's "film flow" (soil physics), and example 04's recorded fit.json, which the next calibration rewrites from parameters.toml. Bit for bit. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(changelog): cite EDmodel#59 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

…es and residuals (#60) * fix(output): one cohort order per month, file_chunk honoured, no overwrites The cohort order of a month's records. The stand is re-sorted every slow step (daily growth and mortality), but the output folded cohort slabs by slot, so a monthly cohort record averaged two cohorts in one slot after a mid-month swap, and staged fast-loop rows were reduced with the post-sort weights. Every cohort value of a month is now folded into its cohort's slot of the month (keyed by global_cohort_id: the live state, the diagnostic rows and the staged rows, whose ids and weights the fast loop stages once per step), and every record of the month is written in the month's end order (write_order, before the I/O phase). The axes' bookkeeping is also written on the F tier, read from the stand. No overwrites. A run lists every diagnostic file it would write and stops before its first step if one exists; the writer creates files with NC_NOCLOBBER. RunConfig.remove_diagnostics() removes a configuration's files; the examples' scripts and calibrate_fast's trials call it before they run again. file_chunk honoured. A tier with a cohort or patch variable was cut to monthly files (F: daily) whatever its file_chunk said. The cohort and patch axes are now unlimited and grow within a file; each record names its cohorts and patches. "run" is renamed "all"; a monthly chunk of a day and an annual one shorter than a year are refused. Model state is bit for bit (ifx and gfortran dev builds, two runs). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(changelog): cite #55 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * feat(output): ET's sources: transpiration, canopy and ground evaporation, dew Four polygon outputs split ET by where the vapour comes from, as the canopy air's vapour ledger booked it on each step: - mean_transp_py transpiration - mean_evap_canopy_py evaporation of the intercepted water (leaf and wood films) - mean_evap_ground_py ground evaporation: bare soil and snow pack - mean_dew_py condensate (dew, fog) sent to soil layer 1 ET = transp + evap_canopy + evap_ground - dew - the canopy air's vapour gain. Default tiers D M Y, F on request (they are patch-block fast rows). The films' evaporation is b-weighted into the step's boundary-flux amounts (stage_bflux_t / boundary_flux_t, both integrators); close_ledgers exports the four amounts in step_exports_t, and patch_diag_row turns them into rates. PA_TRANSP held le_flux / L_v (ET) under a transpiration label and no output read it; it now holds transpiration. The patch rows after it move up by three. test_fast_integ_step: on both schemes, each step's sources less the dew less ET equal the canopy air's net vapour inflow (worst 1.5e-16 ESDIRK2, 2.1e-16 RK45, relative), the morning rain's films evaporate, and the day's canopy and ground evaporation agree between ESDIRK2 and RK45. Bit for bit: example 04's calibrated BCI run on the production ifx build matches the archived v0.4.0 run in all 24 variables both write (43,824 hourly and 1,826 daily records). ctest 95/95 on ifx and gfortran dev builds. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * feat(output): the water and energy budgets close from the output; sub-daily values everywhere Every term the whole-column water and energy ledgers book is now on the patch row and an output variable, so both budgets close from the files on any tier (docs/science/diagnostics.md §4, The budgets): shed water, film excess, store gain and the water's paths and stores; bottom heat, advected enthalpy, the stores' energy gains and H by source. Adds Ra, the respiration split and Rh on the patch row, ecosystem carbon, absorbed PAR, the leaves' wet fraction and patch ET. ET and runoff are read from the fast loop's exports (the polygon accumulators are gone), and the formerly sub-daily-only variables get coarse sources, so every fast-loop quantity is on F and on D/M/Y. A per-cohort row reduced to the patch or polygon is reduced on F from the staged rows. New pytest output_budgets closes both budgets from a run's D and F files. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(changelog): the default files' new variables, an upgrading note Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(changelog): cite #56 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * feat(output): diel_, the mean diel cycle of a daily, monthly or yearly record Every mean_<x> with a sub-daily value has a diel_<x>: its mean at each time of day over the record, in bins of the new [output].diel_interval ("1 hour" to "4 hours"), on a diel axis (the last) with the coordinate diel_hour and its bounds. Each staged sub-step is folded into its bin of the coarse tiers' open records as the F tier reads it, before the step closes them; a cohort's cycle goes to its slot of the month. D, M and Y (D and M on the cohort and patch axes), never F; off by default, so the output is bit for bit. Needs the fast loop and a dt_fast that divides the bin. MAX_OUTPUT_VARS rises to 1024 for the ~150 new names. New pytest output_diel: a daily record is, bin by bin, its day's sub-daily records, on the polygon, patch, soil and cohort axes, at 1- and 3-hour bins; a diel_ variable on F is refused. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(changelog): cite #57 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * feat(output): CF time with bounds; clearer names, labels and defaults time is now each record's start in days since the run's start (proleptic Gregorian, standard_name time), with time_bnds: the start and the end of the record's last step, so a partial record carries its true span and xarray and CDO decode the axis. The year/month/day companions stay. mean_transp_leaf_py -> mean_leaf_transp_py (ecophys, off by default) and mean_transp_co -> mean_leaf_transp_co, so the per-leaf-area transpiration has one name at every level apart from mean_transp_py. mean_abs_par_co's long name says it is the incident-equivalent PAR; mean_cas_co2_pa joins its polygon twin in the energy group; mean_soil_temp_top_pa is on D by default like its siblings; diagnostics.md §7 no longer promises an albedo. ex05's plot script and README read the new time axis. Model state is bit for bit. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(changelog): cite #58 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * feat(output): [output].overwrite, an opt-in to replace existing output A run still stops before its first step when a file it would write exists, unless the new [output].overwrite = true (default false) lets it replace them (NC_CLOBBER, no start-up check). The examples' configs and the Python scripts set it, so rerunning an example replaces its own files; RunConfig.remove_diagnostics() and the scripts' clean-up helpers are gone. The output_files test reruns with overwrite on and checks every file is written again. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * refactor(output): veg_surface, ground_evap and raut names; the water's enthalpy split by direction Review of #58: "film" and "canopy" read unclearly, and "rh"/"ra" read as relative humidity. - mean_canopy_water_py -> mean_veg_surface_water_py, mean_evap_canopy_py -> mean_veg_surface_evap_py, mean_film_excess_py -> mean_veg_surface_overflow_py, mean_evap_ground_py -> mean_ground_evap_py, mean_ra_py -> mean_raut_py; the patch rows and exports this PR added follow (PA_RHET, PA_RAUT, ...). - mean_enth_advect_py ("advected" suggests horizontal transport) is split by direction: mean_precip_enth_py (the rain and snow's in) and mean_outflow_enth_py (runoff, drainage, overflow and the exported vapour's beyond LE, out). The budget test closes on the pair. Model state is bit for bit with the previous head. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * build(output): regenerate ed_io_config.toml after the renames Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * refactor(output): mean_rhet_*, and the cohort repeats of the patch's CO2 and pressure removed Review of #58: - mean_rh_py/_pa -> mean_rhet_py/_pa (FLD_P_RH -> FLD_P_RHET): "rh" read as relative humidity. - mean_ca_co and mean_pressure_co are removed with their rows (CO_CA, CO_PRESSURE): they repeated, for every cohort, the patch's canopy-air CO2 at the sub-step's start and the forcing's air pressure. Only edmore.canopy's docstring named them; it now points the offline leaf replay at mean_cas_co2_pa (the previous sub-step's) and mean_psurf_py. 29 fast cohort rows, 39 in all. Model state is bit for bit; mean_rhet_* equals the old mean_rh_* record for record. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * refactor(fast): "film" identifiers become veg_surface / surface water The water the leaves and wood hold on their surfaces was called a "film" in the code; its state, config keys and outputs already say surf_water / veg_surface. Identifiers now follow: frozen%film -> frozen%veg_surface, film_evap_leaf/_wood -> leaf_/wood_surface_evap, cohort_film_water -> cohort_surface_water, limit_film_evaporation -> limit_surface_evaporation, clamp_canopy_film -> clamp_surface_water, leaf_film_coeff -> leaf_surface_evap_coeff, g_film_leaf/_w -> g_surf_leaf/_wood, and the tests' locals and names. No config key, state-file variable or output changes; comments and docs follow in the next commit. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(fast): "film" in comments, messages and docs becomes "surface water" The prose follows the identifiers of the previous commit: the water the leaves and wood hold on their surfaces is "surface water" throughout the comments, the tests' messages, the TOML comments and the science docs (vegetation_energy_dynamics.md's section is now "The canopy's surface water"). Kept: soil_biophysics.md's "film flow" (soil physics), and example 04's recorded fit.json, which the next calibration rewrites from parameters.toml. Bit for bit. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(changelog): cite #59 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * feat(output): one source for the canopy-air depth; per-patch ET sources and residuals The leftovers of #58: - mean_cas_depth_py reads the patch row, the depth the fast loop used, as mean_cas_depth_pa does, so the polygon value is the patches' area-weighted sum; it joins mean_cas_depth_pa in the forcing group. It moves by at most 6 mm; the model is bit for bit. - mean_transp_pa, mean_veg_surface_evap_pa, mean_ground_evap_pa, mean_dew_pa, mean_resid_energy_pa and mean_resid_water_pa, off by default. mean_sw_up_py stays (calibrate_fast scores albedo with it), and example 04's composite stays hand-built (it skips the tower's gaps, which a diel_ record cannot). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(changelog): cite #60 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

xiangtaoxu and others added 10 commits October 29, 2026 17:57

docs(changelog): cite EDmodel#55

924f15f

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

docs(changelog): the default files' new variables, an upgrading note

e98b0fa

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

docs(changelog): cite EDmodel#56

7de1ce3

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

docs(changelog): cite EDmodel#57

a209fc4

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

docs(changelog): cite EDmodel#58

0b85a8e

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

xiangtaoxu requested a review from a team as a code owner October 29, 2026 23:33

xiangtaoxu and others added 4 commits October 29, 2026 20:26

Merge branch 'fix/output-cohort-order' into feat/output-variables

e230185

Merge branch 'feat/output-variables' into feat/output-diel

dfbc6df

Merge branch 'feat/output-diel' into refactor/output-cleanups

8f3c920

xiangtaoxu commented Oct 30, 2026

View reviewed changes

xiangtaoxu and others added 3 commits October 29, 2026 21:24

Merge branch 'feat/output-variables' into feat/output-diel

b6918c6

Merge branch 'feat/output-diel' into refactor/output-cleanups

c0e8e0b

xiangtaoxu mentioned this pull request Oct 30, 2026

feat(output): the water and energy budgets close from the output; sub-daily values everywhere #56

Merged

3 of 10 tasks

xiangtaoxu and others added 3 commits October 29, 2026 21:28

build(output): regenerate ed_io_config.toml after the renames

9c38225

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

Merge branch 'feat/output-diel' into refactor/output-cleanups

bb52e21

This was referenced Oct 30, 2026

refactor(fast): the canopy's "film" is surface water in code and docs #59

Merged

feat(output): one source for the canopy-air depth; per-patch ET sources and residuals #60

Merged

Merge beta (EDmodel#57 squashed) into refactor/output-cleanups

a0e7cca

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

xiangtaoxu merged commit c91a371 into EDmodel:beta Oct 30, 2026
4 checks passed

xiangtaoxu added a commit to xiangtaoxu/ED-MORE that referenced this pull request Oct 30, 2026

Merge beta (EDmodel#58 squashed) into refactor/veg-surface-water

5fbdb10

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

xiangtaoxu deleted the refactor/output-cleanups branch October 30, 2026 02:58

Navigation Menu

Repository navigation

feat(output): CF time with bounds; clearer names, labels and defaults - #58

feat(output): CF time with bounds; clearer names, labels and defaults#58
xiangtaoxu merged 21 commits into
EDmodel:betafrom
xiangtaoxu:refactor/output-cleanups

xiangtaoxu commented Oct 29, 2026 •

edited

Loading

Uh oh!

xiangtaoxu commented Oct 29, 2026

Uh oh!

Uh oh!

xiangtaoxu Oct 30, 2026

Uh oh!

xiangtaoxu Oct 30, 2026

Uh oh!

xiangtaoxu Oct 30, 2026

Uh oh!

xiangtaoxu Oct 30, 2026

Uh oh!

xiangtaoxu Oct 30, 2026

Uh oh!

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Conversation

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

Uh oh!

What and why

Effect on results (tick exactly one; CI checks)

User-facing changes (tick all that apply; each needs an "Upgrading" note in the changelog)

Verification

Credit and AI use

Not in this PR

Uh oh!

xiangtaoxu commented Oct 29, 2026

Uh oh!

Uh oh!

xiangtaoxu Oct 30, 2026

Choose a reason for hiding this comment

Uh oh!

xiangtaoxu Oct 30, 2026

Choose a reason for hiding this comment

Uh oh!

xiangtaoxu Oct 30, 2026

Choose a reason for hiding this comment

Uh oh!

xiangtaoxu Oct 30, 2026

Choose a reason for hiding this comment

Uh oh!

xiangtaoxu Oct 30, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

xiangtaoxu commented Oct 29, 2026 •

edited

Loading