Version: dev

NoahSoilHeat

Soil heat diffusion and the surface energy balance that closes it, in WRF's unified Noah land-surface model: SUBROUTINE TDFCND (soil thermal conductivity), the heat-capacity and flux-divergence parts of SUBROUTINE HRT, the yy/zz1 closure NOPAC forms before calling SHFLX, and the skin temperature and ground heat flux SHFLX returns (phys/module_sf_noahlsm.F, WRF 4.8, sf_surface_physics = 2), written as instantaneous tendencies. THE PDE: hcpct dT/dt = d/dz( df dT/dz ) on a column of NLEV soil layers, written as the flux-form operator D(K_soil D(T_soil) - F_soil) and lowered by the EarthSciDiscretizations column_nonuniform_1d rule varcoeff_face_flux_laplacian_lev, exactly as the slab scheme does. Orientation: `lev` is the soil layer index with layer 1 at the surface and `ze` the interface DEPTH below ground (positive downward, ze[1] = 0, increasing with the index), so ze is the negative of WRF's ZSOIL prefixed by zero. TWO boundary fluxes close the column and both are prescribed, which is why the diffusive part at the two domain faces is dropped: the surface interface carries F_soil[1] = df1 (yy - stc[1])/(0.5 ze[2] zz1), which is (the negative of) the Fortran's SSOIL and is the linearised surface energy balance rather than a temperature gradient; and the bottom interface carries the conduction to the prescribed deep-soil temperature, F_soil[NLEV+1] = df[NLEV] (stc[NLEV] - tbot)/(z_deep - zc[NLEV]). THE CONDUCTIVITY IS NOT FACE-AVERAGED. WRF evaluates TDFCND at the layer centre and then uses that one value for the flux through the layer's LOWER face, so the coefficient at interface kn is the conductivity of the layer ABOVE it, K_soil[kn] = df[kn-1]; a centred average would be the textbook choice and is NOT what the Fortran does, so the component reproduces the Fortran (FORTRAN_BUGS N90). The top layer's conductivity additionally carries the canopy reduction df1 = df[1] exp(sbeta shdfac), which the Fortran applies in NOPAC but not inside HRT's loop, so it appears at interface 2 and in the surface flux and nowhere else. THE SURFACE CLOSURE: yy and zz1 are NOPAC's linearisation of the skin-temperature energy balance about the air temperature; the skin temperature is DIAGNOSTIC, t1 = (yy + (zz1 - 1) stc[1])/zz1, not a prognostic state, and the reported ground heat flux ssoil = df1 (stc[1] - t1)/(0.5 ze[2]) is WRF's sign convention (positive warms the surface) and therefore the NEGATIVE of F_soil[1]. The sensible heat flux is evaluated at that diagnostic skin temperature and uses cp = 1004.5 and rd = 287.04, unlike SUBROUTINE PENMAN's rch (FORTRAN_BUGS N88). NOT TRANSCRIBED, and a consumer must not mount this component where they matter: the soil-water phase change (SNKSRC/TMPAVG/TBND, which adds a latent sink to the tendency of any layer holding ice or lying below 273.15 K), the snow-pack branch (SNOPAC), the urban override (df = 3.24, csoil = 3e6) and the opt_thcnd = 2 McCumber conductivity for silt/clay loam. The reference set is entirely above freezing and snow-free, so the freezing branch is UNTESTED here, not merely unused. Consumer-supplied column fields (ze, stc, smc, sh2o) are rewrite-target ops `input_<name>` lowered by the consumer, in the inline tests through per-test expression_template_imports libraries generated by tools/fill_tests.py.

Domain
land_surface
Subdomain
noah
Type
model
.esm schema
1.2.0
Source
components/land_surface/noah/soil_heat.esm
Tags
#land_surface #noah #model #wrf #land-surface #soil-heat-diffusion #surface-energy-balance #column #pde

Description

Soil heat diffusion and the surface energy balance that closes it, in WRF’s unified Noah land-surface model: SUBROUTINE TDFCND (soil thermal conductivity), the heat-capacity and flux-divergence parts of SUBROUTINE HRT, the yy/zz1 closure NOPAC forms before calling SHFLX, and the skin temperature and ground heat flux SHFLX returns (phys/module_sf_noahlsm.F, WRF 4.8, sf_surface_physics = 2), written as instantaneous tendencies. THE PDE: hcpct dT/dt = d/dz( df dT/dz ) on a column of NLEV soil layers, written as the flux-form operator D(K_soil D(T_soil) - F_soil) and lowered by the EarthSciDiscretizations column_nonuniform_1d rule varcoeff_face_flux_laplacian_lev, exactly as the slab scheme does. Orientation: lev is the soil layer index with layer 1 at the surface and ze the interface DEPTH below ground (positive downward, ze[1] = 0, increasing with the index), so ze is the negative of WRF’s ZSOIL prefixed by zero. TWO boundary fluxes close the column and both are prescribed, which is why the diffusive part at the two domain faces is dropped: the surface interface carries F_soil[1] = df1 (yy - stc[1])/(0.5 ze[2] zz1), which is (the negative of) the Fortran’s SSOIL and is the linearised surface energy balance rather than a temperature gradient; and the bottom interface carries the conduction to the prescribed deep-soil temperature, F_soil[NLEV+1] = df[NLEV] (stc[NLEV] - tbot)/(z_deep - zc[NLEV]). THE CONDUCTIVITY IS NOT FACE-AVERAGED. WRF evaluates TDFCND at the layer centre and then uses that one value for the flux through the layer’s LOWER face, so the coefficient at interface kn is the conductivity of the layer ABOVE it, K_soil[kn] = df[kn-1]; a centred average would be the textbook choice and is NOT what the Fortran does, so the component reproduces the Fortran (FORTRAN_BUGS N90). The top layer’s conductivity additionally carries the canopy reduction df1 = df[1] exp(sbeta shdfac), which the Fortran applies in NOPAC but not inside HRT’s loop, so it appears at interface 2 and in the surface flux and nowhere else. THE SURFACE CLOSURE: yy and zz1 are NOPAC’s linearisation of the skin-temperature energy balance about the air temperature; the skin temperature is DIAGNOSTIC, t1 = (yy + (zz1 - 1) stc[1])/zz1, not a prognostic state, and the reported ground heat flux ssoil = df1 (stc[1] - t1)/(0.5 ze[2]) is WRF’s sign convention (positive warms the surface) and therefore the NEGATIVE of F_soil[1]. The sensible heat flux is evaluated at that diagnostic skin temperature and uses cp = 1004.5 and rd = 287.04, unlike SUBROUTINE PENMAN’s rch (FORTRAN_BUGS N88). NOT TRANSCRIBED, and a consumer must not mount this component where they matter: the soil-water phase change (SNKSRC/TMPAVG/TBND, which adds a latent sink to the tendency of any layer holding ice or lying below 273.15 K), the snow-pack branch (SNOPAC), the urban override (df = 3.24, csoil = 3e6) and the opt_thcnd = 2 McCumber conductivity for silt/clay loam. The reference set is entirely above freezing and snow-free, so the freezing branch is UNTESTED here, not merely unused. Consumer-supplied column fields (ze, stc, smc, sh2o) are rewrite-target ops input_<name> lowered by the consumer, in the inline tests through per-test expression_template_imports libraries generated by tools/fill_tests.py.

Reference

  • Peters-Lidard, C. D., et al. (1998). The effect of soil thermal conductivity parameterization on surface energy fluxes and temperatures. J. Atmos. Sci., 55, 1209-1224.
  • Johansen, O. (1975). Thermal conductivity of soils. Ph.D. thesis, University of Trondheim.
  • Chen, F., and J. Dudhia (2001). Coupling an advanced land surface-hydrology model with the Penn State-NCAR MM5 modeling system. Part I. Mon. Wea. Rev., 129, 569-585.
  • wrf-model/WRF v4.8.0 phys/module_sf_noahlsm.F SUBROUTINE TDFCND lines 4124-4247, HRT lines 1588-1850, SHFLX lines 2600-2666 and the NOPAC closure lines 2168-2180 (instrumented in ctessum-claude/WRF branch earthsciml-instrumented-chem170, commit ac6d4cc; the internal df1/yy/zz1 are published by kernels/noah_esm_*.inc).

Parameters

NameUnitsDefaultDescription
stc_inK0.0Coupling target (assembly input path): a shaped parameter that receives the soil temperature from an assembly’s ODE state through a document-level variable_map (param_to_var), when that document’s mount-edge library lowers input_stc to it (esm-spec 9.7.10, 10.4; couplings/noah_column.esm). Unused by this component’s own inline tests, which lower input_stc to dumped constants.
smc_inm^3/m^30.0Coupling target for the total soil moisture, as stc_in (lowered from input_smc by an assembly’s mount-edge library).
sh2o_inm^3/m^30.0Coupling target for the unfrozen soil water, as stc_in (lowered from input_sh2o by an assembly’s mount-edge library).
t1_inK300.0Skin temperature ENTERING the step, WRF’s TSK on entry to the Noah driver (T1 = TSK(I,J)), read only by ssoil_pre. It is not this component’s t1: SFLX forms the ground heat flux it hands PENMAN from the incoming skin temperature, before SHFLX diagnoses the new one.
quartz10.25Quartz volume fraction of the soil type, SOILPARM.TBL QTZ, which sets the solid-phase conductivity.
smcmaxm^3/m^30.476Soil porosity (saturation moisture) of the soil type, SOILPARM.TBL MAXSMC.
shdfac10.5Green vegetation fraction, WRF’s SHDFAC, which reduces the surface-layer conductivity through exp(sbeta shdfac).
sfctmpK300.0Air temperature at the lowest model level, WRF’s SFCTMP.
sfcprsPa100000.0Air pressure at the lowest model level, WRF’s SFCPRS.
th2K300.0Potential temperature at the lowest model level referenced to the surface pressure, WRF’s TH2.
q2kg/kg0.01Specific humidity at the lowest model level, WRF’s Q2, used for the virtual temperature in the sensible-heat flux.
chm/s0.01Surface exchange coefficient for heat and moisture, WRF’s CH.
fdownW/m^2600.0Total absorbed downward radiation at the surface, WRF’s FDOWN = solnet + lwdn.
emissi10.985Surface emissivity, WRF’s EMISSI.
rchW/(m^2*K)10.0Bulk heat-transfer coefficient in energy units from the Penman component, WRF’s RCH. Carries PENMAN’s cp = 1004.6.
rr12.0Resistance ratio from the Penman component, WRF’s RR.
epscaK10.0Penman potential evaporation expressed as a temperature, WRF’s EPSCA.
beta10.6Ratio of actual to potential evaporation from the evapotranspiration component, WRF’s BETA. The product beta epsca is the latent-heat term of the surface energy balance.
tbotK300.0Prescribed deep-soil temperature at depth z_deep, WRF’s TBOT.
z_deepm8.0Depth at which tbot is prescribed, |ZBOT_DATA| = 8 m. Positive downward, matching ze.

Observed

NameUnitsDefaultDescription
zemSoil layer interface DEPTHS below the surface, positive downward, ze[1] = 0; the negative of WRF’s ZSOIL prefixed by zero.
zcmSoil layer centre depths, from the grid’s column_cell_center template. Reproduces the midpoints WRF forms implicitly in HRT’s 0.5 (zsoil(k-1) - zsoil(k+1)) denominators.
dzmSoil layer thicknesses, from the grid’s column_cell_thickness template. Equals WRF’s SLDPTH.
stcKSoil temperature of each layer, WRF’s STC. Read as a plain algebraic input: the consumer owns the state and integrates dstc_dt.
smcm^3/m^3TOTAL volumetric soil moisture of each layer, WRF’s SMC.
sh2om^3/m^3UNFROZEN volumetric soil water of each layer, WRF’s SH2O. Equals smc wherever the soil is above freezing.
df_rawW/(m*K)Johansen/Peters-Lidard soil thermal conductivity of each layer from TDFCND (opt_thcnd = 1): the dry value thkdry raised towards the saturated value thksat by the Kersten number ake.
df_layerW/(m*K)df_raw with the canopy reduction applied to the SURFACE LAYER ONLY, df[1] exp(sbeta shdfac). WRF applies it in NOPAC, after HRT’s interior loop has already used the unreduced values, so only the top layer and the surface flux see it.
satratio1Degree of saturation, smc/smcmax.
thksW/(m*K)Solid-phase conductivity, 7.7^quartz 2.0^(1-quartz) (quartz against ‘other minerals’).
xum^3/m^3Unfrozen water volume at saturation, (sh2o/smc) smcmax, the exponent that splits the saturated conductivity between water and ice.
thksatW/(m*K)Saturated conductivity, thks^(1-smcmax) 2.2^(smcmax-xu) 0.57^xu (ice 2.2, water 0.57 W/(m K)).
gammdkg/m^3Dry soil bulk density, (1 - smcmax) 2700.
thkdryW/(m*K)Dry soil conductivity, (0.135 gammd + 64.7)/(2700 - 0.947 gammd).
ake1Kersten number, the ice- and water-weighted blend ((smc - sh2o) akei + sh2o akel)/smc, where the frozen part uses akei = satratio and the liquid part the logarithmic akel = log10(satratio) + 1, floored at zero below satratio = 0.1.
hcpctJ/(m^3*K)Volumetric heat capacity of each layer, sh2o ch2o_vol + (1 - smcmax) csoil + (smcmax - smc) cair_vol + (smc - sh2o) cice_vol.
t24K^4Fourth power of the air temperature, the linearisation point of the surface emission.
yynumW/m^2Net radiation at the linearisation temperature, fdown - emissi sigma t24.
yyKEffective forcing temperature of the linearised surface energy balance, sfctmp + (yynum/rch + th2 - sfctmp - beta epsca)/rr. The skin temperature relaxes towards it.
zz11Coupling strength of the skin temperature to the top soil layer, df1/(0.5 ze[2] rch rr) + 1. zz1 = 1 would mean no conduction and t1 = yy.
t1KDIAGNOSTIC skin temperature, (yy + (zz1 - 1) stc[1])/zz1: a weighted mean of the atmospheric forcing temperature and the top soil layer, not a prognostic state.
K_soilW/(m*K)Thermal conductivity at each layer interface: the conductivity of the layer ABOVE the interface, K_soil[kn] = df_layer[kn-1], which is what the Fortran uses and not a centred average (FORTRAN_BUGS N90). The entries at the two domain faces are never read by the lowering, which takes the prescribed F_soil there instead.
F_soilW/m^2Prescribed heat flux at each interface, positive DOWNWARD (the +lev direction): the linearised surface energy balance at kn = 1, the conduction towards tbot at kn = NLEV+1, and zero in the interior, where the flux is diffusive.
ssoilW/m^2Ground heat flux as WRF reports it, df1 (stc[1] - t1)/(0.5 ze[2]). WRF flips the sign at the end of SFLX so that a positive value WARMS the surface, which makes it the negative of F_soil[1].
ssoil_preW/m^2The ground heat flux SFLX forms BEFORE calling PENMAN, df1 (t1_in - stc[1])/(0.5 ze[2]), from the INCOMING skin temperature t1_in and with the opposite sign to ssoil (positive into the soil): it is the SSOIL argument of PENMAN (the Penman component’s ssoil). df1 is the same canopy-reduced top-layer conductivity df_layer[1], since in the instantaneous form the soil moisture it is evaluated at is the one state. WRF’s lag: the incoming skin temperature is the previous step’s diagnosed t1; in the dt -> 0 limit t1_in = t1, which closes an algebraic loop (t1 depends on epsca through yy, epsca on ssoil_pre) that WRF breaks by lagging and an assembly must either lag the same way or solve.
t2vKVirtual temperature at the lowest model level, sfctmp (1 + 0.61 q2).
sheatW/m^2Sensible heat flux, -(ch cp sfcprs/(rd t2v)) (th2 - t1), evaluated at the DIAGNOSTIC skin temperature and with cp = 1004.5, not PENMAN’s 1004.6.
dstc_dtK/sInstantaneous soil-temperature tendency of each layer, the flux divergence divided by the layer heat capacity. This is the dt -> 0 limit of WRF’s (STC_out - STC_in)/dt, whose finite-dt value differs because HSTEP solves the same operator implicitly.

Observed expressions

$$ ze = \mathrm{input_ze}\left(\right) $$

Soil layer interface DEPTHS below the surface, positive downward, ze[1] = 0; the negative of WRF’s ZSOIL prefixed by zero.

$$ zc = \mathrm{column_cell_center}\left(ze=ze\right) $$

Soil layer centre depths, from the grid’s column_cell_center template. Reproduces the midpoints WRF forms implicitly in HRT’s 0.5 (zsoil(k-1) - zsoil(k+1)) denominators.

$$ dz = \mathrm{column_cell_thickness}\left(ze=ze\right) $$

Soil layer thicknesses, from the grid’s column_cell_thickness template. Equals WRF’s SLDPTH.

$$ stc = \mathrm{input_stc}\left(\right) $$

Soil temperature of each layer, WRF’s STC. Read as a plain algebraic input: the consumer owns the state and integrates dstc_dt.

$$ smc = \mathrm{input_smc}\left(\right) $$

TOTAL volumetric soil moisture of each layer, WRF’s SMC.

$$ sh2o = \mathrm{input_sh2o}\left(\right) $$

UNFROZEN volumetric soil water of each layer, WRF’s SH2O. Equals smc wherever the soil is above freezing.

$$ df_{raw} = ake \cdot \left(thksat - thkdry\right) + thkdry $$

Johansen/Peters-Lidard soil thermal conductivity of each layer from TDFCND (opt_thcnd = 1): the dry value thkdry raised towards the saturated value thksat by the Kersten number ake.

$$ df_{layer} = \mathrm{faq}\left(df_{raw}\right) $$

df_raw with the canopy reduction applied to the SURFACE LAYER ONLY, df[1] exp(sbeta shdfac). WRF applies it in NOPAC, after HRT’s interior loop has already used the unreduced values, so only the top layer and the surface flux see it.

$$ satratio = \frac{smc}{smcmax} $$

Degree of saturation, smc/smcmax.

$$ thks = 7.7^{quartz} \cdot 2.0^{1.0 - quartz} $$

Solid-phase conductivity, 7.7^quartz 2.0^(1-quartz) (quartz against ‘other minerals’).

$$ xu = \frac{sh2o}{smc} \cdot smcmax $$

Unfrozen water volume at saturation, (sh2o/smc) smcmax, the exponent that splits the saturated conductivity between water and ice.

$$ thksat = thks^{1.0 - smcmax} \cdot 2.2^{smcmax - xu} \cdot 0.57^{xu} $$

Saturated conductivity, thks^(1-smcmax) 2.2^(smcmax-xu) 0.57^xu (ice 2.2, water 0.57 W/(m K)).

$$ gammd = \left(1.0 - smcmax\right) \cdot 2700.0 $$

Dry soil bulk density, (1 - smcmax) 2700.

$$ thkdry = \frac{0.135 \cdot gammd + 64.7}{2700.0 - 0.947 \cdot gammd} $$

Dry soil conductivity, (0.135 gammd + 64.7)/(2700 - 0.947 gammd).

$$ ake = \frac{\left(smc - sh2o\right) \cdot satratio + sh2o \cdot \mathrm{ifelse}\left(\mathrm{>}\left(satratio, 0.1\right), \frac{\ln\left(satratio\right)}{\ln\left(10.0\right)} + 1.0, 0.0\right)}{smc} $$

Kersten number, the ice- and water-weighted blend ((smc - sh2o) akei + sh2o akel)/smc, where the frozen part uses akei = satratio and the liquid part the logarithmic akel = log10(satratio) + 1, floored at zero below satratio = 0.1.

$$ hcpct = sh2o \cdot par.ch2o_{vol} + \left(1.0 - smcmax\right) \cdot par.csoil + \left(smcmax - smc\right) \cdot par.cair_{vol} + \left(smc - sh2o\right) \cdot par.cice_{vol} $$

Volumetric heat capacity of each layer, sh2o ch2o_vol + (1 - smcmax) csoil + (smcmax - smc) cair_vol + (smc - sh2o) cice_vol.

$$ t24 = sfctmp^{4.0} $$

Fourth power of the air temperature, the linearisation point of the surface emission.

$$ yynum = fdown - emissi \cdot par.sigma \cdot t24 $$

Net radiation at the linearisation temperature, fdown - emissi sigma t24.

$$ yy = sfctmp + \frac{\frac{yynum}{rch} + th2 - sfctmp - beta \cdot epsca}{rr} $$

Effective forcing temperature of the linearised surface energy balance, sfctmp + (yynum/rch + th2 - sfctmp - beta epsca)/rr. The skin temperature relaxes towards it.

$$ zz1 = \frac{\mathrm{index}\left(df_{layer}, 1\right)}{0.5 \cdot \mathrm{index}\left(ze, 2\right) \cdot rch \cdot rr} + 1.0 $$

Coupling strength of the skin temperature to the top soil layer, df1/(0.5 ze[2] rch rr) + 1. zz1 = 1 would mean no conduction and t1 = yy.

$$ t1 = \frac{yy + \left(zz1 - 1.0\right) \cdot \mathrm{index}\left(stc, 1\right)}{zz1} $$

DIAGNOSTIC skin temperature, (yy + (zz1 - 1) stc[1])/zz1: a weighted mean of the atmospheric forcing temperature and the top soil layer, not a prognostic state.

$$ K_{soil} = \mathrm{faq}\left(df_{layer}\right) $$

Thermal conductivity at each layer interface: the conductivity of the layer ABOVE the interface, K_soil[kn] = df_layer[kn-1], which is what the Fortran uses and not a centred average (FORTRAN_BUGS N90). The entries at the two domain faces are never read by the lowering, which takes the prescribed F_soil there instead.

$$ F_{soil} = \mathrm{faq}\left(df_{layer}, stc, zc, ze\right) $$

Prescribed heat flux at each interface, positive DOWNWARD (the +lev direction): the linearised surface energy balance at kn = 1, the conduction towards tbot at kn = NLEV+1, and zero in the interior, where the flux is diffusive.

$$ ssoil = \frac{\mathrm{index}\left(df_{layer}, 1\right) \cdot \left(\mathrm{index}\left(stc, 1\right) - t1\right)}{0.5 \cdot \mathrm{index}\left(ze, 2\right)} $$

Ground heat flux as WRF reports it, df1 (stc[1] - t1)/(0.5 ze[2]). WRF flips the sign at the end of SFLX so that a positive value WARMS the surface, which makes it the negative of F_soil[1].

$$ ssoil_{pre} = \frac{\mathrm{index}\left(df_{layer}, 1\right) \cdot \left(t1_{in} - \mathrm{index}\left(stc, 1\right)\right)}{0.5 \cdot \mathrm{index}\left(ze, 2\right)} $$

The ground heat flux SFLX forms BEFORE calling PENMAN, df1 (t1_in - stc[1])/(0.5 ze[2]), from the INCOMING skin temperature t1_in and with the opposite sign to ssoil (positive into the soil): it is the SSOIL argument of PENMAN (the Penman component’s ssoil). df1 is the same canopy-reduced top-layer conductivity df_layer[1], since in the instantaneous form the soil moisture it is evaluated at is the one state. WRF’s lag: the incoming skin temperature is the previous step’s diagnosed t1; in the dt -> 0 limit t1_in = t1, which closes an algebraic loop (t1 depends on epsca through yy, epsca on ssoil_pre) that WRF breaks by lagging and an assembly must either lag the same way or solve.

$$ t2v = sfctmp \cdot \left(1.0 + 0.61 \cdot q2\right) $$

Virtual temperature at the lowest model level, sfctmp (1 + 0.61 q2).

$$ sheat = 0.0 - \frac{ch \cdot wrf.cp \cdot sfcprs}{par.rd \cdot t2v} \cdot \left(th2 - t1\right) $$

Sensible heat flux, -(ch cp sfcprs/(rd t2v)) (th2 - t1), evaluated at the DIAGNOSTIC skin temperature and with cp = 1004.5, not PENMAN’s 1004.6.

$$ dstc_{dt} = \frac{\mathrm{D}\left(K_{soil} \cdot \mathrm{D}\left(stc\right) - F_{soil}\right)}{hcpct} $$

Instantaneous soil-temperature tendency of each layer, the flux divergence divided by the layer heat capacity. This is the dt -> 0 limit of WRF’s (STC_out - STC_in)/dt, whose finite-dt value differs because HSTEP solves the same operator implicitly.

Raw .esm

Raw .esm JSON (this component)
{
  "expression_template_imports": [
    {
      "ref": "${ESD_ROOT}/grids/column_nonuniform_1d/grid.esm",
      "bindings": {
        "NLEV": "NLEV"
      }
    },
    {
      "ref": "${ESD_ROOT}/grids/column_nonuniform_1d/rules/varcoeff_face_flux_laplacian_lev.esm",
      "bindings": {
        "NLEV": "NLEV"
      }
    }
  ],
  "subsystems": {
    "wrf": {
      "ref": "../../../lib/wrf_constants.esm"
    },
    "par": {
      "ref": "./noah_parameters.esm"
    }
  },
  "variables": {
    "ze": {
      "type": "unknown",
      "units": "m",
      "shape": [
        "lev_nodes"
      ],
      "description": "Soil layer interface DEPTHS below the surface, positive downward, ze[1] = 0; the negative of WRF's ZSOIL prefixed by zero."
    },
    "zc": {
      "type": "unknown",
      "units": "m",
      "shape": [
        "lev"
      ],
      "description": "Soil layer centre depths, from the grid's column_cell_center template. Reproduces the midpoints WRF forms implicitly in HRT's 0.5 (zsoil(k-1) - zsoil(k+1)) denominators."
    },
    "dz": {
      "type": "unknown",
      "units": "m",
      "shape": [
        "lev"
      ],
      "description": "Soil layer thicknesses, from the grid's column_cell_thickness template. Equals WRF's SLDPTH."
    },
    "stc": {
      "type": "unknown",
      "units": "K",
      "shape": [
        "lev"
      ],
      "description": "Soil temperature of each layer, WRF's STC. Read as a plain algebraic input: the consumer owns the state and integrates dstc_dt."
    },
    "smc": {
      "type": "unknown",
      "units": "m^3/m^3",
      "shape": [
        "lev"
      ],
      "description": "TOTAL volumetric soil moisture of each layer, WRF's SMC."
    },
    "sh2o": {
      "type": "unknown",
      "units": "m^3/m^3",
      "shape": [
        "lev"
      ],
      "description": "UNFROZEN volumetric soil water of each layer, WRF's SH2O. Equals smc wherever the soil is above freezing."
    },
    "stc_in": {
      "type": "parameter",
      "units": "K",
      "shape": [
        "lev"
      ],
      "default": 0.0,
      "description": "Coupling target (assembly input path): a shaped parameter that receives the soil temperature from an assembly's ODE state through a document-level variable_map (param_to_var), when that document's mount-edge library lowers `input_stc` to it (esm-spec 9.7.10, 10.4; couplings/noah_column.esm). Unused by this component's own inline tests, which lower `input_stc` to dumped constants."
    },
    "smc_in": {
      "type": "parameter",
      "units": "m^3/m^3",
      "shape": [
        "lev"
      ],
      "default": 0.0,
      "description": "Coupling target for the total soil moisture, as stc_in (lowered from `input_smc` by an assembly's mount-edge library)."
    },
    "sh2o_in": {
      "type": "parameter",
      "units": "m^3/m^3",
      "shape": [
        "lev"
      ],
      "default": 0.0,
      "description": "Coupling target for the unfrozen soil water, as stc_in (lowered from `input_sh2o` by an assembly's mount-edge library)."
    },
    "t1_in": {
      "type": "parameter",
      "units": "K",
      "default": 300.0,
      "description": "Skin temperature ENTERING the step, WRF's TSK on entry to the Noah driver (T1 = TSK(I,J)), read only by ssoil_pre. It is not this component's t1: SFLX forms the ground heat flux it hands PENMAN from the incoming skin temperature, before SHFLX diagnoses the new one."
    },
    "quartz": {
      "type": "parameter",
      "units": "1",
      "default": 0.25,
      "description": "Quartz volume fraction of the soil type, SOILPARM.TBL QTZ, which sets the solid-phase conductivity."
    },
    "smcmax": {
      "type": "parameter",
      "units": "m^3/m^3",
      "default": 0.476,
      "description": "Soil porosity (saturation moisture) of the soil type, SOILPARM.TBL MAXSMC."
    },
    "shdfac": {
      "type": "parameter",
      "units": "1",
      "default": 0.5,
      "description": "Green vegetation fraction, WRF's SHDFAC, which reduces the surface-layer conductivity through exp(sbeta shdfac)."
    },
    "sfctmp": {
      "type": "parameter",
      "units": "K",
      "default": 300.0,
      "description": "Air temperature at the lowest model level, WRF's SFCTMP."
    },
    "sfcprs": {
      "type": "parameter",
      "units": "Pa",
      "default": 100000.0,
      "description": "Air pressure at the lowest model level, WRF's SFCPRS."
    },
    "th2": {
      "type": "parameter",
      "units": "K",
      "default": 300.0,
      "description": "Potential temperature at the lowest model level referenced to the surface pressure, WRF's TH2."
    },
    "q2": {
      "type": "parameter",
      "units": "kg/kg",
      "default": 0.01,
      "description": "Specific humidity at the lowest model level, WRF's Q2, used for the virtual temperature in the sensible-heat flux."
    },
    "ch": {
      "type": "parameter",
      "units": "m/s",
      "default": 0.01,
      "description": "Surface exchange coefficient for heat and moisture, WRF's CH."
    },
    "fdown": {
      "type": "parameter",
      "units": "W/m^2",
      "default": 600.0,
      "description": "Total absorbed downward radiation at the surface, WRF's FDOWN = solnet + lwdn."
    },
    "emissi": {
      "type": "parameter",
      "units": "1",
      "default": 0.985,
      "description": "Surface emissivity, WRF's EMISSI."
    },
    "rch": {
      "type": "parameter",
      "units": "W/(m^2*K)",
      "default": 10.0,
      "description": "Bulk heat-transfer coefficient in energy units from the Penman component, WRF's RCH. Carries PENMAN's cp = 1004.6."
    },
    "rr": {
      "type": "parameter",
      "units": "1",
      "default": 2.0,
      "description": "Resistance ratio from the Penman component, WRF's RR."
    },
    "epsca": {
      "type": "parameter",
      "units": "K",
      "default": 10.0,
      "description": "Penman potential evaporation expressed as a temperature, WRF's EPSCA."
    },
    "beta": {
      "type": "parameter",
      "units": "1",
      "default": 0.6,
      "description": "Ratio of actual to potential evaporation from the evapotranspiration component, WRF's BETA. The product beta epsca is the latent-heat term of the surface energy balance."
    },
    "tbot": {
      "type": "parameter",
      "units": "K",
      "default": 300.0,
      "description": "Prescribed deep-soil temperature at depth z_deep, WRF's TBOT."
    },
    "z_deep": {
      "type": "parameter",
      "units": "m",
      "default": 8.0,
      "description": "Depth at which tbot is prescribed, |ZBOT_DATA| = 8 m. Positive downward, matching ze."
    },
    "df_raw": {
      "type": "unknown",
      "units": "W/(m*K)",
      "shape": [
        "lev"
      ],
      "description": "Johansen/Peters-Lidard soil thermal conductivity of each layer from TDFCND (opt_thcnd = 1): the dry value thkdry raised towards the saturated value thksat by the Kersten number ake."
    },
    "df_layer": {
      "type": "unknown",
      "units": "W/(m*K)",
      "shape": [
        "lev"
      ],
      "description": "df_raw with the canopy reduction applied to the SURFACE LAYER ONLY, df[1] exp(sbeta shdfac). WRF applies it in NOPAC, after HRT's interior loop has already used the unreduced values, so only the top layer and the surface flux see it."
    },
    "satratio": {
      "type": "unknown",
      "units": "1",
      "shape": [
        "lev"
      ],
      "description": "Degree of saturation, smc/smcmax."
    },
    "thks": {
      "type": "unknown",
      "units": "W/(m*K)",
      "description": "Solid-phase conductivity, 7.7^quartz 2.0^(1-quartz) (quartz against 'other minerals')."
    },
    "xu": {
      "type": "unknown",
      "units": "m^3/m^3",
      "shape": [
        "lev"
      ],
      "description": "Unfrozen water volume at saturation, (sh2o/smc) smcmax, the exponent that splits the saturated conductivity between water and ice."
    },
    "thksat": {
      "type": "unknown",
      "units": "W/(m*K)",
      "shape": [
        "lev"
      ],
      "description": "Saturated conductivity, thks^(1-smcmax) 2.2^(smcmax-xu) 0.57^xu (ice 2.2, water 0.57 W/(m K))."
    },
    "gammd": {
      "type": "unknown",
      "units": "kg/m^3",
      "description": "Dry soil bulk density, (1 - smcmax) 2700."
    },
    "thkdry": {
      "type": "unknown",
      "units": "W/(m*K)",
      "description": "Dry soil conductivity, (0.135 gammd + 64.7)/(2700 - 0.947 gammd)."
    },
    "ake": {
      "type": "unknown",
      "units": "1",
      "shape": [
        "lev"
      ],
      "description": "Kersten number, the ice- and water-weighted blend ((smc - sh2o) akei + sh2o akel)/smc, where the frozen part uses akei = satratio and the liquid part the logarithmic akel = log10(satratio) + 1, floored at zero below satratio = 0.1."
    },
    "hcpct": {
      "type": "unknown",
      "units": "J/(m^3*K)",
      "shape": [
        "lev"
      ],
      "description": "Volumetric heat capacity of each layer, sh2o ch2o_vol + (1 - smcmax) csoil + (smcmax - smc) cair_vol + (smc - sh2o) cice_vol."
    },
    "t24": {
      "type": "unknown",
      "units": "K^4",
      "description": "Fourth power of the air temperature, the linearisation point of the surface emission."
    },
    "yynum": {
      "type": "unknown",
      "units": "W/m^2",
      "description": "Net radiation at the linearisation temperature, fdown - emissi sigma t24."
    },
    "yy": {
      "type": "unknown",
      "units": "K",
      "description": "Effective forcing temperature of the linearised surface energy balance, sfctmp + (yynum/rch + th2 - sfctmp - beta epsca)/rr. The skin temperature relaxes towards it."
    },
    "zz1": {
      "type": "unknown",
      "units": "1",
      "description": "Coupling strength of the skin temperature to the top soil layer, df1/(0.5 ze[2] rch rr) + 1. zz1 = 1 would mean no conduction and t1 = yy."
    },
    "t1": {
      "type": "unknown",
      "units": "K",
      "description": "DIAGNOSTIC skin temperature, (yy + (zz1 - 1) stc[1])/zz1: a weighted mean of the atmospheric forcing temperature and the top soil layer, not a prognostic state."
    },
    "K_soil": {
      "type": "unknown",
      "units": "W/(m*K)",
      "shape": [
        "lev_nodes"
      ],
      "description": "Thermal conductivity at each layer interface: the conductivity of the layer ABOVE the interface, K_soil[kn] = df_layer[kn-1], which is what the Fortran uses and not a centred average (FORTRAN_BUGS N90). The entries at the two domain faces are never read by the lowering, which takes the prescribed F_soil there instead."
    },
    "F_soil": {
      "type": "unknown",
      "units": "W/m^2",
      "shape": [
        "lev_nodes"
      ],
      "description": "Prescribed heat flux at each interface, positive DOWNWARD (the +lev direction): the linearised surface energy balance at kn = 1, the conduction towards tbot at kn = NLEV+1, and zero in the interior, where the flux is diffusive."
    },
    "ssoil": {
      "type": "unknown",
      "units": "W/m^2",
      "description": "Ground heat flux as WRF reports it, df1 (stc[1] - t1)/(0.5 ze[2]). WRF flips the sign at the end of SFLX so that a positive value WARMS the surface, which makes it the negative of F_soil[1]."
    },
    "ssoil_pre": {
      "type": "unknown",
      "units": "W/m^2",
      "description": "The ground heat flux SFLX forms BEFORE calling PENMAN, df1 (t1_in - stc[1])/(0.5 ze[2]), from the INCOMING skin temperature t1_in and with the opposite sign to ssoil (positive into the soil): it is the SSOIL argument of PENMAN (the Penman component's ssoil). df1 is the same canopy-reduced top-layer conductivity df_layer[1], since in the instantaneous form the soil moisture it is evaluated at is the one state. WRF's lag: the incoming skin temperature is the previous step's diagnosed t1; in the dt -> 0 limit t1_in = t1, which closes an algebraic loop (t1 depends on epsca through yy, epsca on ssoil_pre) that WRF breaks by lagging and an assembly must either lag the same way or solve."
    },
    "t2v": {
      "type": "unknown",
      "units": "K",
      "description": "Virtual temperature at the lowest model level, sfctmp (1 + 0.61 q2)."
    },
    "sheat": {
      "type": "unknown",
      "units": "W/m^2",
      "description": "Sensible heat flux, -(ch cp sfcprs/(rd t2v)) (th2 - t1), evaluated at the DIAGNOSTIC skin temperature and with cp = 1004.5, not PENMAN's 1004.6."
    },
    "dstc_dt": {
      "type": "unknown",
      "units": "K/s",
      "shape": [
        "lev"
      ],
      "description": "Instantaneous soil-temperature tendency of each layer, the flux divergence divided by the layer heat capacity. This is the dt -> 0 limit of WRF's (STC_out - STC_in)/dt, whose finite-dt value differs because HSTEP solves the same operator implicitly."
    }
  },
  "equations": [
    {
      "lhs": "ze",
      "rhs": {
        "op": "input_ze",
        "args": []
      }
    },
    {
      "lhs": "zc",
      "rhs": {
        "op": "apply_expression_template",
        "args": [],
        "name": "column_cell_center",
        "bindings": {
          "ze": "ze"
        }
      }
    },
    {
      "lhs": "dz",
      "rhs": {
        "op": "apply_expression_template",
        "args": [],
        "name": "column_cell_thickness",
        "bindings": {
          "ze": "ze"
        }
      }
    },
    {
      "lhs": "stc",
      "rhs": {
        "op": "input_stc",
        "args": []
      }
    },
    {
      "lhs": "smc",
      "rhs": {
        "op": "input_smc",
        "args": []
      }
    },
    {
      "lhs": "sh2o",
      "rhs": {
        "op": "input_sh2o",
        "args": []
      }
    },
    {
      "lhs": "satratio",
      "rhs": {
        "op": "/",
        "args": [
          "smc",
          "smcmax"
        ]
      }
    },
    {
      "lhs": "thks",
      "rhs": {
        "op": "*",
        "args": [
          {
            "op": "^",
            "args": [
              7.7,
              "quartz"
            ]
          },
          {
            "op": "^",
            "args": [
              2.0,
              {
                "op": "-",
                "args": [
                  1.0,
                  "quartz"
                ]
              }
            ]
          }
        ]
      }
    },
    {
      "lhs": "xu",
      "rhs": {
        "op": "*",
        "args": [
          {
            "op": "/",
            "args": [
              "sh2o",
              "smc"
            ]
          },
          "smcmax"
        ]
      }
    },
    {
      "lhs": "thksat",
      "rhs": {
        "op": "*",
        "args": [
          {
            "op": "*",
            "args": [
              {
                "op": "^",
                "args": [
                  "thks",
                  {
                    "op": "-",
                    "args": [
                      1.0,
                      "smcmax"
                    ]
                  }
                ]
              },
              {
                "op": "^",
                "args": [
                  2.2,
                  {
                    "op": "-",
                    "args": [
                      "smcmax",
                      "xu"
                    ]
                  }
                ]
              }
            ]
          },
          {
            "op": "^",
            "args": [
              0.57,
              "xu"
            ]
          }
        ]
      }
    },
    {
      "lhs": "gammd",
      "rhs": {
        "op": "*",
        "args": [
          {
            "op": "-",
            "args": [
              1.0,
              "smcmax"
            ]
          },
          2700.0
        ]
      }
    },
    {
      "lhs": "thkdry",
      "rhs": {
        "op": "/",
        "args": [
          {
            "op": "+",
            "args": [
              {
                "op": "*",
                "args": [
                  0.135,
                  "gammd"
                ]
              },
              64.7
            ]
          },
          {
            "op": "-",
            "args": [
              2700.0,
              {
                "op": "*",
                "args": [
                  0.947,
                  "gammd"
                ]
              }
            ]
          }
        ]
      }
    },
    {
      "lhs": "ake",
      "rhs": {
        "op": "/",
        "args": [
          {
            "op": "+",
            "args": [
              {
                "op": "*",
                "args": [
                  {
                    "op": "-",
                    "args": [
                      "smc",
                      "sh2o"
                    ]
                  },
                  "satratio"
                ]
              },
              {
                "op": "*",
                "args": [
                  "sh2o",
                  {
                    "op": "ifelse",
                    "args": [
                      {
                        "op": ">",
                        "args": [
                          "satratio",
                          0.1
                        ]
                      },
                      {
                        "op": "+",
                        "args": [
                          {
                            "op": "/",
                            "args": [
                              {
                                "op": "log",
                                "args": [
                                  "satratio"
                                ]
                              },
                              {
                                "op": "log",
                                "args": [
                                  10.0
                                ]
                              }
                            ]
                          },
                          1.0
                        ]
                      },
                      0.0
                    ]
                  }
                ]
              }
            ]
          },
          "smc"
        ]
      }
    },
    {
      "lhs": "df_raw",
      "rhs": {
        "op": "+",
        "args": [
          {
            "op": "*",
            "args": [
              "ake",
              {
                "op": "-",
                "args": [
                  "thksat",
                  "thkdry"
                ]
              }
            ]
          },
          "thkdry"
        ]
      }
    },
    {
      "lhs": "df_layer",
      "rhs": {
        "op": "faq",
        "output_idx": [
          "k"
        ],
        "args": [
          "df_raw"
        ],
        "ranges": {
          "k": {
            "from": "lev"
          }
        },
        "expr": {
          "op": "ifelse",
          "args": [
            {
              "op": "==",
              "args": [
                "k",
                1
              ]
            },
            {
              "op": "*",
              "args": [
                {
                  "op": "index",
                  "args": [
                    "df_raw",
                    1
                  ]
                },
                {
                  "op": "exp",
                  "args": [
                    {
                      "op": "*",
                      "args": [
                        "par.sbeta",
                        "shdfac"
                      ]
                    }
                  ]
                }
              ]
            },
            {
              "op": "index",
              "args": [
                "df_raw",
                "k"
              ]
            }
          ]
        }
      }
    },
    {
      "lhs": "hcpct",
      "rhs": {
        "op": "+",
        "args": [
          {
            "op": "+",
            "args": [
              {
                "op": "*",
                "args": [
                  "sh2o",
                  "par.ch2o_vol"
                ]
              },
              {
                "op": "*",
                "args": [
                  {
                    "op": "-",
                    "args": [
                      1.0,
                      "smcmax"
                    ]
                  },
                  "par.csoil"
                ]
              }
            ]
          },
          {
            "op": "+",
            "args": [
              {
                "op": "*",
                "args": [
                  {
                    "op": "-",
                    "args": [
                      "smcmax",
                      "smc"
                    ]
                  },
                  "par.cair_vol"
                ]
              },
              {
                "op": "*",
                "args": [
                  {
                    "op": "-",
                    "args": [
                      "smc",
                      "sh2o"
                    ]
                  },
                  "par.cice_vol"
                ]
              }
            ]
          }
        ]
      }
    },
    {
      "lhs": "t24",
      "rhs": {
        "op": "^",
        "args": [
          "sfctmp",
          4.0
        ]
      }
    },
    {
      "lhs": "yynum",
      "rhs": {
        "op": "-",
        "args": [
          "fdown",
          {
            "op": "*",
            "args": [
              {
                "op": "*",
                "args": [
                  "emissi",
                  "par.sigma"
                ]
              },
              "t24"
            ]
          }
        ]
      }
    },
    {
      "lhs": "yy",
      "rhs": {
        "op": "+",
        "args": [
          "sfctmp",
          {
            "op": "/",
            "args": [
              {
                "op": "-",
                "args": [
                  {
                    "op": "-",
                    "args": [
                      {
                        "op": "+",
                        "args": [
                          {
                            "op": "/",
                            "args": [
                              "yynum",
                              "rch"
                            ]
                          },
                          "th2"
                        ]
                      },
                      "sfctmp"
                    ]
                  },
                  {
                    "op": "*",
                    "args": [
                      "beta",
                      "epsca"
                    ]
                  }
                ]
              },
              "rr"
            ]
          }
        ]
      }
    },
    {
      "lhs": "zz1",
      "rhs": {
        "op": "+",
        "args": [
          {
            "op": "/",
            "args": [
              {
                "op": "index",
                "args": [
                  "df_layer",
                  1
                ]
              },
              {
                "op": "*",
                "args": [
                  {
                    "op": "*",
                    "args": [
                      0.5,
                      {
                        "op": "index",
                        "args": [
                          "ze",
                          2
                        ]
                      }
                    ]
                  },
                  {
                    "op": "*",
                    "args": [
                      "rch",
                      "rr"
                    ]
                  }
                ]
              }
            ]
          },
          1.0
        ]
      }
    },
    {
      "lhs": "t1",
      "rhs": {
        "op": "/",
        "args": [
          {
            "op": "+",
            "args": [
              "yy",
              {
                "op": "*",
                "args": [
                  {
                    "op": "-",
                    "args": [
                      "zz1",
                      1.0
                    ]
                  },
                  {
                    "op": "index",
                    "args": [
                      "stc",
                      1
                    ]
                  }
                ]
              }
            ]
          },
          "zz1"
        ]
      }
    },
    {
      "lhs": "K_soil",
      "rhs": {
        "op": "faq",
        "output_idx": [
          "kn"
        ],
        "args": [
          "df_layer"
        ],
        "ranges": {
          "kn": {
            "from": "lev_nodes"
          }
        },
        "expr": {
          "op": "ifelse",
          "args": [
            {
              "op": "==",
              "args": [
                "kn",
                1
              ]
            },
            {
              "op": "index",
              "args": [
                "df_layer",
                1
              ]
            },
            {
              "op": "index",
              "args": [
                "df_layer",
                {
                  "op": "-",
                  "args": [
                    "kn",
                    1
                  ]
                }
              ]
            }
          ]
        }
      }
    },
    {
      "lhs": "F_soil",
      "rhs": {
        "op": "faq",
        "output_idx": [
          "kn"
        ],
        "args": [
          "df_layer",
          "stc",
          "zc",
          "ze"
        ],
        "ranges": {
          "kn": {
            "from": "lev_nodes"
          }
        },
        "expr": {
          "op": "ifelse",
          "args": [
            {
              "op": "==",
              "args": [
                "kn",
                1
              ]
            },
            {
              "op": "/",
              "args": [
                {
                  "op": "*",
                  "args": [
                    {
                      "op": "index",
                      "args": [
                        "df_layer",
                        1
                      ]
                    },
                    {
                      "op": "-",
                      "args": [
                        "yy",
                        {
                          "op": "index",
                          "args": [
                            "stc",
                            1
                          ]
                        }
                      ]
                    }
                  ]
                },
                {
                  "op": "*",
                  "args": [
                    {
                      "op": "*",
                      "args": [
                        0.5,
                        {
                          "op": "index",
                          "args": [
                            "ze",
                            2
                          ]
                        }
                      ]
                    },
                    "zz1"
                  ]
                }
              ]
            },
            {
              "op": "ifelse",
              "args": [
                {
                  "op": "==",
                  "args": [
                    "kn",
                    {
                      "op": "+",
                      "args": [
                        "NLEV",
                        1
                      ]
                    }
                  ]
                },
                {
                  "op": "/",
                  "args": [
                    {
                      "op": "*",
                      "args": [
                        {
                          "op": "index",
                          "args": [
                            "df_layer",
                            "NLEV"
                          ]
                        },
                        {
                          "op": "-",
                          "args": [
                            {
                              "op": "index",
                              "args": [
                                "stc",
                                "NLEV"
                              ]
                            },
                            "tbot"
                          ]
                        }
                      ]
                    },
                    {
                      "op": "-",
                      "args": [
                        "z_deep",
                        {
                          "op": "index",
                          "args": [
                            "zc",
                            "NLEV"
                          ]
                        }
                      ]
                    }
                  ]
                },
                0.0
              ]
            }
          ]
        }
      }
    },
    {
      "lhs": "dstc_dt",
      "rhs": {
        "op": "/",
        "args": [
          {
            "op": "D",
            "args": [
              {
                "op": "-",
                "args": [
                  {
                    "op": "*",
                    "args": [
                      "K_soil",
                      {
                        "op": "D",
                        "args": [
                          "stc"
                        ],
                        "wrt": "lev"
                      }
                    ]
                  },
                  "F_soil"
                ]
              }
            ],
            "wrt": "lev"
          },
          "hcpct"
        ]
      }
    },
    {
      "lhs": "ssoil_pre",
      "rhs": {
        "op": "/",
        "args": [
          {
            "op": "*",
            "args": [
              {
                "op": "index",
                "args": [
                  "df_layer",
                  1
                ]
              },
              {
                "op": "-",
                "args": [
                  "t1_in",
                  {
                    "op": "index",
                    "args": [
                      "stc",
                      1
                    ]
                  }
                ]
              }
            ]
          },
          {
            "op": "*",
            "args": [
              0.5,
              {
                "op": "index",
                "args": [
                  "ze",
                  2
                ]
              }
            ]
          }
        ]
      }
    },
    {
      "lhs": "ssoil",
      "rhs": {
        "op": "/",
        "args": [
          {
            "op": "*",
            "args": [
              {
                "op": "index",
                "args": [
                  "df_layer",
                  1
                ]
              },
              {
                "op": "-",
                "args": [
                  {
                    "op": "index",
                    "args": [
                      "stc",
                      1
                    ]
                  },
                  "t1"
                ]
              }
            ]
          },
          {
            "op": "*",
            "args": [
              0.5,
              {
                "op": "index",
                "args": [
                  "ze",
                  2
                ]
              }
            ]
          }
        ]
      }
    },
    {
      "lhs": "t2v",
      "rhs": {
        "op": "*",
        "args": [
          "sfctmp",
          {
            "op": "+",
            "args": [
              1.0,
              {
                "op": "*",
                "args": [
                  0.61,
                  "q2"
                ]
              }
            ]
          }
        ]
      }
    },
    {
      "lhs": "sheat",
      "rhs": {
        "op": "-",
        "args": [
          0.0,
          {
            "op": "*",
            "args": [
              {
                "op": "/",
                "args": [
                  {
                    "op": "*",
                    "args": [
                      {
                        "op": "*",
                        "args": [
                          "ch",
                          "wrf.cp"
                        ]
                      },
                      "sfcprs"
                    ]
                  },
                  {
                    "op": "*",
                    "args": [
                      "par.rd",
                      "t2v"
                    ]
                  }
                ]
              },
              {
                "op": "-",
                "args": [
                  "th2",
                  "t1"
                ]
              }
            ]
          }
        ]
      }
    }
  ],
  "tests": [
    {
      "id": "chem170_scm_step1_surface",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SURFACE CLOSURE at pre-dawn, soil warming from below. SFLX recomputes the surface conductivity df1 from the soil moisture that SMFLX has already advanced, so yy, zz1, t1, ssoil and sheat all carry an O(dt) dependence on the replay step; the reference is therefore taken at ESM_DT = 1e-6 s, where that truncation is below 1e-10 relative. At ESM_DT = 0.01 s the same quantities are already off the dt -> 0 limit by up to 5.7e-7 relative, and WRF's own dt = 60 s values by up to 3e-3.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 299.7296142578125,
        "sfcprs": 99642.484375,
        "th2": 300.0362854003906,
        "q2": 0.013806705363094807,
        "ch": 2.107988257193938e-05,
        "fdown": 464.0659408569336,
        "emissi": 0.9850000143051147,
        "rch": 0.024321563028787355,
        "rr": 246.25458052495014,
        "epsca": 48.4970804600707,
        "beta": 0.6509801410881244,
        "tbot": 300.0,
        "z_deep": 8.0,
        "t1_in": 293.0
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_surf_step1_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "yy",
          "time": 0.0,
          "expected": 301.82556840097305,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "zz1",
          "time": 0.0,
          "expected": 2.335787526654383,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "t1",
          "time": 0.0,
          "expected": 299.63780667435327,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil",
          "time": 0.0,
          "expected": -13.103153217511641,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "sheat",
          "time": 0.0,
          "expected": -0.009690660726137421,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil_pre",
          "time": 0.0,
          "expected": -40.002136464007236,
          "tolerance": {
            "rel": 1e-09
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step1_tendency",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SOIL TEMPERATURE TENDENCY at pre-dawn, soil warming from below. SFLX advances the soil column IN PLACE with an implicit tridiagonal solve (HSTEP), so its dt = 60 s increment is NOT the instantaneous tendency: across this reference set WRF's (STC_out - STC_in)/60 differs from the dt -> 0 limit by up to 2.7e-7 K/s, which is the implicit solve and not the physics. The reference here is the replay increment at ESM_DT = 0.01 s, whose own remaining truncation is at most 5e-11 K/s (measured against a Richardson extrapolation from 0.02 and 0.01 s) and therefore an order below the assertion tolerance. A smaller step cannot be used for an INCREMENT: at ESM_DT = 1e-6 s the difference of two ~300 K temperatures loses the tendency to cancellation. The per-level assertions therefore carry an ABSOLUTE bound rather than a tight relative one: the reference's own implicit-solve truncation is up to 1.1e-10 K/s, which is a few times 1e-5 of the QUIETEST deep layer's tendency, so those layers cannot be pinned relatively at this step size while the surface layer is pinned to 1e-6 of its own amplitude. The Linf assertion above is the bound that matters.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 299.7296142578125,
        "sfcprs": 99642.484375,
        "th2": 300.0362854003906,
        "q2": 0.013806705363094807,
        "ch": 2.107988257193938e-05,
        "fdown": 464.0659408569336,
        "emissi": 0.9850000143051147,
        "rch": 0.024321563028787355,
        "rr": 246.25458052495014,
        "epsca": 48.4970804600707,
        "beta": 0.6509801410881244,
        "tbot": 300.0,
        "z_deep": 8.0
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_tend_step1_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "reduce": "Linf_error",
          "expected": 0.0,
          "reference": {
            "op": "faq",
            "output_idx": [
              "k"
            ],
            "args": [],
            "ranges": {
              "k": {
                "from": "lev"
              }
            },
            "expr": {
              "op": "index",
              "args": [
                {
                  "op": "const",
                  "args": [],
                  "value": [
                    3.9437782106688246e-05,
                    -1.2023122053506086e-06,
                    2.5081476451305207e-06,
                    1.0962025953631382e-06
                  ]
                },
                "k"
              ]
            }
          },
          "tolerance": {
            "abs": 1e-09,
            "rel": 0.0
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 1
          },
          "expected": 3.9437782106688246e-05,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 2
          },
          "expected": -1.2023122053506086e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 3
          },
          "expected": 2.5081476451305207e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 4
          },
          "expected": 1.0962025953631382e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step60_surface",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SURFACE CLOSURE at just after sunrise. SFLX recomputes the surface conductivity df1 from the soil moisture that SMFLX has already advanced, so yy, zz1, t1, ssoil and sheat all carry an O(dt) dependence on the replay step; the reference is therefore taken at ESM_DT = 1e-6 s, where that truncation is below 1e-10 relative. At ESM_DT = 0.01 s the same quantities are already off the dt -> 0 limit by up to 5.7e-7 relative, and WRF's own dt = 60 s values by up to 3e-3.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 297.8214111328125,
        "sfcprs": 99616.421875,
        "th2": 298.1263427734375,
        "q2": 0.014198900200426579,
        "ch": 0.00029818981420248747,
        "fdown": 450.068904876709,
        "emissi": 0.9850000143051147,
        "rch": 0.34607737440065345,
        "rr": 17.904844034724064,
        "epsca": 17.58343248369813,
        "beta": 0.6724218277382642,
        "tbot": 300.0,
        "z_deep": 8.0,
        "t1_in": 297.3065490722656
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_surf_step60_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "yy",
          "time": 0.0,
          "expected": 298.9026546553282,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "zz1",
          "time": 0.0,
          "expected": 2.3179702217880642,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "t1",
          "time": 0.0,
          "expected": 298.27310599113633,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil",
          "time": 0.0,
          "expected": -3.9009740050129547,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "sheat",
          "time": 0.0,
          "expected": 0.050786373153974516,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil_pre",
          "time": 0.0,
          "expected": -3.992656280171251,
          "tolerance": {
            "rel": 1e-09
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step60_tendency",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SOIL TEMPERATURE TENDENCY at just after sunrise. SFLX advances the soil column IN PLACE with an implicit tridiagonal solve (HSTEP), so its dt = 60 s increment is NOT the instantaneous tendency: across this reference set WRF's (STC_out - STC_in)/60 differs from the dt -> 0 limit by up to 2.7e-7 K/s, which is the implicit solve and not the physics. The reference here is the replay increment at ESM_DT = 0.01 s, whose own remaining truncation is at most 5e-11 K/s (measured against a Richardson extrapolation from 0.02 and 0.01 s) and therefore an order below the assertion tolerance. A smaller step cannot be used for an INCREMENT: at ESM_DT = 1e-6 s the difference of two ~300 K temperatures loses the tendency to cancellation. The per-level assertions therefore carry an ABSOLUTE bound rather than a tight relative one: the reference's own implicit-solve truncation is up to 1.1e-10 K/s, which is a few times 1e-5 of the QUIETEST deep layer's tendency, so those layers cannot be pinned relatively at this step size while the surface layer is pinned to 1e-6 of its own amplitude. The Linf assertion above is the bound that matters.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 297.8214111328125,
        "sfcprs": 99616.421875,
        "th2": 298.1263427734375,
        "q2": 0.014198900200426579,
        "ch": 0.00029818981420248747,
        "fdown": 450.068904876709,
        "emissi": 0.9850000143051147,
        "rch": 0.34607737440065345,
        "rr": 17.904844034724064,
        "epsca": 17.58343248369813,
        "beta": 0.6724218277382642,
        "tbot": 300.0,
        "z_deep": 8.0
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_tend_step60_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "reduce": "Linf_error",
          "expected": 0.0,
          "reference": {
            "op": "faq",
            "output_idx": [
              "k"
            ],
            "args": [],
            "ranges": {
              "k": {
                "from": "lev"
              }
            },
            "expr": {
              "op": "index",
              "args": [
                {
                  "op": "const",
                  "args": [],
                  "value": [
                    9.513996701571159e-07,
                    -1.6293313365167705e-06,
                    2.4833695988490945e-06,
                    1.0991072940669255e-06
                  ]
                },
                "k"
              ]
            }
          },
          "tolerance": {
            "abs": 1e-09,
            "rel": 0.0
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 1
          },
          "expected": 9.513996701571159e-07,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 2
          },
          "expected": -1.6293313365167705e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 3
          },
          "expected": 2.4833695988490945e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 4
          },
          "expected": 1.0991072940669255e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step230_surface",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SURFACE CLOSURE at strong mid-morning insolation, the surface layer heating fastest. SFLX recomputes the surface conductivity df1 from the soil moisture that SMFLX has already advanced, so yy, zz1, t1, ssoil and sheat all carry an O(dt) dependence on the replay step; the reference is therefore taken at ESM_DT = 1e-6 s, where that truncation is below 1e-10 relative. At ESM_DT = 0.01 s the same quantities are already off the dt -> 0 limit by up to 5.7e-7 relative, and WRF's own dt = 60 s values by up to 3e-3.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 300.14801025390625,
        "sfcprs": 99621.1796875,
        "th2": 300.4555358886719,
        "q2": 0.01483145821839571,
        "ch": 0.012768357992172241,
        "fdown": 976.3091125488281,
        "emissi": 0.9850000143051147,
        "rch": 14.699092998016136,
        "rr": 1.4072549638543315,
        "epsca": 27.09832190762739,
        "beta": 0.6482645238283764,
        "tbot": 300.0,
        "z_deep": 8.0,
        "t1_in": 309.23431396484375
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_surf_step230_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "yy",
          "time": 0.0,
          "expected": 313.16870305574287,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "zz1",
          "time": 0.0,
          "expected": 1.3909433819552584,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "t1",
          "time": 0.0,
          "expected": 309.2848835615129,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil",
          "time": 0.0,
          "expected": -80.33824940960089,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "sheat",
          "time": 0.0,
          "expected": 129.77048364165324,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil_pre",
          "time": 0.0,
          "expected": 79.92930273643769,
          "tolerance": {
            "rel": 1e-09
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step230_tendency",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SOIL TEMPERATURE TENDENCY at strong mid-morning insolation, the surface layer heating fastest. SFLX advances the soil column IN PLACE with an implicit tridiagonal solve (HSTEP), so its dt = 60 s increment is NOT the instantaneous tendency: across this reference set WRF's (STC_out - STC_in)/60 differs from the dt -> 0 limit by up to 2.7e-7 K/s, which is the implicit solve and not the physics. The reference here is the replay increment at ESM_DT = 0.01 s, whose own remaining truncation is at most 5e-11 K/s (measured against a Richardson extrapolation from 0.02 and 0.01 s) and therefore an order below the assertion tolerance. A smaller step cannot be used for an INCREMENT: at ESM_DT = 1e-6 s the difference of two ~300 K temperatures loses the tendency to cancellation. The per-level assertions therefore carry an ABSOLUTE bound rather than a tight relative one: the reference's own implicit-solve truncation is up to 1.1e-10 K/s, which is a few times 1e-5 of the QUIETEST deep layer's tendency, so those layers cannot be pinned relatively at this step size while the surface layer is pinned to 1e-6 of its own amplitude. The Linf assertion above is the bound that matters.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 300.14801025390625,
        "sfcprs": 99621.1796875,
        "th2": 300.4555358886719,
        "q2": 0.01483145821839571,
        "ch": 0.012768357992172241,
        "fdown": 976.3091125488281,
        "emissi": 0.9850000143051147,
        "rch": 14.699092998016136,
        "rr": 1.4072549638543315,
        "epsca": 27.09832190762739,
        "beta": 0.6482645238283764,
        "tbot": 300.0,
        "z_deep": 8.0
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_tend_step230_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "reduce": "Linf_error",
          "expected": 0.0,
          "reference": {
            "op": "faq",
            "output_idx": [
              "k"
            ],
            "args": [],
            "ranges": {
              "k": {
                "from": "lev"
              }
            },
            "expr": {
              "op": "index",
              "args": [
                {
                  "op": "const",
                  "args": [],
                  "value": [
                    0.00031447939932149893,
                    2.931193421318312e-06,
                    2.4268899778689956e-06,
                    1.1074575922975782e-06
                  ]
                },
                "k"
              ]
            }
          },
          "tolerance": {
            "abs": 1e-09,
            "rel": 0.0
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 1
          },
          "expected": 0.00031447939932149893,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 2
          },
          "expected": 2.931193421318312e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 3
          },
          "expected": 2.4268899778689956e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 4
          },
          "expected": 1.1074575922975782e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step364_surface",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SURFACE CLOSURE at local noon. SFLX recomputes the surface conductivity df1 from the soil moisture that SMFLX has already advanced, so yy, zz1, t1, ssoil and sheat all carry an O(dt) dependence on the replay step; the reference is therefore taken at ESM_DT = 1e-6 s, where that truncation is below 1e-10 relative. At ESM_DT = 0.01 s the same quantities are already off the dt -> 0 limit by up to 5.7e-7 relative, and WRF's own dt = 60 s values by up to 3e-3.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 301.68988037109375,
        "sfcprs": 99629.5,
        "th2": 301.9989929199219,
        "q2": 0.015065288171172142,
        "ch": 0.01244852039963007,
        "fdown": 853.9974670410156,
        "emissi": 0.9850000143051147,
        "rch": 14.2568254433333,
        "rr": 1.4270901071047122,
        "epsca": 24.03958869233691,
        "beta": 0.6161406710570033,
        "tbot": 300.0,
        "z_deep": 8.0,
        "t1_in": 308.3092041015625
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_surf_step364_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "yy",
          "time": 0.0,
          "expected": 310.7618034400845,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "zz1",
          "time": 0.0,
          "expected": 1.3950491893101848,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "t1",
          "time": 0.0,
          "expected": 308.2036846590897,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil",
          "time": 0.0,
          "expected": -52.046907987427936,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "sheat",
          "time": 0.0,
          "expected": 88.45040163921459,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil_pre",
          "time": 0.0,
          "expected": 52.89502913304898,
          "tolerance": {
            "rel": 1e-09
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step364_tendency",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SOIL TEMPERATURE TENDENCY at local noon. SFLX advances the soil column IN PLACE with an implicit tridiagonal solve (HSTEP), so its dt = 60 s increment is NOT the instantaneous tendency: across this reference set WRF's (STC_out - STC_in)/60 differs from the dt -> 0 limit by up to 2.7e-7 K/s, which is the implicit solve and not the physics. The reference here is the replay increment at ESM_DT = 0.01 s, whose own remaining truncation is at most 5e-11 K/s (measured against a Richardson extrapolation from 0.02 and 0.01 s) and therefore an order below the assertion tolerance. A smaller step cannot be used for an INCREMENT: at ESM_DT = 1e-6 s the difference of two ~300 K temperatures loses the tendency to cancellation. The per-level assertions therefore carry an ABSOLUTE bound rather than a tight relative one: the reference's own implicit-solve truncation is up to 1.1e-10 K/s, which is a few times 1e-5 of the QUIETEST deep layer's tendency, so those layers cannot be pinned relatively at this step size while the surface layer is pinned to 1e-6 of its own amplitude. The Linf assertion above is the bound that matters.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 301.68988037109375,
        "sfcprs": 99629.5,
        "th2": 301.9989929199219,
        "q2": 0.015065288171172142,
        "ch": 0.01244852039963007,
        "fdown": 853.9974670410156,
        "emissi": 0.9850000143051147,
        "rch": 14.2568254433333,
        "rr": 1.4270901071047122,
        "epsca": 24.03958869233691,
        "beta": 0.6161406710570033,
        "tbot": 300.0,
        "z_deep": 8.0
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_tend_step364_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "reduce": "Linf_error",
          "expected": 0.0,
          "reference": {
            "op": "faq",
            "output_idx": [
              "k"
            ],
            "args": [],
            "ranges": {
              "k": {
                "from": "lev"
              }
            },
            "expr": {
              "op": "index",
              "args": [
                {
                  "op": "const",
                  "args": [],
                  "value": [
                    0.00017502720197626331,
                    9.48526235333702e-06,
                    2.472069127179566e-06,
                    1.1136933153466089e-06
                  ]
                },
                "k"
              ]
            }
          },
          "tolerance": {
            "abs": 1e-09,
            "rel": 0.0
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 1
          },
          "expected": 0.00017502720197626331,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 2
          },
          "expected": 9.48526235333702e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 3
          },
          "expected": 2.472069127179566e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 4
          },
          "expected": 1.1136933153466089e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step601_surface",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SURFACE CLOSURE at afternoon, the surface layer now COOLING while the deep layers still warm. SFLX recomputes the surface conductivity df1 from the soil moisture that SMFLX has already advanced, so yy, zz1, t1, ssoil and sheat all carry an O(dt) dependence on the replay step; the reference is therefore taken at ESM_DT = 1e-6 s, where that truncation is below 1e-10 relative. At ESM_DT = 0.01 s the same quantities are already off the dt -> 0 limit by up to 5.7e-7 relative, and WRF's own dt = 60 s values by up to 3e-3.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 299.3103332519531,
        "sfcprs": 99612.390625,
        "th2": 299.6170959472656,
        "q2": 0.015172002837061882,
        "ch": 0.009639615193009377,
        "fdown": 612.3471374511719,
        "emissi": 0.9850000143051147,
        "rch": 11.125030914697573,
        "rr": 1.7729520881365048,
        "epsca": 15.771016488080244,
        "beta": 0.6748546258311495,
        "tbot": 300.0,
        "z_deep": 8.0,
        "t1_in": 302.1735534667969
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_surf_step601_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "yy",
          "time": 0.0,
          "expected": 301.8006411039044,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "zz1",
          "time": 0.0,
          "expected": 1.4149776378556167,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "t1",
          "time": 0.0,
          "expected": 302.2223814405504,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil",
          "time": 0.0,
          "expected": 8.318468307605414,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "sheat",
          "time": 0.0,
          "expected": 28.980996537777372,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil_pre",
          "time": 0.0,
          "expected": -8.718129172122323,
          "tolerance": {
            "rel": 1e-09
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step601_tendency",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SOIL TEMPERATURE TENDENCY at afternoon, the surface layer now COOLING while the deep layers still warm. SFLX advances the soil column IN PLACE with an implicit tridiagonal solve (HSTEP), so its dt = 60 s increment is NOT the instantaneous tendency: across this reference set WRF's (STC_out - STC_in)/60 differs from the dt -> 0 limit by up to 2.7e-7 K/s, which is the implicit solve and not the physics. The reference here is the replay increment at ESM_DT = 0.01 s, whose own remaining truncation is at most 5e-11 K/s (measured against a Richardson extrapolation from 0.02 and 0.01 s) and therefore an order below the assertion tolerance. A smaller step cannot be used for an INCREMENT: at ESM_DT = 1e-6 s the difference of two ~300 K temperatures loses the tendency to cancellation. The per-level assertions therefore carry an ABSOLUTE bound rather than a tight relative one: the reference's own implicit-solve truncation is up to 1.1e-10 K/s, which is a few times 1e-5 of the QUIETEST deep layer's tendency, so those layers cannot be pinned relatively at this step size while the surface layer is pinned to 1e-6 of its own amplitude. The Linf assertion above is the bound that matters.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 299.3103332519531,
        "sfcprs": 99612.390625,
        "th2": 299.6170959472656,
        "q2": 0.015172002837061882,
        "ch": 0.009639615193009377,
        "fdown": 612.3471374511719,
        "emissi": 0.9850000143051147,
        "rch": 11.125030914697573,
        "rr": 1.7729520881365048,
        "epsca": 15.771016488080244,
        "beta": 0.6748546258311495,
        "tbot": 300.0,
        "z_deep": 8.0
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_tend_step601_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "reduce": "Linf_error",
          "expected": 0.0,
          "reference": {
            "op": "faq",
            "output_idx": [
              "k"
            ],
            "args": [],
            "ranges": {
              "k": {
                "from": "lev"
              }
            },
            "expr": {
              "op": "index",
              "args": [
                {
                  "op": "const",
                  "args": [],
                  "value": [
                    -9.555952829032321e-05,
                    1.3301087165018544e-05,
                    2.682554622879252e-06,
                    1.124817572417669e-06
                  ]
                },
                "k"
              ]
            }
          },
          "tolerance": {
            "abs": 1e-09,
            "rel": 0.0
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 1
          },
          "expected": -9.555952829032321e-05,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 2
          },
          "expected": 1.3301087165018544e-05,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 3
          },
          "expected": 2.682554622879252e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 4
          },
          "expected": 1.124817572417669e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step639_surface",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SURFACE CLOSURE at shallow convection. SFLX recomputes the surface conductivity df1 from the soil moisture that SMFLX has already advanced, so yy, zz1, t1, ssoil and sheat all carry an O(dt) dependence on the replay step; the reference is therefore taken at ESM_DT = 1e-6 s, where that truncation is below 1e-10 relative. At ESM_DT = 0.01 s the same quantities are already off the dt -> 0 limit by up to 5.7e-7 relative, and WRF's own dt = 60 s values by up to 3e-3.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 299.25006103515625,
        "sfcprs": 99611.859375,
        "th2": 299.55670166015625,
        "q2": 0.014930164441466331,
        "ch": 0.009273719042539597,
        "fdown": 587.1937408447266,
        "emissi": 0.9850000143051147,
        "rch": 10.706415528346097,
        "rr": 1.561817766052784,
        "epsca": 14.783276314465839,
        "beta": 0.6757275953158651,
        "tbot": 300.0,
        "z_deep": 8.0,
        "t1_in": 301.93975830078125
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_surf_step639_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "yy",
          "time": 0.0,
          "expected": 301.3821057484072,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "zz1",
          "time": 0.0,
          "expected": 1.49005705152657,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "t1",
          "time": 0.0,
          "expected": 301.92382172848716,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil",
          "time": 0.0,
          "expected": 9.058287500172343,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "sheat",
          "time": 0.0,
          "expected": 25.34084832450164,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil_pre",
          "time": 0.0,
          "expected": -8.927695669247866,
          "tolerance": {
            "rel": 1e-09
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step639_tendency",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SOIL TEMPERATURE TENDENCY at shallow convection. SFLX advances the soil column IN PLACE with an implicit tridiagonal solve (HSTEP), so its dt = 60 s increment is NOT the instantaneous tendency: across this reference set WRF's (STC_out - STC_in)/60 differs from the dt -> 0 limit by up to 2.7e-7 K/s, which is the implicit solve and not the physics. The reference here is the replay increment at ESM_DT = 0.01 s, whose own remaining truncation is at most 5e-11 K/s (measured against a Richardson extrapolation from 0.02 and 0.01 s) and therefore an order below the assertion tolerance. A smaller step cannot be used for an INCREMENT: at ESM_DT = 1e-6 s the difference of two ~300 K temperatures loses the tendency to cancellation. The per-level assertions therefore carry an ABSOLUTE bound rather than a tight relative one: the reference's own implicit-solve truncation is up to 1.1e-10 K/s, which is a few times 1e-5 of the QUIETEST deep layer's tendency, so those layers cannot be pinned relatively at this step size while the surface layer is pinned to 1e-6 of its own amplitude. The Linf assertion above is the bound that matters.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 299.25006103515625,
        "sfcprs": 99611.859375,
        "th2": 299.55670166015625,
        "q2": 0.014930164441466331,
        "ch": 0.009273719042539597,
        "fdown": 587.1937408447266,
        "emissi": 0.9850000143051147,
        "rch": 10.706415528346097,
        "rr": 1.561817766052784,
        "epsca": 14.783276314465839,
        "beta": 0.6757275953158651,
        "tbot": 300.0,
        "z_deep": 8.0
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_tend_step639_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "reduce": "Linf_error",
          "expected": 0.0,
          "reference": {
            "op": "faq",
            "output_idx": [
              "k"
            ],
            "args": [],
            "ranges": {
              "k": {
                "from": "lev"
              }
            },
            "expr": {
              "op": "index",
              "args": [
                {
                  "op": "const",
                  "args": [],
                  "value": [
                    -9.65419758358621e-05,
                    1.2512725788837997e-05,
                    2.724493697314756e-06,
                    1.1267047739238478e-06
                  ]
                },
                "k"
              ]
            }
          },
          "tolerance": {
            "abs": 1e-09,
            "rel": 0.0
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 1
          },
          "expected": -9.65419758358621e-05,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 2
          },
          "expected": 1.2512725788837997e-05,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 3
          },
          "expected": 2.724493697314756e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 4
          },
          "expected": 1.1267047739238478e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step689_surface",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SURFACE CLOSURE at late afternoon. SFLX recomputes the surface conductivity df1 from the soil moisture that SMFLX has already advanced, so yy, zz1, t1, ssoil and sheat all carry an O(dt) dependence on the replay step; the reference is therefore taken at ESM_DT = 1e-6 s, where that truncation is below 1e-10 relative. At ESM_DT = 0.01 s the same quantities are already off the dt -> 0 limit by up to 5.7e-7 relative, and WRF's own dt = 60 s values by up to 3e-3.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 299.45379638671875,
        "sfcprs": 99608.4140625,
        "th2": 299.76068115234375,
        "q2": 0.015095737762749195,
        "ch": 0.008500252850353718,
        "fdown": 566.6586456298828,
        "emissi": 0.9850000143051147,
        "rch": 9.805459324103195,
        "rr": 1.6061814831450427,
        "epsca": 14.211578575008103,
        "beta": 0.6685542680892319,
        "tbot": 300.0,
        "z_deep": 8.0,
        "t1_in": 301.7066650390625
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_surf_step689_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "yy",
          "time": 0.0,
          "expected": 301.19410134654964,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "zz1",
          "time": 0.0,
          "expected": 1.518409287854643,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "t1",
          "time": 0.0,
          "expected": 301.7228928529147,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil",
          "time": 0.0,
          "expected": 8.328121030210477,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "sheat",
          "time": 0.0,
          "expected": 19.238471786578046,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil_pre",
          "time": 0.0,
          "expected": -8.460614767591045,
          "tolerance": {
            "rel": 1e-09
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step689_tendency",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SOIL TEMPERATURE TENDENCY at late afternoon. SFLX advances the soil column IN PLACE with an implicit tridiagonal solve (HSTEP), so its dt = 60 s increment is NOT the instantaneous tendency: across this reference set WRF's (STC_out - STC_in)/60 differs from the dt -> 0 limit by up to 2.7e-7 K/s, which is the implicit solve and not the physics. The reference here is the replay increment at ESM_DT = 0.01 s, whose own remaining truncation is at most 5e-11 K/s (measured against a Richardson extrapolation from 0.02 and 0.01 s) and therefore an order below the assertion tolerance. A smaller step cannot be used for an INCREMENT: at ESM_DT = 1e-6 s the difference of two ~300 K temperatures loses the tendency to cancellation. The per-level assertions therefore carry an ABSOLUTE bound rather than a tight relative one: the reference's own implicit-solve truncation is up to 1.1e-10 K/s, which is a few times 1e-5 of the QUIETEST deep layer's tendency, so those layers cannot be pinned relatively at this step size while the surface layer is pinned to 1e-6 of its own amplitude. The Linf assertion above is the bound that matters.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 299.45379638671875,
        "sfcprs": 99608.4140625,
        "th2": 299.76068115234375,
        "q2": 0.015095737762749195,
        "ch": 0.008500252850353718,
        "fdown": 566.6586456298828,
        "emissi": 0.9850000143051147,
        "rch": 9.805459324103195,
        "rr": 1.6061814831450427,
        "epsca": 14.211578575008103,
        "beta": 0.6685542680892319,
        "tbot": 300.0,
        "z_deep": 8.0
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_tend_step689_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "reduce": "Linf_error",
          "expected": 0.0,
          "reference": {
            "op": "faq",
            "output_idx": [
              "k"
            ],
            "args": [],
            "ranges": {
              "k": {
                "from": "lev"
              }
            },
            "expr": {
              "op": "index",
              "args": [
                {
                  "op": "const",
                  "args": [],
                  "value": [
                    -9.091482979783905e-05,
                    1.1372753760952037e-05,
                    2.7731402951758355e-06,
                    1.1292002000118373e-06
                  ]
                },
                "k"
              ]
            }
          },
          "tolerance": {
            "abs": 1e-09,
            "rel": 0.0
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 1
          },
          "expected": -9.091482979783905e-05,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 2
          },
          "expected": 1.1372753760952037e-05,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 3
          },
          "expected": 2.7731402951758355e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 4
          },
          "expected": 1.1292002000118373e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step918_surface",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SURFACE CLOSURE at night, ground heat flux upward and the surface layer cooling fastest. SFLX recomputes the surface conductivity df1 from the soil moisture that SMFLX has already advanced, so yy, zz1, t1, ssoil and sheat all carry an O(dt) dependence on the replay step; the reference is therefore taken at ESM_DT = 1e-6 s, where that truncation is below 1e-10 relative. At ESM_DT = 0.01 s the same quantities are already off the dt -> 0 limit by up to 5.7e-7 relative, and WRF's own dt = 60 s values by up to 3e-3.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 298.3406066894531,
        "sfcprs": 99612.109375,
        "th2": 298.64642333984375,
        "q2": 0.015314576216042042,
        "ch": 0.00012309601879678667,
        "fdown": 400.2366027832031,
        "emissi": 0.9850000143051147,
        "rch": 0.14251357324362676,
        "rr": 42.23866266329492,
        "epsca": 6.535517036714274,
        "beta": 0.6545212284021527,
        "tbot": 300.0,
        "z_deep": 8.0,
        "t1_in": 296.5185546875
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_surf_step918_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "yy",
          "time": 0.0,
          "expected": 291.23312211475474,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "zz1",
          "time": 0.0,
          "expected": 2.3444891394324054,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "t1",
          "time": 0.0,
          "expected": 296.5046590507855,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil",
          "time": 0.0,
          "expected": 31.732452780701447,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "sheat",
          "time": 0.0,
          "expected": -0.30520009859420977,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil_pre",
          "time": 0.0,
          "expected": -31.61999173092796,
          "tolerance": {
            "rel": 1e-09
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step918_tendency",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SOIL TEMPERATURE TENDENCY at night, ground heat flux upward and the surface layer cooling fastest. SFLX advances the soil column IN PLACE with an implicit tridiagonal solve (HSTEP), so its dt = 60 s increment is NOT the instantaneous tendency: across this reference set WRF's (STC_out - STC_in)/60 differs from the dt -> 0 limit by up to 2.7e-7 K/s, which is the implicit solve and not the physics. The reference here is the replay increment at ESM_DT = 0.01 s, whose own remaining truncation is at most 5e-11 K/s (measured against a Richardson extrapolation from 0.02 and 0.01 s) and therefore an order below the assertion tolerance. A smaller step cannot be used for an INCREMENT: at ESM_DT = 1e-6 s the difference of two ~300 K temperatures loses the tendency to cancellation. The per-level assertions therefore carry an ABSOLUTE bound rather than a tight relative one: the reference's own implicit-solve truncation is up to 1.1e-10 K/s, which is a few times 1e-5 of the QUIETEST deep layer's tendency, so those layers cannot be pinned relatively at this step size while the surface layer is pinned to 1e-6 of its own amplitude. The Linf assertion above is the bound that matters.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 298.3406066894531,
        "sfcprs": 99612.109375,
        "th2": 298.64642333984375,
        "q2": 0.015314576216042042,
        "ch": 0.00012309601879678667,
        "fdown": 400.2366027832031,
        "emissi": 0.9850000143051147,
        "rch": 0.14251357324362676,
        "rr": 42.23866266329492,
        "epsca": 6.535517036714274,
        "beta": 0.6545212284021527,
        "tbot": 300.0,
        "z_deep": 8.0
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_tend_step918_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "reduce": "Linf_error",
          "expected": 0.0,
          "reference": {
            "op": "faq",
            "output_idx": [
              "k"
            ],
            "args": [],
            "ranges": {
              "k": {
                "from": "lev"
              }
            },
            "expr": {
              "op": "index",
              "args": [
                {
                  "op": "const",
                  "args": [],
                  "value": [
                    -0.0001704514545508573,
                    3.908104417860159e-06,
                    2.8671422569459537e-06,
                    1.1442864433774957e-06
                  ]
                },
                "k"
              ]
            }
          },
          "tolerance": {
            "abs": 1e-09,
            "rel": 0.0
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 1
          },
          "expected": -0.0001704514545508573,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 2
          },
          "expected": 3.908104417860159e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 3
          },
          "expected": 2.8671422569459537e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 4
          },
          "expected": 1.1442864433774957e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step1080_surface",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SURFACE CLOSURE at deep night. SFLX recomputes the surface conductivity df1 from the soil moisture that SMFLX has already advanced, so yy, zz1, t1, ssoil and sheat all carry an O(dt) dependence on the replay step; the reference is therefore taken at ESM_DT = 1e-6 s, where that truncation is below 1e-10 relative. At ESM_DT = 0.01 s the same quantities are already off the dt -> 0 limit by up to 5.7e-7 relative, and WRF's own dt = 60 s values by up to 3e-3.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 297.35552978515625,
        "sfcprs": 99613.140625,
        "th2": 297.66033935546875,
        "q2": 0.015274709090590477,
        "ch": 0.00010150901653105393,
        "fdown": 392.67742919921875,
        "emissi": 0.9850000143051147,
        "rch": 0.11791475463723072,
        "rr": 50.35078518825073,
        "epsca": 2.8492893007706352,
        "beta": 0.6531550125641934,
        "tbot": 300.0,
        "z_deep": 8.0,
        "t1_in": 295.1018981933594
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_surf_step1080_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "yy",
          "time": 0.0,
          "expected": 289.91992620402914,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "zz1",
          "time": 0.0,
          "expected": 2.3594109723979004,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "t1",
          "time": 0.0,
          "expected": 295.08404877672245,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil",
          "time": 0.0,
          "expected": 30.659914611646617,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "sheat",
          "time": 0.0,
          "expected": -0.30375243230009996,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil_pre",
          "time": 0.0,
          "expected": -30.51585269342517,
          "tolerance": {
            "rel": 1e-09
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step1080_tendency",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SOIL TEMPERATURE TENDENCY at deep night. SFLX advances the soil column IN PLACE with an implicit tridiagonal solve (HSTEP), so its dt = 60 s increment is NOT the instantaneous tendency: across this reference set WRF's (STC_out - STC_in)/60 differs from the dt -> 0 limit by up to 2.7e-7 K/s, which is the implicit solve and not the physics. The reference here is the replay increment at ESM_DT = 0.01 s, whose own remaining truncation is at most 5e-11 K/s (measured against a Richardson extrapolation from 0.02 and 0.01 s) and therefore an order below the assertion tolerance. A smaller step cannot be used for an INCREMENT: at ESM_DT = 1e-6 s the difference of two ~300 K temperatures loses the tendency to cancellation. The per-level assertions therefore carry an ABSOLUTE bound rather than a tight relative one: the reference's own implicit-solve truncation is up to 1.1e-10 K/s, which is a few times 1e-5 of the QUIETEST deep layer's tendency, so those layers cannot be pinned relatively at this step size while the surface layer is pinned to 1e-6 of its own amplitude. The Linf assertion above is the bound that matters.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 297.35552978515625,
        "sfcprs": 99613.140625,
        "th2": 297.66033935546875,
        "q2": 0.015274709090590477,
        "ch": 0.00010150901653105393,
        "fdown": 392.67742919921875,
        "emissi": 0.9850000143051147,
        "rch": 0.11791475463723072,
        "rr": 50.35078518825073,
        "epsca": 2.8492893007706352,
        "beta": 0.6531550125641934,
        "tbot": 300.0,
        "z_deep": 8.0
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_tend_step1080_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "reduce": "Linf_error",
          "expected": 0.0,
          "reference": {
            "op": "faq",
            "output_idx": [
              "k"
            ],
            "args": [],
            "ranges": {
              "k": {
                "from": "lev"
              }
            },
            "expr": {
              "op": "index",
              "args": [
                {
                  "op": "const",
                  "args": [],
                  "value": [
                    -0.0001528715756649035,
                    -5.794731805508491e-07,
                    2.8213491987116868e-06,
                    1.1553311196621507e-06
                  ]
                },
                "k"
              ]
            }
          },
          "tolerance": {
            "abs": 1e-09,
            "rel": 0.0
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 1
          },
          "expected": -0.0001528715756649035,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 2
          },
          "expected": -5.794731805508491e-07,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 3
          },
          "expected": 2.8213491987116868e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 4
          },
          "expected": 1.1553311196621507e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step1621_surface",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SURFACE CLOSURE at second-day morning, near-peak heating. SFLX recomputes the surface conductivity df1 from the soil moisture that SMFLX has already advanced, so yy, zz1, t1, ssoil and sheat all carry an O(dt) dependence on the replay step; the reference is therefore taken at ESM_DT = 1e-6 s, where that truncation is below 1e-10 relative. At ESM_DT = 0.01 s the same quantities are already off the dt -> 0 limit by up to 5.7e-7 relative, and WRF's own dt = 60 s values by up to 3e-3.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 299.1268310546875,
        "sfcprs": 99615.1875,
        "th2": 299.43359375,
        "q2": 0.015753425657749176,
        "ch": 0.012447330169379711,
        "fdown": 899.6530456542969,
        "emissi": 0.9850000143051147,
        "rch": 14.369566997367366,
        "rr": 1.4121268947240087,
        "epsca": 22.42602252144387,
        "beta": 0.6396619076399839,
        "tbot": 300.0,
        "z_deep": 8.0,
        "t1_in": 307.5830383300781
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_surf_step1621_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "yy",
          "time": 0.0,
          "expected": 311.4861718295139,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "zz1",
          "time": 0.0,
          "expected": 1.3956692331800986,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "t1",
          "time": 0.0,
          "expected": 307.6424531143974,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil",
          "time": 0.0,
          "expected": -77.9954026395952,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "sheat",
          "time": 0.0,
          "expected": 117.94601284532584,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil_pre",
          "time": 0.0,
          "expected": 77.51837426781903,
          "tolerance": {
            "rel": 1e-09
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step1621_tendency",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SOIL TEMPERATURE TENDENCY at second-day morning, near-peak heating. SFLX advances the soil column IN PLACE with an implicit tridiagonal solve (HSTEP), so its dt = 60 s increment is NOT the instantaneous tendency: across this reference set WRF's (STC_out - STC_in)/60 differs from the dt -> 0 limit by up to 2.7e-7 K/s, which is the implicit solve and not the physics. The reference here is the replay increment at ESM_DT = 0.01 s, whose own remaining truncation is at most 5e-11 K/s (measured against a Richardson extrapolation from 0.02 and 0.01 s) and therefore an order below the assertion tolerance. A smaller step cannot be used for an INCREMENT: at ESM_DT = 1e-6 s the difference of two ~300 K temperatures loses the tendency to cancellation. The per-level assertions therefore carry an ABSOLUTE bound rather than a tight relative one: the reference's own implicit-solve truncation is up to 1.1e-10 K/s, which is a few times 1e-5 of the QUIETEST deep layer's tendency, so those layers cannot be pinned relatively at this step size while the surface layer is pinned to 1e-6 of its own amplitude. The Linf assertion above is the bound that matters.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 299.1268310546875,
        "sfcprs": 99615.1875,
        "th2": 299.43359375,
        "q2": 0.015753425657749176,
        "ch": 0.012447330169379711,
        "fdown": 899.6530456542969,
        "emissi": 0.9850000143051147,
        "rch": 14.369566997367366,
        "rr": 1.4121268947240087,
        "epsca": 22.42602252144387,
        "beta": 0.6396619076399839,
        "tbot": 300.0,
        "z_deep": 8.0
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_tend_step1621_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "reduce": "Linf_error",
          "expected": 0.0,
          "reference": {
            "op": "faq",
            "output_idx": [
              "k"
            ],
            "args": [],
            "ranges": {
              "k": {
                "from": "lev"
              }
            },
            "expr": {
              "op": "index",
              "args": [
                {
                  "op": "const",
                  "args": [],
                  "value": [
                    0.00032187143119699613,
                    -2.0642403342208127e-06,
                    2.340590299354517e-06,
                    1.1830991297756555e-06
                  ]
                },
                "k"
              ]
            }
          },
          "tolerance": {
            "abs": 1e-09,
            "rel": 0.0
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 1
          },
          "expected": 0.00032187143119699613,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 2
          },
          "expected": -2.0642403342208127e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 3
          },
          "expected": 2.340590299354517e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 4
          },
          "expected": 1.1830991297756555e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step1730_surface",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SURFACE CLOSURE at second-day deep convection. SFLX recomputes the surface conductivity df1 from the soil moisture that SMFLX has already advanced, so yy, zz1, t1, ssoil and sheat all carry an O(dt) dependence on the replay step; the reference is therefore taken at ESM_DT = 1e-6 s, where that truncation is below 1e-10 relative. At ESM_DT = 0.01 s the same quantities are already off the dt -> 0 limit by up to 5.7e-7 relative, and WRF's own dt = 60 s values by up to 3e-3.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 300.72509765625,
        "sfcprs": 99620.203125,
        "th2": 301.03338623046875,
        "q2": 0.01549688819795847,
        "ch": 0.011664578691124916,
        "fdown": 785.117431640625,
        "emissi": 0.9850000143051147,
        "rch": 13.397118883033619,
        "rr": 1.4518772881912183,
        "epsca": 20.568927059559382,
        "beta": 0.6190404799088226,
        "tbot": 300.0,
        "z_deep": 8.0,
        "t1_in": 306.3863830566406
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_surf_step1730_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "yy",
          "time": 0.0,
          "expected": 309.0481607483091,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "zz1",
          "time": 0.0,
          "expected": 1.4119028220808052,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "t1",
          "time": 0.0,
          "expected": 306.41454099730134,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil",
          "time": 0.0,
          "expected": -51.22646570382922,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "sheat",
          "time": 0.0,
          "expected": 72.08479395270355,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil_pre",
          "time": 0.0,
          "expected": 51.00086680435393,
          "tolerance": {
            "rel": 1e-09
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step1730_tendency",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SOIL TEMPERATURE TENDENCY at second-day deep convection. SFLX advances the soil column IN PLACE with an implicit tridiagonal solve (HSTEP), so its dt = 60 s increment is NOT the instantaneous tendency: across this reference set WRF's (STC_out - STC_in)/60 differs from the dt -> 0 limit by up to 2.7e-7 K/s, which is the implicit solve and not the physics. The reference here is the replay increment at ESM_DT = 0.01 s, whose own remaining truncation is at most 5e-11 K/s (measured against a Richardson extrapolation from 0.02 and 0.01 s) and therefore an order below the assertion tolerance. A smaller step cannot be used for an INCREMENT: at ESM_DT = 1e-6 s the difference of two ~300 K temperatures loses the tendency to cancellation. The per-level assertions therefore carry an ABSOLUTE bound rather than a tight relative one: the reference's own implicit-solve truncation is up to 1.1e-10 K/s, which is a few times 1e-5 of the QUIETEST deep layer's tendency, so those layers cannot be pinned relatively at this step size while the surface layer is pinned to 1e-6 of its own amplitude. The Linf assertion above is the bound that matters.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 300.72509765625,
        "sfcprs": 99620.203125,
        "th2": 301.03338623046875,
        "q2": 0.01549688819795847,
        "ch": 0.011664578691124916,
        "fdown": 785.117431640625,
        "emissi": 0.9850000143051147,
        "rch": 13.397118883033619,
        "rr": 1.4518772881912183,
        "epsca": 20.568927059559382,
        "beta": 0.6190404799088226,
        "tbot": 300.0,
        "z_deep": 8.0
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_tend_step1730_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "reduce": "Linf_error",
          "expected": 0.0,
          "reference": {
            "op": "faq",
            "output_idx": [
              "k"
            ],
            "args": [],
            "ranges": {
              "k": {
                "from": "lev"
              }
            },
            "expr": {
              "op": "index",
              "args": [
                {
                  "op": "const",
                  "args": [],
                  "value": [
                    0.0001888967005925224,
                    3.959820560339722e-06,
                    2.3136067284212913e-06,
                    1.1882434591825586e-06
                  ]
                },
                "k"
              ]
            }
          },
          "tolerance": {
            "abs": 1e-09,
            "rel": 0.0
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 1
          },
          "expected": 0.0001888967005925224,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 2
          },
          "expected": 3.959820560339722e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 3
          },
          "expected": 2.3136067284212913e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 4
          },
          "expected": 1.1882434591825586e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step1801_surface",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SURFACE CLOSURE at second-day noon. SFLX recomputes the surface conductivity df1 from the soil moisture that SMFLX has already advanced, so yy, zz1, t1, ssoil and sheat all carry an O(dt) dependence on the replay step; the reference is therefore taken at ESM_DT = 1e-6 s, where that truncation is below 1e-10 relative. At ESM_DT = 0.01 s the same quantities are already off the dt -> 0 limit by up to 5.7e-7 relative, and WRF's own dt = 60 s values by up to 3e-3.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 301.22955322265625,
        "sfcprs": 99622.421875,
        "th2": 301.538330078125,
        "q2": 0.015355565585196018,
        "ch": 0.011747344397008419,
        "fdown": 771.9682312011719,
        "emissi": 0.9850000143051147,
        "rch": 13.471033454304882,
        "rr": 1.450614920682188,
        "epsca": 20.54472686350452,
        "beta": 0.611660291585341,
        "tbot": 300.0,
        "z_deep": 8.0,
        "t1_in": 306.5345764160156
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_surf_step1801_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "yy",
          "time": 0.0,
          "expected": 308.75221404801084,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "zz1",
          "time": 0.0,
          "expected": 1.4095683753568364,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "t1",
          "time": 0.0,
          "expected": 306.4485080361292,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil",
          "time": 0.0,
          "expected": -45.017369113134336,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "sheat",
          "time": 0.0,
          "expected": 66.13858730916807,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil_pre",
          "time": 0.0,
          "expected": 45.70621663305579,
          "tolerance": {
            "rel": 1e-09
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step1801_tendency",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SOIL TEMPERATURE TENDENCY at second-day noon. SFLX advances the soil column IN PLACE with an implicit tridiagonal solve (HSTEP), so its dt = 60 s increment is NOT the instantaneous tendency: across this reference set WRF's (STC_out - STC_in)/60 differs from the dt -> 0 limit by up to 2.7e-7 K/s, which is the implicit solve and not the physics. The reference here is the replay increment at ESM_DT = 0.01 s, whose own remaining truncation is at most 5e-11 K/s (measured against a Richardson extrapolation from 0.02 and 0.01 s) and therefore an order below the assertion tolerance. A smaller step cannot be used for an INCREMENT: at ESM_DT = 1e-6 s the difference of two ~300 K temperatures loses the tendency to cancellation. The per-level assertions therefore carry an ABSOLUTE bound rather than a tight relative one: the reference's own implicit-solve truncation is up to 1.1e-10 K/s, which is a few times 1e-5 of the QUIETEST deep layer's tendency, so those layers cannot be pinned relatively at this step size while the surface layer is pinned to 1e-6 of its own amplitude. The Linf assertion above is the bound that matters.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 301.22955322265625,
        "sfcprs": 99622.421875,
        "th2": 301.538330078125,
        "q2": 0.015355565585196018,
        "ch": 0.011747344397008419,
        "fdown": 771.9682312011719,
        "emissi": 0.9850000143051147,
        "rch": 13.471033454304882,
        "rr": 1.450614920682188,
        "epsca": 20.54472686350452,
        "beta": 0.611660291585341,
        "tbot": 300.0,
        "z_deep": 8.0
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_tend_step1801_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "reduce": "Linf_error",
          "expected": 0.0,
          "reference": {
            "op": "faq",
            "output_idx": [
              "k"
            ],
            "args": [],
            "ranges": {
              "k": {
                "from": "lev"
              }
            },
            "expr": {
              "op": "index",
              "args": [
                {
                  "op": "const",
                  "args": [],
                  "value": [
                    0.00015551444789707602,
                    6.1659591210627696e-06,
                    2.3249299374583643e-06,
                    1.1915687991859159e-06
                  ]
                },
                "k"
              ]
            }
          },
          "tolerance": {
            "abs": 1e-09,
            "rel": 0.0
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 1
          },
          "expected": 0.00015551444789707602,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 2
          },
          "expected": 6.1659591210627696e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 3
          },
          "expected": 2.3249299374583643e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 4
          },
          "expected": 1.1915687991859159e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step2101_surface",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SURFACE CLOSURE at second-day afternoon, caught mid-transition with the ground heat flux near zero. SFLX recomputes the surface conductivity df1 from the soil moisture that SMFLX has already advanced, so yy, zz1, t1, ssoil and sheat all carry an O(dt) dependence on the replay step; the reference is therefore taken at ESM_DT = 1e-6 s, where that truncation is below 1e-10 relative. At ESM_DT = 0.01 s the same quantities are already off the dt -> 0 limit by up to 5.7e-7 relative, and WRF's own dt = 60 s values by up to 3e-3.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 299.83258056640625,
        "sfcprs": 99627.78125,
        "th2": 300.13995361328125,
        "q2": 0.015452257357537746,
        "ch": 0.009156059473752975,
        "fdown": 608.0508422851562,
        "emissi": 0.9850000143051147,
        "rch": 10.548399258128587,
        "rr": 1.579219377373885,
        "epsca": 15.372540941022166,
        "beta": 0.6636605670134896,
        "tbot": 300.0,
        "z_deep": 8.0,
        "t1_in": 302.6982116699219
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_surf_step2101_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "yy",
          "time": 0.0,
          "expected": 302.9724580895632,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "zz1",
          "time": 0.0,
          "expected": 1.4861370447176665,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "t1",
          "time": 0.0,
          "expected": 302.71167381746176,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil",
          "time": 0.0,
          "expected": -4.344206082419125,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "sheat",
          "time": 0.0,
          "expected": 27.124831162268052,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil_pre",
          "time": 0.0,
          "expected": 4.235187109228536,
          "tolerance": {
            "rel": 1e-09
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step2101_tendency",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SOIL TEMPERATURE TENDENCY at second-day afternoon, caught mid-transition with the ground heat flux near zero. SFLX advances the soil column IN PLACE with an implicit tridiagonal solve (HSTEP), so its dt = 60 s increment is NOT the instantaneous tendency: across this reference set WRF's (STC_out - STC_in)/60 differs from the dt -> 0 limit by up to 2.7e-7 K/s, which is the implicit solve and not the physics. The reference here is the replay increment at ESM_DT = 0.01 s, whose own remaining truncation is at most 5e-11 K/s (measured against a Richardson extrapolation from 0.02 and 0.01 s) and therefore an order below the assertion tolerance. A smaller step cannot be used for an INCREMENT: at ESM_DT = 1e-6 s the difference of two ~300 K temperatures loses the tendency to cancellation. The per-level assertions therefore carry an ABSOLUTE bound rather than a tight relative one: the reference's own implicit-solve truncation is up to 1.1e-10 K/s, which is a few times 1e-5 of the QUIETEST deep layer's tendency, so those layers cannot be pinned relatively at this step size while the surface layer is pinned to 1e-6 of its own amplitude. The Linf assertion above is the bound that matters.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 299.83258056640625,
        "sfcprs": 99627.78125,
        "th2": 300.13995361328125,
        "q2": 0.015452257357537746,
        "ch": 0.009156059473752975,
        "fdown": 608.0508422851562,
        "emissi": 0.9850000143051147,
        "rch": 10.548399258128587,
        "rr": 1.579219377373885,
        "epsca": 15.372540941022166,
        "beta": 0.6636605670134896,
        "tbot": 300.0,
        "z_deep": 8.0
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_tend_step2101_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "reduce": "Linf_error",
          "expected": 0.0,
          "reference": {
            "op": "faq",
            "output_idx": [
              "k"
            ],
            "args": [],
            "ranges": {
              "k": {
                "from": "lev"
              }
            },
            "expr": {
              "op": "index",
              "args": [
                {
                  "op": "const",
                  "args": [],
                  "value": [
                    -3.109837507508928e-05,
                    9.438099368708208e-06,
                    2.4904636575229233e-06,
                    1.2058251286362065e-06
                  ]
                },
                "k"
              ]
            }
          },
          "tolerance": {
            "abs": 1e-09,
            "rel": 0.0
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 1
          },
          "expected": -3.109837507508928e-05,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 2
          },
          "expected": 9.438099368708208e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 3
          },
          "expected": 2.4904636575229233e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 4
          },
          "expected": 1.2058251286362065e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step2134_surface",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SURFACE CLOSURE at second-day shallow convection. SFLX recomputes the surface conductivity df1 from the soil moisture that SMFLX has already advanced, so yy, zz1, t1, ssoil and sheat all carry an O(dt) dependence on the replay step; the reference is therefore taken at ESM_DT = 1e-6 s, where that truncation is below 1e-10 relative. At ESM_DT = 0.01 s the same quantities are already off the dt -> 0 limit by up to 5.7e-7 relative, and WRF's own dt = 60 s values by up to 3e-3.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 299.842041015625,
        "sfcprs": 99618.640625,
        "th2": 300.1493835449219,
        "q2": 0.015339565463364124,
        "ch": 0.008039824664592743,
        "fdown": 559.7423706054688,
        "emissi": 0.9850000143051147,
        "rch": 9.261910040717257,
        "rr": 1.6441605970148188,
        "epsca": 13.870400320951912,
        "beta": 0.662733492941429,
        "tbot": 300.0,
        "z_deep": 8.0,
        "t1_in": 301.7196350097656
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_surf_step2134_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "yy",
          "time": 0.0,
          "expected": 301.5507951357138,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "zz1",
          "time": 0.0,
          "expected": 1.5315128961985285,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "t1",
          "time": 0.0,
          "expected": 301.7324815343629,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil",
          "time": 0.0,
          "expected": 2.766732750097489,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "sheat",
          "time": 0.0,
          "expected": 14.661051626597851,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil_pre",
          "time": 0.0,
          "expected": -2.8707114195102568,
          "tolerance": {
            "rel": 1e-09
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step2134_tendency",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SOIL TEMPERATURE TENDENCY at second-day shallow convection. SFLX advances the soil column IN PLACE with an implicit tridiagonal solve (HSTEP), so its dt = 60 s increment is NOT the instantaneous tendency: across this reference set WRF's (STC_out - STC_in)/60 differs from the dt -> 0 limit by up to 2.7e-7 K/s, which is the implicit solve and not the physics. The reference here is the replay increment at ESM_DT = 0.01 s, whose own remaining truncation is at most 5e-11 K/s (measured against a Richardson extrapolation from 0.02 and 0.01 s) and therefore an order below the assertion tolerance. A smaller step cannot be used for an INCREMENT: at ESM_DT = 1e-6 s the difference of two ~300 K temperatures loses the tendency to cancellation. The per-level assertions therefore carry an ABSOLUTE bound rather than a tight relative one: the reference's own implicit-solve truncation is up to 1.1e-10 K/s, which is a few times 1e-5 of the QUIETEST deep layer's tendency, so those layers cannot be pinned relatively at this step size while the surface layer is pinned to 1e-6 of its own amplitude. The Linf assertion above is the bound that matters.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 299.842041015625,
        "sfcprs": 99618.640625,
        "th2": 300.1493835449219,
        "q2": 0.015339565463364124,
        "ch": 0.008039824664592743,
        "fdown": 559.7423706054688,
        "emissi": 0.9850000143051147,
        "rch": 9.261910040717257,
        "rr": 1.6441605970148188,
        "epsca": 13.870400320951912,
        "beta": 0.662733492941429,
        "tbot": 300.0,
        "z_deep": 8.0
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_tend_step2134_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "reduce": "Linf_error",
          "expected": 0.0,
          "reference": {
            "op": "faq",
            "output_idx": [
              "k"
            ],
            "args": [],
            "ranges": {
              "k": {
                "from": "lev"
              }
            },
            "expr": {
              "op": "index",
              "args": [
                {
                  "op": "const",
                  "args": [],
                  "value": [
                    -6.043405846867245e-05,
                    9.030782166519202e-06,
                    2.51148435381765e-06,
                    1.2074622190993978e-06
                  ]
                },
                "k"
              ]
            }
          },
          "tolerance": {
            "abs": 1e-09,
            "rel": 0.0
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 1
          },
          "expected": -6.043405846867245e-05,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 2
          },
          "expected": 9.030782166519202e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 3
          },
          "expected": 2.51148435381765e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 4
          },
          "expected": 1.2074622190993978e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step2400_surface",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SURFACE CLOSURE at second-day night. SFLX recomputes the surface conductivity df1 from the soil moisture that SMFLX has already advanced, so yy, zz1, t1, ssoil and sheat all carry an O(dt) dependence on the replay step; the reference is therefore taken at ESM_DT = 1e-6 s, where that truncation is below 1e-10 relative. At ESM_DT = 0.01 s the same quantities are already off the dt -> 0 limit by up to 5.7e-7 relative, and WRF's own dt = 60 s values by up to 3e-3.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 298.59136962890625,
        "sfcprs": 99621.1875,
        "th2": 298.8974304199219,
        "q2": 0.015424687415361404,
        "ch": 0.00010980122169712558,
        "fdown": 390.8979187011719,
        "emissi": 0.9850000143051147,
        "rch": 0.12701797655912284,
        "rr": 47.38326964753335,
        "epsca": 3.0628941884611938,
        "beta": 0.6521043911314539,
        "tbot": 300.0,
        "z_deep": 8.0,
        "t1_in": 295.437255859375
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_surf_step2400_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "yy",
          "time": 0.0,
          "expected": 289.7418599291049,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "zz1",
          "time": 0.0,
          "expected": 2.337198989139916,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "t1",
          "time": 0.0,
          "expected": 295.43358193519975,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil",
          "time": 0.0,
          "expected": 34.25578276019409,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "sheat",
          "time": 0.0,
          "expected": -0.4399272299941798,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil_pre",
          "time": 0.0,
          "expected": -34.226215136435634,
          "tolerance": {
            "rel": 1e-09
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step2400_tendency",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SOIL TEMPERATURE TENDENCY at second-day night. SFLX advances the soil column IN PLACE with an implicit tridiagonal solve (HSTEP), so its dt = 60 s increment is NOT the instantaneous tendency: across this reference set WRF's (STC_out - STC_in)/60 differs from the dt -> 0 limit by up to 2.7e-7 K/s, which is the implicit solve and not the physics. The reference here is the replay increment at ESM_DT = 0.01 s, whose own remaining truncation is at most 5e-11 K/s (measured against a Richardson extrapolation from 0.02 and 0.01 s) and therefore an order below the assertion tolerance. A smaller step cannot be used for an INCREMENT: at ESM_DT = 1e-6 s the difference of two ~300 K temperatures loses the tendency to cancellation. The per-level assertions therefore carry an ABSOLUTE bound rather than a tight relative one: the reference's own implicit-solve truncation is up to 1.1e-10 K/s, which is a few times 1e-5 of the QUIETEST deep layer's tendency, so those layers cannot be pinned relatively at this step size while the surface layer is pinned to 1e-6 of its own amplitude. The Linf assertion above is the bound that matters.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 298.59136962890625,
        "sfcprs": 99621.1875,
        "th2": 298.8974304199219,
        "q2": 0.015424687415361404,
        "ch": 0.00010980122169712558,
        "fdown": 390.8979187011719,
        "emissi": 0.9850000143051147,
        "rch": 0.12701797655912284,
        "rr": 47.38326964753335,
        "epsca": 3.0628941884611938,
        "beta": 0.6521043911314539,
        "tbot": 300.0,
        "z_deep": 8.0
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_tend_step2400_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "reduce": "Linf_error",
          "expected": 0.0,
          "reference": {
            "op": "faq",
            "output_idx": [
              "k"
            ],
            "args": [],
            "ranges": {
              "k": {
                "from": "lev"
              }
            },
            "expr": {
              "op": "index",
              "args": [
                {
                  "op": "const",
                  "args": [],
                  "value": [
                    -0.00017459577748013544,
                    1.5755006188555853e-06,
                    2.5797305625019362e-06,
                    1.2207181043777382e-06
                  ]
                },
                "k"
              ]
            }
          },
          "tolerance": {
            "abs": 1e-09,
            "rel": 0.0
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 1
          },
          "expected": -0.00017459577748013544,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 2
          },
          "expected": 1.5755006188555853e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 3
          },
          "expected": 2.5797305625019362e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 4
          },
          "expected": 1.2207181043777382e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step2521_surface",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SURFACE CLOSURE at third-day midnight, the dew step. SFLX recomputes the surface conductivity df1 from the soil moisture that SMFLX has already advanced, so yy, zz1, t1, ssoil and sheat all carry an O(dt) dependence on the replay step; the reference is therefore taken at ESM_DT = 1e-6 s, where that truncation is below 1e-10 relative. At ESM_DT = 0.01 s the same quantities are already off the dt -> 0 limit by up to 5.7e-7 relative, and WRF's own dt = 60 s values by up to 3e-3.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 297.6307373046875,
        "sfcprs": 99621.1953125,
        "th2": 297.93585205078125,
        "q2": 0.015445591881871223,
        "ch": 9.970904648071155e-05,
        "fdown": 387.54791259765625,
        "emissi": 0.9850000143051147,
        "rch": 0.11571418455619041,
        "rr": 51.423855790133636,
        "epsca": 0.7811832179322069,
        "beta": 0.6514180396233441,
        "tbot": 300.0,
        "z_deep": 8.0,
        "t1_in": 294.5173034667969
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_surf_step2521_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "yy",
          "time": 0.0,
          "expected": 289.10464798962715,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "zz1",
          "time": 0.0,
          "expected": 2.350777532559683,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "t1",
          "time": 0.0,
          "expected": 294.51457567134287,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil",
          "time": 0.0,
          "expected": 32.191609880895996,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "sheat",
          "time": 0.0,
          "expected": -0.39585079864305606,
          "tolerance": {
            "rel": 1e-09
          }
        },
        {
          "variable": "ssoil_pre",
          "time": 0.0,
          "expected": -32.169684515431285,
          "tolerance": {
            "rel": 1e-09
          }
        }
      ]
    },
    {
      "id": "chem170_scm_step2521_tendency",
      "description": "Reference: WRF 4.8 em_scm_xy single-column run at chem_opt = 170 with Noah (sf_surface_physics = 2, num_soil_layers = 4, silt loam, dryland cropland), data/eqwefic/chem170_scm, dumped by ctessum-claude/WRF branch earthsciml-instrumented-chem170 (commit ac6d4cc) and replayed through the real64 (-fdefault-real-8) kernels/noah_driver. Inputs are the dumped SFLX arguments (real32 values widened). df1, yy, zz1 and the pre-step ground heat flux are not Registry variables; they are published by the kernels/noah_esm_*.inc splices into a BUILD-TIME COPY of module_sf_noahlsm.F (the WRF source is untouched). The column is above freezing and snow-free throughout, so the SNKSRC phase-change sink and the SNOPAC branch are inactive here and therefore UNTESTED, not merely unused. SOIL TEMPERATURE TENDENCY at third-day midnight, the dew step. SFLX advances the soil column IN PLACE with an implicit tridiagonal solve (HSTEP), so its dt = 60 s increment is NOT the instantaneous tendency: across this reference set WRF's (STC_out - STC_in)/60 differs from the dt -> 0 limit by up to 2.7e-7 K/s, which is the implicit solve and not the physics. The reference here is the replay increment at ESM_DT = 0.01 s, whose own remaining truncation is at most 5e-11 K/s (measured against a Richardson extrapolation from 0.02 and 0.01 s) and therefore an order below the assertion tolerance. A smaller step cannot be used for an INCREMENT: at ESM_DT = 1e-6 s the difference of two ~300 K temperatures loses the tendency to cancellation. The per-level assertions therefore carry an ABSOLUTE bound rather than a tight relative one: the reference's own implicit-solve truncation is up to 1.1e-10 K/s, which is a few times 1e-5 of the QUIETEST deep layer's tendency, so those layers cannot be pinned relatively at this step size while the surface layer is pinned to 1e-6 of its own amplitude. The Linf assertion above is the bound that matters.",
      "parameter_overrides": {
        "quartz": 0.25,
        "smcmax": 0.47600001096725464,
        "shdfac": 0.5,
        "sfctmp": 297.6307373046875,
        "sfcprs": 99621.1953125,
        "th2": 297.93585205078125,
        "q2": 0.015445591881871223,
        "ch": 9.970904648071155e-05,
        "fdown": 387.54791259765625,
        "emissi": 0.9850000143051147,
        "rch": 0.11571418455619041,
        "rr": 51.423855790133636,
        "epsca": 0.7811832179322069,
        "beta": 0.6514180396233441,
        "tbot": 300.0,
        "z_deep": 8.0
      },
      "expression_template_imports": [
        {
          "ref": "./tests/heat_tend_step2521_inputs.esm"
        }
      ],
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "reduce": "Linf_error",
          "expected": 0.0,
          "reference": {
            "op": "faq",
            "output_idx": [
              "k"
            ],
            "args": [],
            "ranges": {
              "k": {
                "from": "lev"
              }
            },
            "expr": {
              "op": "index",
              "args": [
                {
                  "op": "const",
                  "args": [],
                  "value": [
                    -0.00015581836123601533,
                    -1.7483330339018721e-06,
                    2.5342956178064924e-06,
                    1.226754875460756e-06
                  ]
                },
                "k"
              ]
            }
          },
          "tolerance": {
            "abs": 1e-09,
            "rel": 0.0
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 1
          },
          "expected": -0.00015581836123601533,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 2
          },
          "expected": -1.7483330339018721e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 3
          },
          "expected": 2.5342956178064924e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        },
        {
          "variable": "dstc_dt",
          "time": 0.0,
          "coords": {
            "lev": 4
          },
          "expected": 1.226754875460756e-06,
          "tolerance": {
            "rel": 0.001,
            "abs": 5e-10
          }
        }
      ]
    }
  ]
}