Simplified Land-surface Model (SLM)

The original formulation and evaluation of the Simplified Land Model (SLM) are documented in Lee, J. M., and M. Khairoutdinov (2015), A Simplified Land Model (SLM) for use in cloud-resolving models: Formulation and evaluation, Journal of Advances in Modeling Earth Systems, 7, 1368–1392, doi:10.1002/2014MS000419. Readers should consult that publication for the complete original model formulation.

This page documents only the recent changes made to improve SLM and make its behavior more comparable to other contemporary land-surface models. These changes include selected formulations adapted from NoahMP, updates inherited from gSAM-SLM, and SLM-specific corrections to the surface energy and water budgets. SLM remains a simplified land model; it is not a complete implementation of NoahMP.

The Simplified Land-surface Model (SLM) supplies lower-boundary sensible-heat, latent-heat, and momentum fluxes for land cells. Select it with erf.land_surface_model = "SLM". SLM uses the slm. input prefix.

SLM can initialize a horizontally uniform land surface from the input file or read land-surface fields from WRFInput or from metgrid (met_em) files. The soil layers are ordered from the surface downward. slm.soil_dz gives the thickness of each layer in m and must contain exactly slm.nsoil values. An idealized run must set it. A run initialized from WRFInput or metgrid files may leave slm.nsoil and slm.soil_dz unset, and SLM then takes the layers of the data: the file’s DZS for WRFInput, the Noah layers (0.1, 0.3, 0.6 and 1.0 m) for metgrid. If they are set, they must match those layers, and the run stops with a message giving the data’s layers if they do not.

Build and coupling requirements

SLM is part of the standard ERF build. It is coupled through the ERF surface layer and is normally used with a land lower boundary. Terrain-fitted coordinates can be used; SLM receives the physical-height field from ERF when terrain is active.

Basic configuration

The minimum idealized uniform-initialization configuration supplies the soil-layer thicknesses and initial soil/vegetation properties. For example:

erf.land_surface_model = "SLM"
slm.nsoil = 7
slm.soil_dz = 0.02 0.04 0.08 0.16 0.32 0.64 1.28
slm.landtype0 = 1
slm.LAI0 = 3.0
slm.clay0 = 34.0
slm.sand0 = 10.0
slm.sw0 = 0.50
slm.st0 = 300.0

For real cases using erf.init_type == WRFInput or Metgrid, the minimal configuration is

erf.land_surface_model = "SLM"

with the soil layers taken from the data, as above; see also Metgrid initialization.

Core soil and surface options

Parameter	Definition	Acceptable Values	Default
slm.nsoil	number of soil layers	Integer >= 2	7; the data’s for WRFInput and Metgrid
slm.soil_dz	SLM layer layout: thickness of each soil layer [m], from the surface downward; the number of values must equal `nsoil`. For WRFInput and Metgrid it must match the data’s layers (`DZS`, or the Noah layers).	Real > 0; exactly `nsoil` values	must be set; the data’s for WRFInput and Metgrid
slm.landtype0	initial land-use category applied over the domain	Integer land type	16
slm.LAI0	initial leaf-area index	Real >= 0	0.0
slm.clay0	clay fraction/content for each soil layer [%]; not read when initializing from `WRFInput` or `Metgrid`	One Real in [0,100] `nsoil` Reals	required unless
slm.sand0	sand fraction/content for each soil layer [%]; not read when initializing from `WRFInput` or `Metgrid`	One Real in [0,100] `nsoil` Reals	required unless
slm.sw0	initial soil wetness fraction for each layer; not read when initializing from `WRFInput` or `Metgrid`	One Real in [0,1] `nsoil` Reals	must be set
slm.st0	initial soil temperature [K] for each layer; not read when initializing from `WRFInput` or `Metgrid`	One Real or `nsoil` Reals	must be set
slm.relax_hgt	depth-dependent soil nudging weights; read when either soil nudging option is enabled	One Real in [0,1] or `nsoil` Reals	required when nudging is on
slm.soiltnudging	nudge soil temperature toward reference values	Boolean	false
slm.soilwnudging	nudge soil wetness toward reference values	Boolean	false
slm.tausoil	soil nudging time scale [s]	Real > 0	86400.0
slm.tabs_s	prescribed/fallback surface temperature [K]; WRFInput land-cell surface temperature comes from `TSK`	Real	0.0
slm.t00	constant temperature offset used in the surface temperature field	Real [K]	300.0
slm.z0_soil	bare-soil roughness length [m]	Real > 0	0.01
slm.mws_mx0	maximum puddle-water storage [mm]	Real >= 0	50.0
slm.Rc_max	maximum stomatal resistance [s/m]	Real > 0	5000.0
slm.T_opt	optimum temperature for transpiration [K]	Real	298.0
slm.zref	atmospheric reference height above the surface [m]	Real > 0	0.5

Parameter tables and external forcing

Parameter	Definition	Acceptable Values	Default
slm.use_parameter_file	use values from the parameter file; otherwise use built-in MODIS/STAS radiation values	Boolean	false
slm.parameter_file	parameter-table filename used when `use_parameter_file` is true	String	NoahmpTable.TBL
slm.veg_dataset	vegetation dataset within the parameter file	`usgs`, `modis`	`modis`
slm.soil_dataset	soil dataset within the parameter file	`stas`, `stas_ruc`	`stas`
slm.interpolate_lai	interpolate monthly LAI and SAI tables between months	Boolean	false
slm.lai	monthly LAI table, one row per month and one column per land type; requires 12 rows	Table of Reals; 12 rows	required when interpolation is on
slm.sai	monthly SAI table, one row per month and one column per land type; must match the LAI table shape	Table of Reals; 12 rows	required when interpolation is on
slm.use_param_tbl	use WRF/Noah-style vegetation parameter table	Boolean	false
slm.vegparam	vegetation parameter table used when `use_param_tbl` is true	Table of Reals	required when use_param_tbl is true
slm.rad_input_file	NetCDF radiation-forcing file containing SW/LW and cosine-zenith-angle fields; requires NetCDF support	String	empty

When use_parameter_file is false, SLM uses built-in typical MODIS/STAS values for vegetation optical properties, canopy parameters, and soil albedos. When it is true, values from the parameter file are used directly.

use_parameter_file and use_param_tbl are mutually exclusive. When use_param_tbl is true, SLM reads the slm.vegparam table for WRF/Noah-style LAI and vegetation-fraction handling. The parameter-file datasets must be usgs or modis for vegetation and stas or stas_ruc for soil. The radiation file is expected to contain the variables SWVIS, SWNIR, SWVISD, SWNIRD, COSZRS, and LWDS with dimensions matching the horizontal SLM grid.

WRFInput initialization

With erf.init_type = WRFInput, SLM reads the mapped land-surface fields from the WRF input data, including soil thickness and temperature/moisture, LAI, vegetation and soil type, skin temperature, and vegetation fractions. The uniform clay0, sand0, sw0, and st0 values are therefore not used in this mode; clay and sand are derived from the WRF soil type.

slm.nsoil and slm.soil_dz may be left unset, and SLM then takes its layers from the file’s DZS, which is also the soil thickness its physics uses. If they are set they must match DZS.

The active WRF-to-SLM field mapping is:

WRF name	SLM variable and description
DZS	`soil_thickness`: thickness of each soil layer
ZS	`node_z`: depth/elevation of soil-layer nodes
TSLB	`tsoil`: soil temperature
SMOIS	`wsoil`: soil wetness
LAI	`lai`: leaf-area index
IVGTYP	`vegtype`: vegetation type
ISLTYP	`soiltype`: soil type
TSK	`tsurf`: surface/skin temperature
VEGFRA	`veg_frac`: vegetation fraction
SHDMIN	`veg_frac_min`: yearly minimum vegetation fraction
SHDMAX	`veg_frac_max`: yearly maximum vegetation fraction

The WRFInput values are post-processed on the first SLM advance. The topmost soil layer—the layer immediately below the atmospheric surface—is indexed internally by d_khi_lsm and supplies the soil-side surface values. For each horizontal cell, SLM performs the following operations:

WRF soil moisture is divided by the SLM soil porosity. Thus the SLM wsoil value is a normalized wetness fraction, with saturation represented relative to poro_soil.

For WRF ISLTYP values 1–16, SLM assigns the following sand and clay contents to every soil layer in the cell.

ISLTYP	Soil type	Sand [%]	Clay [%]
1	Sand	92.0	3.0
2	Loamy sand	82.0	6.0
3	Sandy loam	65.0	10.0
4	Silt loam	20.0	15.0
5	Silt	8.0	12.0
6	Loam	40.0	20.0
7	Sandy clay loam	60.0	30.0
8	Clay loam	32.0	34.0
9	Silty clay loam	20.0	40.0
10	Sandy clay	52.0	42.0
11	Silty clay	6.0	47.0
12	Clay	20.0	60.0
13	Organic material	0.1	0.1
14	Water	0.1	0.1
15	Bedrock	0.1	0.1
16	Other/urban	0.1	0.1

The category constants and their WRFInput treatment are:

Constant	Value	Treatment
`isnature`	14	Urban-category cells are remapped to this nature category for the non-building SLM treatment.
`isurban`	13	The urban constant is retained; its assignment is currently commented out.
`islake`	21	Water category with WRF land type 17.

The category-specific defaults applied by this branch are:

Category	Values assigned by SLM
Water (land type 17 or `islake` = 21)	`landmask = 0`; `landtype = 0`; `vegetype = 0`; `vege_YES = 0`; soil temperature `= 273.16 K`; surface emissivity `= 0.98`; and all visible/NIR direct and diffuse albedos `= 0.06`.
Urban (land type greater than 16)	`landtype` is changed to `isnature` (14); when `use_param_tbl` is enabled, the nature-category vegetation fraction from `vegparam` is used.
Other land	WRF land type and LAI are retained; the cell remains eligible for land-mask and porosity-based soil-moisture processing.

The local surface fields are initially allocated with vegetation fraction one. WRF vegetation fractions replace that default before category handling; water-category cells subsequently receive the zero vegetation values listed above. The configured slm.landtype0 and slm.LAI0 defaults are used by the non-WRFInput path and are overwritten by WRF fields in this path.

Metgrid initialization

With erf.init_type = Metgrid, ERF builds from each level’s met_em file for the start time the same fields that WRF’s real.exe writes to a wrfinput file, and hands them to SLM, which then initializes exactly as described above for WRFInput. The fields are built as follows:

WRF name	Built from the met_em file
TSLB, SMOIS	the `ST` and `SM` soil layers (e.g. `ST000010` for 0–10 cm), interpolated linearly in depth from the layer midpoints to the Noah layer midpoints and held constant beyond the shallowest and deepest; over water, `TSK` and 1
DZS, ZS	the Noah layers: thicknesses 0.1, 0.3, 0.6, 1.0 m
IVGTYP	`LU_INDEX`
ISLTYP	over land, the dominant `SOILCTOP` category other than soil water (clay loam, 8, where `SOILCTOP` is all zero); over water, `ISOILWATER` (14)
TSK	`SKINTEMP`, replaced over water by `SST` where the file has it
VEGFRA, LAI	the monthly `GREENFRAC` and `LAI12M` interpolated linearly in time to the start date, each monthly value applying on the 15th of its month
SHDMIN, SHDMAX	the smallest and largest monthly `GREENFRAC`

Land and water are distinguished by LANDMASK. The soil layers in a met_em file depend on the source dataset (GFS gives 0–10, 10–40, 40–100 and 100–200 cm, for which the depth interpolation is an exact copy; ERA5 gives 0–7, 7–28, 28–100 and 100–289 cm), so they are found from the variable names. Soil temperature is not adjusted for the difference between the source model’s terrain height (SOILHGT) and the met_em terrain (HGT_M), as real.exe can do; in steep terrain the initial soil temperature may be off by the lapse rate times that difference.

Requirements:

The Noah layers, slm.nsoil = 4 and slm.soil_dz = 0.1 0.3 0.6 1.0, which SLM takes when the two are left unset.
The met_em file must contain LANDMASK, SKINTEMP, LU_INDEX, SOILCTOP, GREENFRAC, LAI12M, and matching ST*/SM* soil layers; SST is optional.
A land soil temperature outside 170–400 K, or a moisture outside 0–1, is taken to be missing (metgrid can leave such values where its source data had none, typically along coastlines). The temperature is replaced by TSK there, and the moisture by the mean of that layer over the other land points, and the number of points filled is printed. If more than 1% of the land points need filling, the run stops.
SLM’s tables assume the MODIS land-use categories; a warning is printed if the file’s MMINLU is not a MODIS dataset.

Every level reads its land state from its own met_em file (erf.nc_init_file_<lev>), so a finer level gets that file’s higher-resolution land use and soil, and the same coastline as its LANDMASK.

Each rank reads only the columns of its own grids. At most erf.metgrid_max_readers ranks (default 256) read the file at the same time.

An example is Exec/RegTests/MetGrid/inputs_metgrid_slm.

Diagnostics and coupling

SLM exports surface temperature, albedos, emissivity, and surface fluxes to the ERF surface-layer and radiation couplings. Its internal soil and canopy fields are available through SLM plotfile/checkpoint output. Reference-file testing additionally updates the atmospheric reference state and surface forcing from the files listed above.

Recent physics changes

1. Scope and relationship to the original SLM

The updated implementation retains the multilayer soil-temperature and soil-water framework of SLM while replacing or extending selected surface and vegetation processes. The principal changes are:

NoahMP-format soil, vegetation, and radiation parameters;
NoahMP-derived canopy and ground radiation;
NoahMP-derived undercanopy, leaf-boundary, stomatal, and soil-vapor resistances;
revised root-zone water stress and layer-specific transpiration extraction;
a prognostic canopy energy and water balance;
a massless ground-skin temperature and an explicit ground surface-energy balance; and
revised interception, pond storage, infiltration, evaporation, and saturation corrections.

Not all of these changes are direct ports from NoahMP. Canopy-temperature subcycling and rain cooling follow gSAM-SLM, while several water-budget and initialization changes are specific corrections in the ERF SLM implementation.

2. Summary of the updated time step

The ordering of the surface calculations is important because each process uses state produced by an earlier process. On each SLM advance, the updated implementation:

applies optional soil-temperature and soil-water nudging;
updates time-varying LAI and SAI;
computes canopy and ground radiation;
intercepts precipitation and updates canopy water storage;
computes surface-layer transfer coefficients;
computes undercanopy, leaf-boundary, stomatal, and soil resistances;
subcycles the canopy energy and moisture balance;
solves the massless ground-skin energy balance;
advances soil water, including evaporation and transpiration sinks; and
advances soil temperature using the skin-to-soil conductive heat flux.

The canopy and ground fluxes are subsequently combined to form the surface fluxes supplied to the atmosphere. This ordering is an SLM implementation choice and should not be interpreted as the time-integration sequence of the full NoahMP model.

3. NoahMP parameter tables

The parameter-file options and accepted datasets are listed above. The recent physics changes use additional values from the selected NoahMP blocks.

The selected NoahMP soil table supplies the following SLM properties:

NoahMP name	SLM use	Conversion
`MAXSMC`	Soil porosity or saturated moisture content	None
`REFSMC`	Volumetric moisture at field capacity	None
`WLTSMC`	Volumetric moisture at the wilting point	None
`SATPSI`	Saturated matric potential	Positive m to negative mm
`BB`	Clapp–Hornberger hydraulic exponent	None
`SATDK`	Saturated hydraulic conductivity	m/s to mm/s

The vegetation table can replace SLM defaults for minimum stomatal resistance (RS), radiation stress (RGL), leaf orientation (XL), vapor-pressure deficit sensitivity (HS), rooted-layer count (NROOT), canopy height (HVT), roughness length (Z0MVT), canopy biomass heat capacity (CBIOM), and characteristic leaf dimension (DLEAF). The radiation calculation additionally uses the leaf and stem visible/NIR reflectances and transmittances, crown radius RC, canopy-bottom height HVB, tree density DEN, and canopy wind-extinction parameter CWPVT.

After table values replace SLM defaults, all dependent soil and vegetation properties are recomputed. This prevents derived conductivities, moisture thresholds, emissivities, interception capacities, or radiation coefficients from retaining values calculated from the old defaults.

The reader validates category counts, required parameter blocks, duplicate or missing values, optical properties outside [0,1], invalid canopy geometry, and land or soil categories outside the selected table. NROOT must be an integer between zero and slm.nsoil.

4. NoahMP-derived canopy and ground radiation

The updated SLM radiation follows the NoahMP canopy-radiation formulation. It uses separate direct and diffuse fluxes in visible and near-infrared bands:

direct visible and near-infrared shortwave;
diffuse visible and near-infrared shortwave;
atmospheric downwelling longwave; and
cosine of the solar zenith angle.

These quantities normally come from the atmospheric radiation model. They may also be supplied from slm.rad_input_file, a NetCDF file containing SWVIS, SWNIR, SWVISD, SWNIRD, COSZRS, and LWDS on the SLM horizontal grid.

The canopy optical calculation retains LAI and SAI as separate quantities. Leaf and stem optical properties are weighted by their fractions of the total vegetation area index

\[VAI = LAI + SAI.\]

A modified two-stream calculation partitions incoming solar energy into reflected radiation, radiation absorbed by vegetation, and radiation absorbed by the ground. It accounts for canopy orientation, direct and diffuse beams, vegetation fraction, and canopy gaps. The shortwave terms obey the budget

\[SW_{in} = SW_{reflected} + SW_{canopy,abs} + SW_{ground,abs}.\]

Bare-soil cells bypass the canopy two-stream calculation and use the moisture-dependent ground albedo directly.

The longwave calculation treats the canopy and ground as separate emitting surfaces. With canopy emissivity \(\epsilon_v\), ground emissivity \(\epsilon_g\), canopy temperature \(T_v\), ground-skin temperature \(T_g\), and atmospheric downwelling longwave \(LW_\downarrow\), the calculation includes atmospheric absorption, canopy and ground emission, and multiple reflection between the canopy and ground. The SLM radiation budget uses positive absorbed energy, whereas the corresponding NoahMP intermediate longwave terms use positive upward energy loss; the implementation changes the sign when storing absorbed longwave energy.

The effective surface emissivity of the canopy-ground system is

\[\epsilon_{sfc} = \epsilon_v + \epsilon_g(1-\epsilon_v) + \epsilon_v(1-\epsilon_v)(1-\epsilon_g).\]

The temperature supplied as the radiometric skin temperature is diagnosed from the upward longwave flux and this effective emissivity. It is therefore not, in general, equal to canopy temperature, ground-skin temperature, or top-soil temperature.

5. Radiation assumptions and limitations

The current updated SLM radiation is limited to snow-free and ice-free land surfaces. The source contains NoahMP snow-aging and BATS snow-albedo routines, but snow cover, snow depth, snow water equivalent, and snowfall are currently set to zero in the radiation driver. Lake and ice radiation paths are likewise not active. Complete snow and ice treatment in SLM will be added in a separate pull request.

The ground-albedo calculation currently uses a fixed soil-color category from the NoahMP albedo table. The modified two-stream and BATS option identifiers are also fixed internally rather than exposed as user inputs. These restrictions should be considered when comparing SLM albedo or surface-energy budgets with a complete NoahMP simulation.

6. Vegetation structure and seasonal state

Recent changes keep LAI and SAI separate instead of accumulating SAI into LAI. Their sum is used only where a process depends on total vegetation area. This distinction is required by the NoahMP optical properties, because leaves and stems have different reflectances and transmittances.

When monthly tables are enabled, SLM linearly interpolates LAI and SAI between the current and following month. Bare-soil cells retain zero LAI and SAI. After each update, SLM recomputes the leaf-angle coefficients, vegetation emissivity, precipitation-extinction coefficient, and canopy water-storage capacity.

Vegetation emissivity is now based on the leaf-angle extinction coefficients \(\phi_1\) and \(\phi_2\):

\[\epsilon_v = 0.97\left[1- \exp\left(-\left(\phi_1+\phi_2\right)LAI\right)\right].\]

The canopy water-storage capacity includes contributions from foliage and woody area. In addition to the original LAI-dependent storage, the update uses canopy height and basal area to represent water retained on trunks. Initial surface albedos and emissivity are weighted by vegetation fraction so that partially vegetated cells do not initially behave as closed canopy.

7. Aerodynamic and surface resistances

The updated surface exchange uses a resistance network connecting the atmospheric reference level, canopy air, vegetation, and ground:

r_a: Aerodynamic resistance between the effective surface and atmospheric reference level. The SLM/gSAM-SLM surface-layer calculation retains stability-dependent similarity functions, a wind-dependent roughness adjustment, and a small minimum turbulent velocity scale.
r_d: NoahMP-derived aerodynamic resistance between the ground and canopy air. It uses canopy height, displacement height, vegetation area, friction velocity, and CWPVT to represent exponential wind attenuation through the canopy. For bare soil, r_d is set equal to r_a.
r_b: NoahMP-derived leaf boundary-layer resistance. It depends on canopy-top wind, wind extinction, and the characteristic leaf dimension DLEAF.
r_c: Leaf-level stomatal resistance. LAI is applied when leaf conductance is converted to canopy conductance, avoiding a second LAI factor in r_c.
r_soil: Soil-pore vapor resistance between the evaporating soil and the overlying air. The revised formulation is described in Section 10.

For a vegetated cell, the heat conductance from vegetation to canopy air is proportional to \(2LAI/r_b\). Vapor conductance distinguishes wet-canopy evaporation from dry-leaf transpiration and combines these paths with ground evaporation through r_d and r_soil.

8. Stomatal stress, root-zone stress, and transpiration

The stomatal resistance uses a Noah/NoahMP Jarvis-style bulk formulation:

\[r_c = \min\left(r_{c,max}, \frac{r_{c,min}} {f_{rad} f_{vpd} f_T f_{soil}}\right).\]

The four nondimensional factors represent incoming radiation, humidity deficit, canopy-air temperature, and root-zone soil moisture. The radiation factor uses incoming shortwave and RGL. The vapor-pressure-deficit factor uses the Noah-style mixing-ratio deficit,

\[f_{vpd} = \frac{1}{1 + HS\,(q_{sat}(T_{cas})-q_{cas})},\]

and the temperature factor is a quadratic response around slm.T_opt.

Root-zone stress follows the NoahMP BTR_OPTION=1 concept. NROOT gives the number of soil layers that participate. In each rooted layer, liquid water between wilting point and field capacity gives a linearly varying stress factor. Layer thickness and moisture stress determine a normalized soil_transp_frac. This fraction both contributes to the bulk soil stress and distributes the transpiration sink among soil layers.

Transpiration is limited by water available above the wilting point in every participating layer. This prevents a single time step from extracting more water than the rooted soil profile can provide.

9. Canopy energy and water balance

The canopy temperature is prognostic. The dry canopy heat capacity follows the NoahMP biomass parameterization,

\[C_{veg} = CBIOM\,\min(6,LAI+SAI)\,C_w,\]

where \(C_w\) is the volumetric heat capacity of water. Water stored on the canopy adds to the effective canopy heat capacity.

Precipitation is partitioned into intercepted water and throughfall using the canopy extinction coefficient. Intercepted rain is assumed to arrive at the atmospheric reference temperature and changes canopy temperature through mixing. This rain-cooling treatment follows gSAM-SLM. Water in excess of the canopy storage capacity drains to the ground.

Canopy sensible heat, wet-canopy evaporation or dew, and dry-leaf transpiration are calculated from the resistance network. Wet-canopy evaporation cannot remove more water than is stored on the canopy, and dry-leaf transpiration cannot remove more water than is available in the rooted soil. The canopy energy tendency is

\[C_{canopy}\frac{dT_v}{dt} = R_{n,v} - H_v - LE_v.\]

To control temperature oscillations when canopy heat capacity is small, this balance is subcycled with approximately one-second substeps. The subcycling is an SLM/gSAM-SLM numerical treatment rather than a NoahMP time-integration scheme. Canopy temperature is limited to 343 K.

10. Ground-skin energy balance and soil evaporation

The updated SLM distinguishes four surface-related temperatures:

t_ground_skin: Massless temperature at the ground-atmosphere interface.
t_canop: Prognostic vegetation temperature.
t_skin: Longwave-equivalent radiometric temperature of the combined canopy-ground system.
tsoil: Temperature of the finite-volume soil layers, including the top soil layer immediately below the massless skin.

At each time step, SLM solves the ground surface-energy balance

\[0 = SW_{ground,abs} - LW_{ground,net} - H_g - L_v E_g - G,\]

where \(H_g\) is ground sensible heat, \(E_g\) is evaporation or dew, and \(G\) is conduction from the skin into the top soil layer. A bounded Newton iteration updates t_ground_skin. The converged conductive flux then provides the upper-boundary forcing for the multilayer soil-temperature solve.

Ground evaporation uses the NoahMP option-1 soil resistance based on the dry surface-layer formulation of Sakaguchi and Zeng. The effective dry-layer thickness increases nonlinearly as normalized top-layer wetness decreases. The vapor diffusivity depends on porosity, wilting-point wetness, and the soil hydraulic exponent. The exponent controlling dry-layer growth is read from the NoahMP RSURF_EXP parameter.

Very dry soil, urban surfaces, and invalid soil hydraulic states receive a large soil resistance. Evaporation is further limited by the sum of ponded water and water available in the top soil layer after transpiration. Urban ground evaporation is multiplied by the pervious fraction. Condensation is added to surface-water storage when the soil humidity state permits dew.

11. Soil and surface water budget

The revised water budget separately tracks canopy evaporation, transpiration, pond evaporation, and direct soil evaporation. Water reaching the ground is the sum of canopy throughfall and canopy drainage.

Positive ground evaporation first removes water from surface pond storage. Any remaining demand is applied to the top soil layer. Dew adds water to pond storage. This ordering prevents pond and soil water from both supplying the same evaporation flux.

Infiltration into unfrozen soil is limited by the top-layer saturated hydraulic conductivity. It is reduced by the urban impervious fraction and set to zero when all soil layers are saturated or the top layer is frozen. Transpiration is removed from each rooted layer using soil_transp_frac.

After the implicit soil-water solve, water above saturation is redistributed toward deeper unsaturated layers while preserving total water. If the entire profile cannot accept the excess, the rejected amount is returned to surface pond storage by reducing the accepted infiltration. Pond water above slm.mws_mx0 becomes surface drainage. This correction prevents rainwater from disappearing when infiltration encounters a saturated profile.

Wetland cells are maintained at field-capacity wetness. Their surface water is implicit in the wetland treatment, so explicit pond storage is cleared after the soil-temperature update. Snow and ice hydrology are outside the scope of the current update and will be addressed together with the separate snow/ice SLM development.

Questions regarding SLM may be addressed to Jungmin Lee at Lawrence Livermore National Laboratory (lee1046@llnl.gov).