Section 2.1.4.5 Modelling soil structure

Modelling the soil

The soil is important since it represents the main land storage of heat and water that is available for subsequent release into the atmosphere.  The multi-layer soil model has a fairly realistic representation of the vertical density and temperature profiles of the soil which allows a good representation of its thermal properties.  

Structure of the soil

The structure of the soil is not usually uniform throughout the top layers of the earth but for a given location is normally similar within the top ~1.3m of soil.  Variations of density of the soil and fluxes of heat and moisture are more related to the texture of the soil and the water or ice content.  Vegetation roots also have an impact on the retention of water.

There is an exchange of heat, moisture and momentum between the atmosphere and underlying surface according to the vegetation type.   

The multi-layer soil model

The IFS multi-layer soil model uses four layers to represent the top ~1.3m of soil and the complex heat fluxes and interactions between them.   These are sufficient to represent correctly all timescales from one day to one year.  The soil model represents the vertical structure of the soil and the evolution of soil temperature and liquid water content in each layer.  The heat and moisture energy flux is represented by the model:

• at the top of the surface model soil layer (Layer 1):
  • thermal coupling across the soil-atmosphere interface is the balance of upward and downward energy fluxes at the soil surface.  These are computed as a weighted average over the 'tiles' representing vegetation type, and include the appropriate albedo and the effect of any water evaporation.   Upward heat fluxes cross the snow-soil interface of any overlying snow - this allows a representation of changes in permafrost.

  • an interception layer collects water from precipitation and dew fall.  Infiltration and run-off are represented depending on soil texture and standard deviation of subgrid orography.   A fraction of the water flux (rain or snow melt) is considered runoff according to the soil texture, soil water content and the standard deviation of orography (runoff can be up to 30% of rainfall in complex orography or mountainous regions).

• between all layers:
  • heat transfers upwards or downwards.  The effects of frozen water, or freezing and melting of water has an effect upon the transfer, release and absorption of heat in each layer.
  • water percolates downwards but soil water transfer is dependent on soil water potential.  Water is removed by the roots at all levels according to the root depth and transpiration of surface vegetation.  Transpiration is suppressed in frozen ground.
• at the base of the lowest model soil layer (Layer 4):
  • there is no flux of heat.
  • free drainage is assumed with no modelling of bedrock (geographical distribution of bedrock depth not used).

The fluxes are illustrated and explained in Fig2.1.4.5-1.  

The characteristics of each grid box are updated through the forecast period (e.g. model snowfall might increase the area or depth of snow cover; model rainfall might increase soil moisture rather than be removed by run-off).  The areal extent of each land surface tile type (listed above) can vary in a rapid, interactive way during the model run, as rain falls then evaporates or snow accumulates then melts, etc.  The slope and aspect of orography within each grid box (e.g. south-facing, steepness) is not taken into account and HTESSEL may consequently under- or over-estimate solar heating and runoff.

The soil type for each land grid box is defined by an offline dataset and this soil type is used for all the layers.   Each soil type has its own physical characteristics:

• soil texture (defines water retention and hydraulic conductivity in the soil):
  • Coarse.
  • Medium.
  • Medium-Fine.
  • Fine.
  • Very Fine.
  • Extra-tropical Organic.
  • Tropical Organic.
• hydraulic properties (define amount of water in the soil and availability for vegetation):
  • saturation.
  • field capacity.
  • permanent wilting point.
  • residual moisture.
  • plant available soil moisture.
• infiltration capacity:
  • ability of water to percolate downwards from one layer to another (rain or dew at the surface).
  • surface evaporative fluxes consider separately the contributions from snow cover, wet and dry vegetation, and bare soil.
  • base of lowest layer is considered as free draining.
• surface run-off:

  • when the water flux at the surface exceeds the maximum infiltration rate the excess water is considered surface runoff.

• extraction of water by plant root:
  • root depth varies according to plant species as defined by the "tile".

Fig2.1.4.5-1: Schematic of four-level soil model with land surface tiles.  Surface heat and moisture are illustrated in the schematic using the maximum selection of six different surfaces ('tiles').  The four layers of soil have differing moisture contents which vary:

• as water transfers directly (mostly downwards through gravity).
• by evapotranspiration (upwards) via the roots which penetrate to different soil depths.
• by freezing or melting.
• by runoff on the surface (assumed to be "lost" via river flow).
• by free drainage at the base of the model soil.

Table2.1.4.5-1: List of symbols for parameters shown in Fig2.1.4.5-1.

Soil temperature

Soil temperature is a forecast variable in IFS.  It needs to be initialised at each analysis cycle but there are relatively few directly measured observations.  Soil surface (skin) temperature is derived from the expected air temperature structure in the lowest 2 m together with energy fluxes (from HTESSEL) and an analysis of observed screen level (2 m) temperatures.  

Soil moisture

Soil moisture is a measure of the water content within the ground.  It is commonly expressed as a percentage of the soil water content compared with the water that the ground could hold when fully saturated.  The evaluation and prediction of soil moisture is important as this governs the efficiency of evapotranspiration from vegetation.  Thus:   

• If soil moisture content is too little (below the Permanent Wilting Point, PWP) the soil is dry.  The plant cannot extract any more water and dies.
• Higher soil moisture implies greater evapotranspiration efficiency.  This reaches a maximum at Field Capacity (CAP) when the soil is wet and contains all the water it can hold against gravity.  Not all water drains through the soil and some moisture is retained within the soil pores and cavities.  The soil is said to be at Field Capacity when large soil pores are filled with both air and water while the smaller pores are full of water.  These conditions are considered ideal for crop growth and plants flourish best. 
• As soil moisture increases beyond Field Capacity the large soil pores are increasingly filled with water.  However, the efficiency of plant evapotranspiration remains the same.
• The soil is said to be at Full Saturation (SAT) when all soil pores, large and small, are filled with water.   Flooding is possible as a result of additional precipitation. 

For each soil type and location there is a pre-defined value of the ability to hold moisture and this is used to assess the impact of model rainfall.  The HTESSEL system includes allowance for water capture by interception of precipitation and dew fall, and at the same time, there are infiltration and run-off schemes that take account of soil texture and the standard deviation of sub-grid scale orography.

Measurement of soil moisture

Soil moisture is a forecast variable in IFS.  It needs to be initialised at each analysis cycle but there are very few directly measured observations.  Soil surface (skin) moisture is derived from:

• the expected air temperature and moisture structure in the lowest 2 m together with energy fluxes (from HTESSEL) and an analysis of observed screen level (2 m) humidities.
• satellite soil moisture data from the ASCAT sensor on the MetOp satellites
• data from the Soil Moisture and Ocean Salinity satellite mission (SMOS) is used for operational monitoring (see Fig2.1.4.5-2).

The 2m temperature and humidity are diagnostic parameters of the model, so their analysis only has an indirect effect on atmosphere through the soil and snow variables. 

Fig2.1.4.5-2: Measurements from the Soil Moisture and Ocean Salinity satellite mission (SMOS)  polar orbiter satellite data.  At L-band frequency (1.4 GHz) the surface emission is strongly related to soil moisture over continental surfaces. Surface radiation at this frequency is influenced by the vegetation layer (and hence soil moisture if the vegetation type is known), but proximity of lakes etc cause difficulties with interpretation.

Soil moisture charts

Fig2.1.4.5-3: Examples of Soil Moisture at T+00 and T+192 DT 00UTC 06 March 2023.  

Example soil moisture chart VT 00UTC 06 March 2023 showing moisture in soil level 1, the surface layer.  The legend shows:

• Sandy shades: Soil moisture SM < Permanent wilting point PWP.  Living vegetation cannot be sustained. Values show soil moisture as a percentage of the permanent wilting point value.
• Yellow/Green shades: Permanent wilting point PWP < soil moisture SM < field capacity CAP.  Evapotranspiration efficiency in percent increases as soil moisture increases.
• Blue shades: Capacity CAP < Saturation SAT.  Soil moisture super-saturation. Dark blue (>90%) suggests flooding (in the model).

Note the change in soil moisture over France from ~60% of field capacity (greens) to above 60% of saturation (blues). This is largely due to rain exceeding evaporation in these areas during the forecast period.  Conversely, parts of northern Morocco, northern Algeria and northern Tunisia have become a little drier.

See the current soil moisture chart.  Select "Layer 1 2 3" from the drop down menu for the average moisture in the top metre of the earth.

Rarely in moist areas there are some soil moisture plots (except over Europe) indicating the soil is exceptionally dry.

Grid point data is plotted for Europe.  Elsewhere, for (most) other parts of the world, soil moisture is interpolated from surrounding grids points.  Field capacity, saturation, wilting point etc. depend on the soil type so can consequently be affected.  Users should check nearby soil moisture before accepting misleading soil moisture actual and forecast data.

Contrasting examples of surface and soil water budgets

Surface water budget in a typical mid-latitude agricultural landscape reacting to high rainfall in the model.

Recent periods of persistent rain over Britain over the winter of 2023/24 increased the soil moisture content in the river valleys and countryside around Reading.  Soil water storage in all model soil layers had been consistently between 120% and 150% of field capacity but generally below saturation.  Nevertheless there were areas of standing water in low-lying areas.

• On 12 Feb: There was some minor depletion of water in levels 1 and 2 due to evaporation.  
• On 13 Feb: Rain caused an increase in the rate of storage in the already high water content in soil levels 1 and 2.  There is only a small change in the already high fraction of field capacity in these levels.
• On 14 Feb: Rain also showed a small rise in the rate of storage in level 1 but the fraction of field capacity remains constant.   Water storage from the rain is partially offset by evaporation.  

Fig2.1.4.5-5: Example of surface and soil water budget.  DT12UTC 12 Feb 2024, VT12-14 Feb 20-24.  Temperate mid-latitudes.

Surface water budget in desert soil reacting to extreme rainfall in the model.

A tropical system moved over the Northern Territories, Australia depositing a period of significant rainfall.  

The coarse soil type allows the water to penetrate instead of creating runoff even after the heavy rain.  

• On 21 Jan: The fraction of field capacity in level 1 is virtually zero in the desert.  
• On 22 Jan: The very heavy rain caused significant rate of water penetrations into levels 1, 2 and 3.  The rate of water storage penetrates steadily down from level 1 through level 2 percolates down into level 3.  The total soil water storage within all layers rises through the day and approaches, but does not reach, field capacity.  However, this hides the indication of fraction of field capacity in level 1 and level 2 reaches saturation for a time and even soil level 3 approaches field capacity for a while.  All water is soaked up by the soil and the rainfall does not create an immediate increase in evaporation.

Recycling of moisture by evaporation often has an impact on maintaining cyclones over the dessert.

Fig2.1.4.5-4: Example of surface and soil water budget.  DT00UTC 21 Jan 2024, VT21-23 Jan 20-24.  Desert areas.

Surface water budget in a dry desert

The model soil moisture charts sometimes show moisture layers below the surface in dry desert areas.  There is very little ground truth so there must be some uncertainty.

Soil moisture charts consistently give an indication of water below the surface in mid-Sahara (near 23N 7E).  This should not be relied on.  However, it may well be correct as the area is around the oasis town of Tamanrasset in Algeria.  There is indication of water in layers 2 and 3 in Arabia, locally as high as 60%, but there is little data to confirm this.

 Fig2.1.4.5-6: Example of soil moisture in desert areas.  DT and VT 12UTC 07 Sep 2023.  In the area around Tamanrasset (Algeria) and much of Saudi Arabia, soil moisture charts show dry surface layers (level 1, orange) and about 20% moisture in lower layers (levels 2 and 3, green) and locally as high as 60% in Saudi Arabia.

Fig2.1.4.5-7: Afilal Oasis lies within the area where the model indicates soil with moderate moisture content in level 2 (~25% of Field Capacity) and level 3 (~40% of Field Capacity).  Level 1 has water content below the wilting point) and remains so as there is no vegetation and roots to bring water upwards from lower layers.  Nevertheless, subterranean water is locally sufficient to reach the surface at Afilal Oasis as springs. Soil moisture is a mean over a grid square.  Local details and individual oases are unlikely to be captured.  Soil moisture and soil type is not necessarily representative of an individual location.

Considerations

• Actual soil characteristics can vary widely within a grid box.  Users and forecasters should take into account the peculiarities of a location when interpreting model output.
• The assigned average soil type for a grid box is not necessarily representative of an individual location.
• Runoff can be up to 30% of rainfall in complex orography or mountainous regions.
• Recycling of moisture by evaporation from surface often has an impact on maintaining cyclones over the dessert.
• Impacts of errors associated with soil moisture. 

Additional sources of information

(Note: In older material there may be references to issues that have subsequently been addressed)

• Read more on Soil Moisture and Evapotranspiration Efficiency with example chart.