SOILR: A Model to Estimate Soil Moisture and Temperature Regimes


M. Donatelli1, C. Stockle2, E.A.Costantini3, R. Nelson2


1 ISA Modena, Viale C. in Guerra 134, 41100 Modena, Italy
2 BSE, L.J. Smith Hall, Washington State University, Pullman, WA 99164-6120, USA
3 ISSDS, Piazza M. D'Azeglio 30, 50121 Firenze, Italy


Introduction

Soil moisture and temperature are essential to address pedogenetic processes and to determine soil suitability for different uses. Thus, they are important parameters for Soil Taxonomy and Classification (Soil Survey Staff, 1975). A new classification criteria was proposed by the International Committee on Soil Moisture and Temperature Regimes (ICOMMOTR, 1991) with the intent of classifying soil properties instead of location climate, as it is done by using the "control-section" approach (Soil Survey Staff, 1975).
Attempts to classify soils on the basis on their water content and temperature following the Soil Taxonomy criteria have been made using static and empirical water budgeting models that account for weather variables through the use of monthly values (Thornthwaite and Mather, 1957; Newhall, 1972; Billaux, 1978). Such models fail at times to properly estimate soil moisture and temperature due to the limited amount of information they use to describe the system under evaluation. They do not account dynamically for the key processes which determine water and heat budgets in the soil, including plant growth and water uptake. Finally, they use descriptions of climate which do not include the stochastic component of weather variability.
Dynamic models developed to estimate water budgets for cropping systems have been extensively tested in various conditions. Such models describe key processes of the system soil-plant as disturbed by weather. Attempts to use one such models, EPIC (Sharpley and Williams, 1990), to classify soils on the basis of their water regime provided a more reliable estimate of soil moisture content (Calì et al., 1995). However, subroutines dedicated to soil classification on the basis of soil moisture and temperature were not explicitly available. Moreover, in EPIC the submodel for crop growth can overestimate biomass production in water stressful conditions (Cabelguenne et al., 1988; Ceotto et al., 1993), hence overestimating and subsequently underestimating water uptake during the growing season. Finally, the use of EPIC was difficult because of the lack of a dedicated user interface. To overcome such problems while taking advantage of the ability of dynamic simulation models of estimating water and temperature budgets, a new model and the corresponding software to classify soils is presented in this paper.



Model description

SOILR was developed from the model CropSyst (Stockle et al., 1994; Stockle and Donatelli, 1997; Stockle and Nelson, 1998) to evaluate soil moisture and temperature regimes for a reference grass. The model requires daily weather inputs such us global solar radiation, minimum and maximum air temperature, precipitation, and, optionally, wind and minimum and maximum air relative humidity. Daily weather data are generated by the weather generator ClimGen (Ndlovu et al., 1994; Stockle and Nelson, 1998b), including a new model to estimate daily solar radiation (Donatelli and Campbell, 1998). Daily data are generated from location parameters, which are estimated from long-term weather data of the site.SOILR runs using a daily time step.
The key processes implemented in SOILR are shown in Fig. 1.


Figure 1. Driving forces, daily time step processes and state variables of interest in SOILR.


Potential evapotranspiration can be estimated using either the Penman-Monteith (Monteith, 1965) or the Priestley-Taylor (1972) methods. In the latter method, the a constant is modeled as proposed by Steiner et al. (1991), accounting for vapor pressure deficit. Potential evapotranspiration is partitioned between potential evaporation and potential transpiration as a function of grass leaf area index. The function is the same used to estimate solar radiation interception by the grass canopy. Potential biomass increase is estimated by two functions:



where:
GTR = water-limited growth (kg m-2 d-1)
KBT = biomass-transpiration Coefficient (kPa)
T = transpiration (kg m-2 d-1)
VPD = Vapor Pressure Deficit (kPa)


and

where:
GIPAR = radiation-limited growth (kg m-2 d-1)
e = radiation conversion efficiency (kg MJ-1)
fint = fraction of radiation intercepted
PAR = photosinthetically active radiation (MJ m-2 d-1)


Potential biomass growth is then estimated as:



While GTR provides more reliable estimates of growth in arid conditions (VPD>1), GIPAR provides a better estimate in conditions of VPD<1. Actual water uptake is set as the minimum between the demand and the availability of water in the soil profile explored by roots. When the water supply is smaller than the demand, water stress accelerates leaf area senescence, thus reducing both the transpiring and the photosynthetic capacity of the grass canopy. The grass growth is regulated by a set of parameters; three parameter files are provided for cool, temperate, and hot weather conditions.
Evaporation occurs conventionally in the upper 0.05 m layer of the soil. Water evaporates at the potential rate until the volumetric soil moisture content reaches the permanent wilting point (qPWP ; yPWP = -1500 J kg-1); then it decreases linearly to zero at a water content equal to 1/3 qPWP.
The water input in the system is given by precipitation. Water can be intercepted by the grass and grass residues on the soil surface. Water not intercepted may be subject to run off, which is calculated using the curve number approach of the SCS/USDA (1972). After subtracting interception and runoff, the remainder of the water infiltrates and it is redistributed in the soil profile using either a capacity approach or a numerical solution of Richard's equation for water flux in the soil. The model currently does not account for preferential flux in soil cracks. At each soil moisture content and for each layer, soil moisture potential is calculated using the method proposed by Campbell (1985).
Surface soil temperature is modeled as a function of air temperature and grass biomass according to Parton (1984). The temperature in the soil profile is modeled by calculating the damping depth as a function of soil moisture content (Kenneth et al., 1994). Soil temperature exponentially varies as a funtion of depth reaching the costant value at the damping depth, which is estimated as the yearly average air temperature. Soil mineralogic characteristics do not affect the estimate of heat flux given that in general the differences in heat conduction due to mineralogy are much smaller compared to the ones due to moisture content.
Further details on the equations used in the model can be found in the CropSyst documentation (Stockle and Nelson, 1998); the model documentation of SOILR is currently under development.
Soil moisture and temperature regimes are currently classified using the criteria presented by ICOMMOTR (1991), and described as follows.
The soil temperature is selected by computing the following variables:
  • Mean Annual Soil Temperature (MAST) is defined as the normal mean annual soil temperature determined at a depth of 0.5 m from soil surface. The mean soil temperature for an individual year is taken to be the arithmetic average of soil temperature. This temperature is used in table 1 to do the main classification of the soil temperature regime.
  • Mean Summer Soil Temperature (MSST) is defined as the normal mean soil temperature of 90-day period of the greatest soil temperature values determined at a depth of 0.5 m from soil surface.
  • Mean Winter Soil Temperature (MWST) is defined as the normal mean soil temperature of 90-day period of the lowest soil temperature values determined at a depth of 0.5 m from soil surface.
The three variables are used to further detail the soil temperature regime classification.

Temperature regime at
0.5 m depth
Yearly Tempearature
°C
Pergelic
-5<
Gelic
-5 - 0
Cryic
0 - 10 (6)
Frigid
0 - 8 (10)
Mesic
8 - 15
Thermic
15 - 22
Hypertermic
22 - 28
Megathermic
>28

The soil moisture regime is based on the computation, for each year, of three variables, which are:
  • Mean Annual Soil Water State (MAWS), which is defind as the normal geometric mean (logaritmic average) annual soil water tension (negative matric potential) determined at a 0.75 m depth from the soil surface
  • Mean Dry Season Soil Water State (MDWS), which is defind as the normal geometric mean (logaritmic average) dry season soil water tension (negative matric potential) determined at a 0.75 m depth from the soil surface. The dry season is defined as the 90 days period of greatest soil water tension. The dates of this 90 days period may vary from year to year
  • Mean Dry Season Soil Water State (MDWS), which is defind as the normal geometric mean (logaritmic average) dry season soil water tension (negative matric potential) determined at a 0.75 m depth from the soil surface. The dry season is defined as the 90 days period of greatest soil water tension. The dates of this 90 days period may vary from year to year
The yearly values of the variables described are used calculate the site value by classifying soil moisture regime according to figure 2. A soil is classified when at least in 6/10 of the years it falls into one group.



Figure 2. Soil classification based on soil water potential according to ICOMMOTR (1991), redrawn.



Software description

The software runs under Windows 95/NT. The diagram of figure 3 shows the structure of the software suite SOILR. Two main programs are installed, SOILR and ClimGen. The model runs 50 years of daily data without reinitializing state variables, allowing the classification of soils based on either water content or potential, and soil temperature. Location inputs are provided using two files: the soil file and the location file.


Figure 3. The SOILR software structure


The description of the soil profile in the soil file is entered by layers, corresponding to soil horizons. Data required are layer thickness, bulk density and volumetric soil moisture content at field capacity (qFC; yFC -10/-33 J kg-1) and at permanent wilting point (qPWP; yPWP -1500 J kg-1). When bulk density and/or characteristic water content were not available, the accessory software SOILPAR (Donatelli et al., 1996; Donatelli et al., 1997) can be used to estimate such parameters. Other data entered in the soil file allow the automatic selection of the proper curve number for the run off submodel.
The location is described in a dedicated file where site specific inputs are stored. Such data are essentially the location latitude and inputs to select and characterize the evapotranspiration methods.
A third file is the simulation control, which defines the soil, the location, and the grass parameter files used to simulate the sequence of 50 years.
The model outputs are the run-time graphic, which allows monitoring the key state variables for grass growth, soil moisture and soil temperature, and the text files, which contain yearly values of variables such us biomass produced, water percolation, runoff, cumulated evapotranspiration, water stress index, and the values of the variables previously described and used to classify both soil temperature and moisture regimes. Another summary files contains the classification of the soil based on the 50 years simulation.


Conclusions

The model SOILR allows classifying soil moisture and temperature regimes based on the simulation of key processes which describe the system soil-grass as affected by weather. Such a sound basis, and the information used to classify a soil, give more confidence in using the model in different environments.
The daily step simulation of a reference grass offers a better insight into the soil moisture and temperature regimes of different soils, alloing a better understanding of the system soil-grass as influenced by weather. The model is currently being calibrated for a range of conditions, showing a good sensitivity to changing conditions. The user frendly interface allows users with an average PC knowledge to easily use the model for soil classification.


References

Billaux P. (1978). Estimation du "regime hydrique" des sols au moyen des données climatiques. La méthode graphique: son utilisation dans le cadre de la Taxonomie Americaine des sols. ORSTOM, ser. Pedol. Vol. XVI, 3.
Cabelguenne M., Jones C.A., Marty J.R. and Quinones H. (1988). Contribution à l' étude des rotations culturales: tentative d'utilisation d'un modèle. Agronomie, 8: 549-556.
Campbell G.S. (1985). Soil Physics with BASIC: Transport Models for Soil-Plant Systems. Elsevier, Amsterdam, Netherlands. 150pp.
Ceotto E., Donatelli M., Castelli F., Quaranta F., Rinaldi M. , Spallacci P. (1993). Il modello EPIC nella simulazione di sistemi colturali attuati in ambienti italiani: II Validazione per dati produttivi. Agricoltura Ricerca, XV, n. 151/152: 209-228.
Cali’ A., Ceotto E. , Costantini E.A.C., Donatelli M. (1996). Applicazione del modello EPIC per la classificazione del pedoclima e confronto con altri metodi e con indici climatici. Boll. Soc. It. Scienza del suolo, 6, p. 61-86.
Costantini E.A.C., Cali’ A., Castelli F., Lorenzoni P., Raimondi S., Rustici L.(1996). Esperienze di applicazione e validazione di alcuni modelli tradizionali ed innovativi per la classificazione del pedoclima. Agricoltura Ricerca, XVIII, 164-165-166 P.7-24.
Donatelli M., Acutis M., Laruccia N. (1996). Evaluation of methods to estimate soil water content at field capacity and wilting point. Proc. 4th European Society of Agronomy Congress, Veldhoven, The Netherlands, 86-87
Donatelli M., Acutis M., Francaviglia R.. (1997). SOILPAR, un software per la stima e la validazione dei parametri idrologici del suolo. In: Proceedings of the annual congress of the Italian Soil Science Society, Roma 2-5 giugno 1997.
Donatelli M., Campbell G. (1998). A model to estimate global solar radiation using daily air temperature. In: Proceedings of the 5th ESA Congress, Nitra, Slovak Republic, June 1998.
Kenneth N. P., Williams J.R. (1994). Predicting Daily Mean Soil Temperature in the EPIC Simulation Model. Agron. J. 86:1006-1011.
ICOMMOTR, International Committee on Soil Moisture and Temperature Regimes. (1991). Circular Letter NO.2. National Soil Survey Center, Soil Conservation Service, USDA, Lincon, Nebraska, USA
Monteith, J.L. (1965) Evaporation and Environment. 19th Symposia of the Society for Experimental Biology, University Press, Cambridge, 19:205-234.
Ndlovu L.S. (1994). Weather generation and its use in estimating evapotranspiration. Ph.D. dissertation, Biological Systems Engineering Department, Washington State University, Pullman, WA. 143 pp.
Newhall F. (1972). Calculation of soil moisture regimes from climatic record. Rev. 4 mimeographed, soil conservation service, USDA, Washington DC.
Parton, W.J. (1984). Predicting soil temperature in a shortgrass steppe. Soil Sci. 138:93-101.
Priestley, C.H.B. , Taylor, R.J. (1972). On the assessment of surface heat flux and evaporation using large scale parameters. Mon. Weath. Rev. 100: 81-92.
Sharpley, A.N. , and Williams, J.R. (1990). EPIC, Erosion Productivity Impact Calculator: 1. Model Documentation. U.S. Dept. of Agriculture Technical bulletin No. 1768. 235 pp. U.S. Dept. of Agriculture, Soil Conservation Service. 1972. National Engineering Handbook, Hydrology Section4, Chapters 4 - 10.
SCS/USDA (Soil Conservation Service, U.S. Department of Agriculture. (1972). National Engineering Handbook, Hydrology Section 4, Chapters 4-10.
Soil Survey Staff. (1975). Soil Taxonomy: A basic system for making and interpreting soil surveys. USDA Handbook 436, 754 pp. Washington DC.
Steiner J.L., Howell T.A., Schneider A.D. (1991). Lysimetric evaluation of daily potential evapotranspiration models for grain sorghum. Agron. J. 83:240-247.
Stockle C.O., Donatelli M. (1997). The CropSyst model: a brief description. In: Plentinger, M.C. & F.W.T. Penning de Vries (Ed.). Rotation models for ecological farming. CAMASE/PE workshop report. Quantitative Approaches in Systems Analysis No. X. AB-DLO, P.O.Box 14, 6700 AA Wageningen, The Netherlands p 35-43
Stockle C. O., S. Martin and G. S. Campbell. (1994). CropSyst, a cropping systems model: water/nitrogen budgets and crop yield. Agricultural Systems 46:335-359.
Stockle C.O. , Nelson R. (1998). CropSyst, user manual. Biological Systems Engineering Dept., Washington State University, Pullman, WA.
Stockle C.O. , Nelson R. (1998). ClimGen, user manual. Biological Systems Engineering Dept., Washington State University, Pullman, WA.
Thornthwaite C.W., Mather J.R. (1957). Instructions and tables for computing potential evapotranspiration and the water balance. Climatology, x, 3. Centerton n.y. USA.


top