Soil Science Society of America Journal 66:1889-1896 (2002)
© 2002 Soil Science Society of America
DIVISION S-5—PEDOLOGY
Assessing Soil Moisture Regimes with Traditional and New Methods
Edoardo A. C. Costantini*,a,
Fabio Castellib,
Salvatore Raimondic and
Paolo Lorenzonid
a Istituto Sperimentale per lo Studio e la Difesa del Suolo, 50121 Firenze, Italy
b Istituto Sperimentale per il Tabacco, 37051 Bovolone VR, Italy
c Istituto di Agronomia, Università di Palermo, 90128 Palermo, Italy
d Istituto Sperimentale per lo Studio e la Difesa del Suolo, 02100 Rieti, Italy
* Corresponding author (costantini{at}issds.it)
 |
ABSTRACT
|
|---|
Soil moisture regime classes are required by U.S. soil taxonomy and other classification systems. Soil moisture regimes are based on long-term daily data of soil water content, which are as a rule estimated by means of models. International Commitee on Soil Moisture and Temperature Regimes (ICOMMOTR) has proposed classifying pedoclimate on the basis of biweekly water potential. This study was conducted to validate the use of the Erosion-Productivity Impact Calculator (EPIC) model in assessing soil water content of experimental fields placed in different European pedoclimatic conditions, to compare the pedoclimatic classification obtained with EPIC with those produced by the traditional Billaux and Newhall models, and to evaluate the results attained with the ICOMMOTR methodology. The trial was carried out over a 5-yr period in four experimental farms. The soil water content of a meadow was measured weekly or biweekly, at 0.15- and 0.75-m (or 0.45-m) depths. The EPIC model results were compared with measured data and submitted to statistical analysis of accuracy. Predicted daily soil water from EPIC was utilized to classify the soil moisture regimes following the requirements of U.S. soil taxonomy. The traditional Billaux methodology led to an overestimation of the presence of the xeric class whereas the Newhall method overrated the ustic soil moisture regime. The ICOMMOTR classification methodology was less affected by crop and by year variability than all the other methods and performed better in differentiating the soil moisture regimes of the four study sites.
Abbreviations: AE, average error • AWC, available water capacity • EPIC, Erosion-Productivity Impact Calculator • ICOMMOTR, International Commitee on Soil Moisture and Temperature Regimes • IoA, index of agreement • RMSE, root mean square error
|
INTRODUCTION
|
|---|
THE CHALLENGE OF CLASSIFYING PEDOCLIMATES, which are a direct result of the interaction between atmospheric climate and soil, has been solved in a number of ways. Some soil classification systems, such as the World Reference Base for Soil Resources (Food and Agricultural Organization [FAO] et al., 1998), only indirectly account for soil water status and, as a result, they do not classify soil moisture regimes properly. On the other hand, Soil Taxonomy (Soil Survey Staff, 1999) is a worldwide soil classification system which requires a soil moisture regime classification. The recent manual for a Georeferenced Soil Database of Europe (Finke et al., 1998) also includes a soil moisture classification. In both of these systems, the soil moisture regime is assessed in relation to the daily water content of a specific soil moisture control section. As it is very difficult to determine such values, the ICOMMOTR (1991) proposed to classify soil moisture regime based on the soil water potential at a fixed depth with biweekly measurements. This depth should be deep enough to dampen individual precipitation events, but shallow enough to show seasonal variations.
The lack of data on the actual long-term daily soil moisture requires the characterization of the pedoclimate on the basis of climatic data processed by mathematical models. The most widely used models for classifying soil moisture regime according to U.S. soil taxonomy are those of Billaux (Billaux, 1978) and Newhall (Newhall, 1972). Both systems work on the basis of a precipitation-evapotranspiration balance, with monthly mean air temperature and total rainfall as main inputs, and use the Thornthwaite and Mather evapotranspiration formula (Thornthwaite and Mather, 1957). The total monthly precipitation is assumed to fall at the end of the month in the Billaux model, while Newhall divides it in two parts, a half which falls at the middle of the month, and another half which falls during all days of the month. The Billaux method determines the water content of the moisture control section graphically, while the Newhall system represents the soil profile as a matrix of buckets with different water holding capacities. The only soil parameter taken into account is the available water capacity (AWC). The two methods provide the total amount of days when soil is moist, dry, or partially dry, but they do not calculate the soil water content at specific dates. The supplied responses should be ambiguous or conflict with experimental data, as well as with climatic and bioclimatic classifications (Calì et al., 1995, Costantini et al., 1996). Clearly there is the need to work out a more reliable methodology for assessing the soil moisture regime.
The EPIC model (Williams et al., 1989; Sharpley and Williams, 1990) might be used for soil moisture regime classification. The EPIC model estimates quantitative water content of soil layers, works on a daily time step, can be calibrated with many climatic, soil and crop parameters, and permits the simulation for time periods longer than the available meteorological data.
The objectives of this research were (i) to verify the reliability of the EPIC soil water content outputs in comparison to measured data from four Italian sites, each representative of different European pedoclimatic conditions, (ii) to compare the classification resulting from the simulated soil moisture in the control section using EPIC with those obtained using the Billaux and Newhall models, and (iii) to evaluate the results attained with the ICOMMOTR methodology.
|
MATERIALS AND METHODS
|
|---|
Soil moisture classification, according to Soil Taxonomy (Soil Survey Staff, 1999), is based on a yearly assessment of the number of days in which the soil moisture control section is moist, partially dry or completely dry. Older versions of Soil Taxonomy evaluated pedoclimate by considering "most years" or "6 or more out of 10 years". The latest edition of Soil Taxonomy replaces this approach with the concept of "normal year", in which the precipitation is plus or minus one standard deviation of the long-term (
30 yr) mean annual precipitation. A normal year also has monthly precipitation that is plus or minus one standard deviation of the long-term monthly precipitation for at least 8 mo of the year (Soil Survey Staff, 1999, p. 94). In the present study, the term "normal year" only reflected the annual precipitation because, if we also considered the monthly precipitation, none of the studied years can be defined as "normal".
The experimental sites chosen (Fig. 1
, Table 1) can be considered as representative of four different European climates (Finke et al., 1998), namely temperate-suboceanic (Bovolone), mediterranean-oceanic to mediterranean-suboceanic (Cesa), warm-temperate-subcontinental (Rieti), and mediterranean to subtropical (Sparacia). Each experimental field is equipped with a meteorological station, which has been collecting daily climatic data for periods ranging from 19 to 43 yr. Experimental fields were equipped to evaluate pedoclimate and validate estimations derived from the EPIC model, only model with daily outputs of soil water content.
Benchmark soils at each site occur on level fields, do not have shallow groundwater or well developed vertic properties, and have more than 100-mm AWC (Table 1). Benchmark soil properties are summarized in Table 2. The fields were covered by a stable dry meadow vegetation, which was maintained at a height of about 0.2 m. Soil moisture data were collected weekly or biweekly at 0.15 and 0.75 m below the surface (as recommended by ICOMMOTR, 1991) from 1994 through 1998 (1997 at Rieti). At Rieti, soil moisture depths were 0.15 and 0.45 m because of the presence of a fragipan at 1.2 m.
At Cesa and Sparacia, the gravimetric soil water content was obtained with the oven-dry method (sampling with manual auger with three replications). At Bovolone and Rieti, a TRASE instrument (Soil Moisture Equipment Corp., Goleta, CA) employing time domain reflectometry technology (TDR) was used. The measurements were calibrated with oven-dry method measurements. The moist soil bulk density, attained with the core method on replicated samples taken during profile description, was used to calculate volumetric soil water contents.
The Newhall model was run with the software produced by Van Wambeke (1986). This model initializes every year with soil moisture at field capacity. The Billaux method was implemented using a spreadsheet. This method initializes first year soil moisture at field capacity. Successive years are initialized with the soil water content that was present at the end of the previous year.
The new methodology relies on a water content simulation obtained by means of the EPIC model run on a daily step. Daily climatic inputs to the EPIC model's weather generator were minimum and maximum air temperature, relative humidity, rainfall, and radiation. The Priestley-Taylor (Priestley and Taylor, 1972) method was used to estimate potential evapotranspiration (PET), because of lack of wind velocity data at the Rieti site. The reference crop was a stable meadow, cultivated without irrigation, and with the actual management of the experimental fields. Other crop parameters were those provided by EPIC as default values, with optimal temperature for plant growth of 20°C for Sparacia and 15°C for the other stations.
Soil input data included horizon depth, texture, water content at field capacity and wilting point estimated by the Richards method (Richards, 1949), bulk density, organic C, pH, total carbonate, exchangeable bases, and cation-exchange capacity (Tables 1 and 2). Since EPIC allowed us to estimate the water content at different depths, according to the given soil horizons, we structured the input data to include a layer corresponding to the moisture control section of each studied soil. The EPIC model was initialized at field capacity and a simulation was run with the actual climatic data of the years under study (1994–1997 for Rieti and 1994–1998 for the other localities). Then, using the long-term climatic data and the weather generator, the EPIC model ran for a 50-yr period of time. We used a spreadsheet to classify, according to the U.S. soil taxonomy requirements, the soil moisture regime of the studied years and of the 50 simulated years, using the daily predicted values of soil moisture in the control section.
The weekly to biweekly soil moisture measurements taken in each experimental field were used to characterize the different environments year by year and to validate the EPIC soil water content outputs at the corresponding depth and date of the measurements. The validation was made using soil moisture data at 0.15 m, which always fell within the soil moisture control section of the studied soils (Table 1).
The statistical indices of accuracy utilized were those used in other similar experiences (Janssen and Heuberg, 1995) and included: the correlation coefficient (r), average error (AE), root mean square error (RMSE), and index of agreement (IoA) (Table 3). The AE compares the mean of predicted and observed values for the entire studied period. The RMSE measures the difference between predicted and observed values in quadratic terms, and therefore is sensitive to the extreme values. The IoA is a standardized RMSE. It can vary from 0, total disagreement, to 1, total agreement, between predicted and observed values.
Volumetric water content measured at 0.15 and 0.75 m (0.45 m at Rieti) was converted into soil matric potential using the multiple linear regression approach proposed by Rawls (1992). This method considers total sand, silt and clay, organic matter contents, bulk density, wilting point, and field capacity.
The ICOMMOTR methodology for assessing soil moisture regime is based on the determination of the geometric mean annual and seasonal water tensions (negative matric potentials) at a 0.75-m depth. This method uses a minimum of 12 values, taken at intervals of 25 to 35 d, for mean annual state, and six values taken at 14 d intervals for mean seasonal states (Fig. 2)
. Note that seasonal measurements are taken in both wet and dry seasons and that the xeric soil moisture regime of U.S. soil taxonomy is replaced by seasonal classes of ustic and aridic.
View larger version (28K):
[in this window]
[in a new window]
|
Fig. 2. Soil moisture regimes following the ICOMMOTR methodology. Mean soil water state is defined as the geometric mean of the negative matric potential at a depth of 0.75 m from the surface. MDWS, mean dry season soil water state; MWWS, mean wet season soil water state; MAWS, mean annual soil water state.
|
|
|
RESULTS AND DISCUSSION
|
|---|
Precipitation and Soil Water Content Measurements
Following the U.S. soil taxonomy criteria for a normal year, the precipitation at all sites during 1994 through 1998 (or 1994–1997 at Rieti) were normal with the following exceptions. At Bovolone, 1997 had lower than normal precipitation. In 1994, Cesa and Rieti had below-normal year precipitation. At Sparacia, 1996 and 1997 had greater than normal precipitation.
At Bovolone (Fig. 3)
, the measured soil water content at 0.75 m during the period from October 1994 through December 1998 decreased below the wilting point (0.095 m3 m-3) only during the summer of 1998 for five weekly measurements. In contrast, water content at 0.15 m was below the wilting point (0.105 m3 m-3) every year for an average of 12 weekly measurements.
View larger version (23K):
[in this window]
[in a new window]
|
Fig. 3. Volumetric soil water content at the 0.15-m (dot) and the 0.75-m (circle) depths at Bovolone during the period of 1994 through 1998. Arrows indicate summer solstices. Dashed lines indicate average field capacity (F.C.) and wilting point (W.P.) for the two depths.
|
|
At Cesa, the soil moisture profile was more homogeneous (Fig. 4) . During 1994, the dry year, the soil water content decreased below the wilting point 13 times at 0.15 m (0.184 m3 m-3) and for 12 weekly measurements at 0.75 m (0.220 m3 m-3). In the normal years, soil water content decreased below the wilting point for only two to six weekly measurements at both depths.
View larger version (28K):
[in this window]
[in a new window]
|
Fig. 4. Volumetric soil water content at the 0.15-m (dot) and the 0.75-m (circle) depths at Cesa during the period 1994–1998. Arrows indicate summer solstices. Dashed lines indicate average field capacity (F.C.) and wilting point (W.P.) for the two depths.
|
|
At Rieti, the soil was dry at 0.15 m (water content lower than 0.235 m3 m-3) for only three weekly measurements on average during summer (Fig. 5)
. At 0.45 m, the water content was lower than the wilting point (0.327 m3 m-3) for more than 10 measurements, on average, for every summer of normal years. At the same depth, during the rainy periods, the water content was frequently greater than the field capacity (0.445 m3 m-3).
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 5. Volumetric soil water content at the 0.15-m (dot) and 0.45-m (circle) depths at Rieti during the period of 1994 through 1997. Arrows indicate summer solstices. Dashed lines indicate average field capacity (F.C.) and wilting point (W.P.) for the two depths.
|
|
Finally, at Sparacia, the 0.75-m soil moisture content was often lower than at 0.15 m, reaching values lower than the wilting point (0.250 m3 m-3) in most of the measurements (Fig. 6)
. Rains were apparently unable to interrupt the water deficit of the deep horizons, apart from during the winter season when evapotranspiration decreased considerably. At 0.15 m, soil water content was less than the wilting point (0.150 m3 m-3) at least from the beginning of June to the end of September in all normal years. But, during the 2 yr with greater precipitation, the continuous dry period finished at the end of July (1997) or was limited to the month of August (1996).
View larger version (22K):
[in this window]
[in a new window]
|
Fig. 6. Volumetric soil water content at the 0.15-m (dot) and 0.75-m (circle) depths at Sparacia during the period of 1994 through 1998. Arrows indicate summer solstices. Dashed lines indicate average field capacity (F.C.) and wilting point (W.P.) for the two depths.
|
|
The Use of the EPIC Model to Estimate Soil Water Content
The reliability of the EPIC model to estimate water content in the soil moisture control section could be appraised through the examination of the results of the statistical analysis reported in Table 4.
The correlation coefficient between estimated and observed soil water contents at 0.15 m was highly significant and similar in all places, although the mean correlation coefficient for the four sites was not very high (r = 0.71). The average error indicated that, as already observed by other authors (Roloff et al., 1998), EPIC tended to slightly overestimate water content, which was overestimated at all but the Cesa station. Estimated soil moisture content at Cesa was on average slightly lower (-0.07 m3 m-3) than the measured value. The main cause of error, explaining the reported RMSE values (0.07 on average), could be related to the days where the observed soil water content was either higher than the field capacity or lower than the wilting point, and when the EPIC estimate did not go beyond those values. The IoA was quite satisfactory (0.68-0.81) with the exception of Rieti (0.37). We note that the lowest AEs and RMSE of the estimates were observed for Sparacia and Bovolone, where the highest values of the IoA were also obtained. These two experimental fields showed the driest conditions inside the soil moisture control section during the studied years. Summarizing, we can conclude that the estimated values of water content provided by EPIC were fairly close to the observed values, especially in dry and partially dry soil conditions, which are the most important ones for the pedoclimatic classification according to U.S. soil taxonomy.
Soil Moisture Regime Classification Using Different Models
On a yearly basis we matched the soil moisture classification resulting from the EPIC model outputs, with those obtained through the traditional methods of Newhall and Billaux (Table 5). The different methods gave conflicting responses. With the EPIC model, Bovolone had 3 yr that were ustic and 2 yr that were udic soil moisture regimes. At the same site, the Billaux method always indicated a xeric regime, whereas Newhall showed ustic or xeric regimes. At Cesa, EPIC gave 4 yr of udic soil moisture regime and only 1 yr of xeric. In contrast, the pedoclimate was always xeric using Billaux and mainly ustic using Newhall. The soil moisture regime at Rieti was mainly udic using EPIC and xeric only in 1995. Using Billaux, Rieti was udic in 1995 and 1996 and xeric in 1994 and 1997. Newhall, on the other hand, gave ustic in 1996 and 1997, and the other years xeric or udic. For Sparacia, EPIC placed the soil moisture regime in the xeric class in 1995, 1997, and 1998, then ustic in 1994 and udic in 1996. Billaux always indicated xeric and Newhall showed ustic for 4 yr and xeric in 1994.
View this table:
[in this window]
[in a new window]
|
Table 5. Soil moisture regime classification, following Soil Taxonomy, carried out by applying different models to observed locations for each examined year.
|
|
The dissimilar results, obtained from the different methods, suggest caution in interpretation. The Newhall and Billaux methods run on the basis of a limited number of parameters and simplified functional assumptions, which might affect their annual estimation of the pedoclimate by errors that could be offset in the long-term. Therefore, a comparison between the three systems was also made on a long-term basis. Soil water content was simulated with Newhall and Billaux for the average year and for all the years with available climatic data (19–43 yr), as well as for the average year and for the simulated period of 50 yr with EPIC (Table 6). The Billaux method, in comparison to the other two models, continued to overestimate the presence of the xeric soil moisture regime, while the Newhall system model gave results more similar to that obtained with the EPIC model.
It is remarkable that all three methods allocated the experimental sites of Cesa and Rieti in the same U.S. soil taxonomy soil moisture regime (udic with Newhall and EPIC, xeric with Billaux) although the two sites have different soil characteristics and belong to different Italian and European climates.
Soil Moisture Regime Classification Based on Water Potential
The soil water potentials in the four sites showed greater differences at lower depths than in the upper layers, considering both seasonal (Fig. 7 and 8)
and annual values (Fig. 9)
. This result might be explained by the fact that all the fields had the same cover and husbandry, and therefore crop water suction capacity in upper layers was similar in the four experimental plots.
Comparing the classification of the soil moisture regimes according to water potential, as suggested by ICOMMOTR (Table 7), with the classification obtained with the three models that predict water content in the soil moisture control section (Table 5), we can recognize several important differences. The two classification methods (ICOMMOTR vs. U.S. soil taxonomy) do not always give an equivalent pedoclimatic classification. For example, Bovolone had a typic ustic soil moisture regime for all 5 yr studied using ICOMMOTR classification, whereas U.S. soil taxonomy soil moisture regimes ranged from xeric, ustic, to udic. At Rieti and Sparacia, where there is a high contrast between the mean water potential in the upper and lower horizons, the pedoclimates of these two sites were better differentiated with the ICOMMOTR methodology than with U.S. soil taxonomy. In particular, the ICOMMOTR classifications better reflected the climate and soil differences between Cesa (udic ustic in normal years) and Rieti (seasonal ustic). Finally, utilizing the water potential at the 0.75-m depth (0.45 m at Rieti), ICOMMOTR soil moisture regime classification seems to be less affected by annual variations, particularly if we consider the years with normal precipitation.
View this table:
[in this window]
[in a new window]
|
Table 7. Geometric mean soil matric potentials and pedoclimatic classification in accordance with ICOMMOTR (1991).
|
|
|
CONCLUSIONS
|
|---|
The measured soil water content of the experimental fields during the studied years differed notably and confirmed the differences in climate and soil properties between the four study sites, which are representative of distinctive climates within Italy and Europe. The validation of the EPIC model predictions of the water content in the control section of these soils, maintained in meadow vegetation, produced fairly good outcomes and permitted an appropriate classification of soil moisture regime according to U.S. soil taxonomy requirements. The characteristics of the Newhall and Billaux methods, on the other hand, did not allow a proper validation without daily and long-term measurements, which are always very difficult to obtain. The limited soil and climatic parameters considered by these models and the simplified assumptions made for the simulations produced results that, for a specific soil and year, could be very different from those obtained by a more sophisticated model, such as EPIC.
With respect to the U.S. soil taxonomy pedoclimatic classification obtained with EPIC, Billaux appeared to exaggerate the presence of the xeric soil moisture regime, while Newhall overestimated the ustic class. The use of EPIC should be preferred to classify pedoclimate following U.S. soil taxonomy whenever a year and soil specific evaluation is desired, because it can be calibrated with the data coming from a standard soil survey and validated with a reasonable set of field measurements. The main limitation to this method is the need for a long record of daily meteorological data.
The Newhall method, as well as the Billaux method in the drier climates, could be used to classify the pedoclimate of a specific site, but only on a long-term basis. Our experience, limited to these cases, suggests a need for at least 19–43 yr of meteorological data.
Soil moisture classification based on an at-depth (0.75 or 0.45 m) seasonal and annual water potential (ICOMMOTR, 1991) was less affected by crop presence and by year-to-year variability than all the other methods, which instead considered the daily water content of the control section, as currently required by U.S. soil taxonomy. The ICOMMOTR methodology also resulted in a more accurate characterization of the different soil moisture regimes of the four sites, in spite of the fact that it envisaged biweekly instead of daily values.
The mean annual water potential at a depth of 0.75 m (or 0.45 m) easily distinguished the four soil climates, and may be a potential soil quality index or an indicator of drought and groundwater potential pollution risks.
The ICOMMOTR method furnished outcomes which differed considerably from those obtained using the three models run to fulfill present U.S. soil taxonomy criteria. Although the ICOMMOTR pedoclimatic definitions used are similar to those provided by U.S. soil taxonomy, its adoption would entail a radical change in the classification of many current soil series.
|
ACKNOWLEDGMENTS
|
|---|
Authors thank the anonymous referees and Janis L. Boettinger of the Utah State University for their useful suggestions, and Douglas A. Miller of the Pennsylvania State University for the precious help in improving the style of the manuscript. S. Raimondi was responsible for the Sparacia experimental field, P. Lorenzoni for Rieti, F. Castelli for Bovolone and E.A.C. Costantini for Cesa; E.A.C. Costantini and F. Castelli analyzed the data and wrote the paper.
|
NOTES
|
|---|
Research under the auspices of the project PANDA, subproject 1, series 1, paper no. 27. Project supported by the Agricultural Ministry of Italy.
Received for publication February 19, 2001.
|
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. (In French, with English abstract.) Editions de L'Orstom, Ser. Pedol. 3:317–318.
- Calì, A., E. Ceotto, E.A.C. Costantini, and M. Donatelli. 1995. Applicazione del modello EPIC per la classificazione del pedoclima e confronto con altri metodi e con indici climatici. (In Italian, with English abstract.) Boll. Soc. It. Scienza del Suolo 6:61–86.
- Costantini, E.A.C., A. Calì, F. Castelli, P. Lorenzoni, S. Raimondi, and L. Rustici. 1996. Esperienze di applicazione e validazione di alcuni modelli tradizionali ed innovativi per la classificazione del pedoclima. (In Italian, with English abstract.) Agricoltura Ricerca 164-165-166:7–24.
- FAO, ISRIC, and ISSS. 1998. World reference base for soil resources. World Soils Resources Rep. 84, FAO, Rome, Italy.
- Finke, P., R. Hartwich, R. Dudal, J. Ibanez, M. Jamagne, D. King, L. Montanarella, and N. Yassoglu. 1998. Georeferenced soil database for Europe. EUR 18092, Ispra, Italy.
- ICOMMOTR. 1991. Circular letter no. 2. USDA-SCS, Lincoln, NE.
- Janssen, P.H.M., and P.S.C. Heuberg. 1995. Calibration of process-oriented models. Ecol. Modell. 83:55–60.
- Munsell Color Co. 1954. Munsell Soil Color Charts. Munsell Color Co., Baltimore, MD.
- Newhall, F. 1972. Calculation of soil moisture regimes from climatic record. Rev. 4 Mimeographed, USDA-SCS, Washington, DC.
- Priestley, C.H.B., and R.J. Taylor. 1972. On the assessment of surface heat flux and evaporation using large scale parameters. Mon. Weath. Rev. 100:81–82.
- Rawls, W.J. 1992. Estimating soil hydraulic properties from soil data. p. 329–340. In M.T. Van Genuchten et al. (ed.) Proc. Int. Worksh. Indirect Methods for Estimating the Hydraulic Properties of Unsaturated Soils, Riverside, CA.
- Richards, L.A. 1949. Methods of measuring soil moisture tension. Soil Sci. 68:95–112.
- Roloff, G., R. de Jong, C.A. Campbell, R.P. Zentner, and V.M. Benson. 1998. EPIC estimates of soil water, nitrogen and carbon under semiarid temperate conditions. Can. J. Soil Sci. 78:551–562.
- Sharpley, A.N., and J.R. Williams. 1990. EPIC-Erosion/Productivity Impact Calculator. USDA Tech. Bull. 1768. NTIS, Springfield, VA.
- Soil Survey Staff. 1999. Soil taxonomy: A basic system of soil classification for making and interpreting soil surveys. 2nd ed. USDA-NRCS Agric. Handb. 436. U.S. Gov. Print. Office, Washington, DC.
- Thornthwaite, C.W., and J.R. Mather. 1957. Instructions and tables for computing potential evapotranspiration and the water balance. Climatology, X, 3.
- Van Wambeke, A. 1986. Newhall simulation model, a basic program for the IBM PC [floppy disk computer file]. Dep. of Agron., Cornell University, Ithaca, NY.
- Williams, J.R., C.A. Jones, J.R. Kiniry, and D.A. Spanel. 1989. The EPIC crop growth model. Trans. ASAE 32:497–511.
This article has been cited by other articles:
|
|
|
|
 
X. He, R. C. Izaurralde, M. B. Vanotti, J. R. Williams, and A. M. Thomson
Simulating long-term and residual effects of nitrogen fertilization on corn yields, soil carbon sequestration, and soil nitrogen dynamics.
J. Environ. Qual.,
July 1, 2006;
35(4):
1608 - 1619.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
R. Marchetti, F. Castelli, and R. Contillo
Nitrogen Requirements for Flue-Cured Tobacco
Agron. J.,
May 3, 2006;
98(3):
666 - 674.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
