CHAPTER 5 OUTLINE
5.1 Description of Model Inputs
Soil Water Inputs
Genetic Inputs
Nitrogen Inputs
Management Inputs
Weather Inputs
5.2 Measurement of Model Inputs
5.3 Methods for Approximating Model Inputs
Soil Water Parameters
Solar Radiation
Weather Generator
Genetic Coefficients
Soil Nitrogen Inputs
5.4 Structure of Model Input Files
Overview
Structure for Model Directory Files
Structure for Model Input Data Files
5.5 Methods for Generating Model Input Files
The standard input and output files were developed after careful study of the three existing models being adapted for IBSNAT use: CERES-Maize (Jones and Kiniry 1986), CERES-Wheat, and SOYGRO V5.0 (Wilkerson et al. 1985). Each of thses models uses daily weather data, the same soil water balance model (Ritchie 1985), and similarly detailed descriptions of crop phenological development, growth, and yield (Jones and Kiniry 1986; Wilkerson et al. 1985). However, inputs and outputs were initially different and changes were required in each model as it was adapted for use in the DSSAT.
The input and output files for the CERES-Wheat model is organized into four types. a user-friendly interface for the wheat model allows users to select an experiment and then select any or all treatments from the experiment for simulation. Thus, experiments reported in the MDS can easily be simulated for comparison with observed data. Users may also elect to modify treatment conditions to evaluate "what-if" questions. For example, different weather, soil, cultivar, planting date, irrigation management, row spacing, and nitrogen fertilizer management can be changed interactively. Simulated results can then be plotted from any of the runs for comparison with real experimental treatments or for evaluation of hypothetical treatments.
The first type of file has information which identifies experimental data (EXP.DIR) and weather data (WTH.DIR). A second group of files provides input data for crop genetic coefficients, weather, soil, and management information for all of the treatments of an experiment (FILE1,FILE2,...,FILE0). The third type of file contains field-measured data extracted from the MDS for comparison with simulated results for all experimental treatments (FILEA, FILEB). The fourth file type contains output results
Determining the Genetic Coefficients
Unless the six genetic coefficients for the varieties to be used have already been determined and are available from the genetics file (see Table 2), they will need to be estimated from experimental data. If no data are available and you wish to run the model, some approximate suggestions will be made. Phasic development in the model must be approximately right before other parts of the model can be expected to work properly, because the duration of crop growth is usually proportional to productivity. The date of anthesis or a similar phenological event, such as ear fully emerged, must come from a test data set in which air temperature was measured. Because the date of anthesis for plants in a field varies considerably, recorded dates of anthesis are subject to error. Thus, having the date of 50% anthesis will help to make accurate determinations of model coefficients. Also, temperature values used in evaluating the phasic development at a site could be incorrect because of a bias in the measuring equipment. For example, the instrument may be some distance from the experimental site, or our assumption regarding the average temperature of the air, being similar to the plant growing point temperature, may be inaccurate under certain circumstances.
Table 2. Genetic Coefficients Ranges and First Try Values.
Parameter Range Notes
P1V 1 = Winter wheat These values usually work for varieties
3 = Spring wheat characterized as winter/spring type except
2 = Intermediate type in regions such as subtropical areas where vernalization of fall sown winter wheat may be marginal. For those areas and a determined intermediate type such as Ralle, a value of 2 can be chosen.
P1D 1 - 5 Use smaller values for varieties with less photoperiod sensitivity; larger values for more sensitive varieties.
P5 1 - 5 Parameters are converted within the model to values which correspond to the length of the grain filling period in oCd, which ranges from 430 - 520 oC. If this is measured, use: P5 = (X oC1 - 430) / 20 to find the correct parameter value.
G1 1 - 5 Use larger values if the variety has above average number of kernels per ear; use smaller values for below average kernel number varieties.
G2 1 - 5 Use larger values for varieties with heavier kernels and smaller values for smaller kernel weight.
G3 1 - 5 Use larger values for varieties with heavier stems; smaller values for smaller varieties.
How to Determine the Parameters
1. Run the model with first try values, which are 3 for P1D, P5, G1, G2, G3. Determine whether the variety is a winter or spring type, or choose the values of a variety which is close to the employed variety or one adapted to the same environment.
2. P1V and P1D are to be checked first, because they govern the phasic development of the crop. If the dates of anthesis and maturity, or other major events are simulated correctly, P1V and P1D have the correct values. If predicted dates are early, P1D should be increased for another trial run. Likewise, if the model date is late, the P1D value should be decreased. The values should be only those as given in the table. If the measured dates still contain an error when limiting value is tried, then some other factor, probably the value for PHINT, is contributing to the problem of simulating the anthesis date. The PHINT value can be changed if there is uncertainty about it, but it must also not go out of the range of reality. A user should check the accuracy of the temperature data entries if these procedures do not allow proper estimation of time of anthesis using the proper genetic parameters.
3. To check P5 one must measure the duration of grain fill by sampling individual kernels. Unless the variety under consideration is not known for unusually short or long grain filling duration, the value of 480 degree days should be satisfactory. Often the approximated date of maturity from visual observations is a few days later than the end of grain filling, so that if a date of maturity is recorded from phenological observations, the dates shown in the output for maturity will likely be a few days earlier.
4. The values of G2 and G3 can be changed realistically only if the duration of the entire growth cycle, final biomass and grain yield are approximately correct in the trial run and there was not much stress during the crop growth cycle. If the value of simulated GPSM (grains per square meter) is small, G3 should be increased; if it is large, G3 should be decreased. The new G3 value can be obtained by multiplying the original value used for G3 in the simulation by the ratio of the measured GPSM to the simulated GPSM.
5. For modern varieties of wheat, G2 and G3 are inversely related. G3 is proportional to final kernel weight, thus larger values should produce larger grains in simulations. However, because of the possible problems of assimilate supply to fill grains toward the end of grain filling period, it is often not possible to obtain larger grain size by increasing G3. As with other genetic coefficients, the values of G2 and G3 should not be taken out of range to try to obtain a fit with experimental data. Some other problem could be causing the error.
6. The value of G4 is related to size of individual stems. Thus larger values are related to larger stem varieties. The best strategy for obtaining a good value for G4 is to check the number of ears obtained in the first simulation with the number of ears measured. If the simulated value is too large, the stem weight parameter G4 is too small; if the simulated number of ears is too small, G4 should be decreased.
Accuracy for the G2, G3 and G4 parameters is necessary if yield components are to be successfully simulated. However, if getting ear number, grain number, and grain weight are not too critical, the yield usually will not be strongly influenced by the choice of the parameters, at least for the more modern varieties, unless the value chosen for G2 is too small.
When any of the genetic parameters are fit from experimental data, the accuracy of the model cannot be validated with the same data set. However, if only the duration constants (P1, P2, P3) are fit, the biomass and yield parts of the model can be properly tested. The most desirable strategy is first to grow a crop to obtain the proper genetic coefficients, and then test the model with another year's crop or another sowing date or location.
Description
The experiment file directory was developed to allow great flexibility in retrieving data needed to simulate various experiments from different locations and different years. This file contains the names of all input and output data files associated with a particular experiment for a crop. For each experiment in the file, three lines of information are required, and there must be a blank space before each field, except before the first field on each line, to ensure readability of this file. On the first line, the experiment identifier (8 characters) specifies the institute code, site code, year of experiment, and experiment number. After skipping one space, the next 40 characters briefly describe the experiment. The next two 12-character fields on line one identify the weather file name associated with this experiment (FILE1) and the name of the soil profile file (FILE2). On the second line of the experiment directory file, there are six 12-character fields which identify the names of files FILE4 through FILE9 for this experiment. On the third line, two 12-character fields identify the names of the validation files FILEA and FILEB, and four 7-character fields identify output files OUT1 through OUT4. If more than one experiment is to be simulated, three lines, equivalent in content to the first three described above, are appended to the EXP.DIR file for each experiment.
An example of EXP.DIR follows for a 1981 experiment in Ashland, Kansas. An EXP.DIR is needed for each experiment and the wheat identifier is appended to the front of this file to signify this is for wheat.
Data Formats
Variable Name FORTRAN Format Description
_________________________________________________________________________
Format for line 1
EXPID A8 Experiment identifier.
EXPDE 1X,A40 Experiment description.
FILE1 1X,A12 Daily weather data file name.
FILE2 1X,A12 Soil profile file name.
Format for line 2
FILE4 A12 Soil nitrogen dynamics properties file name.
FILE5 1X,A12 Soil profile initial conditions file name.
FILE6 1X,A12 Irrigation management data file name.
FILE7 1X,A12 Nitrogen fertilizer management data file name.
FILE8 1X,A12 Crop management data file name.
FILE9 1X,A12 Genetic coefficients file name.
Format for line 3
FILEA A12 Measured summary data file name.
FILEB 1X,A12 Measured seasonal data file name.
OUT1 1X,A12 File name for output file 1.
OUT2 1X,A12 File name for output file 2.
OUT3 1X,A12 File name for output file 3.
OUT4 1X,A12 File name for output file 4.
Description
This file has a list of all weather data file names along with information on location, and beginning and ending month of weather data in the file. Weather file names include institute and site code identifiers, beginning month of weather, number of months of weather records, and the year in which the data starts. For example, KSAS0517.W81 is the name of the weather data file for Kansas State University (KS), at the Ashland site (AS), beginning with May data (05), and containing 17 months of data (17) beginning in 1981 (81). Weather data in a file can start in one year and go into a second year. KSAS0517.W81 has data starting in May, 1981 and continuing for 17 months through 1982. This WTH.DIR file should contain the file names of all weather data that a user would need to simulate actual or hypothetical experiments. An example of this file is given below with reference to two weather data sets.
Data Formats
Variable Name FORTRAN Format Description
_________________________________________________________________________
WTHID A4 Weather station ID.
WTHDES 1X,A40 Weather station description.
BEGDATE A8 Beginning date in weather file.
ENDDATE 1X,A8 Ending date in weather file.
FILE1 1X,A12 Weather file name.
Example of the Weather Directory File, WTH.DIR
Description
Daily weather data must be available in FILE1 for all days of the growing season (minimum requirement), beginning with day of planting and ending at crop maturity. Ideally, the file should contain weather data collected both before planting and after crop maturity. Then, the simulation could start before planting so that soil processes would be simulated. Initial conditions for the soil should coincide with the first day of simulation. Additional weather data also allows users to select alternate planting dates or longer duration crop varieties for model sensitivity analysis. On the first line of this file, the institute and weather station site code identifiers are listed, followed by latitude. Provision is made for compatibility with other models utilized by the IBSNAT Project. Some require as extra inputs the following:
-longitude
-a conversion coefficient (PARFAC) to convert total radiation to photosynthetically active radiation (PAR)
-an indicator as to whether PAR is available in the data file (PARDAT).
For further details, refer to IBSNAT Technical Report 5.
Data Formats
Variable Name FORTRAN Format Description
_________________________________________________________________________
Format for line 1 of weather data file
INSTW A2 Code for institute ID.
STATW A2 Code for weather station ID.
XLAT 1X,F6.2 Latitude of station.
Format for all other lines of weather data
INSTW A2 Code for institute ID.
STATW A2 Code for weather station ID.
IYR 1X,I2 Year for which weather data is being read.
JUL 1X,I3 Julian date of weather record in data file.
SOLRAD 1X,F5.2 Daily total solar radiation, MJ/m2.
XTMAX 1X,F5.1 Daily value of maximum temperature, oC.
XTMIN 1X,F5.1 Daily value of minimum temperature, oC.
XRAIN 1X,F5.1 Daily total precipitation, m/day.
Example
This example has weather data from an Ashland experiment:
Description
Soil profile properties are used in the soil water, nitrogen, and root growth sections of the crop models. The first line of data in this file contains space for a soil number, a pedon number, and a soil classification name. With the exception of the soil number, this can be left blank and not affect the functioning of the model. The second line of data contains soil properties that do not vary with depth, such as surface albedo, runoff curve number, etc. Starting with line 3, one line of data is used for each layer in the profile. After the lines with properties for all layers, a line with a "-1" in the first field is required to indicate the end of data for a soil. The number of layers in this file and the thickness of each layer must be consistent with the initial conditions in FILE5. Default initial conditions for the soil are in FILE2 and will be used if FILE5 is not available. Properties for several soils are input into this file by appending data from each available soil, each with its own sequence number and pedon number. This file may contain properties for several soils with the same soil classification. Model users can use this format and manually input their own values for a soil. LL(L), DUL(L), SAT(L), can be determined from other soil properties if they have not been determined directly. Procedures for doing this are described in 5.XXXX__________.
Data Formats
Variable Name FORTRAN Format Description
_________________________________________________________________________
Format for line 1
IDUMSL 1X,I2 Number assigned to a soil type.
PEDON 1X,A12 Pedon number.
TAXON 1X,A60 Soil classification name.
Format for line 2
SALB F6.2 Bare soil albedo, no units.
U 1X,F5.2 Upper limit of stage 1 soil evaporation, mm.
SWCON 1X,F6.2 Soil water drainage constant, fraction drained per day.
CN2 1X,F6.2 SCS curve number used to calculate daily runoff.
TAV 1X,F5.1 Annual average ambient temperature, oC.
AMP 1X,F5.1 Annual amplitude in mean monthly temperature, oC.
DMOD 1X,F3.1 Zero-to-unity factor whichreduces the rate constant for mineralization of the humus pool for soils which are poor mineralizers due to chemical or physical protection of the organic matter (default = 1).
Format for line 3 through n
DLAYR (L) F6.0 Thickness of soil layer L, cm.
LL (L) 1X,F6.3 Lower limit of plant-extractable soil water for soil layer L, cm3/cm3.
DUL (L) 1X,F6.3 Drained upper limit soil water content for soil layer L, cm3/cm3.
SAT (L) 1X,F6.3 Saturated water content for soil layer L, cm3/cm3.
SW (L) 1X,F6.3 Default soil water contentfor soil layer L, cm3/cm3.
WR (L) 1X,F6.3 Weighting factor for soil depth L to determine new root growth distribution, no units.
BD (L) 1X,F5.2 Moist bulk density of soil in soil layer L, g/cm3.
OC (L) 1X,F5.2 Organic carbon concentration in soil layer L, %.
NH4 (L) 1X,F4.1 Default soil ammonium in soil layer L, mg elemental N/kg soil.
NO3 (L) 1X,F4.1 Default soil nitrate in soil layer L, mg elemental N/kg soil.
PH (L) 1X,F4.1 Default pH of soil in soil layer L in a 1:1 soil: water slurry.
_________________________________________________________________________
These are not required by the model for calculation purposes but function in the labelling of outputs and sections of the files.
Example
Description
These are treatment-specific parameters required by the wheat model that use the nitrogen dynamics component. For this file, the parameters may also depend on the treatments of the experiment. Therefore, one set of these data is needed for each treatment of an experiment and they must be recorded consecutively. This file is not needed if the model is to be run with the assumption that nitrogen is nonlimiting.
Data Formats
Variable Name FORTRAN Format Description
_________________________________________________________________________
INSTS A2 Code for institute ID.
SITES A2 Code for site ID.
YEAR A2 Year number, last two digits.
EXPTNO I2 Experiment number.
TRTNO 1X,I2 Treatment number.
STRAW 1X,F5.0 Weight of organic residue of previous crop and/or added green manure, kg/ha.
SDEP 1X,F5.0 Depth of surface residue incorporation, cm.
SCN 1X,F5.0 C:N ratio of surface residue of previous crop, kg C/kg N (default = 75.0).
ROOT 1X,F5.0 Dry weight of root residue of previous crop, kg/ha (default = 500).
Example
Description
FILE5 contains initial conditions for soil profile water and nitrogen dynamics submodels. These initial conditions specify the values of water content, ammonium, nitrate, and pH in each layer at the start of the first day of the simulation. Thus, the simulation must be started on the day for which the initial conditions are specified, even if the planting date is later. Soil profile initial conditions must be specified for a date before planting, or at the latest, on the date of planting which is input in FILE8. The thickness of each layer and the number of layers in this file must correspond exactly with those in FILE2. The first line of data in FILE5 consists of treatment number and an experiment code identifier. Then, there will be one line of data for each soil layer and a "-1" on the line immediately following the data for the last soil layer. This file will have data for each treatment of an experiment at a site, with the treatment being identified on the top line of each consecutive set.
Data Formats
Variable Name FORTRAN Format Description
_________________________________________________________________________
Format for line 1
TRTNO I2 Treatment number.
INSTS 1X,A2 Code for institute ID.
SITES A2 Code for site ID.
YEAR A2 Year number, last two digits.
EXPTNO I2 Experiment number.
Format for all other lines
DLAYR (L) F6.0 Depth of layer L, cm.
SW (L) 1X,F6.3 Soil water content of soil layer L, cm3/cm3.
NH4 (L) 1X,F4.1 Soil ammonium in soil layer L, mg elemental N/kg soil.
NO3 (L) 1X,F4.1 Soil nitrate in soil layer L, mg elemental N/kg soil.
PH (L) 1X,F4.1 pH of soil in soil layer L in a 1:1 soil: water slurry.
Example
Description
For each treatment in an experiment at a site, the amounts and dates of irrigation events are contained in FILE6. The first line of data for each treatment in the file must contain the treatment number and the experiment code identifier. Then, one line of data is required for each irrigation event. After all irrigation events have been entered for a treatment, a "-1" is entered in each field to signal the end of data for the treatment. Data for the second treatment and subsequent treatments are stacked below that of the first treatment, and data for all treatments are thus contained in this file.
Data Formats
Variable Name FORTRAN Format Description
_________________________________________________________________________
Format for line 1
TRTNO I2 Treatment number.
INSTE 1X,A2 Code for institute ID.
SITEE A2 Code for site ID.
YEAR A2 Year number, last two digits.
EXPTNO I2 Experiment number.
Format for all other lines
IDAY (J) I4 Day of year of irrigation event J.
AIRR (J) 1X,F4.0 Amount of irrigation added on IDAY (J), mm.
Example
Description
This file is organized similarly to FILE6. For each fertilizer application, one line of data with the four variables listed below must be supplied to FILE7. Since fertilizer applications may vary among treatments, data for each treatment will be sequentially listed in this file. Each set will have the treatment number and experiment code identifier on its top line of data. Following the last entry for each treatment, a "-1" in each field will be used to signal the end of that treatment's data. This file is not required if the model is being run with the assumption that nitrogen is nonlimiting.
Data Formats
Variable Name FORTRAN Format Description
_________________________________________________________________________
Format for line 1 of each treatment
TRTNO A2 Treatment number.
INSTE 1X,A2 Code for institute ID.
SITEE A2 Code for site ID.
YEAR A2 Year number, last two digits.
EXPTNO I2 Experiment number.
Formats for all fertilizer application events
FDAY (J) I4 Day of year of nitrogen fertilizer application J.
AFERT (J) 1X,F5.1 Amount of fertilizer nitrogen added on FDAY (J), kg N/ha.
DFERT (J) 1X,F5.1 Depth of incorporation of fertilizer application on FDAY (J), cm.
IFTYPE (J) 1X,I2 Code number for type of fertilizer as specified in Appendix (xxxx)......
Example
Description
FILE8 contains crop management data for each treatment averaged over all replications. Two lines of data are required for each treatment of an experiment and must be in consecutive order. On the first line, the experiment code identifier, a brief description of the treatment, the soil number for the treatment, and the cultivar used in the treatment are designated. On the second line, day to begin simulation, planting date, row spacing, and other management data for the treatment are specified. The first pair of lines in this file are for treatment 1 of the experiment, the second pair are for treatment 2, and so on.
Data Formats
Variable Name FORTRAN Format Description
_________________________________________________________________
Format for line 1
INSTE A2 Code for institute ID.
SITEE A2 Code for site ID.
YEAR A2 Year number, last two digits.
EXPTNO I2 Experiment number.
TRTNO 1X,I2 Treatment number.
TITLET 1X,A40 Title of treatment.
ISOILT 1X,I4 Soil number for this treatment as described in FILE2.
IVARTY 1X,I4 Cultivar number for this treatment as described in FILE2.
Format for line 2 of each treatment in this file
ISIM I4 Date simulation begins.
ISOW 1X,I3 Sowing date, day of the year.
PLANTS 1X,F6.2 Plant population, plants/m2.
ROWSPC * 1X,F6.3 Row spacing, m.
SDEPTH 1X,F5.2 Sowing depth, cm.
IIRR 1X,I2 Switch describing irrigation (default = 1).
1: no irrigation applied
2: irrigation applied using field schedule
3: automatically irrigated at threshold soil water
4: assume no water stress, water balance not used
ISWNIT 1X,I2 Switch to indicate if nitrogen routines are used (default = 0).
0: nitrogen subroutines are not used, assumes adequate nitrogen
1: nitrogen subroutines are used
EFFIRR 1X,F6.2 Irrigation system efficiency, fraction.
DSOIL 1X,F5.2 Irrigation management depth, m.
THETAC 1X,F6.1 Available water triggering irrigation, %.
PHINT 1X,F6.2 Phyllochron interval (day degree). Default = 95 (CERES models).
_________________________________________________________________________
* This is not required by the model but is included for compatibility with IBSNAT models.
Example
Description
This file has the name GENETICS.WH9 and contains the genetic coefficients which describe specific characteristics of a cultivar. The file has one line for each cultivar and the complete file is listed in Appendix NN. The file consists of a code number for each cultivar, the cultivar name and the list of genetic coefficients. The appropriate cultivar is selected in a model run by entering the cultivar code number in FILE8. Alternatively a different cultivar can be selected by invoking the menu options. Provision has also been made within the menus to add a new cultivar to a genetics file.
Data Formats
Variable Name FORTRAN Format Description
_________________________________________________________________________
IVAR I4 Cultivar code number.
VARTY A16 Name of cultivar.
P1V F3.1 Relative vernalization sensitivity (1,2 or 3).
P1D 5X,F4.1 Relative photoperiod sensitivity (0 to 5 scale).
P5 5X,F7.1 Relative duration of grain filling Phase (0 to 5 scale).
G1 F4.1 Relative value of conversion factor for grain number per
unit stem weight (0 to 5 scale).
G2 3X,F4.1 Relative value of maximum possible daily growth rate of a kernel (0 to 5 scale).
G3 3X,F4.1 Relative value for potential dry weight of a single stem and ear at anthesis (0 to 5 scale).
Description
For each treatment of each experiment, crop experimental data may differ. FILEA contains crop measured field data for each treatment averaged over all replications. The measured field data are needed for the standard outputs which list simulated and measured data side-by-side.
Data Formats
Variable Name FORTRAN Format Description
_________________________________________________________________
Wheat
INSTE A2 Code for institute ID.
SITEE A2 Code for site ID.
YEAR A2 Year number, last two digits.
EXPTNO I2 Experiment number.
TRTNO 1X,I2 Treatment number.
XYIELD 1X,F7.0 Actual field-measured grain yield dry weight basis, kg/ha.
XGRWT 1X,F7.4 Field-measured kernel dry weight, g/kernel.
XGPSM 1X,F6.0 Field-measured grain number, grains/m2.
XGPE 1X,F4.0 Field-measured grain number, grains/ear.
XLAI 1X,F5.2 Field-measured leaf area index at anthesis, m2/m2.
XBIOM 1X,F6.0 Field-measured aboveground dry biomass at maturity, kg/ha.
XSTRAW 1X,F6.0 Measured straw and chaff dry weight at maturity kg/ha.
IANTD 1X,I3 Field-measured anthesis date, day of year.
MATD 1X,I3 Field-measured physiological maturity date, day of year.
Start line 2
GRPCTN F6.2 Measured nitrogen concentration in grain at maturity, %.
XTOTNP 1X,F5.1 Measured crop nitrogen content at maturity, kg/ha.
XAPTNP 1X,F5.1 Measured stover nitrogen content at maturity, kg/ha.
XGNUP 1X,F5.1 Measured grain nitrogen content at maturity, kg/ha.
Description
FILEB is reserved for storage outputs of various parameters for use in graphical displays.