SimpleCrop.jl

SimpleCrop.jl is an implementation of SimpleCrop model which was introduced as an example of how DSSAT models were structured in a modular approach. The original model was coded in Fortran and here reimplemented in a domain-specific language based on Julia using Cropbox framework.

Installation

using Pkg
Pkg.add("SimpleCrop")

Getting Started

using Cropbox
using SimpleCrop

The model is implemented as a system named Model defined in SimpleCrop module.

parameters(SimpleCrop.Model; alias = true)

Config for 1 system:

Model
weather_data	=	nothing
planting_day_of_year	=	121 d
LAI_coeff1	=	0.104 m²
LAI_coeff2	=	0.64
canopy_fraction	=	0.85
duration_of_reproductive_stage	=	300.0 d K
initial_leaf_area_index	=	0.013
maximum_leaf_number	=	12.0
initial_leaf_number	=	2.0
LAI_coeff	=	5.3
leaf_senescence_rate	=	0.03 g K⁻¹
plant_density	=	5.0 m⁻²
maximum_leaf_appearance_rate	=	0.1 d⁻¹
base_temperature	=	10.0 °C
initial_plant_dry_matter	=	0.3 g m⁻²
initial_canopy_dry_matter	=	0.045 g m⁻²
initial_root_dry_matter	=	0.255 g m⁻²
radiation_use_efficiency	=	2.1 g MJ⁻¹
row_spacing	=	60.0 cm
CH2O_conversion_efficiency	=	1.0
irrigation_data	=	nothing
runoff_curve_number	=	55
soil_profile_depth	=	145 cm
daily_drainage	=	0.1 d⁻¹
soil_water_portion_at_wilting_point	=	0.06
soil_water_portion_at_field_capacity	=	0.17
soil_water_portion_at_saturation	=	0.28
initial_soil_water_content	=	246.5 mm
soil_albedo	=	0.1
crop_albedo	=	0.2
water_table_depth_threshold	=	250 mm

As many parameters are already defined in the model, we only need to prepare time-series data for daily weather and irrigation, which are included in the package for convenience.

using CSV
using DataFrames
using Dates
using TimeZones

loaddata(f) = CSV.File(joinpath(dirname(pathof(SimpleCrop)), "../test/data", f)) |> DataFrame

config = @config (
    :Clock => :step => 1u"d",
    :Calendar => :init => ZonedDateTime(1987, 1, 1, tz"UTC"),
    :Weather => :weather_data => loaddata("weather.csv"),
    :SoilWater => :irrigation_data => loaddata("irrigation.csv"),
)

Let's run simulation with the model using configuration we just created. Stop condition for simulation is defined in a flag variable named endsim which coincides with plant maturity or the end of reproductive stage.

r = simulate(SimpleCrop.Model; config, stop = :endsim)

The output of simulation is now contained in a data frame from which we generate multiple plots. The number of leaf (N) went from initial_leaf_number (= 2) to maximum_leaf_number (= 12) as indicated in the default set of parameters.

visualize(r, :DATE, :N; ylim = (0, 15), kind = :line)

Thermal degree days (INT) started accumulating from mid-August with the onset of reproductive stage until late-October when it reaches the maturity indicated by duration_of_reproductive_stage (= 300 K d).

visualize(r, :DATE, :INT; kind = :line)

Assimilated carbon (W) was partitioned into multiple parts of the plant as shown in the plot of dry biomass.

visualize(r, :DATE, [:W, :Wc, :Wr, :Wf];
    names = ["Total", "Canopy", "Root", "Fruit"], kind = :line)

Leaf area index (LAI) reached its peak at the end of vegetative stage then began declining throughout reproductive stage.

visualize(r, :DATE, :LAI; kind = :line)

For soil water balance, here is a plot showing water runoff (ROF), infiltration (INF), and vertical drainage (DRN).

visualize(r, :DATE, [:ROF, :INF, :DRN]; kind = :line)

Soil water status has influence on potential evapotranspiration (ETp), actual soil evaporation (ESa), and actual plant transpiration (ESp).

visualize(r, :DATE, [:ETp, :ESa, :EPa]; kind = :line)

The resulting soil water content (SWC) is shown here.

visualize(r, :DATE, :SWC; ylim = (0, 400), kind = :line)

Which, in turn, determines soil water stress factor (SWFAC) in this model.

visualize(r, :DATE, [:SWFAC, :SWFAC1, :SWFAC2]; ylim = (0, 1), kind = :line)

For more information about using the framework such as simulate() and visualize() functions, please refer to the Cropbox documentation.