Parameters and forcings (input series)
A parameter is a user-provided value that is held constant through each model run, and can be used to e.g. set up the model for a specific geographical location, tune the model to fit observed data, or to run scenarios. In Mobius2, parameters are always put inside parameter groups.
A forcing (input series) is another value that is not computed by the model, but unlike a parameter its value varies over time. In Mobius2 any property state variable without a math expression is assumed to be an input series.
model("The first model") {
soil : compartment("Soil")
temp : property("Temperature")
growth_rate : property("Growth rate")
module("A module", version(0, 0, 1)) {
par_group("Growing parameters") {
# A par_real is a parameter taking values in the real number line.
# The third argument is the default value. This is not always the
# value that is used when the model is run, instead the value from
# the data_set is used
br : par_real("Base growth rate", [k g, day-1], 1)
# This parameter has a dimensionless unit:
td : par_real("Growth rate temperature dependence", [], 2)
}
# This variable does not have a math expression, and since this
# is the only module, it can't be declared with one anywhere else either.
# It must therefore be a forcing.
var(soil.temp, [deg_c], "Soil temperature")
# We can refer to other values in the math expression for this variable:
var(soil.growth_rate, [k g, day-1], "Plant growth rate") {
br * td^((soil.temp - 20[deg_c])/10[deg_c])
}
}
}
Parameters and series are provided in the data_set. If a parameter is not provided, it is given its default value when it is loaded. If you then save the data_set again (from MobiView2), the parameter will be in it. Note how the structure of the data_set mimics the structure of the model.
data_set {
# The "System" par_group always exists, and is not explicitly
# declared in the model.
par_group("System") {
par_datetime("Start date")
[ 2000-01-01 ]
par_datetime("End date")
[ 2000-12-31 ]
}
# The module - par_group - parameter nested scoping matches
# how these are declared in the model.
module("A module", version(0, 0, 1)) {
par_group("Growing parameters") {
par_real("Base growth rate")
[ 4 ]
par_real("Growth rate temperature dependence")
[ 2.1 ]
}
}
# The series() declaration instructs the model to load series
# data from the provided file. You can also use .xlsx files.
series("data.csv")
}
The data.csv file (see the data format documentation):
"Soil temperature" [spline_interpolate]
2000-01-01 -5
2000-02-01 -4
2000-03-01 -1
2000-04-01 2
2000-05-01 7
2000-06-01 11
2000-07-01 14
2000-08-01 12
2000-09-01 9
2000-10-01 4
2000-11-01 1
2000-12-01 -3
2001-01-01 -5
Here Mobius2 is instructed to interpolate for missing values, but you could just provide a value per time step if you have that.
Currently we are running with a daily time step. That is configurable, but we will worry about that later.
Exercises
-Run the model in MobiView2, plot the “Plant growth rate” series, and see what happens if you edit a parameter value and run the model again.
- Try also to experiment with changing the equations or adding more complexity to the model.
- Rembember that if you add a new parameter, or parameter group, you don’t need to manually add these to the
data_set. Mobius2 will auto-generate parameter values for these for you.
- Rembember that if you add a new parameter, or parameter group, you don’t need to manually add these to the