Dynamic simulation models are increasingly used in environmental and agricultural science. Here we present a method that allows models to be used to determine optimum timing of sampling for field trials. The model is used to decide when to concentrate sampling effort before the field trial begins. The model chosen to design sampling strategy should include an appropriately sensitive description of all processes that influence measurements significantly. The simulation is run, using predicted weather data, to generate the full time series before the trial begins. Every point in the simulation is considered initially to be a potential sampling point. The potential error due to not including a measurement at each point is calculated using the 'dot-to-dot' method of Smith et al. (2002) by omitting simulated values consecutively. The calculated potential error provides a measure of the priority that should be given to sampling at each point. Where the error introduced by omitting the simulated value exceeds an acceptable error, the value at the last discernible time step should be measured so that all statistically significant changes in the system can be observed. The output from the calculation is a plan of sampling times needed to capture all statistically significant events that are likely to occur over the course of the trial.
- simulation models