Randomly Sample Historical Dataset by Year and Month

The purpose of this model is to demonstrate how GoldSim can randomly sample a historical dataset by a random year and change this random year on each month of the simulation. This is different than simply time shifting annually in a Time Series like this example demonstrates: Time Shifting Historic Time Series Data

Instead, this model will sample daily values from the historic record using a randomly selected year of the record and then resample the year on each month of the simulation time. This method provides a way to introduce more short-term variability of the data over time. This method does require that the time series record is reformatted to a 3-D lookup table, which is best done beforehand using a pivot table in Excel. Below is a screen capture of a pivot table I created in Excel using a daily time series record of rainfall. This table shows the years in the rows and Days of the Month in the columns and only shown for the month of January (see cell B1). You can convert this data to a 3D lookup table by copy-pasting the values from each month, one at a time.

The example model below is a GoldSim model demonstrating this functionality.