Precipitation Time Series Imputation

This model uses a simple Markov Chain model to impute replacement values for missing records within a time series of daily precipitation data. The time series "Precipitation_TS" contains error data (with values of -9) and also is missing rows of date-value pairs. This model will flag both cases as missing data just before they occur and replace them with modeled precipitation that is already calibrated for the location (see input data for Markov_Precipitation).

To prepare to use this for your application, you must first generate a new time series in a spreadsheet ("DT_TS") that calculates the duration between dates on each row of the time series. This must be done prior to using in GoldSim. Once you calculate these values, put them into the DT_TS time series. For rows that are not missing, you will see the value is zero. The Markov_Precipitation model must also be prepared with input data using the Markov Precipitation PAR model, found here: https://support.goldsim.com/hc/en-us/articles/115012794028-Markov-Process-Precipitation-Simulator.

Download the model file: