Toolkit Home
Toolkit Banner


Phone: 1300 5 WATER (1300 5 92837)

Innovation Centre
University of Canberra
ACT 2601



NSFM (Non-parametric Seasonal Forecast Model) is a non-parametric seasonal forecasting model that forecasts continuous exceedance probabilities of streamflow (or any other hydroclimate variable). NSFM forecasts the exceedance probabilities of streamflow several months ahead by exploiting the lag relationship between streamflow and ENSO (El Nino/Southern Oscillation) and the serial correlation in streamflow.

Target user group

NSFM is designed for hydrologists, managers of storage systems, water users, consultants and researchers to facilitate probabilistic forecasting of streamflow several months ahead.


NSFM is relatively easy to use. Some basic understanding of probability is required to interpret the results from NSFM.

Example applications

NSFM can be used to forecast streamflow (or reservoir inflow or any other hydroclimate variable) several months ahead. Probabilistic streamflow forecasts can be used to provide probabilistic indication of future water allocation or water availability to make better informed risk-based decisions for farm and crop management, and to make better operational decisions on management of storage systems and water allocation for competing users.

Overview of features, advantage and benefits

NSFM uses a robust non-parametric approach to forecast streamflow several months ahead from predictor (explanatory) variables. NSFM allows the use of two predictor variables - antecedent streamflow (streamflow in previous months) and antecedent ENSO indicator (ENSO value in previous months).

The basic data required to run NSFM are historical time series of monthly streamflow and ENSO indicator.

NSFM provides streamflow forecasts for exceedance probabilities of 10% to 99%.

Features of NSFM include:

  • NSFM is relatively easy to understand and use
  • NSFM supports various time series input data formats
  • NSFM allows the user to choose the length of forecast period and the lead time of forecast
  • NSFM provides model calibration and model cross verification diagnostics (as tables and plots)
  • NSFM provides forecast skill statistics to assess the quality of forecast
  • NSFM provides exceedance probability forecasts (as tables and plots)
  • NSFM allows easy retrieval of results.

About Contact Us | Privacy Statement | Site Map | This site brought to you by