SeFo: A Package for Generating Probabilistic Forecasts from NMME Predictive Ensembles

Nir Y Krakauer

doi:10.5334/jors.112

Abstract

Long-range weather forecasts based on output from ensembles of computer simulations are attracting increasing interest. A variety of methods have been proposed to convert the ensemble outputs to calibrated probabilistic forecasts. The package presented here (SeFo, for Seasonal Forecasting) implements a number of methods for producing forecasts of monthly surface air temperature anomalies up to 9 months in advance using output from the North American Multi-Model Ensemble (NMME). The package contains modules for downloading and reading past observations and ensemble output; producing forecast probability distributions; and verifying and calibrating a user-determined subset of methods using arbitrary past periods. By changing individual modules, the package could be extended to use other model ensembles, forecast other weather variables, or apply other forecast methods. SeFo is written in the numerical computing language Octave and is available on Bitbucket under the GNU General Public License (Version 3 or later).

Highlights

Long-range weather forecasts based on output from ensembles of computer simulations are attracting increasing interest
A variety of methods have been proposed to convert the ensemble outputs to calibrated probabilistic forecasts
The workflow for applying, assessing and comparing the performance of different methods involves many steps: downloading historic observations and ensemble outputs; applying the selected methods to produce probabilistic forecasts; comparing the performance of the selected methods over a common past period using any of several possible metrics; and producing summary graphics

Summary

Introduction

Long-range weather forecasts based on output from ensembles of computer simulations are attracting increasing interest. SeFo: A Package for Generating Probabilistic Forecasts from NMME Predictive Ensembles The package presented here (SeFo, for Seasonal Forecasting) implements a number of methods for producing forecasts of monthly surface air temperature anomalies up to 9 months in advance using output from the North American Multi-Model Ensemble (NMME).

Results

Conclusion