Week 50 Influenza Forecast for the 2012-2013 U.S. Season

TL;DR

This paper evaluates influenza forecasts for the 2012-2013 U.S. season, comparing models with and without humidity forcing across multiple weeks to assess prediction accuracy.

Contribution

It introduces a forecast methodology for influenza that incorporates observation assimilation and compares models with and without humidity forcing.

Findings

Forecast accuracy varies across weeks and models

Humidity forcing impacts forecast performance

Multiple weeks' forecasts demonstrate model robustness

Abstract

We present results of a forecast initiated following assimilation of observations for week Week 50 (i.e. the forecast begins December 16, 2012) of the 2012-2013 influenza season for municipalities in the United States. The forecast was made on December 21, 2012. Results from forecasts initiated the three previous weeks (Weeks 47-49) are also presented. Also results from forecasts generated with an SIRS model without absolute humidity forcing (no AH) are shown.