Proceedings of the
European Safety and Reliability Conference (ESREL2026)
14 – 19 June 2026, Braga, Portugal

Non-stationary analysis of the date of ice breaking in the ice classic of Nenana river (Alaska)

Patricia Mares-Nasarre

Department of Hydraulic Engineering, Delft University of Technology, the Netherlands.

p.maresnasarre@tudelft.nl

Stuart G. Pearson

Department of Hydraulic Engineering, Delft University of Technology, the Netherlands.

s.g.pearson@tudelft.nl

Kieran B.J. Dunne

Department of Hydraulic Engineering, Delft University of Technology, the Netherlands.

k.b.j.dunne@tudelft.nl

Robert Lanzafame

GEI Consultants, 180 Grand Ave Suite 950, Oakland, California USA, 94612.

rlanzafame@geiconsultants.com

ABSTRACT

The Nenana Ice Classic is an annual event in Nenana, Alaska, which functions as a community fundraiser where participants wager on the precise date and time (hour and minute) of the spring breakup of the Tanana River ice at Nenana. Beyond its cultural and fundraising role, ice breakup in northern rivers is closely related with the formation of ice jams, which can rapidly dam river flow and lead to severe flooding in adjacent communities. Historical records of the ice breakup of the Tanana River ice are available since 1917. Here, the time of ice breakup relative to 1st January every year is analyzed. A significant linear decreasing trend is found while the residuals are found to be homoscedastic. A non-stationary distribution using time as a covariate is used to model uncertainty in the data and estimate probabilities to find the breakup date before the wagering day date for different time scenarios. In addition, several climatic covariates are analyzed as possible contributors to uncertainty in the breakup date, including air temperature, the discharges in the river or the cloud coverage, amongst others.

Keywords: non-stationary, ice breakup, Nenana Ice Classic, covariates.



Download PDF