River Ice Explorer App

Tomás E. Tecce
January 31, 2016

Introduction

Available data suggests that regions in the far north of the Earth are warming more quickly than other parts of the world. We can explore this situation using ice breakup data available for some rivers in Alaska, USA.

  • Local citizens have been recording ice breakup data for the Nenana River since 1917
  • Ice breakup is an important time of transition for communities that rely on these rivers for transportation, subsistence hunting and fishing, and other needs
  • Early thawing can lead to severe ice movement, jamming, damage to infrastructure, and destructive floods

The River Ice Explorer App

With this app you can explore the Nenana River ice breakup data for any range of years from 1917 to 2010. You can:

  • Browse the data: sort by columns, search
  • Plot the data
  • Fit two types of models:
    • simple linear model
    • segmented linear model (to explore possible shifts in the trends)
  • Compare the results of the fits to select the best model

Explore Data and Fit Models

plot of chunk data-and-fit-figure

Available data contains ice breakup dates (Julian dates) for all years between 1917 and 2010.

The app allows you to select a range of years, plot the data and fit linear (blue) and segmented (red) models. In the latter case, you can specify the break point (1970 in this example).

Compare Models


Model selection based on AICc :

          K   AICc Delta_AICc AICcWt Cum.Wt      LL
Segmented 4 594.36       0.00   0.72   0.72 -292.95
Linear    3 596.21       1.86   0.28   1.00 -294.97

The app prints the summaries of the fitted models. Additionally, if two models are selected, you can compare them based on the AIC.

In this example, the segmented model gets first place (the output is always sorted to show the lower AIC model first).