False alarms: How early warning signals falsely predict abrupt sea ice loss

Abstract

Uncovering universal early warning signals for critical transitions has become a coveted goal in diverse scientific disciplines, ranging from climate science to financial mathematics. There has been a flurry of recent research proposing such signals, with increasing autocorrelation and increasing variance being among the most widely discussed candidates. A number of studies have suggested that increasing autocorrelation alone may suffice to signal an impending transition, although some others have questioned this. Here we consider variance and autocorrelation in the context of sea ice loss in an idealized model of the global climate system. The model features no bifurcation, nor increased rate of retreat, as the ice disappears. Nonetheless, the autocorrelation of summer sea ice area is found to increase in a global warming scenario. The variance, by contrast, decreases. A simple physical mechanism is proposed to explain the occurrence of increasing autocorrelation but not variance when there is no approaching bifurcation. Additionally, a similar mechanism is shown to allow an increase in both indicators with no physically attainable bifurcation. This implies that relying on autocorrelation and variance as early warning signals can raise false alarms in the climate system, warning of “tipping points” that are not actually there.

Statistical indicators of Arctic sea-ice stability – prospects and limitations

“…2b-d) or in fact to any positive feedback. This specific result has also been obtained in the latitudinally resolved version of E12 by Wagner and Eisenman (2015b).…”

“…Moreover, the relatively large timescale of warming or cooling of the ocean's mixed layer becomes important once sea ice is not present during a substantial part of the year. Using a latitudinally explicit version of the model by Wettlaufer (2009), Wagner andEisenman (2015b) therefore argue that the mixed-layer effect can raise false alarms of abrupt ice loss.…”

Wave inhibition by sea ice enables trans-Atlantic ice rafting of debris during Heinrich events

“…Alternatively, in the case where iceberg meltwater is concentrated within 10% of the IRD belt, 2 months' worth of freshwater discharge would stabilize the water column with a 25 m deep surface layer, and in the case where iceberg meltwater is concentrated within 50% of the IRD belt, 2 months' worth of freshwater discharge would stabilize the water column with a 5 m deep surface layer. The sea ice that formed would then insulate the ocean from the atmosphere and thereby substantially reduce the effective surface heat capacity (e.g., Wagner and Eisenman, 2015). We make the somewhat extreme approximation that this stabilization would cause seasonal temperature variations to be as large as they are over land.…”

Times are changing : Implications of climate change for the occurrence and predictability of tipping points

“…There can be different kinds of drivers for such systematic change in climate memory. The slowing down of Pacific fluctuations is likely due to deepening of the ocean mixed layer, increasing the heat capacity of this element (Boulton & Lenton, 2015), and a similar (but reversed) effect occurs for Arctic sea ice loss (Bathiany et al, 2016;Wagner & Eisenman, 2015). Such volume effects on slowness may well be common in the climate and are accompanied by a decrease in the variance of the fluctuations.…”

“…Importantly, the autocorrelation in climatic variables may change over time (Lenton et al, 2017). For instance, the Pacific Decadal Oscillation and North Pacific sea surface temperatures have become more autocorrelated in the period from 1900 to 2015 (Boulton & Lenton, 2015;Huntingford et al, 2013), and large changes in climate variability are to be expected in the Arctic where sea ice loss leads to larger persistence and smaller variance in temperature variability (Bathiany et al, 2016;Wagner & Eisenman, 2015).…”