Uncertainties in the timing of unprecedented climates

Abstract

The question of when the signal of climate change will emerge from the background noise of climate variability -the 'time of emergence' -is potentially important for adaptation planning. Mora et al. 1 (M13) presented precise projections of the time of emergence of unprecedented regional climates. However, their methodology produces artificially early dates at which specific regions will permanently experience unprecedented climates and artificially low uncertainty in those dates everywhere. This overconfidence could impair the effectiveness of climate risk management decisions 2 .Any human-induced changes in climate will be modulated by natural fluctuations of the oceans and atmosphere (e.g. El Niño events). These fluctuations occur randomly and independently, in both reality and individual modelbased projections, and act to obscure the climate change signal 3,4,5 . M13 discuss projections of when changes in climate emerge permanently above the levels of such fluctuations (a metric first considered by ref. 6). However, by ignoring the irreducible limits imposed by these same random fluctuations, M13 express their emergence dates with too much certainty.Several methodological oversights contribute to the erroneous uncertainty quantification. Firstly, M13 ignore the possibility that emergence dates before the end of the simulations are not permanent deviations from the historical range 6 (termed 'pseudo-emergence'). In many regions where emergence has not occurred by the year 2100, M13 even artificially set the emergence date to equal 2100. This oversight produces several effects, including: (i) early and overconfident estimates of regional temperature emergence, and (ii) implausible emergence dates for precipitation of exactly 2100 with zero uncertainty almost everywhere.Secondly, M13 estimate precision of regional emergence timing using the standard error of the ensemble mean (σ/√N), where N(=39) is the number of simulations and σ is their standard deviation. While the estimate of the ensemble-mean becomes more precise with larger ensemble size, natural fluctuations of the climate (such as El Niño) dictate that the future evolution of climate will not behave like the mean, but as a single realization from a range of outcomes 5,7 . The use of σ/√N greatly underestimates 8 this irreducible uncertainty, as well as the climate-response uncertainty given by the inter-model spread, and is therefore inappropriate for use in

Extreme temperatures in Southeast Asia caused by El Niño and worsened by global warming

“…We find that the latter subperiods show more extreme post-Niño April SATs because of a significant shift in the mean of the distribution (two-sample t -test; P <0.01) but statistically indistinguishable variance (f-test; P >0.05). Changes in the tails of the distribution caused by shifts in the mean, similar to the ones identified here, are the most straightforward way in which global warming can lead to increased frequency of extremes1629. Although increased extremes can also arise from increased variability, we do not observe statistically distinguishable variance between the two subperiods across any of these distributions ( f -test; P >0.05).…”

“…These studies indicate that on the spatial scales of MSA, long-term warming is expected to emerge sooner in the lower latitudes relative to anywhere else on the planet. This occurs because lower latitudes experience naturally lower year-to-year variability compared to higher latitude continental regions, thus allowing for an earlier detection of climate departures from a reference state121617. For this reason the Intergovernmental Panel on Climate Change 5th Assessment Report concluded with high confidence that relative to natural variability, near-term increases in seasonal and annual mean temperatures are expected to be larger in the tropics and subtropics than in mid-latitude regions8.…”

Global Emergence of Anthropogenic Climate Change in Fire Weather Indices

“…This 50-year period was also used to define the local 95th percentile threshold value for calculating FWI 95d and midpoint range (average of maximum and minimum value of FWI) for calculating FWI fwsl . Additionally, we restrict emergence to occurring no later than 2050 to ensure persistence for at least 2 decades after initial detection, following previous methods (Hawkins et al, 2014;King et al, 2015). This was done using moving 30-year windows starting in the year 1980 on the basis of the signal of change, defined as the difference of means between a given 30-year period and the baseline period, exceeding the noise of natural variability, defined herein as one standard deviation (calculated for the baseline period).…”

Estimating daily climatologies for climate indices derived from climate model data and observations

“…Expectedly, the LOESS‐fit and the true climate are not identical and can differ from each other. These differences are due to the model's internal variability and are therefore irreducible [ Hawkins et al ., ].…”