Ok, so I think we converge pretty much then—essentially what I am saying is that people concerned about compounding risks would argue that these are not modeled correctly in GBD and that there is much more uncertainty there (and that the estimate is probably an underestimate, from the perspective of taking the compounding risk view seriously).
Makes sense. Just to clarify, the data on deaths and disease burden from non-optimal temperature until now are from GBD, but the projections for the future death rates from non-optimal temperature are from Human Climate Horizons.
I’m not sure it would change your underlying conclusions, but for what it’s worth I took a quick look at the data/methods from Human Climate Horizons and think that it is likely an underestimate of future heat-related mortality.
This paper underlies the Human Climate Horizons/OWID estimates. A few quick assumptions I think are worth highlighting:
1. The authors (by admission) do not consider the effects of humidity. Accurate humidity data is much harder to come by than temperature, and it’s often not included in the down-scaled versions of the climate models, though some exceptions apply. However, it’s potentially quite relevant for heat-related mortality, though there is some debate in the literature. Given that it is unlikely that incorporating humidity would decrease heat-related mortality, my own view here is that this pushes current estimates towards a lower bound.
2. The authors use climate model data that are downscaled with a technique called Bias-Correction Spatial Disaggregation (BCSD). This approach has two key assumptions in it: (1) it assumes that the relative spatial patterns in the training data will remain constant under future climate change, and (2) it is calibrated to monthly data, not day to day changes, which has the effect of dampening extreme values. In practice, these assumptions limit the ability to model things like extreme heat waves and heat domes, which can cause large fatality spikes (e.g. figure below from Washington State in 2021). Missing these features in some locations might be akin to missing almost all the possible heat related mortality in cooler climates.
Again, I don’t think these are significant enough to fundamentally change your conclusions, but I do think it’s worth highlighting that these types of results can be quite sensitive to specifics of the climate modeling approaches that are used.
Thanks, Bradley, and welcome to the EA Forum! Strongly upvoted.
Given that it is unlikely that incorporating humidity would decrease heat-related mortality, my own view here is that this pushes current estimates towards a lower bound.
If adequately modelling humidity would increase heat deaths, I wonder whether it would also decrease cold deaths, such that the net effects is unclear.
In practice, these assumptions limit the ability to model things like extreme heat waves and heat domes, which can cause large fatality spikes (e.g. figure below from Washington State in 2021). Missing these features in some locations might be akin to missing almost all the possible heat related mortality in cooler climates.
As illustrated below, deaths from extreme cold and heat accounted for only a tiny fraction of the deaths from non-optimal temperature in 2015 in some countries, which attenuates the effect you are describing.
Ok, so I think we converge pretty much then—essentially what I am saying is that people concerned about compounding risks would argue that these are not modeled correctly in GBD and that there is much more uncertainty there (and that the estimate is probably an underestimate, from the perspective of taking the compounding risk view seriously).
Makes sense. Just to clarify, the data on deaths and disease burden from non-optimal temperature until now are from GBD, but the projections for the future death rates from non-optimal temperature are from Human Climate Horizons.
Thanks, Bradley, and welcome to the EA Forum! Strongly upvoted.
If adequately modelling humidity would increase heat deaths, I wonder whether it would also decrease cold deaths, such that the net effects is unclear.
As illustrated below, deaths from extreme cold and heat accounted for only a tiny fraction of the deaths from non-optimal temperature in 2015 in some countries, which attenuates the effect you are describing.