Thursday, August 12, 2021

IPCC sea level rise hot take

 The 6th IPCC Assessment report was released yesterday (find the full report here), which provides a good motivation for digging into any new sea level insights they summarize, and also to start to work out what the IPCC's new global projections mean for Washington State (and how they compare to the projections we released in 2018).  

Time-series of ice mass loss in Greenland, published by NASA 

First off, a few interesting nuggets from the section on observed changes in the Summary for Policy Makers (mostly pages SPM-5 and SPM-6):  Contributions to sea level change from melting ice became the dominant contributor sometime around the beginning of the 21st century - prior to that contributions from thermal expansion were about equal to those from melting ice.  This is a big change, as sea level contributions from melting ice are more uncertain, and potentially far larger as we move through the 21st century, than those from thermal expansion.  The IPCC report ascribes a strong confidence ("very likely" in IPCC parlance) that melting in Greenland (see figure above or here) is associated with anthropogenic forcing, whereas finds "only limited evidence, with medium agreement, of human influence on the Antarctic Ice Sheet mass loss" (which again, NASA visualizes here).  Also, observed global average sea level rise since 1900 is pegged at about 8 inches...if it were to stay there for the 21st century we would be doing quite well.  The problem is that the IPCC assigns "high confidence" to rates of global average sea level rise having accelerated already in the 21st century.   

On to the projections:

Summary global average sea level projections from the IPCC 6th AR.

There aren't huge changes to the overall sea level change picture in this AR.  There is a new family of emissions scenarios but each is still tied to a particular level of radiative forcing, so its easy to do a quick comparison to projections associated with the RCP family of emissions scenarios.  For each emission scenario the IPCC focuses on a "likely range" around their best estimate, that is intended to represent the zone of uncertainty between the 17th and 83rd percentile of a probability distribution - and I found this acknowledgement of uncertainty in a probabilistic framework to be useful.  Generally we see projections in the ~close-to-1-meter zone for the higher emissions scenarios, and somewhere in the 0.5 meter zone by 2100 for the lower emissions scenarios...about what we've been working with for a while. 

They also state, "only likely ranges are assessed...due to difficulties in estimating the distribution of deeply uncertain processes".  The IPCC has NOT adopted the expert elicitation process that was used to define the tails of the uncertainty distribution that we used in our 2018 report, which I think is fine.  Interestingly, though, the IPCC sort of gives a nod to the upper part of the uncertainty distribution, and also includes a "low-likelihood, high impact scenario" in this AR, which is included presumably because of the hints that Antarctica is giving to the world about possibly bigger-than-expected contributions to sea level rise in the future.  In our 2018 report these high magnitude low likelihood scenarios were built into the model (again, using expert elicitation), but the message from both of these approaches is the same...there is a "can't be excluded as a possibility" chance of some pretty significant sea level rise this century, that will represent a significant strain on coastal communities and landscapes everywhere:  whether you think of it as having a 1% chance of happening, or you think of it as a "low-likelihood" scenario perhaps doesn't matter so much.  In this AR the IPCC also provides a 2300 projection, presumably to emphasize that, after the 2100 time horizon, sea level rise isn't going to slow down, regardless of the chosen emission scenario.   

Projections from our 2018 SLR assessment for Washington State, localized for Seattle (and available here).

To take things a bit further, the NASA sea level change team worked with the IPCC to develop a localized sea level viewer for the 6th AR projections, and they include projections for Seattle:

Seattle projections from the IPCC 6th AR for SSP-8.5 out to 2150, available through the NASA Sea Level Change portal.

which provides an easy way to do an apple-to-apples (or close to) comparison with our 2018 projections (shown above).  For 2100 for Seattle for the SSP-8.5 scenario they project sea level 0.69 meters above present, with a likely range between 0.51 and 0.98 meters.  Our projections suggest 0.70 meters by 2100, with a likely range between 0.51 and 0.94 meters...basically dead on equivalent.  

The differences, of course, are in the upper part of the uncertainty distribution.  We, for example, allow users to access a 1% chance sea level scenario, which comes in at 1.55 meters by 2100 for Seattle, whereas the IPCC opted to include a "low confidence" scenario, which they peg at 0.81 meters by 2100.  They DO though include a "likely range" for this "low confidence" scenario (which I'm having a hard time wrapping my brain around), which ranges between 0.51 and 1.45 meters.  Its also worth noting that the 0.81 meter 2100 "low confidence" scenario projection falls within their "likely range" for the "normal confidence" projections...again something that I can't quite figure out how to interpret.  Clearly, the upper part of the sea level rise uncertainty distribution remains a bit of a bugaboo.