Ben, BK, Corinne, Alexey meeting - 2018-01-18

[ashiklom]

Based on Ben and Alexey's discussion on 2018-01-15

• Two papers
• Paper 1 -- Sensitivity of climate to permafrost emissions using Hector
  • Showcase the utility of Hector (and simple climate models more generally) to explore a really wide range of parameters
  • Already done: Run Hector with exogenous permafrost emissions scenarios. Useful as an exploratory analysis, but will need to do something a little more sophisticated to publish with Hector alone.
  • First pass at dynamics: Use Hector's existing structure. Maybe split the Arctic (or just active layer permafrost) into its own "biome" with a dedicated warming multiplier and custom respiration Q10 and CO2 fertilization parameters.
    • We can basically do this already -- Alexey debugged and ran some preliminary tests of Hector's multi-biome capabilities (see this branch).
    • Main challenge is to define the parameter space and run conditions
    • Consider "calibrating" (Bayesian) parameter ranges by trying to match Hector permafrost predictions against existing permafrost emissions scenarios.
      • There are lots of numbers in the literature. Potentially one piece of this could be a section that matches specific scenarios to specific Hector parameter combinations (e.g. scenario in paper X corresponds to respiration Q10 Y and fertilization effect Z). This would then make it possible to consider other implications of specific literature estimates, including economic impacts in paper 2 (see below).
  • Advanced: Build a simple permafrost model in Hector
    • 2-3 parameters max -- keep it simple!
    • Maybe just use the existing respiration model, but add a single parameter that controls CO2-CH4 fraction. Or if we built a dedicated module, CO2-CH4 fraction would be interesting and novel and should be included.
    • Burke et al. 2017 (Biogeosciences) define a "Frozen Carbon Residence Time" parameter. Could be an interesting approach.
      • More generally, lots of good ideas and references about permafrost modeling in that paper.
  • Might be a good idea to try to involve somebody from the permafrost C community.
  • Target journal -- JGR? Biogeosciences? ERL?
• Paper 2 -- Economic impact of permafrost emissions using GCAM-Hector
  • Others have done this with IAMs that have "climate damages" like ΔGDP = f(ΔT). The novelty here would be the nuance that GCAM allows -- e.g. How much harder is it to meet a 2°C warming target if we account for permafrost emissions? How different is the shadow carbon price? How different are the energy, agriculture, transportation, etc. sectors?
  • Key here is that this paper would be informed by the results of paper 1. This way, we can also present a narrative about the modeling process (i.e. first, use a simple model to identify key parameters and sensitivities; then, sample that space intelligently for running a more complex model).
  • Modeling-wise, shouldn't be that hard to do. GCAM already works with current version of Hector, and incorporating any Hector changes from paper 1 should be trivial.
  • Main challenges are (1) how to sample the parameter space, and (2) how to configure GCAM to answer the questions we're interested in. Lots of JGCRI people could help with 2.
    • Maybe bring on others from JGCRI -- Kate, Steve, others?
  • Target journal -- shoot high! Nature? Science? PNAS?

Also, here is the link to Alexey's slides from JGCRI Earth System Modeling presentation on permafrost.

[ashiklom]

@bpbond @cahartin @bkravitz

[bpbond]

e.g. scenario in paper X corresponds to respiration Q10 Y and fertilization effect Z).

I think this would be really interesting.

[ashiklom]

Summary of next steps:

• Hector development
  • Add CO2-CH4 ratio parameter (α) to simpleNbox respiration calculation
  • Add simpleNbox CH4 calculation to natural CH4 emissions at each timestep
  • Fix biome warming multiplier
• Set parameters and inputs
  • Parameter list:
    • Q10 (arctic, global)
    • β (arctic, global)
    • α (arctic; set global to 0)
    • Arctic warming multiplier
  • Initial conditions: Vegetation, detritus, and soil pools
• Writing/organization
  • Over next few weeks, put together a more detailed paper outline and project task list. This will be useful for determining personnel.

General notes:

• Parameterize CH4-CO2 ratio – single input
• CH4 only comes from soil
• Soil moisture?
  • Model independently
  • Use as covariance between CH4-CO2 ratio (let's call it α) and Q10[arctic]
• Research question: Tipping point
• Discussion section
  • Link between soil moisture and precipitation (downscaling?)
  • Scale-aware processes will go into land surface models
  • NGEE-Arctic
• Maybe calculate a turnover time / stockover / flux-type term for permafrost C
• "Second-order" sensitivity analysis idea – How does the strength of the covariance between α and Q10 affect uncertainty in climate output?
• Hector development
  • Start with just a "global" box and a "permafrost" box
  • Think about what we mean by permafrost – maybe first X (2?) meters?
• Pools should be straightforward to initialize
• Arctic is warming faster – define warming coefficient
  • Based on observations and/or CMIP?
  • Steve Smith should have a number
• Teach simpleNbox to compute extra methane.
  • Track this methane independently
  • Need to add to forcing calculation, which is complicated because atmospheric CH4 interacts with NO2
• Vegetation-albedo feedback?
  • Mention in discussion
  • BK: Should be internalized as part of the Arctic warming coefficient
• Vegetation-temperature feedback?
  • Can capture through parameter covariance between Arctic CO2 fertilization effect (β) and warming effect, similar to soil moisture
  • This should be a reasonable approximation because CO2 and temperature are roughly linearly related
• Add others to project? Someone with permafrost knowledge?
  • BK suggested Doug MacMartin (Caltech)
  • BBL suggested Claire Treat – she's done a lot of synthesis work on CO2-CH4 sensitivity w.r.t. soil moisture

[bpbond]

Also I'm working at the moment with Avni Malhotra (Stanford) on a synthesis on peatland CO2/CH4 fluxes right now, and we've collected a lot of data for this effort. This may be useful.

[bpbond]

Key papers off the top of my head:

Schuur et al. (2015)

The new northern permafrost zone carbon inventory reports the surface permafrost carbon pool (0–3 m) to be 1,035 ± 150 Pg carbon (mean ± 95% confidence interval, CI)

Koven et al. 2015

To calculate potential CH4 emissions in addition to the CO2 losses, we assume that a constant fraction of these anoxic C losses are emitted as CH4 to the atmospher

Schädel et al. 2016

Potential carbon emissions dominated by carbon dioxide from thawed permafrost soils

Treat et al. 2015

Our results demonstrate that as climate warms in arctic and boreal regions, rates of anaerobic CO2 and CH4 production will increase, not only as a result of increased temperature, but also from shifts in vegetation and increased ground saturation that will accompany permafrost thaw.