Shorthand

In this blogpost, in order to specify between the original EBM and the second version of the EBM, I will denote:

• D1= original, faster driver, with the simplified radiative transfer
  • more accurate for small changes in radiation
• D2= second, larger, slower driver, which allows explicit pCO2 values
  • by including a polynomial parameterization of outgoing longwave radiation and planetary albedo

Questions

1. in D2 there is no stochastic functionality/variables (as there is in D1)
   • Will the addition of stochastic calculations in D2 help or hurt the accuracy of our model?
2. in D1 there is a variable called soladj=solar forcing adjustment…
   • What is its purpose?
   • It is not in D2, will that considerably affect the accuracy of the results?
3. in the version of D2 you sent me, is there any built-in, uncommented, functionality which dictates how pCO2 changes with time?
   • I've added an ODE for pCO2 which allows me to specify as an input into D2 a civilizations annual, per-capita addition of atmospheric CO2: \frac{dP_{co2}}{dt}=\epsilon N
     1. Should I leave this be, and turn on the programs built-in reduction of pCO2 by weathering?
        • if yes, how?
     2. Should I reduce a term from my existing pCO2 ODE (\frac{dP_{co2}}{dt}=\epsilon N-\omega) and turn off the programs built-in functionality?
        • allowing me to have an input in D2, (\omega=annual reduction of pCO2 by Earths natural weathering processes)

Summary of Current State of Equations

\frac{dN}{dt}=r(T)*N
\frac{dP}{dt}=\epsilon N - \omega
\frac{dT}{dt}=EBM(pCO2, T_{0})
​Link to REU Paper describing Summary of Work Done This Summer (so far)
Progress: Calibrated our model with population and CO2 data from the past two centuries, summed up in the table below.

\Delta t (yrs)	0	140	190	198
t=time (yrs)	1820	1960	2010	2018
N=Population (10^{9} ppl)	1.042	3.032	6.834	7.594
P=pCO2 (ppm)	284	316	390	413

Poster Showing Result of Modelling Earth

​Link to My Poster

Next Steps

1. increase the accuracy of modeling earth
   • find true temperature dependence on the growth rate
   • find more accurate value for \omega(=reduction of CO2 by earth's natural weathering processes)
2. determine the climate sensitivity of various energy resources
   • can let us determine the inevitability of global warming
3. put a dTdt dependence on the deathrate
   • to quantify acclimation
4. introduce parameter E as a proxy for technology
   • will raise peak of relative growth rate (r)
   • will increase the average carbon footprint (\epsilon)

Radiation Pressure Paper

Believe almost every comment has been addressed. Two things remaining:

• Neutral speed plot: better version below?

• Timescale for bulging of wind. It may actually be a box effect (cycle is very close to 2 crossing times), though I'm not yet sure physically how reaching the edge of the box might increase the effect of radiation pressure.

Charge Exchange

See ​my previous blog post for the latest thoughts on charge exchange. It's recently started again on Stampede and is moving fairly quickly (~6 days left as of this morning). Here are the latest frames:

Species proportions

Charge exchange potential

Non-rotating Charge Exchange

Using the terminal velocity and mass loss rate from ​this thesis to approximate the wind of an O6 star at the orbit of Jupiter, I get a wind density of ~2x10^-18 g/cm³, which is within an order of magnitude of our current values of 4.5-6.5x10^-19 g/cm³. Given this, I've queued the rotating tests with the same wind as HD209458b (though technically they should be corrected for a different planetary and stellar mass).

AMR line transfer

Mixed news on this front. Bad news: The (current on the repo) version of the AMR line transfer is taking about double the amount of time on BlueHive as the previous. Good news: Based on one frame, it appears to be working for a full problem (rather than the simplified tests).

I'd like to run this in a profiler, but I haven't yet gotten it working (getting errors when attempting to run through slurm). I may try a run on an interactive session on a single node, though that will be far fewer cores, which may affect the hotspots…

For reference, here's the updated table:

CEJet

Current run on Stampede

In queue, expected start time is tomorrow morning at 6:30.

Same as the 1st run, but with all the total quantities defined as the current CE settings. Use refinement radius as CE model B in paper II, 4d12 for frames 0-45.

Parameters of the Jet

• mass loss

• launching radius

• speeds

• others

—-

PN

Flows during initial wind period

The total amount of gas flow into the wind radius is 0.19~M_{\odot} . This amount of gas has been removed from the simulation. The inflow through the outer boundaries is almost negligible (1e-7 solar mass).

This means almost half the initial gas (0.4 solar mass) has been removed. There is a stream of inflow through the funnel with a slow speed, but since it’s at the region of highest density, it takes a lot mass in.

​velocity plots

​density plots

​python notebook

​test without IW

The SPH CEE simulation contains

• Primary of mass 0.88 M_sun, where
  • the core mass is 0.392 M_sun, and
  • the envelope is 0.488 M_sun
• Secondary (point particle) of mass 0.6 M_sun
• so the amount of gas for us to start the PN simulation is about the envelope mass of the SPH primary

Frame 0 has

• total mass of the gas about 16 solar mass — M_tot =16.136 M_sun
• if set a density threshold at 1e-8 g/cc, then
  • the high density region sum up to — M_high=15.712 M_sun, while
  • the low density region sum up to — M_low = 0.424 M_sun

NOTE

1. The density threshold is chosen as such, because since frame 1, the maximum density in the simulation is on the order of 1e-8 g/cc.
2. The total mass of the low density gas is about right to the total amount of gas we start with.
3. The high density region is produced from mapping the SPH data onto the AMR grids. The SPH data has local maximum density on the order of 0.1 g/cc, while frame 0 has local maximum of 0.05 g/cc.
4. The mass of the high density gas is confined at a point close to the simulation center with a radius of 7.6 R_sun (about the size of our finest grid). These 4 extremely high density grids are within the wind initialization radius, so once the simulation starts they are removed, and won't affect the simulation any further.
5. I interprete this 16 M_sun region as the final location of the binary core from the SPH simulation. Maybe the core is more confined in the SPH simulation but spreads out in AMR grids and produce this artifect of a huge amount of mass.

Between Frame 0 and Frame 1, there are few sources and sinks of mass:

1. Sink - Lowering the floor density for density protection, the maximum effect has the change of -1.4e-3 M_sun
2. Source-Sink - carving out the central region and replace with a 1 solar mass point particle and fill the rest of the region with initial wind density.

From Frame 1 to Frame 149 (before fast wind starts):

1. Source - inflow from the outer boundary of the simulation. During the entire initial wind period (3000 days), the total amount of inflow gas is 1.11E-7 M_sun, almost negligible
2. Sink - gas flows into the wind initilization radius will be removed from the simulation, and the total change is -0.19 M_sun