SPECIAL NOTICE: As part of RAND's sponsorship, INFER will now be the RAND Forecasting Initiative (RFI). Rebranding will be in progress over the next couple days.
I am decreasing my probability estimate a bit based on that and the fact that since February till now we only had 1 exoplanet added to the catalog. I will return to the case of Gliese 12 b in the moment. Of course, for our resolution only additions to the catalog count, we do not have to care about the removed exoplanets, so this is a bit different since, as you can see, exoplanets were indeed removed from the list of potentially habitable ones, and that happened a few times as my research confirms (see the comments' column in the spreadsheet).
4 months passed we still have 8 months which will count towards the resolution. We need to remember that a) Gliese 12 b might be added to the catalog soon, and that this year 7 exoplanets were added to the catalog already (including one which counts towards the resolution). At the same time Gliese 12 b might not qualify. I know that there are different formulas to count if the planet is in habitable zone, an it was reported as being in habitable zone, so likely scientists were right. But I use formulas to calculate this using the same variables which they used in their own old calculator https://web.archive.org/web/20120121043647/http://phl.upr.edu/projects/habitable-exoplanets-catalog/calculators (unfortunately it doesn't work anymore) I get negative result.
Please enter the star's luminosity (Lsun)
0.00728
Please enter the object's semi-major axis (AU)
0.0668
Please enter the star's temperature (K)
3296
This object's Seff: 1.6314676037147264
*** This systems HZ stats: *** DISTANCES IN AU Recent Venus (1 Me): 0.0701834972212403 Runaway Greenhouse (1 Me): 0.08859776188221559 Maximum Greenhouse (1 Me): 0.1738302893179218 STELLAR FLUX (EFFECTIVE) Recent Venus (1 Me): 1.477955537271435 Runaway Greenhouse (1 Me): 0.9274401159976421 Maximum Greenhouse (1 Me): 0.24092422649365566 Early Mars (1 Me): 0.2165999781463119
STELLAR FLUX values are considered more important than distance, so we compare 1.6314676037147264 value with recent Venus (1.477955537271435) and it is higher.
I used their own python library https://github.com/planetaryhablab/PHL-Library to check stellar flux for Gliese 12 b, and even when we include orbital eccentricity (based on NASA database it is <0.50, we can only get higher values, not lower.
import phl
"""
Definitions:
a - orbital semi-major axis (AU)
e - orbital eccentricity
A - planetary bond albedo
L - stellar luminosity (solar units)
"""
a = 0.0668
e = 0.0 # I have also tested 0.49 but since NASA exoplanets catalog says that eccentricity is <0.50, but it gives even higher value
L = 0.00728
print(phl.flux(a,e,L)) result: 1.6314676037147262
So scientists likely used more advanced and/or more recent formulas or more recent input data. There is one planet in the HEC catalog with a bit higher value of stellar flux than this one, but I have not yet run the full calculation of recent Venus etc. on that star system data. Also thanks to @ctsats we know that the catalog is updated irregularly - form my reconstruction it looks like the website was updated in recent years much less often than at the beginning. I was wondering if we could have kind of an early warning system based on a python script which would take data from NASA catalog (there is already a library to do exactly that) and run a classification task which would be trained on the NASA data with information about the classification from the catalog. I found some articles and jupyter notebooks like this one https://github.com/alpercakr/Planet-Hunting-and-Habitability-Detection-with-Machine-Learning and I am now learning how to use xgboost in python to approach this. I think I would concentrate on calculating the variables which we know that are used in the catalog, like Earth Similarity Index etc. If we were able to create something like that we will not be surprised with massive but infrequent update (like with 6 exoplanets added this year before the question has opened).
@ctsats it would be super cool if we could find out about the exact criteria for classifying planet as habitable - which variables are used, what are the cut off values (I have some idea but it is not the same as knowing) or what formulas are used to calculate each metric (especially useful in case when there are a few approaches possible - I saw some articles about it on the webpage and its archived versions but the problem is it could have changed and it was not complete set of criterias). Do you think we could get another reply from the professor if we ask about it?
We were wondering about the history of updates to the catalog, so I spent many hours analyzing the archived versions of the catalog on Archieve.org Wayback Machine (both archived website and csv files) + on archive.ph Here is the result of my work: https://docs.google.com/spreadsheets/d/1RHV22QmSWF6m9YpVCkXLspZ4I688j36feD6Ra3OFMsw/edit?usp=sharing
I am decreasing my probability estimate a bit based on that and the fact that since February till now we only had 1 exoplanet added to the catalog. I will return to the case of Gliese 12 b in the moment. Of course, for our resolution only additions to the catalog count, we do not have to care about the removed exoplanets, so this is a bit different since, as you can see, exoplanets were indeed removed from the list of potentially habitable ones, and that happened a few times as my research confirms (see the comments' column in the spreadsheet).
4 months passed we still have 8 months which will count towards the resolution. We need to remember that a) Gliese 12 b might be added to the catalog soon, and that this year 7 exoplanets were added to the catalog already (including one which counts towards the resolution). At the same time Gliese 12 b might not qualify. I know that there are different formulas to count if the planet is in habitable zone, an it was reported as being in habitable zone, so likely scientists were right. But I use formulas to calculate this using the same variables which they used in their own old calculator https://web.archive.org/web/20120121043647/http://phl.upr.edu/projects/habitable-exoplanets-catalog/calculators (unfortunately it doesn't work anymore) I get negative result.
I used this code to calculate Habitable Zone https://github.com/Eelt/HabitableZoneCalculator and checked the results with this calculator https://web.archive.org/web/20200613190344/http://depts.washington.edu/naivpl/sites/default/files/hz.shtml (I get similar values) and it looks like the stellar flux value for Gliese 12 b is to high (higher than the value for recent Venus which is considered the inner limit of the zone supporting liquid water).
STELLAR FLUX values are considered more important than distance, so we compare 1.6314676037147264 value with recent Venus (1.477955537271435) and it is higher.
I used their own python library https://github.com/planetaryhablab/PHL-Library to check stellar flux for Gliese 12 b, and even when we include orbital eccentricity (based on NASA database it is <0.50, we can only get higher values, not lower.
So scientists likely used more advanced and/or more recent formulas or more recent input data. There is one planet in the HEC catalog with a bit higher value of stellar flux than this one, but I have not yet run the full calculation of recent Venus etc. on that star system data.
Also thanks to @ctsats we know that the catalog is updated irregularly - form my reconstruction it looks like the website was updated in recent years much less often than at the beginning.
I was wondering if we could have kind of an early warning system based on a python script which would take data from NASA catalog (there is already a library to do exactly that) and run a classification task which would be trained on the NASA data with information about the classification from the catalog. I found some articles and jupyter notebooks like this one https://github.com/alpercakr/Planet-Hunting-and-Habitability-Detection-with-Machine-Learning and I am now learning how to use xgboost in python to approach this. I think I would concentrate on calculating the variables which we know that are used in the catalog, like Earth Similarity Index etc. If we were able to create something like that we will not be surprised with massive but infrequent update (like with 6 exoplanets added this year before the question has opened).