{ "id": 388724, "title": "Understanding Ice Transparency factors", "url": "/competitions/icecube-neutrinos-in-deep-ice/discussion/388724", "author_name": "", "post_date": "2023-02-19T07:20:03.953938500Z", "votes": 9, "comment_count": 4, "views": 0, "content": "

Notebook by @anjum48 employes the ice transparency data to train DynEdge from scratch.
\nSample of ice_transparency data

\n
depth scattering_len absorption_len\n1398.4 13.2 45.1\n1408.4 14.0 48.6\n1418.4 14.7 53.2\n
\n

Let's understand what are these values :

\n

Cherenkov photons are emitted with a characteristic wavelength dependence of 1 λ2 in the wavelength range of 300-600 nm, which includes the relevant sensitivity region of the photosensors. Photons are emitted in a cone around the direction of particle motion with an opening angle, determined by the speed of the particle and refractive index of the ice, of about 41◦ for relativistic particles.
\nAs the photons propagate from the point of emission to the receiving sensor, they are affected by absorption and scattering in the ice. These propagation effects must be considered for both simulation and reconstruction of IceCube data and thus need to be carefully modeled. The important parameters to describe photon propagation in a transparent medium are: the average distance to absorption, the average distance between successive scatters of photons, and the angular distribution of the new direction of a photon at each given scattering point.
\nPhotons are emitted by the LEDs in DOMs and recorded by other DOMs, as shown :
\n
\nThe geometrical scattering coefficient b determines the average distance between successive scatters (as 1/b). It is often more convenient to quote the effective scattering coefficient, be = b · (1 − {cos θ}), where θ is the deflection angle at each scatter.
\nThe absorption coefficient a determines the average distance traveled by a photon before it is absorbed (as 1/a).
\nThere scattering_len and absorption_len are the average distance between successive scatters and average distance traveled by a photon before it is absorbed respectively at particular depth.

\n

References :

\n
    \n
  1. Measurement of South Pole ice transparency with the IceCube\nLED calibration system
    2. \n
  2. Early Sharing Prize - DynEdge - 1.046
    3. \n
", "messages": [ { "id": "2150356", "postDate": "02/19/2023 07:20:03", "content": "

Notebook by @anjum48 employes the ice transparency data to train DynEdge from scratch.
\nSample of ice_transparency data

\n
depth scattering_len absorption_len\n1398.4 13.2 45.1\n1408.4 14.0 48.6\n1418.4 14.7 53.2\n
\n

Let's understand what are these values :

\n

Cherenkov photons are emitted with a characteristic wavelength dependence of 1 λ2 in the wavelength range of 300-600 nm, which includes the relevant sensitivity region of the photosensors. Photons are emitted in a cone around the direction of particle motion with an opening angle, determined by the speed of the particle and refractive index of the ice, of about 41◦ for relativistic particles.
\nAs the photons propagate from the point of emission to the receiving sensor, they are affected by absorption and scattering in the ice. These propagation effects must be considered for both simulation and reconstruction of IceCube data and thus need to be carefully modeled. The important parameters to describe photon propagation in a transparent medium are: the average distance to absorption, the average distance between successive scatters of photons, and the angular distribution of the new direction of a photon at each given scattering point.
\nPhotons are emitted by the LEDs in DOMs and recorded by other DOMs, as shown :
\n
\nThe geometrical scattering coefficient b determines the average distance between successive scatters (as 1/b). It is often more convenient to quote the effective scattering coefficient, be = b · (1 − {cos θ}), where θ is the deflection angle at each scatter.
\nThe absorption coefficient a determines the average distance traveled by a photon before it is absorbed (as 1/a).
\nThere scattering_len and absorption_len are the average distance between successive scatters and average distance traveled by a photon before it is absorbed respectively at particular depth.

\n

References :

\n
    \n
  1. Measurement of South Pole ice transparency with the IceCube\nLED calibration system
    2. \n
  2. Early Sharing Prize - DynEdge - 1.046
    3. \n
", "rawMarkdown": "Notebook by @anjum48 employes the ice transparency data to train DynEdge from scratch.\n**Sample of ice_transparency data**\n```\ndepth scattering_len absorption_len\n1398.4 13.2 45.1\n1408.4 14.0 48.6\n1418.4 14.7 53.2\n```\nLet's understand what are these values : \n\nCherenkov photons are emitted with a characteristic wavelength dependence of 1 λ2 in the wavelength range of 300-600 nm, which includes the relevant sensitivity region of the photosensors. Photons are emitted in a cone around the direction of particle motion with an opening angle, determined by the speed of the particle and refractive index of the ice, of about 41◦ for relativistic particles. \nAs the photons propagate from the point of emission to the receiving sensor, they are affected by absorption and scattering in the ice. These propagation effects must be considered for both simulation and reconstruction of IceCube data and thus need to be carefully modeled. The important parameters to describe photon propagation in a transparent medium are: the average distance to absorption, the average distance between successive scatters of photons, and the angular distribution of the new direction of a photon at each given scattering point.\nPhotons are emitted by the LEDs in DOMs and recorded by other DOMs, as shown :\n\nThe geometrical scattering coefficient b determines the average distance between successive scatters (as 1/b). It is often more convenient to quote the effective scattering coefficient, be = b · (1 − {cos θ}), where θ is the deflection angle at each scatter. \nThe absorption coefficient a determines the average distance traveled by a photon before it is absorbed (as 1/a).\nThere `scattering_len` and `absorption_len` are the **average distance between successive scatters** and **average distance traveled by a photon before it is absorbed** respectively at particular depth.\n\nReferences : \n1. [Measurement of South Pole ice transparency with the IceCube\nLED calibration system](https://arxiv.org/pdf/1301.5361.pdf)\n2. [Early Sharing Prize - DynEdge - 1.046](https://www.kaggle.com/code/anjum48/early-sharing-prize-dynedge-1-046)", "votes": null }, { "id": "2151607", "postDate": "02/20/2023 07:39:20", "content": "

Thank you for the clarification.
\nI note that, as expected, these two characteristics are quite strongly correlated.

\n

\"\"

\n

Along the way, I have a question - what is the depth of the origin of the coordinates of the sensors from the dataset?

", "rawMarkdown": "Thank you for the clarification.\nI note that, as expected, these two characteristics are quite strongly correlated.\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F1394009%2Faa3a830835ae58b70e3cb62f08fba5af%2Fice_transparency.png?generation=1676878527673899&alt=media)\n\nAlong the way, I have a question - what is the depth of the origin of the coordinates of the sensors from the dataset?", "votes": null }, { "id": "2151654", "postDate": "02/20/2023 08:38:13", "content": "

Center of IceCube (depth of 1950 m, temperature −30.4◦ C); cf. ρ = 0.9167 at 1 atm. at 0◦ according to the resource

", "rawMarkdown": "`Center of IceCube (depth of 1950 m, temperature −30.4◦ C); cf. ρ = 0.9167 at 1 atm. at 0◦ ` according to the resource", "votes": null }, { "id": "2152234", "postDate": "02/20/2023 16:55:20", "content": "

There is one more interesting graph in the cited paper, describing the angular dependence of the DOM sensitivity. It was already discussed here that it's lower for the from-the-top direction of photons than for from-the-bottom one.
\nHere we can see this quantitatively, and what's interesting, the maximum photon acceptance is at angles close to cos η = 0.75, that is η close to 45° to vertical.
\n\"\"
\nIt seems to be the result of the optical properties of the ice in the holes containing the DOMs, which are quite different from the ancient ice that makes the bulk of the detector volume.

", "rawMarkdown": "There is one more interesting graph in the cited paper, describing the angular dependence of the DOM sensitivity. It was already discussed here that it's lower for the from-the-top direction of photons than for from-the-bottom one.\nHere we can see this quantitatively, and what's interesting, the maximum photon acceptance is at angles close to cos η = 0.75, that is η close to 45° to vertical. \n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F1581556%2Fd6eedfc91597033283f428473cd9ebc4%2Fdom_sens_angl.png?generation=1676911910335398&alt=media)\nIt seems to be the result of the optical properties of the ice in the holes containing the DOMs, which are quite different from the ancient ice that makes the bulk of the detector volume.", "votes": null }, { "id": "2153087", "postDate": "02/21/2023 07:25:27", "content": "

Thanks for sharing it @alexz0

", "rawMarkdown": "Thanks for sharing it @alexz0", "votes": null } ], "comments": [ { "id": 2151607, "author_name": "synset", "author_url": "", "post_date": "02/20/2023 07:39:20", "content": "

Thank you for the clarification.
\nI note that, as expected, these two characteristics are quite strongly correlated.

\n

\"\"

\n

Along the way, I have a question - what is the depth of the origin of the coordinates of the sensors from the dataset?

", "votes": null, "replies": [ { "id": 2151654, "author_name": "himanshuwagh", "author_url": "", "post_date": "02/20/2023 08:38:13", "content": "

Center of IceCube (depth of 1950 m, temperature −30.4◦ C); cf. ρ = 0.9167 at 1 atm. at 0◦ according to the resource

", "votes": null, "replies": [] } ] }, { "id": 2152234, "author_name": "alexz0", "author_url": "", "post_date": "02/20/2023 16:55:20", "content": "

There is one more interesting graph in the cited paper, describing the angular dependence of the DOM sensitivity. It was already discussed here that it's lower for the from-the-top direction of photons than for from-the-bottom one.
\nHere we can see this quantitatively, and what's interesting, the maximum photon acceptance is at angles close to cos η = 0.75, that is η close to 45° to vertical.
\n\"\"
\nIt seems to be the result of the optical properties of the ice in the holes containing the DOMs, which are quite different from the ancient ice that makes the bulk of the detector volume.

", "votes": null, "replies": [ { "id": 2153087, "author_name": "himanshuwagh", "author_url": "", "post_date": "02/21/2023 07:25:27", "content": "

Thanks for sharing it @alexz0

", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "2150356": "Notebook by @anjum48 employes the ice transparency data to train DynEdge from scratch.\n**Sample of ice_transparency data**\n```\ndepth scattering_len absorption_len\n1398.4 13.2 45.1\n1408.4 14.0 48.6\n1418.4 14.7 53.2\n```\nLet's understand what are these values : \n\nCherenkov photons are emitted with a characteristic wavelength dependence of 1 λ2 in the wavelength range of 300-600 nm, which includes the relevant sensitivity region of the photosensors. Photons are emitted in a cone around the direction of particle motion with an opening angle, determined by the speed of the particle and refractive index of the ice, of about 41◦ for relativistic particles. \nAs the photons propagate from the point of emission to the receiving sensor, they are affected by absorption and scattering in the ice. These propagation effects must be considered for both simulation and reconstruction of IceCube data and thus need to be carefully modeled. The important parameters to describe photon propagation in a transparent medium are: the average distance to absorption, the average distance between successive scatters of photons, and the angular distribution of the new direction of a photon at each given scattering point.\nPhotons are emitted by the LEDs in DOMs and recorded by other DOMs, as shown :\n\nThe geometrical scattering coefficient b determines the average distance between successive scatters (as 1/b). It is often more convenient to quote the effective scattering coefficient, be = b · (1 − {cos θ}), where θ is the deflection angle at each scatter. \nThe absorption coefficient a determines the average distance traveled by a photon before it is absorbed (as 1/a).\nThere `scattering_len` and `absorption_len` are the **average distance between successive scatters** and **average distance traveled by a photon before it is absorbed** respectively at particular depth.\n\nReferences : \n1. [Measurement of South Pole ice transparency with the IceCube\nLED calibration system](https://arxiv.org/pdf/1301.5361.pdf)\n2. [Early Sharing Prize - DynEdge - 1.046](https://www.kaggle.com/code/anjum48/early-sharing-prize-dynedge-1-046)", "2151607": "Thank you for the clarification.\nI note that, as expected, these two characteristics are quite strongly correlated.\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F1394009%2Faa3a830835ae58b70e3cb62f08fba5af%2Fice_transparency.png?generation=1676878527673899&alt=media)\n\nAlong the way, I have a question - what is the depth of the origin of the coordinates of the sensors from the dataset?", "2151654": "`Center of IceCube (depth of 1950 m, temperature −30.4◦ C); cf. ρ = 0.9167 at 1 atm. at 0◦ ` according to the resource", "2152234": "There is one more interesting graph in the cited paper, describing the angular dependence of the DOM sensitivity. It was already discussed here that it's lower for the from-the-top direction of photons than for from-the-bottom one.\nHere we can see this quantitatively, and what's interesting, the maximum photon acceptance is at angles close to cos η = 0.75, that is η close to 45° to vertical. \n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F1581556%2Fd6eedfc91597033283f428473cd9ebc4%2Fdom_sens_angl.png?generation=1676911910335398&alt=media)\nIt seems to be the result of the optical properties of the ice in the holes containing the DOMs, which are quite different from the ancient ice that makes the bulk of the detector volume.", "2153087": "Thanks for sharing it @alexz0" }, "source": "meta" }