Question about the calculation of J moments

[chen-ying-ying]

Hi, I have some questions for you. Does the J-plots algorithm have any special requirements on the format of the input real observation data? The data I need to analyse now is 2D PP data (3D PPV data obtained by integrating in the velocity direction). According to what is stated in J-plots: a new method for characterising structures in the interstellar medium, the molecular cloud core should belong to the centrally concentrated type of shapes, and then most of the calculated J-values should be in the first quadrant (i.e., J1 is greater than 0 and J2 is greater than 0). However, I have tried a lot of data and the value of J2 is less than 0, but I don't know what is wrong.

[chen-ying-ying]

I see that you mentioned in the code notes that you can use the mask obtained after detection as an input to the J-plots, so how should I calculate the J moments?

[SJaffa]

Hi @chen-ying-ying If J1 is positive and J2 is negative, this indicates that your input shape is elongated, like a filament. If J2 is only slightly negative it may also have some central concentration like a core. It isn't strict cut-off at J2=0, more of a gradient, so shapes will be more centrally concentrated or more filamentary as they move around the J-space. If you expect your shapes to be cloud cores and are getting J-values that don't indicate that, you might want to try a different segmentation method, using a higher threshold, or using the RJ-plot method that rotates the J-values to give a better differentiation of filamentary structure.

[chen-ying-ying]

Thank you for your reply! After receiving your reply I tried to use RJ-plots to analyse the molecular cloud clumps detected by the algorithm, and the results are somewhat improved compared to J-plots. Meanwhile, I have another question for you, using the parabola in RJ-plots to distinguish between quasi-circular and elongated structures, does this parabolic equation need to be fitted based on the actual observation data? If it needs to be fitted, what approach should be taken to fit the parabola. I hope to get your help and reply!

捂风挽笑 @.***

 

------------------ 原始邮件 ------------------ 发件人: "SJaffa/Jplots" @.>; 发送时间: 2024年9月27日(星期五) 下午4:34 @.>; @.**@.>; 主题: Re: [SJaffa/Jplots] Question about the calculation of J moments (Issue #5)

Hi @chen-ying-ying If J1 is positive and J2 is negative, this indicates that your input shape is elongated, like a filament. If J2 is only slightly negative it may also have some central concentration like a core. It isn't strict cut-off at J2=0, more of a gradient, so shapes will be more centrally concentrated or more filamentary as they move around the J-space. If you expect your shapes to be cloud cores and are getting J-values that don't indicate that, you might want to try a different segmentation method, using a higher threshold, or using the RJ-plot method that rotates the J-values to give a better differentiation of filamentary structure.

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: @.***>

[SJaffa]

Section 2.2 in the RJ-plots paper (linked in my previos comment) explains how the parabola was derivied - it was fitted to shapes with and aspect ratio (A) of 2, as this is a commonly used cut-off to separate filamentary structures from just slightly elongated clouds. It is an arbitrary definition and you could use a different cut-off if your data contains different kinds of structures with different aspect ratios, but it is probably best to stick with the published definition unless you have a strong reason to change.

The original motivation for that choice was to isolate filamentary structure so if that is not your aim then this method might not be appropriate.

[SJaffa]

If you want to discuss RJ-plots further I would recommend you contact the corresponding author of that paper. He might have more time to help you than me as I am no longer working in the field.

[chen-ying-ying]

Thank you for your answer. My main research focus is on molecular cloud nucleation on small scales, and for the data currently used, the theory is that there is a majority of central concentration (visual observation). I'll continue to use the definitions disclosed in the paper for my analyses. Again, many thanks for taking the time to reply.

捂风挽笑 @.***

 

------------------ 原始邮件 ------------------ 发件人: "SJaffa/Jplots" @.>; 发送时间: 2024年10月26日(星期六) 凌晨2:12 @.>; @.**@.>; 主题: Re: [SJaffa/Jplots] Question about the calculation of J moments (Issue #5)

Section 2.2 in the RJ-plots paper (linked in my previos comment) explains how the parabola was derivied - it was fitted to shapes with and aspect ratio (A) of 2, as this is a commonly used cut-off to separate filamentary structures from just slightly elongated clouds. It is an arbitrary definition and you could use a different cut-off if your data contains different kinds of structures with different aspect ratios, but it is probably best to stick with the published definition unless you have a strong reason to change.

The original motivation for that choice was to isolate filamentary structure so if that is not your aim then this method might not be appropriate.

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: @.***>