Problem 2611

Summary:	Anisotropic angle distribution in isotropic flux
Product:	Geant4	Reporter:	Yaroslav Strilchuk <ditegen>
Component:	event	Assignee:	Makoto Asai <asai>
Status:	CLOSED INVALID
Severity:	minor
Priority:	P4
Version:	11.2
Hardware:	All
OS:	All

Description Yaroslav Strilchuk 2024-05-01 20:53:21 CEST

In G4SPSAngDistribution::GenerateIsotropicFlux(), cosine of a uniformly distributed random angle is not calculated in the proper way:

https://github.com/Geant4/geant4/blob/master/source/event/src/G4SPSAngDistribution.cc#L348

costheta = std::cos(MinTheta) - rndm * (std::cos(MinTheta) - std::cos(MaxTheta));

Suggested fix:

Theta = MinTheta - rndm * (MinTheta - MaxTheta);
costheta = std::cos(Theta);

The bug affects at least GPS's sources with angular distribution set to "iso" and results in primaries of the expected isotropic flux being less likely oriented towards the poles.

Comment 1 Makoto Asai 2024-05-01 21:17:29 CEST

Theta is azimuth angle. If it is not biased, it should distribute between 0 to pi/2, not evenly but in cosine shape. 

The current implementation is correct. If theta distributes evenly as you suggest, tracks would concentrate to the polar region.

Comment 2 Makoto Asai 2024-05-01 21:20:49 CEST

Sorry, "0 to pi/2" should read "0 to pi".

Comment 3 Yaroslav Strilchuk 2024-05-01 23:49:29 CEST

Thank you, my bad.