Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Run Dependent Monte Carlo at Belle II

Giovanni Gaudino

TL;DR

This paper presents run-dependent Monte Carlo (MCrd) as a solution to accurately model time-dependent detector conditions and beam backgrounds in Belle II data analyses. It details a comprehensive workflow that integrates run-specific detector payloads, data-driven background overlays, and a grid-based production infrastructure, contrasting MCrd with the traditional run-Independent MC (MCri). Key contributions include the classification of physics channels into Generic and Signal samples, a robust background-overlay strategy, per-run detector configuration management, and a metadata-rich data-management layer that supports reproducible analyses. The approach significantly improves data-simulation fidelity and reduces systematic uncertainties, demonstrated by Belle II’s large data set and substantial MC production, with MCrd samples reaching approximately $1.7×10^{3}$ fb⁻¹ equivalent for high-multiplicity channels alongside $500$ fb⁻¹ of data collected to date, as reported at EPS-HEP 2025.

Abstract

The Belle II experiment at the SuperKEKB accelerator in Tsukuba, Japan, searches for physics beyond the Standard Model, with a focus on precise measurements of flavor physics observables. Highly accurate Monte Carlo simulations are essential for this endeavor, as they must correctly model the variations in detector conditions and beam backgrounds that occur during data collection. To meet this requirement, the "run-dependent" Monte Carlo has been developed. This approach incorporates time-dependent detector conditions and beam-induced backgrounds collected via random triggers, allowing for different conditions with a granularity of just a few hours. In this document, we will discuss the procedures and challenges associated with producing run-dependent Monte Carlo simulations for Belle II. We will also highlight the improvements these simulations offer over traditional "run-independent" Monte Carlo methods.

Run Dependent Monte Carlo at Belle II

TL;DR

This paper presents run-dependent Monte Carlo (MCrd) as a solution to accurately model time-dependent detector conditions and beam backgrounds in Belle II data analyses. It details a comprehensive workflow that integrates run-specific detector payloads, data-driven background overlays, and a grid-based production infrastructure, contrasting MCrd with the traditional run-Independent MC (MCri). Key contributions include the classification of physics channels into Generic and Signal samples, a robust background-overlay strategy, per-run detector configuration management, and a metadata-rich data-management layer that supports reproducible analyses. The approach significantly improves data-simulation fidelity and reduces systematic uncertainties, demonstrated by Belle II’s large data set and substantial MC production, with MCrd samples reaching approximately fb⁻¹ equivalent for high-multiplicity channels alongside fb⁻¹ of data collected to date, as reported at EPS-HEP 2025.

Abstract

The Belle II experiment at the SuperKEKB accelerator in Tsukuba, Japan, searches for physics beyond the Standard Model, with a focus on precise measurements of flavor physics observables. Highly accurate Monte Carlo simulations are essential for this endeavor, as they must correctly model the variations in detector conditions and beam backgrounds that occur during data collection. To meet this requirement, the "run-dependent" Monte Carlo has been developed. This approach incorporates time-dependent detector conditions and beam-induced backgrounds collected via random triggers, allowing for different conditions with a granularity of just a few hours. In this document, we will discuss the procedures and challenges associated with producing run-dependent Monte Carlo simulations for Belle II. We will also highlight the improvements these simulations offer over traditional "run-independent" Monte Carlo methods.
Paper Structure (18 sections, 1 figure, 1 table)

This paper contains 18 sections, 1 figure, 1 table.

Table of Contents

  1. Introduction
  2. Data Production Workflow
  3. Data Transfer and Grid Registration
  4. Reconstruction and Calibration Streams
  5. Monte Carlo Production Framework
  6. Run-Independent Monte Carlo (MCri):
  7. Run-Dependent Monte Carlo (MCrd):
  8. Run-Dependent Monte Carlo Production
  9. Physics Channel Classification
  10. Generic Samples
  11. Signal Samples
  12. Background Modeling and Overlay
  13. Detector Configuration Management
  14. Grid-Based Production Infrastructure
  15. Data Management and Book-keeping
  16. ...and 3 more sections

Figures (1)

  • Figure 1: Schematic overview of the Belle II Grid computing infrastructure, illustrating the data flow from raw acquisition through reconstruction to analysis-ready datasets.