Forecasting LSST Cosmology: Building Pipelines in the Era of Systematics

Abstract

The Vera C. Rubin Observatory’s LSST promises unprecedented cosmological constraints, but achieving them requires more than just statistical power—it demands forecasting pipelines that can account for complex astrophysical systematics and modeling challenges on small scales. In this talk, I present work within the LSST Dark Energy Science Collaboration (DESC) to develop a modular forecasting framework that connects realistic data modeling with infrastructure built for extensibility and validation. Drawing on my role as forecasting group lead, I will outline key challenges in pipeline design, the role of validation in maintaining forecast credibility, and the use of good coding practices—such as modularization and model registries—to ensure long-term adaptability. I’ll close with a look at new directions, including forecasts at high redshift and multi-probe combinations, underscoring how thoughtful infrastructure enables reliable science in the LSST era.