METEORv1.6: Spatial climate variability and integrated impact emulation

Abstract. Climate impact assessment increasingly requires spatially explicit projections with realistic temporal variability at sub-annual resolution. METEORv1.6 extends the established METEOR spatial multi-timescale, multi-forcer climate emulation framework with two major capabilities: (1) a monthly climate variability model that generates realistic sub-annual climate sequences with seasonal cycles and inter-annual variability, enabling the generation of ensemble projections and (2) a modular impact assessment framework that translates climate projections into impact metrics. The monthly climate model represents seasonal harmonics and noise, while preserving covariance structures from the source model. Seasonal cycles are represented through harmonic analysis with temperature-dependent parameterization, enabling non-stationary simulation of seasonal timing shifts under warming. Principal Component Analysis is used to decompose monthly anomalies into spatial modes, then their temporal evolution and climate variability is modeled using Vector Autoregressive with eXogenous variables (VARX) processes. The impact assessment framework provides a standardized interface for ensemble processing and uncertainty quantification through a modular system of impact calculators. The initial case implementation includes heating and cooling degree days calculations which are key drivers in estimating energy sector demand, demonstrating ensemble-based uncertainty propagation from climate projections to impact metrics. Validation against CMIP6 data demonstrates that METEORv1.6 accurately reproduces statistical properties of monthly climate variability for a range of future scenarios when trained on a single scenario from an Earth System Model (together with a CO₂ quadrupling idealized experiment). The integrated impact framework enables rapid generation of probabilistic climate risk assessments suitable for sectoral applications, bridging the gap between global climate projections and local decision-making needs. The open-source implementation supports broad adoption and continued expansion to additional impact domains.