Spatiotemporal variability and uncertainty in wood decay estimates across Europe and Scandinavia using a data-driven moisture content model with multi-scale weather datasets

• Reanalysis data driven moisture content model enabled high-resolution wood decay analyses • Point to grid comparing of decay indicated biases variable by micro to macro climate conditions • High-resolution decay hazard maps reveal significant spatiotemporal variability • Vertical boards show higher spatial uncertainty due to wind-driven rain exposure • Uncertainty mapping enables uncertainty propagation for service life assessment The use of remote sensing and reanalysis weather data has not been adequately acknowledged for large-scale built environment applications, primarily regarding engineering analyses that traditionally rely on computationally expensive numerical models. This study employs a parsimonious data-driven approach concerning service life assessment of rain-exposed wooden components, enabling mapping and assessment of related risks across Europe and Scandinavia from gridded data. Wood moisture content is the primary vector for fungal decay and accurate modeling is essential for the design of wooden commodities. Numerical models of moisture flow within woody material provide accuracy but are impractically slow for large-scale applications. We explore high-resolution decay risk assessments using a neural network (NN) trained on existing numerical predictions of moisture content. Specifically, the model used time-lagged gridded weather data as inputs to calculate moisture content at grids of 5.5 (1.0) km resolution over 21-36 (6) years across Europe (Scandinavia), compared with similar calculations from 890 weather stations. The moisture data were subsequently used to predict decay rates and quantify uncertainties stemming from temporal and spatial variation in weather. The results present high-resolution decay hazard maps alongside the associated spatiotemporal variations. Both spatial and temporal variations were identified as significant contributors to uncertainty in decay assessments, with their impacts varying considerably across regions in Europe. The study is part of an effort to integrate uncertainty propagation in service life prediction of wood, to enhance the resilience and reliability of wood in outdoor applications and to optimize its potential for sustainable and enduring construction solutions.