Integrating Historical Records and Climate Models for Temperature Reconstruction in East Asia (1403-1911)
Conference
Format: IPS Abstract
Keywords: bayesian approach, kalman_filter, kriging, penalized_maximum_likelihood, quantile_mapping
Session: Invited Session 11A - Advanced Statistical Methods for Climate Data Analysis
Thursday 5 December 3:30 p.m. - 5 p.m. (Australia/Adelaide)
Abstract
This study introduces a novel methodology for reconstructing annual temperature profiles in East Asia from 1403 to 1911, utilizing the Reconstructed East Asian Climate Historical Encoded Series (REACHES), a database derived from historical Chinese documents. The lack of instrumental data for this period has long posed a significant challenge, which we address by leveraging the climatic data recorded at four ordinal levels within the REACHES database. These index-based records are primarily biased toward extreme weather events, resulting in data gaps during normal conditions. To overcome this, we propose a three-tiered statistical framework. First, we apply annual kriging to interpolate temperature data across East Asia, assuming a zero mean to account for missing information. Next, we use quantile mapping to align the kriged REACHES data with Celsius scales, employing the Last Millennium Ensemble (LME) reanalysis data. Finally, we introduce an innovative Bayesian data assimilation technique that integrates the kriged Celsius REACHES data with LME simulations, enhancing the accuracy of the reconstructions. This comprehensive integration of historical records, contemporary climate models, and advanced statistical methods not only improves the precision of historical temperature reconstructions but also provides a valuable resource for future climate research.