Researchers have digitized more than 9,000 handwritten pages of daily meteorological records from 37 weather stations in the Democratic Republic of Congo, revealing important new insights into regional temperature changes and historical warming in the congo. Covering the period from 1908 to 2023, the data was recovered from paper archives stored in Yangambi using a specialized machine learning transcription tool called MeteoSaver. This comprehensive climate data rescue helps scientists better understand historical temperature shifts in a previously data-sparse region.
How did researchers recover the historical climate data?
Researchers utilized a machine learning tool named MeteoSaver to transcribe more than 9,000 handwritten pages of daily meteorological records stored on paper in Yangambi. These documents covered 37 weather stations across the Democratic Republic of Congo and spanned from 1908 to 2023, overcoming significant archiving challenges to build a continuous digital climate history.
What do the new temperature findings show?
The digitized records indicate that between 1961 and 1990, average temperatures in the Democratic Republic of Congo rose by 0.5 degrees Celsius, while maximum temperatures increased by about 0.7 degrees Celsius. Furthermore, data analysis revealed that extremely hot days were nearly twice as frequent in the 1980s as they were in the 1960s.
