A Semi-Quantitative, Regional Precipitation Data Set for Studying African Climates of the Nineteenth Century, Part I. Overview of the Data Set

Abstract

This article describes a historical archive of proxy and actual precipitation data that extends the African climate record back to the early nineteenth century. The `proxy' archive includes verbal, documentary references which contain information related to rainfall conditions, such as references to famine, drought, agriculture or the nature of the rainy season. The precipitation archive includes all observations made in Africa during the nineteenth century. It consists of records for 60 stations in Algeria, 87 stations in South Africa and 304 stations scattered over the rest of Africa. Information is particularly plentiful for the 1880s and 1890s. The two parts have been be combined into a semi-quantitative regional data set indicating annual rainfall conditions in terms of anomaly classes (e.g., normal, dry, wet). This data set extends from the early nineteenth century to 1900 and distinguishes seven anomaly classes, using numbers ranging from –3 to +3 to represent very wet, wet, good rains, normal, dry, drought, and severe drought. The regionalization is based on 90 geographical regions shown via studies of the modern precipitation record to be climatically homogeneous with respect to the interannual variability of rainfall. The regional aggregation allows the voluminous fragmentary information available in historical sources to be used systematically to produce multi-year time series that can be directly integrated into the modern record for each region. The resultant time series can also be subjected to statistical analysis, in order to investigate nineteenth century climate over Africa. Spatial detail is added to the data set by utilizing a unique methodology based on climatic teleconnections established from studies of rainfall variability over Africa. The historical information and station records have been combined into a file containing a regional anomaly value for up to 90 geographic regions and the years 1801–1900. Gaps necessarily remain in the matrix, but as early as the 1820s over 40 regions are represented. By the 1880s generally around 70 regions or more are represented.