Towards a historical precipitation database for West Africa: Overview, quality control and harmonization. (29th December 2021)

• Record Type: Journal Article
• Authors: Bliefernicht, Jan; Salack, Seyni; Waongo, Moussa; Annor, Thompson; Laux, Patrick; Kunstmann, Harald
• Is Part Of: International journal of climatology. Volume 42:Number 7(2022)
• Page Start: 4001
• Page End: 4023
• Publication Date: 2021-12-29
• Subjects: climate observations -- geostatistics -- precipitation -- quality control -- West Africa; Climatology -- Periodicals; Climat -- Périodiques; Climatologie -- Périodiques; 551.605
• DOI: 10.1002/joc.7467
• Languages: English
• ISSNs: 0899-8418

Abstract

Abstract: Reliable long‐term observations from precipitation stations are often required for climatological studies but are strongly limited in many regions of the world. To improve this limitation for West Africa, we compiled daily and monthly observations from more than 20 national, continental and global databases, to establish a historical precipitation archive with a focus on four countries (Burkina Faso, Ghana, Benin and Togo). The new archive contains long‐term daily and monthly precipitation measurements from 1819 to 2013 for more than 1, 000 sites. It is, therefore, the most comprehensive historical dataset with daily and monthly precipitation observations for this region. To produce a quality‐controlled and harmonized precipitation dataset for the focal region, various statistical algorithms have been implemented. These algorithms rely on straightforward geostatistical approaches (e.g., spatial correlograms) and corresponding statistical tests for identification and elimination of unreliable time series, in addition to various standard approaches used by global data centers. Although the quality control revealed various data errors and uncertainties for measurements and meta‐information (e.g., unit conversion errors, temporal offsets, frequent and long data gaps), a spatial interpolation using the quality‐controlled and harmonized dataset produced relatively reliable precipitation patterns for different target variables (e.g., monthly precipitation amount and daily precipitation probability). A major remaining challenge is providing free access to this database for research and other noncommercial purposes, due to national data protection regulations. However, several further tasks have been initiated and implemented (e.g., free provision of gridded precipitation datasets and point statistics) to improve the access and availability of station‐based precipitation observations and related data products for this climatologically challenging region. Abstract : This study presents the most comprehensive station‐based precipitation database for West Africa, with more than 1, 000 long‐term daily and monthly time series (1819–2013) compiled from national data archives, among others. Various statistical algorithms have been implemented, using novel geostatistical and standard approaches, in order to produce an initial quality‐controlled and harmonized dataset. Although the quality control revealed various uncertainties, a spatial interpolation using the new database produced relatively reliable precipitation patterns for the main monsoon period.