Theoretical and Applied Climatology

, Volume 119, Issue 1, pp 255–272

Assessing the capability of CORDEX models in simulating onset of rainfall in West Africa

  • Moussa S. Mounkaila
  • Babatunde J. Abiodun
  • J. ‘Bayo Omotosho
Original Paper

DOI: 10.1007/s00704-014-1104-4

Cite this article as:
Mounkaila, M.S., Abiodun, B.J. & ‘Bayo Omotosho, J. Theor Appl Climatol (2015) 119: 255. doi:10.1007/s00704-014-1104-4

Abstract

Reliable forecasts of rainfall-onset dates (RODs) are crucial for agricultural planning and food security in West Africa. This study evaluates the ability of nine CORDEX regional climate models (RCMs: ARPEGE, CRCM5, RACMO, RCA35, REMO, RegCM3, PRECIS, CCLM and WRF) in simulating RODs over the region. Four definitions are used to compute RODs, and two observation datasets (GPCP and TRMM) are used in the model evaluation. The evaluation considers how well the RCMs, driven by ERA-Interim reanalysis (ERAIN), simulate the observed mean, standard deviation and inter-annual variability of RODs over West Africa. It also investigates how well the models link RODs with the northward movement of the monsoon system over the region. The model performances are compared to that of the driving reanalysis—ERAIN. Observations show that the mean RODs in West Africa have a zonal distribution, and the dates increase from the Guinea coast northward. ERAIN fails to reproduce the spatial distribution of the RODs as observed. The performance of some RCMs in simulating the RODs depends on the ROD definition used. For instance, ARPEGE, RACMO, PRECIS and CCLM produce a better ROD distribution than that of ERAIN when three of the ROD definitions are used, but give a worse ROD distribution than that of ERAIN when the fourth definition is used. However, regardless of the definition used, CCRM5, RCA35, REMO, RegCM3 and WRF show a remarkable improvement over ERAIN. The study shows that the ability of the RCMs in simulating RODs over West Africa strongly depends on how well the models reproduce the northward movement of the monsoon system and the associated features. The results show that there are some differences in the RODs obtained between the two observation datasets and RCMs, and the differences are magnified by differences in the ROD definitions. However, the study shows that most CORDEX RCMs have remarkable skills in predicting the RODs in West Africa.

Supplementary material

704_2014_1104_MOESM1_ESM.doc (1.6 mb)
ESM 1(DOC 1,589 kb)

Copyright information

© Springer-Verlag Wien 2014

Authors and Affiliations

  • Moussa S. Mounkaila
    • 1
    • 2
  • Babatunde J. Abiodun
    • 1
  • J. ‘Bayo Omotosho
    • 2
  1. 1.Climate Systems Analysis Group, Department of Environmental and Geographical ScienceUniversity of Cape TownCape TownSouth Africa
  2. 2.West African Science Service Center for Climate Change and Adapted Land Use (WASCAL) Graduate Research ProgramFederal University of Technology AkureAkureNigeria