On the variations in the frequency of 25–70-day intraseasonal oscillations in Central Africa using wavelet-based indices

Research Article On the variations in the frequency of 25–70‑day intraseasonal oscillations in Central Africa using wavelet‑based indices Alain Tchakoutio Sandjon1,2,3 · Angennes Lucie Djiotang Tchotchou2 · Derbetini Appolinaire Vondou2 · Armand Joel Komkoua Mbienda2,4 · Guy Merlin Guenang2,4 · Roméo Stève Tanessong2,5 · Armand Nzeukou Takougang3 Received: 28 October 2020 / Accepted: 27 January 2021 © The Author(s) 2021  OPEN Abstract Atmospheric variability at the intraseasonal timescale remains of great concern in tropical Africa because of the vulnerability of the population to variations in the distribution and amount of rainfall within a season. Then, the parameterization of the processes that induce the intraseasonal variability of the rainfall is still a challenge for the sub-seasonal-to-seasonal forecast in the tropics. In the study of intraseasonal oscillations (ISOs) in Central Africa, almost all of the authors focused only on the amplitude of the oscillations, even though the frequency is also very important because it also undergoes strong spatiotemporal variations. The novelty of this study is that we applied wavelet transform on the 2.5° × 2.5° daily Outgoing Long-wave Radiation (OLR) to extract the frequency (period) of intraseasonal oscillations (ISO) and then study its spatiotemporal variations over Central Africa (CA) within the period 1981–2015 (35 years). By the algorithm used, we obtained a dataset of daily ISO Period Indices (ISOPI) within the study period, with the same dimensions as the original OLR datasets. The analyses showed that the mean ISOPI globally fluctuates between 32 and 52 days, but undergoes strong day-to-day variations. The ISO frequency is highly seasonal, with high ISOPI (low frequency) during December–February and June–August, and short low ISOPI (high frequency) during March–May and September–November. The composites of OLR and 850 hpa zonal winds revealed that the low-frequency ISOs (LFISOs) are predominant in Eastern Central Africa and around the Cameroon Volcanic Line, while the long-frequency events (HFISOs) are mostly found in Western Central Africa, especially around the Congo basin. The plots of yearly mean ISOPI showed that the ISO period exhibits strong interannual variations with years of very high ISOPI such as 1983, 1985, 1987, 1989, 1999, 2002 and 2009, and years of lower ISOPI as 1988, 1994, 1995. Finally, it was proved in this study that there is an enhancement of rainfall during LFISOs, especially in northern hemisphere, while HFISOs are generally associated with normal or suppressed rainfall regime. Keywords Rainfall · Intraseasonal oscillations · Frequency · Central Africa · Wavelet analysis * Alain Tchakoutio Sandjon, | 1Department of Computer Science Including Basic Sciences, Higher Technical Teacher’s Training College Kumba, University of Buea, Buea Road, P.O box 249, Kumba, Cameroon. 2Laboratory of Environmental Modelling and Atmospheric Physics, Department of Physics, Faculty of Science, University of Yaoundé I, Yaoundé, Cameroon. 3Laboratory of Industrial Systems and Environmental Engineering, Fotso Victor University Institute of Technology, University of Dschang, Bandjoun, Cameroon. 4Laboratory of Mechanics and Modeling of Physical Systems (L2MPS), Department of Physics, Faculty of Science, University of Dschang, Dschang, Cameroon. 5School of Wood, Water and Natural Resources, Faculty of Agronomy and Agricultural Sciences, University of Dschang, Ebolowa, Cameroon. SN Applied Sciences (2021) 3:304 | https://doi.org/10.1007/s42452-021-04285-1 Vol.:(0123456789) Research Article SN Applied Sciences (2021) 3:304 | https://doi.org/10.1007/s42452-021-04285-1 1 Introduction In Central Africa (CA), rainfall is a crucial resource because the socioeconomic activities of the population such as agriculture, livestock and energy are highly rainfall dependent. In recent years, for instance, many African countries have been affected by rainfall variability and long-term changes, in terms of amount and spatial and temporal distributions. The distribution of rainfall within the season is then of great interest to the local socioeconomic actors since it allows them to plan their activities throughout the year. In terms of climate, CA (15°S–15°N; 0–50°E) is a particular region because of its geographical location, topography and surface cover. It extends mainly over the land and part of Atlantic and Indian oceans on its edges. The topography of the region is also diversified, including highlands, mountains and plateaus. The gradient in the surface elevation between the East and West boundaries can reach up to 3000 m (Fig. 1a). CA features varied vegetation types ranging from desert landscapes to humid tropical forest (Fig. 1b). Rainfall distribution is also varied, exhibiting a strong gradient from western to the eastern boundaries of the region (Fig. 1c). The Congo basin embedded in CA was proved to be the third most extensive region of deep convection, globally, after the West Pacific warm pool region and Amazonia [1, 2]. Rainfall variability in CA is complex, ranging from diurnal cycle of rainfall to the interannual variability [3–5]. However, the intraseasonal timescale is of relatively great interest because it provides the information about the distribution of rainfall within the season, allowing the farmers to plan their activities throughout the year. Intraseasonal variability (ISV) refers to the atmospheric oscillations with that a period less than the length of a season (less than 90 days). But practically, to make a difference with the synoptic scale (less than about 10 days), the intraseasonal variability is more often referred to the cycle with a timescale between the synoptic scale and a season (10–90 days). Until the years 1960s, in many climate variability studies, the authors focused mainly on the annual cycle of rainfall as well as the interannual variability because they were limited by the poor resolution of the data available those days [6, 7]. But since the beginning of 1970s, the advent of satellite-based rainfall products and advanced mathematical tools used in signal processing allowed the investigation of rainfall variability in the tropics at shorter timescales. The notion of ISV raised in the scientific community in the early 1970s, and since then, many authors used different datasets and techniques to document the ISV of some atmospheric variables in different geographical areas around the world (e.g., [8–12]). Vol:.(1234567890) (a) (b) (c) Fig. 1  a Surface elevation over the study area based on 30-min topographic data (meter) from digital elevation model (DEM) of the US Geological Survey. The study area is shown as solid box. b Mean distribution of vegetation (NDVI), the density of the vegetation here can be interpreted as the green level of the surface. c Annual mean rainfall (mm), the plot is based 2.5° × 2.5° pentad GPCP precipitation for the period 1981–2000 Unfortunat (...truncated)