Press Releases
JAMSTEC
On the predictability of the extreme drought
in East Africa during the short rains season
~Key roles of the negative Indian Ocean Dipole~
1. Key Points
- ◆
- East Africa suffered from an extreme drought during the short rains season of October–December 2021.
- ◆
- The 2021 negative Indian Ocean Dipole was responsible for the devastating drought and played a key role in its seasonal predictability.
- ◆
- A hybrid statistical-dynamical prediction of drought in East Africa on longer lead time may help people to take mitigation measures.
2. Abstract
Many parts of East Africa experienced extremely dry conditions during the short rains season (October–December) of 2021. Such a devastating drought in East Africa leads to unsafe drinking water, food insecurity, and other socio-economic issues. Takeshi Doi (Senior researcher at Application Laboratory, VAiG, JAMSTEC) and his research team showed that the extreme drought in East Africa was predicted a few months earlier by the large-ensemble seasonal prediction system based on the SINTEX-F climate model (※1). They also found the 2021 negative Indian Ocean Dipole (IOD) was responsible for these unusually dry conditions over East Africa. The IOD, an intrinsic ocean–atmosphere coupled climate phenomenon in the tropical Indian Ocean, is characterized by cold (warm) sea surface temperature anomalies and reduced (enhanced) rainfall in the western (eastern) tropical Indian Ocean during its negative phase (Fig. A). Some of the devastating droughts (severe floods) in East Africa are observed during negative (positive) IOD years, respectively. In addition, they demonstrated that a hybrid statistical-dynamical framework is more skillful than the SINTEX-F model at predicting drought in East Africa over a longer lead time. This skillful prediction will help people to take necessary mitigation measures to reduce the devastating impact of a drought in future.
This paper is published in Geophysical Research Letters on November 22, 2022.
- Title：
- On the predictability of the extreme drought in East Africa during the short rains season
- Authors：
- Takeshi Doi1, Swadhin K. Behera1, and Toshio Yamagata1,2
- Affiliation：
- 1Application Laboratory (APL)/Research Institute for Value-Added-Information Generation (VAiG)/Japan Agency for Marine-Earth Science and Technology (JAMSTEC), Yokohama 236-0001, Japan
2Institute of Climate and Application Research (ICAR), Nanjing University of Information Science and Technology (NUIST), Nanjing 210044, China - DOI：
- 10.1029/2022GL100905
Fig. A: Schematic of a negative IOD event. Sea surface temperature (SST) anomalies are shaded (red color is for warm anomalies and blue is for cold). White patches indicate increased convective activities and arrows indicate anomalous wind directions during negative IOD events.
3. Background
Many parts of East Africa experienced extremely dry conditions during the short rains season (October–December) of 2021, and suffered severe food insecurity outcomes afterwards (e.g. the Famine Early Warning Systems Network at https://fews.net/east-africa/). Regarding the time series of East African short rains anomalies within 10°S–10°N; 30°E–45°E (hereafter, the EASR index) during October–December, the 2021 drought reached a level that is comparable to the extreme drought events in 1996, 1998, 2005, and 2016 in the past 40 years. Such devastating droughts in East Africa always led to unsafe drinking water, food insecurity, and resurgence of infectious diseases. Thus, successful prediction of such an extreme drought at least a few months ahead may contribute to reducing the socio-economic losses by taking the necessary mitigation measures. Such a research stream is critically important because extreme impacts due to natural climate variability are becoming more serious with ongoing global warming.
4. Results
The observed drier-than-normal condition in East Africa during October–December of 2021 was successfully predicted using the SINTEX–F model from early September initialization (Fig. 1). Although the ensemble mean prediction has captured only about 25% of the observed EASR index, the successful prediction of the dry phase itself gives us an opportunity to explore potential sources of predictability.
The SINTEX-F ensemble prediction system with 108-member has an advantage in finding possible processes influencing noisy predictions of precipitation. We found that ensemble members that predicted stronger negative IODs tended to predict severe droughts in East Africa (Fig. 2). However, the co-occurrence of La Niña in the prediction did not seem to be related to the predictions of the drought in East Africa. Therefore, it suggests that the 2021 negative Indian Ocean Dipole (IOD) was responsible for these unusually dry conditions over East Africa.
A longer-lead prediction of anomalous climate conditions is a prerequisite for early warning systems that prepare society for extreme weather events. However, it is still challenging to predict the short rains with the available state-of-the-art dynamical prediction systems. But there is a practical way to compensate for this shortcoming. This is because we observe a high correlation between the Indian Ocean Dipole Mode index (DMI) and the EASR. Utilizing this aspect, we have proposed a hybrid prediction system by using the combination of the dynamical DMI prediction model and the regression model based on the DMI-EASR relationship. This new hybrid system turns out to be quite promising (Fig. 3).
5. Future Perspectives
The outcome of the present study may help society take the necessary mitigation measures to reduce the devastating impact of the drought on the human population in East Africa. However, it is not enough; we definitely need improvements to adaptive capacity, infrastructure, and forecast communications. For example, Early Warning of Food Insecurity (https://fews.net/sectors-topics/approach/early-warning-food-insecurity) estimates food security outcomes for the coming eight months and reports are issued in February, June and October. The information may help governments and other humanitarian stakeholders quickly identify an evolving crisis (or potential crisis) and take necessary mitigation measures. We believe that collaboration with such an existing dissemination system will be a good step forward.
Persistent droughts in East Africa are very damaging to the local socio-economic conditions. Based on this research, we have recently begun to provide a more detailed forecast for East African rainfall during the upcoming short rains season on the SINTEX-F website (see https://www.jamstec.go.jp/aplinfo/sintexf/e/seasonal/outlook.html). Going forward, we have noted that East Africa may experience drier than normal conditions again during the short rains season of 2022, owing to the negative IOD, as similar to the 2021 drought (Fig. 4).
【Supplementary Information】
- ※1
- The large-ensemble seasonal prediction system based on the SINTEX-F climate model: The SINTEX-F is an ocean-atmosphere-sea ice-land coupled general circulation model developed on the JAMSTEC super-computer “Earth Simulator” (https://www.jamstec.go.jp/es/en/) under the international research collaboration with European researchers. The model is written with a series of numerical program codes that describe physical and dynamical processes of the climate systems including ocean-atmosphere-sea ice-land. As the initialization scheme, the modeled sea surface temperatures are strongly nudged toward the satellite observations in the coupled run with subsurface ocean data assimilation using all types of ocean profiling instruments that provide temperature and salinity (when available) from the expendable bathythermographs (XBTs), mooring buoys, sea stations, Argo floats, etc. We employ 108 ensemble members to reduce the prediction uncertainties associated with different initial conditions and physical schemes.
Figure 1. (a) Precipitation anomaly from the observational data averaged in October–December 2021 (mm day-1). The East African short rains (EASR) index and the Indian Ocean Dipole Mode index regions are shown by blue and black boxes, respectively. (b) Same as (a), but for the prediction issued in early September 2021 (108-ensemble mean). Although the prediction underestimated the observed EASR index, the successful prediction of the dry phase deserves to explore potential sources of predictability.
Figure 2. Scatter plot of 108-members ensemble predictions by the SINTEX-F between the East African short rains (EASR) index (x-direction, mm day-1) and the Indian Ocean Dipole Mode index (y-direction, °C) for the 2021 October-December average issued on early September (black cross). The ensemble mean prediction and the observation are shown by blue square and red square, respectively. The linear regression line is shown by a black line. The inter-ensemble correlation is shown on the lower-right corner. Ensemble members which predicted stronger negative Indian Ocean Dipole tend to predict severe drought in East Africa (Fig. 2). It suggests that the 2021 negative Indian Ocean Dipole was responsible for these unusually dry conditions over East Africa.
Figure 3 Time series of October-December average of the East African short rains (EASR) index from the observational data (mm day-1; black) and the statistical-dynamical hybrid predictions using the ensemble mean predictions of October-December averaged the Indian Ocean Dipole Mode index issued on 1st June by the SINTEX-F system as the input on the linear regression equation are shown by the dashed red line. The standard deviation from the observational data (0.83 mm day-1) is shown by the dashed line. The correlation skill is 0.47. This new hybrid prediction system turns out to be quite promising.
Figure 4 Prediction issued on June 1st 2022 of precipitation anomaly averaged in October–December (OND) 2022 (mm day-1). The East African short rains (EASR) index region is shown by a black box. A dynamical prediction by the SINTEX-F system and a hybrid statistical-dynamical (S.-D.) prediction of the EASR index are also shown in the right respectively.
Contacts:
- (For this study)
- Takeshi Doi, Senior Researcher,
Research Institute for Value-Added-Information Generation (VAiG) Application Laboratory (APL) Climate Variability Prediction and Application Research Group, JAMSTEC - (For press release)
- Press Office, Marine Science and Technology Strategy Department, JAMSTEC