Abstract

Drought is one of the leading causes of death in Africa because of its impact on access to sanitary water and food. This challenge has mobilized the international community to develop famine early warning systems (FEWS) to bring safe food and water to populations in need. Over the past several decades, much attention has focused on advance risk planning in agriculture and water and, more recently, on health. These initiatives require updates of weather and climate outlooks. This paper describes the active role of NOAA’s African Desk in FEWS and in enhancing the capacity of African institutions to improve forecasts. The African Desk was established in 1994 to provide services to U.S. agencies and African institutions. Emphasis is on the operational products across all time scales from weather to climate forecasts in support of humanitarian relief programs. Tools to provide access to real-time weather and climate information to the public are described. These include the downscaling of the U.S. National Multimodel Ensemble (NMME) to improve seasonal forecasts. The subseasonal time scale has emerged as extremely important to many socioeconomic sectors. Drawing from advances in numerical models, operational subseasonal forecasts are included in the African Desk product suite. These capabilities along with forecast skill assessment, verifications, and regional hazards outlooks for food security are discussed. Finally, the African Desk residency training program, an effort aimed at enhancing the capacity of African institutions to improve forecasts, and supported by this seamless approach to operational forecasting, is described.

NOAA’s African Desk, established in 1994, provides decision support services and contributes to the development of capacity in weather and climate forecasting through knowledge and technology transfer.

Drought is one of the leading causes of death in Africa because of its impact on access to sanitary water and food. This challenge has mobilized the international community to develop famine early warning systems (FEWS) to bring safe food and water to populations in need. Over the past several decades, much attention has focused on advance risk planning in agriculture and water and, more recently, on health. These initiatives require updates of weather and climate outlooks. This paper describes the active role of NOAA’s African Desk in FEWS and in enhancing the capacity of African institutions to improve forecasts. The African Desk was established in 1994 to provide services to U.S. agencies and African institutions. Emphasis is on the operational products across all time scales from weather to climate forecasts in support of humanitarian relief programs. Tools to provide access to real-time weather and climate information to the public are described. These include the downscaling of the U.S. National Multimodel Ensemble (NMME) to improve seasonal forecasts. The subseasonal time scale has emerged as extremely important to many socioeconomic sectors. Drawing from advances in numerical models, operational subseasonal forecasts are included in the African Desk product suite. These capabilities along with forecast skill assessment, verifications, and regional hazards outlooks for food security are discussed. Finally, the African Desk residency training program, an effort aimed at enhancing the capacity of African institutions to improve forecasts, and supported by this seamless approach to operational forecasting, is described.

Global climate fluctuations and their impacts on economic development have been a worldwide concern over the past several decades. The problem is even more acute in Africa, where alternating severe droughts and floods have been persistent causes of severe economic hardships. Recurrent and severe droughts along the Sahel band from the eastern Atlantic coast of West Africa to Ethiopia and Somalia, which started in the late 1960s (Lamb 1978; Nicholson 1980) and extending into the 1970s and the 1980s, caused demand for water for human and agricultural use to exceed reliable supply. African National Meteorological Services (NMSs) were extremely challenged to deliver timely and reliable climate information and services to respond to the needs of governments and the public for informed decision making.

The Comité Inter-Etat de Lutte Contre la Sécheresse au Sahel (CILCSS) headquartered in Burkina Faso and its technical body the Centre d’Agriculture, d’Hydrologie et de Météorologie (AgrHymet) located in Niger were formed in partnership between the Sahel countries and international donors to address the challenges associated with the water deficits.

The U.S. Agency for International Development (USAID) started a famine early warning systems (FEWS) project in the mid-1980s and coordinated activities with the Centre AgrHyMet. The objective was to provide the United States and Sahel countries with timely food security information to enhance decision support tools for humanitarian relief planning and response in the areas most affected by the droughts. The USAID FEWS project recognized quickly the importance of weather and climate as a key factor in agricultural production and river flow. The National Oceanic and Atmospheric Administration (NOAA)’s Climate Prediction Center (CPC) began to provide USAID FEWS with real-time gauge-based 10-day weather summaries to enable operational monitoring of crop conditions. However, given the difficulty to access rain gauge measurements from Africa, CPC also began to provide USAID FEWS with access to real-time satellite rainfall estimates to improve the weather summaries and to enable effective monitoring of drought conditions.

While the Sahel water shortages continued to increase, the severe droughts in southern Africa during the 1991/92 El Niño event (Cane et al. 1994) and persistent flooding resulting from enhanced rainfall activity in equatorial Africa during warm El Niño–Southern Oscillation (ENSO) events (Ropelewski and Halpert 1987; Ogallo et al. 1988) further demonstrated the need for improved research, prediction, and monitoring of the African climate system. The African Centre of Meteorological Applications for Development (ACMAD) was established in 1987 by the Conference of Ministers of the African Union (AU). Then the Drought Monitoring Centre Nairobi in Kenya and its subcenter in Harare Zimbabwe were created in 1989 by countries in eastern and southern Africa. Both centers have since evolved to become the Intergovernmental Authority on Development (IGAD) Climate Predictions and Applications Centre (ICPAC) in Nairobi, Kenya, and the Southern African Development Community (SADC) Climate Services Centre (SCSC) in Gaborone, Botswana, respectively. These continental- and regional-scale initiatives were prompted by the numerous challenges faced by African governments to cope with devastating weather related disasters. The Earth Summit in Rio de Janeiro in 1992 called for a mobilization of the international community for sustainable development especially in Africa. USAID FEWS also expanded its mission to eastern and southern Africa.

Recognizing that to assist African institutions meet the challenges of extreme weather events requires access to timely weather and climate information, the National Weather Service (NWS) established the African Desk in 1994 at CPC of the National Centers for Environmental Prediction (NCEP) as part of the U.S. contribution to the World Meteorological Organization (WMO) Voluntary Cooperation Program (VCP). The VCP is funded by the State Department and managed by NWS. The goal of the VCP is to foster collaboration between WMO member countries and to develop capacity in weather and climate wherever needed through transfer of technology and expertise. Hence, the African Desk from the outset focused on the development and the delivery of climate products that meet the requirements for operational monitoring at NMSs and on professional development training as part of the NCEP International Training Desks (ITD). The NCEP ITD includes the South American Desk and the Caribbean Desk established at the Weather Prediction Center (WPC) in 1989 and 1993, respectively. These two desks specialize in operational weather forecasting for South America and the Caribbean.

The African Desk initially focused on climate monitoring and forecasting. A Weather Desk was added in 2006 in support of the WMO Severe Weather Forecasting Demonstration Project (SWFDP) for southern Africa. The expansion included training in weather forecasting and the development of operational weather forecasting products that meet the requirements of NMSs in Africa. The African Desk also further expanded as projects to support the mission of USAID Office of Foreign Disaster Assistance (OFDA) were implemented. These include the provision of 1) real-time tropical cyclone alerts and extremely heavy rainfall forecasts for the Indian Ocean basin drawing from forecasts from the Joint Typhoon Warning Center (JTWC) and NCEP; 2) satellite rainfall estimates for South Asia in support of flash flood forecasting guidance along the Mekong River, working with both the International Center of Integrated Mountain Development (ICIMOD) in Nepal and the Mekong River Commission (MRC) in Cambodia; and 3) a postdoctoral appointment at CPC for young scientists from developing countries. The USAID FEWS became FEWS Network (FEWSNET) and expanded its mission during this time to include Afghanistan, Central America, and Hispaniola. The African Desk was then merged with the FEWSNET unit at CPC and began to provide FEWSNET with regional hazards outlooks for food security in all the regions of interest to FEWSNET.

The residency training program of the African Desk was also further expanded with the establishment of the CPC Monsoon Desk in 2010 to form with the African Desk and other international activities—what is now known as the CPC International Desks (ID). Thus, the core mission of the CPC ID is to provide access to NCEP operational global climate monitoring and forecasting and to develop capacity in the use and interpretation of such products through both a residency and offsite training programs.

In the following, the function of the desk is presented. Then access to real-time NCEP products to support operational forecasting and climate monitoring is discussed. An example of the various expert assessment products with a focus on the regional hazards outlooks for food security is presented. Finally, the professional development training program is described, followed by a summary.

THE FUNCTIONS OF THE AFRICAN DESK.

NCEP is one of very few global weather and climate centers in the world that provide unrestricted access to a suite of weather and climate information from short range forecasts to seasonal and interannual climate predictions globally. These forecasts are accompanied by a set of real-time monitoring products featuring the state of the global climate providing the most up-to-date information on ENSO and the state of the global ocean, the global monsoon, and drought information. As stated earlier, the African Desk was established on one hand to provide NMSs in Africa with timely access to weather and climate information to support decision making and on the other hand to recognize the emphasis on weather and climate within the U.S. government humanitarian assistance programs such as USAID. In essence, the African Desk is the NCEP focal point for real-time dissemination of operational African weather and climate data and products, which are integrated into the process of systematically updating food security outlooks for informed decisions in foreign aid such as food and water supply and assistance in population displacements. The African Desk operations also feed directly into NMSs monitoring and forecasting operations. The primary functions of the desk are to continuously monitor weather and climate patterns in Africa and to disseminate these in the form of bulletins or briefings via the CPC website and through e-mail distribution lists. Specific operational weather and climate products are derived from NCEP observational data and model outputs and seamless forecasts from short range weather to seasonal climate outlooks are issued. These serve as guidance to NMSs operations. The sources for observational data include the NCEP reanalysis (Kalnay et al. 1996), the CPC global land precipitation data (PREC/L) (Chen et al. 2002), the CPC African rainfall climatology version 2 (ARC2) (Novella and Thiaw 2013), the extended reconstructed sea surface temperature version 3 (ERSSTv3b) (Smith et al. 2008), and the National Aeronautic and Space Administration (NASA) Tropical Rainfall Measuring Mission (TRMM) 3B42v7 (Huffman et al. 2007). Anomalous weather and climate are identified and the information communicated to partner agencies. Land surface information is gathered from partner agencies within FEWSNET and integrated with up-to-date weather and climate data to provide regional hazards outlooks. This information provides initial qualitative crop impact assessments that can feed into food security outlooks. In addition, NCEP data are staged and reformatted in a way that NMSs can access and use them. Finally, hands-on training in operational monitoring and forecasting of weather and climate, and the assessment of impacts, are delivered to NMSs staff.

PROVIDING ACCESS TO NCEP PRODUCTS.

The African Desk expanded its mission tremendously over the years to provide access to NCEP data and products globally. Products are made available primarily through a website displayed in Fig. 1. This page, which initially focused on Africa, has now evolved into the CPC ID website. The links at the top of the page provide access to the NCEP Climate Forecast System version 2 (CFSv2) (Saha et al. 2014) and the U.S. NMME (Kirtman et al. 2014) in both global and regionalized formats. Below these links is the main weather website and also home to bulletins and summaries of weather and climate forecasts and evolving conditions for Africa and different regions of the world. In both the climate and weather web pages, the globe is divided into ocean basins and continental regions. The continental regions include Africa, Central America, the Caribbean, the Maritime Continent, Central Asia, East Asia, South Asia, and South America. The weather page features 1) the expert assessment bulletins, forecasts, and summaries; 2) weather and subseasonal forecasts derived from the NCEP Numerical Weather Prediction (NWP) over various regions in Africa and over all of Africa for a wider view; 3) surface observations, analyses, and monitoring products, including SST; and 4) satellite derived rainfall estimates.

Fig. 1.

CPC International Desks website. NCEP GFS and GEFS, GDAS, satellite rainfall estimates, and SST are made available over each geographical region through the clickable maps or through the menu on the left of the page. Expert assessment products are also available for Africa and other regions of special interest. Links to the global and regional seasonal climate forecasts page are indicated at the top of the main page.

Fig. 1.

CPC International Desks website. NCEP GFS and GEFS, GDAS, satellite rainfall estimates, and SST are made available over each geographical region through the clickable maps or through the menu on the left of the page. Expert assessment products are also available for Africa and other regions of special interest. Links to the global and regional seasonal climate forecasts page are indicated at the top of the main page.

NCEP model guidance is staged for the globe as well as continental and regional domains, including East Africa, northern Africa, southern Africa, and West Africa. Both deterministic and ensemble guidance products are displayed. The ensemble page features products critical to forecasting severe weather events. Graphical displays include probability of exceedance of parameters such as precipitation and winds for various threshold values. For precipitation, the range is from 1 mm to more than 100 mm.

The climate forecasting page features CFSv2 and the U.S. NMME model guidance, including SST predictions expressed in the form of anomalies, standardized anomalies, and skill-masked standardized anomalies. Skill mask is applied such that only anomalies at grid points where the correlation coefficient between the model SST forecasts and the observed SST (ERSSTv3b) is equal or greater than 0.3 are displayed. This is similar for the precipitation and air temperature forecasts, which also include maps of departures from climatological probabilities of the three equiprobable categories of above-, near-, and below-average rainfall (Thiaw et al. 1999). Figure 2 shows seasonal rainfall guidance maps for Africa expressed as indicated above. An examination of the maps reinforces the fact that there clearly is a need to express seasonal forecasts in probabilistic form to convey uncertainty in the forecasts and also to remove signal from random forecasts. Deterministic forecasts expressed in terms of standardized anomalies (Fig. 2a) show a very wet Sahel for July–September 2014, with 1–8 May initial conditions. However, the probability forecasts (Fig. 2b) show a much reduced signal in this region, with only a slight tilt in the odds to favor above-average rainfall in the Sahel. The probability forecasts convey the uncertainty in the forecasts and allow for more informed decision based on the forecasts.

Fig. 2.

Seasonal rainfall forecasts from 1 to 8 May 2014 initial conditions for Jul–Sep 2014, from the ensemble mean of the U.S. NMME forecasts, expressed in terms of (a) standardized anomaly and (b) departure from climatological probability of the three equiprobable categories of above-average, average, and below-average rainfall. Shaded in brown–red (green–yellow) indicate a higher chance for below (above) rainfall to occur. Blank indicates no skill in the forecast system. The correlation skill used in the calculation of the probability shift is based on the CPC PREC/L data.

Fig. 2.

Seasonal rainfall forecasts from 1 to 8 May 2014 initial conditions for Jul–Sep 2014, from the ensemble mean of the U.S. NMME forecasts, expressed in terms of (a) standardized anomaly and (b) departure from climatological probability of the three equiprobable categories of above-average, average, and below-average rainfall. Shaded in brown–red (green–yellow) indicate a higher chance for below (above) rainfall to occur. Blank indicates no skill in the forecast system. The correlation skill used in the calculation of the probability shift is based on the CPC PREC/L data.

This real-time weather and climate information accessible through the website can be of tremendous benefit not only to NMSs and USAID programs but also to the public globally and especially the education sector providing the current state of the climate and its evolution at any given time.

EXPERT ASSESSMENT PRODUCTS.

The various bulletins produced in the African Desk include seasonal rainfall outlooks, subseasonal outlooks, weather forecasts, and regional hazards outlooks that enable decision making in food security. Because of space limitation for this article, the latter is discussed here. The hazards outlooks bulletin for food security is prepared weekly in collaboration with FEWSNET partners, including the U.S. Geological Survey (USGS), NASA, USAID, and Chemonix. The bulletin features both long-term (past conditions throughout the season) monitoring of the climate system and outlooks into the near future about one week to a season. The objective is to provide targeted forecasts for areas that are vulnerable to droughts or flooding that might result in adverse impact on crops or pastures. Hence, the hazards outlooks are based on a wide range of products, including rain gauge data and satellite rainfall estimates, rainfall and surface temperature forecasts up to 16 days, and subseasonal and seasonal climate forecasts. Other inputs to the hazards outlooks include USGS’s river flow forecasts and water requirement satisfaction index (WRSI) for crops and rangelands (Senay et al. 2011), NASA’s normalized difference vegetation index (NDVI) (Fung et al. 1987), NOAA’s vegetation health index (VHI) (Kogan 2002), and field observations. Figure 3 shows a schematic of the regional hazards outlooks process designed as a loop that begins with the identification of areas that exhibit consistent rainfall deficits or frequent flooding through routine in-depth monitoring of the climate system. These areas are often faced with a high risk of food insecurity. Then model guidance tools are used to examine both short range and extended range forecasts. The reliability of these forecasts is qualitatively assessed by looking at consistency both within each model and between different models. Then, based on current conditions and forecasts, a geographical information system (GIS) software is employed to draw polygons on a map to highlight areas at risk for food security. The preliminary hazards outlook bulletin is distributed to partners within FEWSNET including the field representatives who have expert knowledge of conditions on the ground. Then a teleconference takes place for a live discussion of current weather and climate conditions and the preliminary hazards outlooks. The feedback received allows for the finalization of the hazards outlooks. An example of a hazard outlook valid for 2–8 January 2014 is displayed in Fig. 4. Polygon shapes are numbered and show areas of concern for flooding or drought. The numbering allows for the long-term monitoring of the conditions in the areas of interest. The color shades determine the nature and severity of the hazards. In the example shown in the figure, delayed onset to the start of the rainfall season and poor temporal distribution of the rains since the beginning of the season resulted in drought in polygon area 1 encompassing much of northeastern Kenya. Developing dryness that may have adverse impacts on crops is highlighted in polygon areas 2, 3, and 4. In contrast, four consecutive weeks of above-average rains caused flooding to persist along the Caprivi Strip in northeastern Namibia and neighboring areas (polygon area 5), while the development of a tropical wave disturbance was expected to result in flooding over west central Madagascar. The hazards outlooks are disseminated through the website and an e-mail distribution list. FEWSNET uses the information together with current climate forecasts and trends and other food security indicators to issue monthly food security outlooks. Finally, this information is provided to USAID for informed decision in humanitarian response planning based on the level of food security threats.

Fig. 3.

Schematic for the process of the regional hazards outlooks for food security.

Fig. 3.

Schematic for the process of the regional hazards outlooks for food security.

Fig. 4.

Regional hazards outlooks for food security, valid 2–8 Jan 2014. Shaded in brown, yellow, and blue, are areas that exhibit drought, abnormal dryness, and flooding, respectively. Polygon shapes are numbered based on the evolution of weather and climate conditions that led to the events. Outlooks are generated every Wednesday of the week and valid for the coming week from Thursday to the following Wednesday. Included in the hazards outlooks are the long-term conditions in the field and the current meteorological and climate forecasts.

Fig. 4.

Regional hazards outlooks for food security, valid 2–8 Jan 2014. Shaded in brown, yellow, and blue, are areas that exhibit drought, abnormal dryness, and flooding, respectively. Polygon shapes are numbered based on the evolution of weather and climate conditions that led to the events. Outlooks are generated every Wednesday of the week and valid for the coming week from Thursday to the following Wednesday. Included in the hazards outlooks are the long-term conditions in the field and the current meteorological and climate forecasts.

PROFESSIONAL DEVELOPMENT TRAINING PROGRAM.

The residency training program.

The African Desk became operational one year after it was established and hosted its first trainee in March 1995. The four-month residency training program is a U.S. contribution to the WMO VCP managed by NWS. Each trainee receives a WMO fellowship. They arrive in staggered intervals of two months for a maximum capacity of 12 trainees per year: six in the climate desk and six in the weather desk, such that at any given time there are two trainees in the climate desk and two trainees in the weather desk. This approach allows the visitors who have been in training the longest to contribute to the training of the new trainees. The objective of the training program is to work with NMSs in Africa to enhance their capacity to deliver improved weather and climate forecasts. Each trainee returns to his or her home institution, equipped with a new set of tools that could be applied to improve forecasts. To date, the African Desk has trained 150 professional meteorologists and scientists from 38 countries in Africa. The programs take into account the diversity of the climate system in Africa. Hence, fellows are selected from each region of Africa on a rotating basis and invited to participate in the training program during the active rainfall seasons of the respective regions. The Climate Desk and the Weather Desk have separate daily schedules. However, the desks share some common activities. Upon arrival at the desk, the trainees are introduced to the CPC online tutorial on major modes of variability including ENSO and the Madden–Julian oscillation (MJO). Then, they spend time learning the use of basic UNIX commands, shell programming, and the use of graphical packages such as the Grid Analysis Display System (GrADS) and applications of the GIS. These basic tools enable the trainees to access and process NCEP data for future use in climate diagnostics studies or in model forecasts verifications. The trainees also spend time practicing the use of GIS to prepare forecast graphics. In the following, we outline the curriculum specific to both desks and describe the day-to-day operations.

Climate desk

The mid-1990s were marked by a strong interest in developing capacity in climate monitoring and forecasting for Africa. The NOAA’s Climate Program Office (CPO) organized an international workshop “Reducing Climate Vulnerability in Southern Africa” in 1996 that concluded that there was a greater need for access to seasonal climate forecasts to enable decision making in various socioeconomic sectors including agriculture, water management, and health. A key recommendation was to initiate RCOFs, whereby experts in climate science and the users of climate information would meet to issue consensus seasonal climate outlooks and to translate this information to the user community in a language that is easy to understand. The first official RCOF focused on southern Africa (SARCOF) and took place in September 1997 in Zimbabwe as the warm ENSO event of 1997/98 became fully developed. RCOFs have been organized ever since several times during the year across Africa and have now expanded to encompass many parts of the world. The African Desk participated in the first SARCOF and has since been contributing to RCOFS around the world. Monthly updates of seasonal rainfall guidance for Africa have been prepared and issued since 1996. At first the CPC developed canonical correlation analysis (CCA) model was employed (Barnston 1994; Barnston et al. 1996; Thiaw et al. 1999). More recently, improvements in climate models provided an opportunity to calibrate the NNME forecasts, including the NCEP CFS, with observations, and to issue probabilistic seasonal forecasts across Africa and the global tropics. These forecasts feed into the RCOFs, providing NMSs and regional climate centers with real-time access to information relevant to the state of the global climate and NCEP outlooks.

The trainees in the climate desk work primarily on operational climate monitoring and forecasting. The trainees learn how to use the International Research Institute for Climate and Society (IRI) Climate Predictability Tool (CPT) (Mason 2011). CPT consists of several advanced statistical packages that are employed to conduct diagnostics studies and to make seasonal climate forecasts and their verifications. It can be used on both Windows and Linux systems. The trainees learn how to prepare historical data to run seasonal prediction experiments in CPT. These can be in situ data, satellite observations, and/or climate model outputs. The trainees learn how to design and run several sets of seasonal forecast experiments and how to interpret the results. This set of skills leads to the development of a statistical seasonal forecasting model for the country or region of interest.

The training desks provide an excellent opportunity to assess the performance of coupled climate models in the tropics and to downscale the forecasts to improve skills in certain regions. In the following, we discuss diagnostics studies performed in the African Desk on seasonal time scales with a focus on East Africa. The region and season were chosen because of the strong ENSO influence and hence increased predictability, but also this is a region where the interannual variability is very strong with alternating floods and drought episodes. The recent availability of hindcasts from the U.S. NMME forecasts has provided an excellent opportunity to compare model performance and to improve forecasts skills in certain regions of the world. The U.S. NMME models in this study include the NCEP CFS versions 1 and 2, the Canadian Climate Model (CCM) versions 1 and 2, the European Centre Hamburg Model (ECHAM versions A and F), the NOAA/Geophysical Fluid Dynamics Laboratory (GFDL) model, the NASA model, the National Center for Atmospheric Research (NCAR) model, and the ensemble mean of these models (ENSM). All model hindcasts cover the period 1981–2010. In an attempt to improve seasonal forecast skill, we employ CCA to downscale the U.S. NMME forecasts. The technique employed is described in Mo and Thiaw (2002) and consists of using U.S. NMME precipitation forecasts in a relatively large domain encompassing the target area as predictor field in the CCA experiment and observed precipitation in the target area as predictand field. In a sense, CCA is applied to correct the U.S. NMME model forecasts by taking into account local conditions. Results from seasonal prediction experiments for the East Africa rains during the October–December rainfall season are presented in Table 1. The model skill is measured by the mean anomaly correlation (AC) for the period 1981–2010. The CFSv1 has the highest skill and NCAR has the lowest skill. The CCA correction improves the forecasts for each individual model as well as the ensemble mean, except for the GFDL model. The geographical distribution of the cross validated skill for the October–December East Africa seasonal rainfall for the ECAHAM-A model (not shown) indicates that correlations 0.6 or higher can be found over the western half of Kenya and northwestern Tanzania.

Table 1.

Mean anomaly correlation for U.S. NMME models and after CCA corrections for equatorial eastern Africa. Bolded values indicate model skill improvement after CCA correction. EC-A and EC-F denote ECHAM-A and ECHAM-F models.

Mean anomaly correlation for U.S. NMME models and after CCA corrections for equatorial eastern Africa. Bolded values indicate model skill improvement after CCA correction. EC-A and EC-F denote ECHAM-A and ECHAM-F models.
Mean anomaly correlation for U.S. NMME models and after CCA corrections for equatorial eastern Africa. Bolded values indicate model skill improvement after CCA correction. EC-A and EC-F denote ECHAM-A and ECHAM-F models.

Weather desk

The devastation suffered by millions of people in southern Africa following tropical cyclones land falling in Mozambique in 1999 and 2000 led the Commission of Basic System (CBS) of WMO to recommend WMO to establish the SWFDP. CBS further recommended that global weather forecasting centers (GWFCs) be identified and be asked to provide support to the SWFDP. NCEP was designated global center in 2006 for the pilot phase of the SWFDP with a focus on southeastern Africa. The objective was to provide NMSs in southeastern Africa with an opportunity to take advantage of recent development in NWP systems to improve early warning systems through effective use of model derived products. The project was structured as a cascading forecasting process (Chen 2010) such that, on a daily basis, operational GWFCs provide regional specialized meteorological centers (RSMCs) and NMSs with access to weather information. RSMCs in turn use high resolution limited area models to prepare severe weather forecasting guidance up to day 5, where day 0 is the initial condition date. NMSs use this guidance together with GWFC products and local weather condition to refine the national forecasts and as necessary to issue warnings and deliver the information to their respective governments.

NCEP provides support to the SWFDP in two important ways: 1) provision of real-time NCEP products staged in ready-to-use format for the weather forecasters in Africa and accessible via the website described in the section on “Providing access to NCEP products” and 2) training of African meteorologists in the use and interpretation of NWP. The trainees in the weather desk work on daily operational weather forecasting. They spend time analyzing operational NWP products up to 7-day weather forecasts from NCEP, the Met Office, and the European Centre for Medium Range Weather Forecasts (ECMWF) as displayed on the NCEP Advanced Weather Interactive Processing System (NAWIPS). Then they draw significant weather patterns using tools on the NAWIPS. The trainees also review additional forecast fields on the CPC ID website. These fields are specifically constructed for the needs of the forecasters in different regions of Africa and include a wide variety of products such as probabilities of exceedance of precipitation and winds, as well as spaghetti diagrams for winds and 500-hPa geopotential heights. The trainees look at several sets of diagnostics and use the information to measure confidence in the models’ forecasts. In particular, they examine the NCEP ensemble forecasts to evaluate the spread in the model. The examination of these charts lead to the preparation of a weather forecast bulletin released every day, except for weekends. The bulletin features 24-h quantitative precipitation forecasts (QPF) for Africa up to day 5 and a 5-day cumulative precipitation forecast map. Preliminary forecasts are discussed with colleagues before finalized. Then GIS is used to draw shaded polygons to highlight areas that are likely to receive significant rainfall during the forecast period. Then the trainee writes the forecast bulletin that includes a model diagnostics section describing how the models depict the most significant events in the forecasting systems. The bulletin is reviewed and posted on the CPC website and disseminated to countries in Africa.

Climate and weather

The subseasonal time scale usually defined as variability between 10 and 90 days is extremely important and is seen as a bridge between short and medium range weather forecasts and seasonal climate predictions. This is also a time scale of interest to many users of climate information to gain advance knowledge of the onset of the seasonal rains, the dry and wet spells frequencies, and the probability for extreme events to occur. Recent research has indicated important potential sources of predictability for this time range. This includes the realization of a better representation of the MJO in numerical weather and climate models (Rashid et al. 2011) and the recent acceleration of the development of tools to monitor and predict the MJO and its impacts on the global climate. The MJO is a coupled ocean–atmosphere planetary wavenumber-1 to -3 pattern that propagates from west to east around the globe (Madden and Julian 1972, 1994). It exhibits an area of enhanced rainfall activity associated with upward motion and divergence in the upper troposphere followed by an area of suppressed rainfall associated with descending motion. It is known to be the dominant mode of variability on the intraseasonal time scale between 30 and 90 days (Zhang 2005). When active, the MJO can have an influence on precipitation in some regions in Africa—most notably in tropical eastern Africa. Figure 5 displays rainfall and wind anomaly composites over Africa south of the equator during the December–February season during two opposite phases of the MJO are portrayed in the Wheeler–Hendon MJO index diagram (Wheeler and Hendon 2004). Figure 5a displays anomaly composites during phase 2 and Fig. 5b shows anomaly composites during phase 6 of the MJO. Enhanced rainfall activity associated with low level esterly wind anomalies prevails over equatorial east Africa when the center of convection associated with the MJO is located over the Indian Ocean (phase 2). Note the reversal in the sign of the anomalies when the MJO is centered over the western Pacific (phase 6). However, the MJO is only one factor among many other complex processes that govern the African climate system, including Rossby waves, Kelvin waves, African waves, thermohaline circulation, etc. Recent development in NWP models has made forecasts more reliable in certain parts of the tropics. The African Desk takes advantage of this advanced modeling capability to prepare and issue outlooks for week 1 and week 2. These can be extremely useful to decision makers who look into the potential for extreme weather to occur within the next few weeks and to assist in short-term advance planning. The approach of the African Desk to operational subseasonal forecasting (week 1 and week 2) is to first assess the state of the MJO and forecasts. When the MJO is active and projected to remain active during the forecasting period, the MJO rainfall anomaly composites and NWP tools up to two weeks are used as guidance to draw polygons of enhanced or suppressed rainfall. The NWP tools that are examined include 850- and 200-hPa winds, 850- and 500-hPa geopotential heights, and mean sea level pressure to locate major weather and climate systems that modulate the regional precipitation patterns. The NWP ensemble precipitation forecasts, including the probability of exceedance maps and bias corrected precipitation anomaly forecasts, are also valuable tools for consideration in the forecasting system. Upon making preliminary forecasts, the forecaster leads a forecast briefing to review the forecasting tools and to discuss the forecasts before they are final. The briefing is very important, especially when there are inconsistencies among the prediction tools. A GIS tool is then used to draw polygons that indicate high chances for above-average or below-average precipitation as displayed in Figs. 6a and 6b. In this case, the forecasts for week 1 are valid for 8–14 January 2013 and for week 2 are valid for 15–21 January 2013. The forecasts for week 1 indicated an area of rainfall surplus over the northern half of Mozambique, Malawi, southern Tanzania, eastern Zambia, and northern Zimbabwe. There was an increased chance for below-average rainfall over portions of Congo, Democratic Republic of Congo, and Angola. For week 2, the area of above-average rainfall was limited to Zambia, while the suppression was over northern Mozambique and southeastern Tanzania. The outlooks are disseminated via a distribution list and posted on the website. Eyeball verifications for the forecasts are displayed in Figs. 6c and 6d for week 1 and week 2, respectively, suggesting that the forecasts were indeed reasonable. Figure 7 shows the Heidke Skill scores (Livezey 1995) for the week-1 and week-2 outlooks. These are based on 150 forecast events. Much of tropical Africa exhibit positive skill scores for week 1 and week 2. As expected, the forecast skill for week 1 is higher than that for week 2. Portions of northern southern Africa and coastal equatorial eastern Africa tend to exhibit the highest skill score for both week-1 and week-2 outlooks. This is probably due to the fact that the MJO influence is strongest in these locations (not shown).

Fig. 5.

MJO composites of daily rainfall anomaly (shaded) from the high-resolution satellite rainfall estimates (ARC2) and 850-hPa wind anomalies (vectors) from NCEP reanalysis for the December–February (DJF) season for (a) phase 2 and (b) phase 6 of the MJO as portrayed in the Wheeler and Hendon (2004) diagram. The period is 1983–2012. Only anomalies significant at the 90% confidence level are plotted.

Fig. 5.

MJO composites of daily rainfall anomaly (shaded) from the high-resolution satellite rainfall estimates (ARC2) and 850-hPa wind anomalies (vectors) from NCEP reanalysis for the December–February (DJF) season for (a) phase 2 and (b) phase 6 of the MJO as portrayed in the Wheeler and Hendon (2004) diagram. The period is 1983–2012. Only anomalies significant at the 90% confidence level are plotted.

Fig. 6.

Subseasonal outlooks for Africa for week 1 (a) valid 8–14 Jan 2013 and week 2 (b) valid 15–21 Jan 2013. Shaded in green (yellow) are areas of predicted enhanced (suppressed) rainfall. Green (brown) shades indicate areas of observed enhanced (suppressed) rainfall. Eyeball verifications for (c) week 1 and for (d) week 2, based on CPC ARC2. Blue and red contours show areas where the forecasts were generally in agreement with the observations.

Fig. 6.

Subseasonal outlooks for Africa for week 1 (a) valid 8–14 Jan 2013 and week 2 (b) valid 15–21 Jan 2013. Shaded in green (yellow) are areas of predicted enhanced (suppressed) rainfall. Green (brown) shades indicate areas of observed enhanced (suppressed) rainfall. Eyeball verifications for (c) week 1 and for (d) week 2, based on CPC ARC2. Blue and red contours show areas where the forecasts were generally in agreement with the observations.

Fig. 7.

Heidke Skill for (a) week-1 and (b) week-2 outlooks over Africa based on 247 week-1 forecasts and 124 week-2 forecasts. Observation is the CPC ARC2.

Fig. 7.

Heidke Skill for (a) week-1 and (b) week-2 outlooks over Africa based on 247 week-1 forecasts and 124 week-2 forecasts. Observation is the CPC ARC2.

In the assessment of the NCEP CFSv2 forecast skill for week-1 and week-2 forecasts, the computation of AC between observations and forecasts with and without bias correction is performed. The methodology outlined by Fan and van den Dool (2011) is used to correct the bias in the forecasting system, where every day the week-1 and week-2 precipitation ensemble forecasts are corrected with the past N days mean forecast errors, defined as follows:

 
formula
 
formula

where Pf is the NCEP ensemble week-1 and week-2 precipitation forecasts, Po is the observed week-1 and week-2 precipitation from CPC daily Global Unified Precipitation Analysis, and N is number of days. Figure 8 shows the AC levels for week-1 and week-2 forecasts. A tremendous improvement in the forecasting system is gained in some areas when the bias is removed from the raw forecasts. The fact that CFSv2 has high correlation skill over equatorial eastern Africa suggests that this model can reasonably be useful in operational week-2 forecasting in this region.

Fig. 8.

Anomaly correlation between observed rainfall (TRMM 3B42v7) and CFSv2 forecasts without bias correction, for (a) week 1 and (b) week 2, and with bias correction for (c) week 1 and (d) week 2 for the period 2002–12.

Fig. 8.

Anomaly correlation between observed rainfall (TRMM 3B42v7) and CFSv2 forecasts without bias correction, for (a) week 1 and (b) week 2, and with bias correction for (c) week 1 and (d) week 2 for the period 2002–12.

The NOAA/USAID Climate Training Workshop Series.

The residency training program is complemented by a NOAA/USAID climate training workshop series (Thiaw et al. 2014) initiated in 2009 that have enabled the training in climate of far more professionals from different regions of the world than NCEP could host in the residency program (Fig. 9). The training workshops have been organized for all the ocean basins of the world. A map of countries that participated in both the residency training program at NCEP and the offsite training in the NOAA/USAID series is displayed in Fig. 10. A total over 300 meteorologists and scientists from countries in Africa, Asia, Caribbean, Central America, South America, and southeast Europe have participated in either or both the NCEP residency training program or the NOAA/USAID series. In these series, emphasis is on practical exercises combined with lectures on recent advances in climate variability and change. The basic state of the global monsoon system and major modes of climate variability including ENSO and the MJO, teleconnections, modeling, and predictions are discussed. The trainees then learn how to set up seasonal climate prediction experiments using CCA and how to downscale model outputs to improve forecast skills. They also learn how to verify the forecasts. GIS is also introduced to enable the digitization of the forecast maps. It is expected that the trainees return to their home institutions with an improved understanding of the global climate system and how modes of variability can influence the climate in their respective regions. They also take home tools to improve forecast operations. The long-term goal is for the trainees to become resource persons in their countries and regions to train other professionals. More detailed information can be found in Thiaw et al. (2014).

Fig. 9.

Snapshots of NCEP residency training and the International Training Workshop Climate Variability and Predictions (ITWCVP) in the NOAA/USAID series. (a) Staff and trainees during a forecasting session in the NCEP’s African Desk. (b) A classroom session during the (4ITWCVP) in San Jose, Costa Rica, 8–17 Aug 2012; instructor is Dr. Carolina Vera of the University of Buenos Aires in Argentina. (c) Participants in the 6ITWCVP, Istanbul, Turkey, 4–16 Aug 2014.

Fig. 9.

Snapshots of NCEP residency training and the International Training Workshop Climate Variability and Predictions (ITWCVP) in the NOAA/USAID series. (a) Staff and trainees during a forecasting session in the NCEP’s African Desk. (b) A classroom session during the (4ITWCVP) in San Jose, Costa Rica, 8–17 Aug 2012; instructor is Dr. Carolina Vera of the University of Buenos Aires in Argentina. (c) Participants in the 6ITWCVP, Istanbul, Turkey, 4–16 Aug 2014.

Fig. 10.

NCEP residency training and NOAA/USAID training workshop coverage. Color shades indicate countries that have participated in either the residency training at NCEP or the NOAA/USAID series—both training workshops.

Fig. 10.

NCEP residency training and NOAA/USAID training workshop coverage. Color shades indicate countries that have participated in either the residency training at NCEP or the NOAA/USAID series—both training workshops.

SUMMARY.

For two decades NOAA has been providing support to African NMSs to develop capacity in weather and climate forecasting through a residency training program and the delivery of products required for operational monitoring and forecasting of weather and climate. NOAA has also been working with sister agencies with specific interest in Africa and the developing world to support decision making in areas that are challenged by natural disasters such as droughts, floods, tropical cyclones, etc., and considered economically vulnerable because of shortage of food or drinking water. NWS established the African Desk at CPC in March 1994 as a part of the NCEP ITD. It became operational and hosted its first trainee from Kenya in March 1995. The desk focused initially on climate monitoring and forecasting. Then a weather desk was established in 2006 to provide support to the WMO SWFDP through training in weather forecasting and the development of products required for severe weather forecasting. One hundred thirty six professional meteorologists from 38 countries in Africa have since been trained in the African Desk. Four former trainees moved up in the chain of command to become WMO Permanent Representatives (PRs) of their respective countries. Dozens moved up to leadership positions. In 2010, the activities of the African Desk were further expanded with the establishment of the Monsoon Desk in 2010. The two desks, combined with the international projects in support of the U.S. government humanitarian mission around the world, now form the CPC ID.

Despite the success of the NCEP residency training, the demand for training in operational climate monitoring and forecasting remains far greater than current NCEP capacity to meet it. Hence, the NOAA/USAID climate training workshop series serve to fill the gap by providing professionals from developing and emerging countries around the world an opportunity to improve their understanding of the climate system and enhance their forecasting skills to improve climate services. The training programs also provide an excellent opportunity to use NCEP and U.S. NMME model data and to evaluate the performance of the models especially in the tropics. Through this process, NMSs can develop effective weather and climate monitoring tools that enable decision making.

Finally, one of the key missions of the CPC ID is to provide domestic and international agencies with access to real-time NCEP operational weather and climate forecasts for any given region of the world. A website has been created and maintained for this purpose. The CPC ID also provides support to many domestic and international programs, including the USAID’s FEWSNET and Disaster Risk Reduction (DRR) Program, RCOFs, and more recently the World Health Organization (WHO) climate and health initiative.

ACKNOWLEDGMENTS

The first author would like to express gratitude to the management of the U.S. National Weather Service since 1994, Dr Louis Uccellini for his leadership in sustaining the NCEP International Desks, Mrs. Laura Furgione, Dr. Jack Hayes. Brig. Gen. David Johnson, Brg. Gen. John Kelly, Dr. Joe Friday, and Mr. John Jones, for unwavering support for the African Desk and the NCEP International Desks; NCEP management, Dr. William Lapenta, Mr. Dennis Staley, Dr. Ronald McPhereson, Col. James Howcroft, Mrs. Sondra Young-Wick, Mr. Michael Halpert, Dr. Arun Kumar, Dr. Wayne Higgins, Mr. James Laver, Dr. Ants Leetmaa, Dr. David Rodenhuis (CPC Director at the time the African Desk was established), and Mr. Alvin Miller, for their constant encouragement. Appreciation also to the CPC International Team, including Endalkachew Bekele, Thomas DiLiberto, Nicholas Novella, Chalump Oonariya, Bradley Pugh, and Miliaritiana Robjhon, for their timely analysis and updates of a number of the African Desk products, and for their contribution to the training and for preparing some of the figures in this article. Acknowledgment to all the instructors and trainees who participated in the various training activities both through the residency program and the NOAA-USAID series. The CPC International Desks is supported by NWS base budget and partially funded primarily by USAID.

APPENDIX

List of Acronyms and Their Definitions.

ACAnomaly correlation

ACMADAfrican Centre of Meteorological Applications for Development

AgrHyMetAgriculture Hydrology and Meteorology

ARCAfrican rainfall climatology

CBSCommission of Basic System

CCACanonical correlation analysis

CCMCanadian Climate Model

CFSClimate Forecast System

CILCSSComité Inter-Etat de Lutte Contre la Sécheresse au Sahel

CPCClimate Prediction Center

CPOClimate Program Office

CPTClimate Predictability Tool

DRRDisaster risk reduction

ECHAMEuropean Centre Hamburg Model

ECMWFEuropean Centre for Medium Range Weather Forecasting

ENSMEnsemble mean

ENSOEl Niño–Southern Oscillation

ERSSTv3bExtended Reconstructed Sea Surface Temperature version 3b

FEWSFamine early warning system

FEWSNETFamine Early Warning System Network

GFDLGeophysical Fluid Dynamic Laboratory

GISGeographical information system

GrADSGrid Analysis Display System

GWFCGlobal Weather Forecasting Center

ICIMODInternational Centre for Integrated Mountain Development

ICPACIGAD Climate Prediction and Application Centre

IDInternational desks

IGADIntergovernmental Authority on Development

IRIInternational Research Institute for Climate and Society

ITDInternational discontinuity

JTWCJoint Typhoon Warning Centre

MJOMadden–Julian oscillation

MRCMekong River Commission

NASANational Aeronautic and Space Administration

NAWIPSNCEP Advanced Weather Interactive Processing System

NCARNational Center for Atmospheric Research

NCEPNational Centers for Environmental Prediction

NDVINormalized difference vegetation index

NMSNational Meteorological Service

NOAANational Oceanic and Atmospheric Administration

NWSNational Weather Service

NWPNumerical Weather Prediction

OFDAOffice of Foreign Disaster Assistance

PREC/LCPC global land precipitation data

QPFQuantitative precipitation forecast

RCOFRegional Climate Outlook Forum

RSMCRegional Specialized Meteorological Centre

SADCSouthern African Development Community

SARCOFSouthern African Regional Climate Outlook Forum

SCSCSADC Climate Services Centre

SSTSea surface temperature

SWFDPSevere Weather Forecasting Demonstration Project

TRMMTropical Rainfall Measuring Mission

USAIDU.S. Agency for International Development

USGS U.S.Geological Survey

VCPVoluntary Cooperation Program

VHIVegetation health index

WHOWorld Health Organization

WMOWorld Meteorological Organization

REFERENCES

REFERENCES
Barnston
,
A. G.
,
1994
:
Linear statistical short-term climate predictive skill in the Northern Hemisphere
.
J. Climate
,
7
,
1513
1564
, doi:.
Barnston
,
A. G.
,
A. G.
,
W.
Thiaw
, and
V.
Kumar
,
1996
:
Long-lead forecasts of seasonal precipitation in Africa using CCA
.
Wea. Forecasting
,
11
,
506
520
, doi:.
Cane
,
M. A.
,
G.
Eshel
, and
R. W.
Buckland
,
1994
:
Forecasting Zimbabwean maize yield using eastern equatorial Pacific sea surface temperature
.
Nature
,
370
,
204
205
, doi:.
Chen
,
M.
,
P.
Xie
,
J.
Janowiak
, and
P. A.
Arkin
,
2002
:
Global land precipitation: A 50-yr analysis based on gauge observations
.
J. Hydrometeor.
,
3
,
249
266
, doi:.
Chen
,
P.
,
2010
: WMO strategy for improving severe weather forecasting in developing countries. WMO Executive Council on Disaster Risk Reduction and Service Delivery, Second Session, 22 pp.
Fan
,
Y.
, and
H.
van den Dool
,
2011
:
Bias correction and forecast skill of NCEP GFS ensemble week-1 and week-2 precipitation, 2-m surface air temperature, and soil moisture forecasts
.
Wea. Forecasting
,
26
,
355
370
, doi:.
Fung
,
I. Y.
,
C. J.
Tucker
, and
K. C.
Prentice
,
1987
:
Application of Advanced Very High Resolution Radiometer vegetation index to study atmosphere-biosphere exchange of CO2
.
J. Geophys. Res.
,
92
,
2999
3015
, doi:.
Huffman
,
G. J.
, and Coauthors
,
2007
:
The TRMM Multisatellite Precipitation Analysis (TMPA): Quasi-global, multiyear, combined-sensor precipitation estimates at fine scales
.
J. Hydrometeor.
,
8
,
38
55
, doi:.
Kalnay
,
E.
, and Coauthors
,
1996
:
The NCEP/NCAR 40-Year Reanalysis Project
.
Bull. Amer. Meteor. Soc.
,
77
,
437
471
, doi:.
Kirtman
, and Coauthors
,
2014
:
The North American Multimodel Ensemble (NMME): Phase-1 seasonal to interannual prediction; phase-2 toward developing intraseasonal prediction
.
Bull. Amer. Meteor. Soc.
,
95
,
585
601
, doi:.
Kogan
,
F.
,
2002
:
World droughts in the new millennium from AVHRR-based vegetation health indices
.
Eos, Trans. Amer. Geophys. Union
,
83
,
557
572
, doi:.
Lamb
,
P.
,
1978
:
Large-scale Tropical Atlantic surface circulation patterns associated with Subsaharan weather anomalies
.
Tellus
,
30
,
240
251
, doi:.
Livezey
,
R. E.
,
1995
: The evaluation of forecasts. Analysis of Climate Variability: Applications of Statistical Techniques in Analysis of Climate Variability, H. von Storch and A. Navarra, Eds., Springer-Verlag, 177–196.
Madden
,
R.
, and
P.
Julian
,
1972
:
Description of global-scale circulation cells in the tropics with a 40–50 day period
.
J. Atmos. Sci.
,
29
,
1109
1123
, doi:.
Madden
,
R.
, and
P.
Julian
,
1994
:
Observations of the 40–50-day tropical oscillation: A review
.
Mon. Wea. Rev.
,
122
,
814
837
, doi:.
Mason
,
S. J.
,
2011
:
Seasonal forecasting using the Climate Predictability Tool (CPT)
.
Proc. 36th NOAA Annual Climate Diagnostics and Prediction Workshop
,
Fort Worth, TX
,
NOAA’s National Weather Service
,
180
182
.
Mo
,
K.
, and
W. M.
Thiaw
,
2002
:
Ensemble canonical correlation prediction of precipitation over the Sahel
.
Geophys. Res. Lett.
,
29
,
1570
, doi:.
Nicholson
,
S. E.
,
1980
:
The nature of rainfall fluctuations in subtropical West Africa
.
Mon. Wea. Rev.
,
108
,
473
487
, doi:.
Novella
,
N. S.
, and
W. M.
Thiaw
,
2013
:
African Rainfall Climatology version 2 for famine early warning systems
.
J. Appl. Meteor. Climatol.
,
52
,
588
606
, doi:.
Ogallo
,
L. J.
,
J. E.
Janowiak
, and
M. S.
Halpert
,
1988
:
Teleconnection between seasonal rainfall over East Africa and global sea surface temperature anomalies
.
J. Meteor. Soc. Japan
,
66
,
807
821
.
Rashid
,
H. A.
,
H.
Hendon
,
M. C.
Wheeler
, and
O.
Alves
,
2011
:
Prediction of the Madden–Julian oscillation with the POAMA dynamical prediction system
.
Climate Dyn.
,
36
(
3–4
),
649
661
, doi:.
Ropelewski
,
C. F.
, and
M. S.
Halpert
,
1987
:
Global and regional scale precipitation patterns associated with the El Niño/Southern Oscillation
.
Mon. Wea. Rev.
,
115
,
1606
1626
, doi:.
Saha
,
S.
, and Coauthors
,
2014
:
The NCEP Climate Forecast System version 2
.
J. Climate
,
27
,
2185
2208
, doi:.
Senay
,
G. B.
,
J. P.
Verdin
, and
J.
Rowland
,
2011
:
Developing an operational rangeland water requirement satisfaction index
.
Int. J. Remote Sens.
,
32
,
6047
6053
, doi:.
Smith
,
T. M.
,
R. W.
Reynolds
,
T. C.
Peterson
, and
J.
Lawrimore
,
2008
:
Improvements to NOAA’s historical merged land–ocean surface temperature analysis (1880–2006)
.
J. Climate
,
21
,
2283
2296
, doi:.
Thiaw
,
W. M.
,
A. G.
Barnston
, and
V.
Kumar
,
1999
:
Predictions of African rainfall on the seasonal timescale
.
J. Geophys. Res.
,
104
(
D24
),
31 598
31 597
, doi:.
Thiaw
,
W. M.
,
A. S.
Tokar
,
R.
Kolli
, and
I.
Gunes
,
2014
:
Climate variability and predictions: A NOAA–USAID Global Climate Training Workshop series
.
Bull. Amer. Meteor. Soc.
,
95
,
ES159
ES162
, doi:.
Wheeler
,
M. C.
, and
H. H.
Hendon
,
2004
:
An all-season real-time multivariate MJO index: Development of an index for monitoring and prediction
.
Mon. Wea. Rev.
,
132
,
1917
1932
, doi:.
Zhang
,
C.
,
2005
:
Madden-Julian Oscillation
.
Rev. Geophys.
,
43
,
RG2003
, doi:.

Footnotes

This article is included in the In Honor of Peter J. Lamb special collection.