First West African Monsoon Modeling and Evaluation
(WAMME) Workshop
in 2008 American Meteorological Society Annual Meeting

(Workshop Abstracts)

Yongkang Xue¹, William Lau, Kerry Cook, David Rowell, Aaron.Boone, Jinming Feng, Fernsndo De Sales, Paul Dirmeyer, Leonard M. Druyan, Matthew Fulakeza, Zhichang Guo, S. M. Hagos, Kyu-Myong Kim, Akio Kitoh, Abdourahamane Konare, Vadlamani Kumar, Benjamin Lamptey, Patrick Lonergan, Natalie Mahowald, Wilfran Moufouma-Okia, Phil Pegion, Jae Schemm, Siegfried D. Schubert, Wassila Thiaw, Augustin Vintzileos, Edward K. Vizy, Steve Williams, Man-Li C. Wu

Semi-arid Sahelian Africa is one region that has experienced significant changes in climate in the past century. Annual rainfall has persistently remained below the long-time average since the late sixties. Starting from the late 1980s, however, there is evidence of some rainfall recovery relative to the very dry period. Despite recent progress in understanding the effects of boundary forcing on West African monsoon (WAM) variability, the monsoon precipitation and associated important features, as well as their strong and complex interactions with oceanic, land processes, and aerosols, are still not well understood, due in part to lack of detailed observations and to the inability of general circulation models (GCMs) to simulate these features at different temporal and spatial scales.

The West African Monsoon Modeling and Evaluation project (WAMME), a CEOP Inter-Monsoon Study (CIMS) initiative, uses GCMs and regional climate models (RCMs) to address issues regarding the role of ocean-land-atmosphere interactions, land-use and water-use changes, vegetation dynamics, as well as aerosols, particularly dust, on the WAM development as well as the long-term drought and partial recovery. WAMME also has close collaborations with the African Monsoon Interdisciplinary Analysis (AMMA) project and will use AMMA field data for validation.

In the first phase, a set of experiments consisting of four spring and summer seasons in the 21th century with different climate characteristics were selected to test the state-of-the-art models’ ability to simulate the WAM’s basic features on diurnal, intraseasonal, seasonal, and interannual scales. The model intercomparison emphasizes the WAM evolution, its onset and demise, precipitation intensity and frequency, and some important modes. The evaluation of the simulated rainfall at diurnal, intraseasonal, and seasonal scales and its link with the large scale dynamics (such as strength of the African Easterly Jet (AEJ) and penetration of the monsoon flow) and external forcing (such as SST, meridional gradients of surface temperature and soil moisture, surface evaporation and energy partitioning) will be the primary focus. Interannual variability will also be documented. The observational data from AMMA will be applied to evaluate the model simulation. In addition, the RCM’s downscaling ability will be evaluated when reanalyses as well as GCM outputs are imposed as lateral boundary conditions. The RCM’s ability to reproduce the sudden jump in rainfall at monsoon onset and in improving the representation of the AEJ and diurnal variability will be the focus.

We believe the results from this first set of experiments could be a good starting point providing benchmarks for further studies to understand the roles of external forcing, and internal dynamics in WAM variability.

Modeling interannual variability of the WAM jump using WAMME regional model simulations

Kerry H. Cook, J. F. Newman, E. K. Vizy, and S. M. Hagos

Despite the smooth northward progression of insolation during boreal spring, the northward progression of the rainfall maximum is discontinuous. In spring and early summer, the precipitation maximum lingers over the Guinean coast. Then, over the course of a few days any time between the middle of June and the middle of July, the rainfall maximum moves into the continental interior. This event is known as the West African monsoon jump, and it can also be termed "monsoon onset" for Sahelian Africa. Hagos and Cook (2008) show that the monsoon jump occurs as a result of the development of inertial instability over the Guinean coast. The purpose of this paper is to use that improved understanding of the basic dynamics of the jump to better understand the interannual variability of rainfall and to benefit intraseasonal prediction.

Regional model simulations produced for WAMME capture the monsoon jump. They are analyzed to understand why the inertial instability develops and to relate its development to known circulation features. In an idealized context, the criterion for inertial instability is that the meridional shear of the zonal wind exceeds the planetary vorticity. We show that this occurs over West Africa when the low-level westerly jet develops, bringing sufficiently strong positive meridional gradients in the zonal wind to the Guinean coast region which induces the inertial instability and the jump.

West African summer monsoon climate: A comparison of RM3 downscaled analyses to downscaled GCM forecasts

Leonard M. Druyan^1,2, Matthew Fulakeza^1,2 and Patrick Lonergan^1,2

1.Center for Climate Systems Research, Columbia University, 2880 Broadway, New York, NY 10025
2.NASA/Goddard Institute for Space Studies, 2880 Broadway, New York, NY 10025

The seasonal evolution of the West African summer monsoon is simulated in two modes for several seasons by the RM3, a regional, limited area climate model run at the Center for Climate Systems Research (Columbia University) and the Goddard Institute for Space Studies (NASA). In Mode 1, the RM3 is driven by lateral boundary data from NCEP reanalysis 2, creating a higher resolution 3-D climate analysis on the RM3 horizontal 0.5_ grid at 28 vertical levels. In Mode 2, the RM3 is driven by lateral boundary data from UK Meteorological Office GCM seasonal predictions, forced by prescribed sea-surface temperatures based on observations. Comparison of the results from these two modes of simulations informs about the potential quality of seasonal climate predictions using the RM3/UK GCM coupled system. Diagnostics of special interest will be presented which relate to the behavior of RM3 simulated African easterly waves for each of the two modes. Validations of RM3 precipitation are made against TRMM and CPC/FEWS data. Validations of temperatures and circulation are made against NCEP reanalysis 2 data. In addition, the presentation will discuss characteristics of the RM3 climate during the SOP-3 and comparisons with AMMA field measurements.

West African Monsoon in a 20km-Mesh Atmospheric GCM

The West African monsoon in the GFS runs for WAMME

Vadlamani B. Kumar, Wassila M. Thiaw, and Jae K, Schemm

Climate Prediction Center
National Centers for Environmental Predictions
National Weather Service/NOAA
Washington DC 20233
 

As part of the West African Monsoon Modeling Experiment (WAMME), NCEP has made runs of the global forecast system (GFS) T126L64 using Hadley Center SST for boundary conditions. We used NCEP Reanalysis 2 for initial conditions and started the runs from April 1, 00 UTC (one member only) through October 31^st for the five years 2000, 2003-2006. Model outputs include both flux and pgb files from surface to top of the troposphere. Preliminary assessments of the GFS indicate that the model has a wet bias in the precipitation field. Seasonal rainfall totals for the period June-September extends well north into the Sahel. The areas of precipitation maximum in the Guinean highlands, the eastern end of the Gulf of Guinea region, and the Ethiopian western Plateau are well depicted but rainfall amounts are much larger than observed. Consistent with the wet bias, the African Easterly Jet (AEJ) in the GFS is shifted northward and the tropical easterly Jet (TEJ) stronger than observed. Low level westerlies are also enhanced in the GFS. The model’s annual cycle is amplified but tends to mimic the observations. Representations of key features of the West African monsoon system are discussed. We also compare results with other NCEP models runs and diagnose causes of model errors.

A. Laing, R. Pandya, R. Bruintjes, B. Lamptey, P. Kucera, F. Semazzi, T. Yoksas, M. Ramamurthy, M. Weingroff, T. Spangler, A. Traore, M. Konate, N. Fall, R. Boger, T. Warner, S. Herrmann, M. Moncrieff and R. Low

The UCAR Africa Initiative (AI, www.africa.ucar.edu) is a collection of efforts aimed at building sustainable partnerships between UCAR and African institutions in support of meteorological research and applications in Africa. Our initial focus on short-term weather prediction represents an intersection of the needs identified by Africans and the expertise of UCAR. The initiative envisioned will:

These principles are realized in three pilot activities, chosen for their high probability of short-term results and alignment with longer-term collaboration: (i) A modest radar network and data-distribution system in Mali and Burkina Faso, including a data-sharing MOU between the Mali and Burkina Faso Weather Services; (ii) Operational forecasts for West Africa using the Weather Research and Forecasting (WRF) model; the output is used by Ghanaian meteorologists. Model output is also part of a demonstration project aimed at allowing humanitarian agencies to share geo-referenced African information via a web portal; (iii) "Sahel 2007: Improving Lives by Understanding Weather" workshop in Ouagadougou, Burkina Faso, 2-6 April 2007. The workshop, which had over 80 participants from 18 countries, included hands-on training on UCAR-developed tools and technology, setting recommendations for collaboration, and a field trip to Lycee Nelson Mandela, where a GLOBE weather station was installed.

Recognizing that universities and agencies are already leading efforts in atmospheric-science capacity building in Africa, we are exploring how UCAR can align with existing efforts. UCAR scientists are involved with these efforts through collaborative research and serving on working groups and committees, such as those of AMMA and THORPEX. Contributions include: rain gauge installations and rainfall analysis in Senegal; an ensemble data assimilation system using WRF over northern Africa; a climatology of the statistics of deep convection in Africa, provided as a web resource, facilitates forecasting in a probabilistic manner; analysis of remotely-sensed vegetation in Burkina Faso to assess vegetation dynamics, rainfall, and climate. Furthermore, COMET and NCEP provided numerical weather prediction (NWP) training at the "WRF modeling and the Use of GEFS NAEFS in Operational Weather Forecasting", South Africa, 3-12 October 2007. Representatives from nine countries attended and received computers to establish their own regional modeling systems. In addition, COMET modules are used in 40 African countries. Finally, Prof. Fred Semazzi, North Carolina State University, who is on sabbatical at UCAR to help lead the initiative, is focusing on building connections between UCAR activities and the university community, in the U.S. and abroad. Our presentation will provide an update of these and other activities being undertaken by the UCAR African Initiative, and explore avenues for sustained collaboration.

Chris Folland, Met Office Hadley Centre, Exeter, UK and Jim Kinter, COLA, Maryland, USA

We describe the purposes of C20C and how C20C evolved from its beginnings in the mid 1990s. Particular emphasis is given to the differences in objectives between AMIP and C20C, and the experimental protocols that have been developed in C20C. There have been four full Workshops of C20C and one Special Meeting. The achievements of C20C from the first three Workshops and the Special Meeting are summarised. C20C has evolved from a purely Atmospheric General Circulation Model based set of experiments to include coupled models and "Pacemaker" models. We discuss the relationship between these approaches, all seen as different types of dynamically based tools for studying climate variability and predictability on seasonal to multidecadal time scales. Such tools of course require high quality observed data of various kinds to complement them. The key global data sets are described, where considerable progress has been made in the last few years.

Finally, particular emphasis is given to the discussions at, and results of, the fourth Workshop during which initial results of Pacemaker experiments were described and plans for further experiments were elaborated. A brief summary of future C20C plans is also provided, including a discussion of the ongoing development of links between C20C and other international projects, such as WAMME.

PM Ruti¹,  F. HOURDIN ²,  S. Janicot³

1. ENEA, Roma, Italy
2. Laboratoire de Météorologie Dynamique, CNRS, IPSL, Paris, France
3. IRD, Paris, France
 

West Africa is characterized by well defined strong meridian surface gradients coupled to specific atmospheric circulations, such as the African Easterly jet (AEJ) which is present during the monsoon season. The location of the AEJ itself is strongly constrained by meridian surface temperature and moisture gradients. Synoptic variability in turn is dominated by African easterly waves (AEW) which are dynamically linked to the AEJ.

The structure and variability of these basic large-scale features involve complex interactions with soil, surface, turbulent and convective processes occurring on different scales. Finally, the WAM exhibits specific seasonal variations, with an abrupt monsoon onset to be compared to a more progressive latitudinal retreat.

While current numerical weather prediction (NWP) analyses seem able to reasonably capture these large-scale atmospheric features, the extent to which large-scale models are able to properly reproduce these observations remains unclear, and likely sensitive to changes in the physical parametrizations.

Here, we present the effort of the AMMA-EU project in order to inter-compare several Global and Regional Climate Models, and to analyze their sensitivity to surface boundary conditions.

Some Thoughts on SST Forcing of Sahel Rainfall

David P. Rowell

Met Office Hadley Centre

It is well known that a large proportion of the recent variance of seasonal Sahel rainfall occurs at decadal timescales, leading to long periods of either mostly abundant or inadequate rainfall. Unfortunately, however, climate models are unable to simulate the full amplitude of this decadal variability – their spectra are too white – even when forced by observed SSTs. Two explanations for this serious shortcoming have been proposed. First, it may be that some of the teleconnection mechanisms from relevant oceanic regions are poorly modelled, and/or that the rainfall response to the resultant West African circulation anomalies is poorly modelled. Second, it is possible that an important process that provides multi-annual memory – and so reddens the spectrum – is missing from the models. Here we concentrate on the former hypothesis, and describe a simple analysis of observational data which suggests that SST influences alone (without a supplementary memory source) have at least the potential to induce a sufficiently red spectrum of seasonal Sahel rainfall. If this reflects reality, then it becomes critical to seek improved modelling of SST-Sahel teleconnections. This further motivates one of the key strands of WAMME, which is to improve model simulations of the West African monsoon. The presentation concludes with a number of questions to stimulate discussion on the role that WAMME might play in describing, understanding and resolving errors in the models’ representation of decadal (and interannual) variations of the West African monsoon.

Chris Thorncroft

University at Albany, SUNY

From a large-scale perspective the West African monsoon (WAM) can be described in terms of the annual march of the ITCZ and its associated regional circulations. On the synoptic and mesoscale, the WAM is comprised of a complex collection of wave patterns, organized weather systems and deep convection. These include synoptic systems such as African easterly waves (AEWs) and mesoscale convective systems (MCSs). Since MCSs provide most of the rainfall over West Africa it may be argued that the WAM is strongly linked to the statistics of these MCSs and that the variability in the WAM in turn is linked to variability in these statistics. In this context it is important that we improve our understanding of the 2-way interactions between the MCSs and the synoptic environment in which they develop including, in particular, the interactions with the African easterly jet and AEWs, features that we hope to explicitly predict in weather and climate models.

Improving our understanding of these scale interactions is hindered by lack of appropriate observations. At present we rely heavily on the use of high resolution models to explore these issues but these need to be evaluated with appropriate observations. Some thoughts on the use of models and observations to explore scale interactions in the WAM will be provided.

Land-atmosphere coupling strength in Africa: results from GLACE multi-model experiments

Zhichang Guo, Center for Ocean-Land-Atmosphere Interactions, USA

Paul Dirmeyer, Center for Ocean-Land-Atmosphere Interactions, USA

Randal D. Koster, NASA Goddard Space Flight Center, USA

The multi-model intercomparison of land-atmosphere coupling strength in Africa is conducted using data sets from the Global Land-Atmosphere Coupling Experiment (GLACE). The general features of the coupling and the extent to which coupling strength varies among the participating GCMs are evaluated. The inter-model differences in coupling strength in Africa are also studied in terms of soil moisture's ability to affect evaporation and evaporation's ability to affect precipitation.

The impact of soil moisture gradient on the Africa Easterly Jet and West Africa Monsoon

Wu, Man-Li C., Siegfried D. Schubert, Oreste Reale, Max J. Suarez, Randy D. Koster, Philip J. Pegion

The impact of soil moisture on surface temperature gradient and thus on the formation of the African Easterly Jet (AEJ) and on the strength and location of the West Africa monsoons (WAMs) will be presented. A number of global atmospheric general circulation (AGCM) model experiments, using the NASA Seasonal-to-Interannual Prediction Project AGCM version 1 (NSIPP1), with different soil moisture configurations are being carried out to reveal the controls of the surface temperature gradient, which consequently affect the formation of the AEJs and affect the strength and location of the WAMs.

Regional modeling study of the effect of Saharan dust on the West African monsoon

A. Konare¹, A.S. Zakey^*,2, F. Solmon³, F. Giorgi², S. Raucher², S. Ibrah⁴, X. Bi²

1. Laboratoire de Physique de l’Atmosphère, Université de Cocody, Abidjan, Côte d’Ivoire.
2. International Centre for Theoretical Physics (ICTP), Trieste-Italy.
3. Laboratoire d’Aérologie, Toulouse, France.
4. Département de Physique, Université de Niamey, Niger.
* On leave from Egyptian Meteorological Authority (EMA), Cairo-Egypt

 

We investigate the effect of the shortwave radiative forcing of Saharan dust on the West Africa monsoon with a regional climate model interactively coupled to a dust model. Towards this purpose we intercompare sets of 38 summer monsoon season simulations (1969-2006) with and without dust effects over a domain encompassing most of the African continent and adjacent regions. We find that the main effect of the dust radiative shortwave forcing is to reduce precipitation over the Sahel region. This is in response to cooling over the Sahara, which decreases the meridional gradient of moist static energy and results in a weakening of the monsoon energy pump. The dust effects also cause a strengthening of the southern branch of the African Easterly Jet and a weakening of Tropical Easterly Jet. Over the Sahel the dust forcing causes climate response patterns that are similar to those found during dry years over the Sahel, which suggests that Saharan dust feedbacks might have a role in maintaining drought events over the region. Overall, the inclusion of dust also tends to improve the model simulation of the West Africa monsoon, as well as African and tropical Easterly jet.

Edward K. Vizy, Cornell University, USA., E. E. Riddle, S. M. Hagos, C. M. Patricola, and K.H. Cook

Careful thought must be partaken when utilizing a regional climate model to study climate on monthly to seasonal timescales. Issues such as domain placement, selection of model parameterizations, and choice of initial and lateral boundary conditions need to be addressed to produce a model simulation that can best simulate the seasonal evolution of any climate system.

In this talk we will highlight some of the important lessons that our group has learned over the past 6 years using MM5 and more recently WRF to simulate the evolution of West African monsoon system. Results will be placed in the context of offering potential design changes that would improve our WAMME contribution using MM5.

On the evolution of African Easterly Waves and precipitation systems over the **Sahel** as a function of horizontal model resolution *

Augustin Vintzileos^{1, 2}

1. EMC/NCEP/NWS/NOAA and 2. UCAR

Realistic simulation and consequently better prediction of precipitation over the Sahel during the West African Monsoon season may lead to significant societal benefits. These benefits concern not only the local population through better forecasts at lead times from synoptic to inter-decadal but also remotely located economic interests as a significant number of named tropical storms is seeded in this area. In fact a good forecast of the statistical characteristics of the African Easterly Waves is one of the key elements for forecasting the statistics of Atlantic Tropical Storms at subseasonal-to-seasonal time scales. In a series of 60-day hindcast that we performed under NOAA’s Climate Test Bed (CTB) we tested the importance of horizontal resolution of the atmospheric component of the Climate Forecasting System (CFS). We ran the CFS at three resolutions T62 (200 km x 200 km), T126 (100 km x 100 km) and T254 (50 km x 50 km) initialized by Reanalysis-2 and by the NCEP operational analysis (GDAS). In this paper we will concentrate in the mean August precipitation characteristics computed over an ensemble of 70 hindcast initialized during the first part of July from 2002 to 2006. We show that the increase of model resolution from T62 to T126 has marginal effects. We then show that a further increase of the resolution to T254 results to a significant northward shift of the mean August precipitation towards the Sahel leading to a more realistic simulation. Here we discuss how the characteristics of the Easterly African Waves are affected by resolution.

Samson M. Hagos, Kerry H. Cook, Edward K. Vizy, Emily E. Riddle** and Christina M. Patricola,

**Presented by Emily E. Riddle

Regional atmospheric models (e.g., WRF, MM5) are widely used to investigate regional climate sensitivity, and understand processes that occur at scales too small to be captured with a GCM. However, many of these processes, even at small scales, involve complex interactions and feedbacks between the atmosphere, oceans and land-surface. To capture these feedbacks, regional climate models must incorporate land-surface and ocean modules. Our research group at Cornell has recently developed a coupled regional climate model for the tropics which interactively links existing atmospheric and landsurface models with a simple mixed-layer model of the ocean. The model allows for dynamic and thermodynamic coupling between the atmospheric boundary layer and an ocean mixed layer with variable prescribed depth. This coupled model is then used to investigate interactions between the West African monsoon and the eastern tropical Atlantic Ocean.

Ibrah Seidou Sanda ¹, Wilfran Moufouma-Okia³, Abdourahamane Konare², Joseph Intsiful³, David Hassell ³, and David Hein ³

1. Faculte des Sciences, Universite Abdou Moumouni, BP 10662 Niamey, Niger
2. LAPAMF, Université de Cocody-Abidjan, BP 231 Abidjan, Cote d’Ivoire.
3. Hadley Centre for Climate Prediction and Research, UK
 

We present an overview of the results from four simulations of the West African Monsoon by the UK met Office Regional Climate Model PRECIS. The chosen domain (10S-40N; 30W-20E) is different from the WAMME standard domain and does not include the East African Highlands. As required by the WAMME protocol, the model is run at a 50 kilometer resolution from April to October for each of the years 2000, 2003, 2004 and 2005. The lateral boundary conditions are provided by the NCEP/R2 and the sea surface temperatures (SST) are the HadISST.