Predictability of the moisture regime associated with the preonset of Sahelian rainfall

The present study seeks to understand the predictability of the primary attributes associated with the moisture regime in the preonset period of Sahelian rainfall. The moisture regime associated with the preonset of the West African Monsoon (WAM) is of primary concern for the population in the region due to its influence in areas such as health. As an example, there exists a robust and actionable weather/meningitis relationship given the strong correlation between the start of the rainy season and the abrupt decline in the transmission of the disease. If the onset of higher humidity can be forecast at the district level scale at which public health managers make decisions, then public health officials can plan to move a vaccine away from high-humidity districts where meningitis transmission will cease to low-humidity areas where meningitis transmission will persist. In order to accomplish this, public health decision makers need reliable information 10-14 days in advance and in a user-centered and user-friendly form. This need is currently unmet in West Africa, is the focus of the Google/ University Corporation for Atmospheric Research (UCAR) project. Our study is devoted to the link between Sahelian climate and the occurrence of meningitis in West Africa, its significance in terms of forecasting as well as aspects relating to the economic value attached to a forecast. Our study will serve as a compliment to work currently under development at UCAR by evaluating the skill of the Weather Research and Forecasting (WRF) model and other model forecasts of weather variables that are relevant to meningitis management in the region.

Given the influence that the onset of monsoon rains has on meningitis epidemics, we have dedicated our resources to diagnose the main characteristics of Sahelian climate leading up to the influx of moisture. The preonset period of the WAM is defined as the arrival in the intertropical front (ITF) at 15 N, where the confluence line between moist southwesterly monsoon winds and dry northeasterly Harmattan occurs. In particular, we investigate how sea surface temperatures adjacent to the region, along with teleconnections with mid-latitude systems and global circulation patterns, modulate the moisture content of the region on an interannual scale. Further, we explore the role that regional dynamics of the Saharan heat low and the African Easterly Jet as well as transient disturbances such as mid-latitude systems and westward-propagating convective systems have in dictating intraseasonal variability of the moisture front dynamics. Initial results using parcel back-trajectory analysis has provided new insight of the pathways for atmospheric humidity over the Sahel during the dry season, the transition period, and the early monsoon. Air masses can be traced to upper levels of the atmosphere from the mid-latitudes during the Harmattan, and to lower level parcels traversing the South Atlantic and Gulf of Guinea regions during the monsoon.

We use the Weather and Research Forecasting (WRF) model as both a predictive and analytical tool to create high resolution information of moisture conditions at a district scale appropriate for local health efforts. We hypothesize that using a regional model like WRF will capture dynamics important for predicting moisture advection into the region such as local moisture convergence and static instability. The model is configured over West Africa with two domains: outer domain at 90km (45W-40E, 16S-40N) and inner domain at 30km resolution (28.5W-21E, 7S-23.5N). We use the WSM3 class simple ice microphysical scheme, CAM radiation scheme, Noah Land Surface and the Kain-Fritsch cumulus scheme. The available hindcast period runs from 2005 to the present, including real-time forecast analysis. As part of the project, data assimilation will be included using both in-situ and satellite information to create a high resolution reanalysis using WRF at 30km for the period 2000-2009, during which more reliable meningitis information for the region of study is available.

Within our downscaling efforts, we employ a recently-developed spectral nudging method that will preserve the large scales in the GFS reanalysis and real-time forecast data. This new technique involves two-way nesting based on Scale-Selective 3 D Variational Data Assimilation (SSDA) and should aid us in determining the relationship between large-scale processes and the moisture regime in the region of study. This technique will be tested against the spectral nudging option using gridded analyses available with WRF Version 3.1. Forcing in the WRFV3.1 spectral nudging technique is not limited to the lateral boundaries but also in the model interior. The addition of nudging terms to the model equations of horizontal winds in the spectral domain maintains maximum efficiency in large scales and no effect in small scales.

Preliminary results for our study include model configuration and optimization for the region of study, real-time ensemble forecasts for the relevant variables during the 2009 season, analysis of model skill using gridded reanalysis and in-situ observations, as well as the analysis of important predictors of surface moisture such as winds at 925mb, SSTs in the Gulf of Guinea and the Eastern Tropical Atlantic Ocean and Mid-Latitude influences. The model is able to capture intraseasonal as well as diurnal variability, both which are highly important in the study of the WAM. Unlike the global model-based reanalysis data, WRF simulations for 2009 accurately reproduce the anticorrelation between onset of observed high humidity and the abrupt decline in the transmission of the disease over Kano, Nigeria. We find that westward-propagating systems have an important influence in the modulation of moisture regime at latitudinal and temporal scales. The implication of transient convective systems on surface moisture is evident in our analysis and WRF allows us to trace a precipitation event on May 7-8, 2009 at Niamey, Niger as well as the associated spike in moisture to a westward-propagating disturbance. This disturbance is also evident in the NCEP/NCAR reanalysis as well as observation reports from Kano, Nigeria and Bamako, Mali. As each one of the transient disturbances progresses, the net amount of moisture at the surface increases to a higher regime. Our continuing efforts are to further explore this mechanism along with other aspects of the intraseasonal variability of the moisture front.