Developing An Explanatory Prediction Model Based On Rainfall Patterns For Cholera Outbreaks In Africa
Abstract
Cholera has become endemic in coastal and inland areas within the tropics as well as areas outside of the tropics in Africa. Climate conditions and weather patterns differ between areas reporting cholera cases in Africa. Some areas experience two rainfall seasons compared to areas with only one rainfall season in a year. Further, climate variability or ENSO events affect local weather conditions differently. La Niña, i.e. cold events lead to higher than normal rainfall in areas in southern Africa compared to areas close to the equator in eastern Africa which report less than normal rainfall. Time series analysis of cholera cases and rainfall data at different spatial resolutions highlight the overlap of the rainfall season with the reporting of cholera cases. Cholera cases are also reported in between rainy seasons in different areas but the incidence is significantly less compared to the rainy season. An increase in the intensity of outbreaks is also noted during the rainy season following a drier than normal 'dry' season. This necessitates the understanding of the reasons for the observed correlation between rainfall season and cholera outbreaks in order to develop a prediction model which can accurately predict the likelihood of an outbreak. Due to the complexities associated with accurately predicting weather data more than seven days ahead of time it is necessary to identify global drivers with a lagged effect on local rainfall patterns. Climate variability, i.e. ENSO is investigated at different temporal scales; spatial locations and time lags. Sea surface temperature anomalies (SSTa) measured closed to the equator and in the southern parts of the Indian Ocean are more closely associated with rainfall anomalies at specific time lags in equatorial, East African, south East African and central African areas compared to SSTa measured in different regions in the Pacific Ocean. An explanatory prediction model is developed for conditions in Mozambique (coastal area) and Uganda (inland area) which is not only based on correlation study results but also on the identification of cause-effect mechanisms. This is done by following an integrative multidisciplinary approach which involves the integration of laboratory and field study results, in situ and satellite data, and modeled data. We conclude that a prediction model for early warning and intervention purposes needs to be based on the identification and understanding of cause-effect mechanisms associated with the correlation between cholera outbreaks and rainfall; be parametrized for local conditions; and be based on a driver(s) or proxy for a driver(s) which allows sufficient time for decision makers to act.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2012
- Bibcode:
- 2012AGUFM.H23I..03V
- Keywords:
-
- 1807 HYDROLOGY / Climate impacts;
- 1872 HYDROLOGY / Time series analysis