HEAT HEALTH IN BANGLADESH DEFINING AND PREDICTING HEAT WAVES ACROSS TIMESCALES Hannah Nissan1, Katrin Burkart2, Simon Mason1
Introduction Predicting heat waves It is now well established that extreme heat poses During the heat wave season a serious health risk, causing many excess deaths (April – June), a zone of each year. Heat early warning systems are known discontinuity separates two to save lives by improving preparedness, and air masses in Bangladesh: dry should form an important component of a climate westerlies from India and change adaptation strategy. However, very little is moist southerlies from the known about heat waves in Bangladesh. Bay of Bengal, which bring early pre-monsoon rainfall to the country. Sanderson & Rafique (2009) In Bangladesh, both Hyd. Sci. Bull. 24:3 hot and cold mortality On heat wave days, the dry westerlies are effects are evident. A stronger, and the moist southerlies are weaker, sharp increase in than normal for this time of year. This reduces mortality at high rainfall below normal pre-monsoon levels. These temperatures shows warning signs can be seen 8-10 days before a that heat waves are a heat wave begins. major public health Log relative risk of mortality with concern. temperature, 2003-2007 (black line and grey range). Average daily temperature distribution (bars).
Defining a heat wave To prepare for a heat wave, decisions must be Anomaly composite of 850hPa wind on heat wave days, 1989-2011.iii taken over when to issue a warning, and this Rainfall is highly variable, but detecting changes requires a heat wave definition that is both in soil moisture would indicate that predictability a) related to human health outcomes of heat waves beyond a few days may be b) forecastable using available weather & climate possible. We found that heat waves are information. associated with unusually dry soil moisture during the whole heat wave season. This suggests that We used regression modelling to test six possible seasonal and sub-seasonal forecasts of heat wave heat wave definitions against mortality data from risk may be possible by monitoring soil moisture 2003-2007.i The six definitions tested combined conditions in Bangladesh. minimum (night) and maximum (day) temperature and heat index. High minimum and maximum Anomaly composite of total soil moisture in ii th temperatures (>95 percentile) for 3 consecutive advance of heat waves in days was the best predictor of mortality in Bangladesh, 1989-2011. Bangladesh.
Heat waves in Seasonal total number Bangladesh occur of heat waves (blue), during the pre- precipitation (green) and soil moisture monsoon season, (orange), 1989-2011. from April to June. Correlations between number of heat wave days and total rain and Seasonality of heat wave occurrence soil moisture were -0.3 and -0.6, respectively. i Mortality data from the Sample Vital Registration System, Bangladesh Bureau of Statistics ii Temperature data from the Bangladesh Meteorological Department Iii Synoptic climate fields taken from the ECMWF ERA Interim Reanalysis from 1989-2011
1 International Research Institute for Climate and Society, Columbia University 2 Mailman School for Public Health, Columbia University