main concepts scientists sections
SimplePopularAdvanced

Warm ocean warned of cholera eight months ahead ✎

Original: "Cholera forecast for Dhaka, Bangladesh, with the 2016 El Niño"
· Pamela P. Martinez, Robert C. Reiner, Manojit Roy, Benjamin A. Cash, Md. Yunus, A. S. G. Faruque, Sayeeda Huq, Aaron A. King, Mercedes Pascual
arXiv:1607.01008 · 2016-07-04 · arXiv non-exclusive · independent analysis · 1 min
q-bio.PE
Water temperature in the Pacific Ocean helped predict a cholera outbreak in Dhaka long before it began.

In January 2016, the Pacific Ocean was 2.6 degrees warmer than usual. That is far from Bangladesh, but it was this warmth that helped predict a large number of cholera cases in Dhaka that autumn.

Scientists took years of disease data for the city and built two models. One looks at how the disease spreads from person to person and through water. The other divides the city into districts and tracks how outbreaks move from one district to another. Both models used the same signal: ocean temperature in January. If the water is warm, expect an outbreak.

The test showed that in calm years from 2011 to 2015, the models gave no false alarms. The probability of a major outbreak was below 29%. But for 2016, both models gave more than 87%.

Comparison with the past confirms this: in January 1998, the ocean was almost as warm, and Dhaka suffered its worst cholera outbreak. Now doctors have months to prepare: stockpile medicines, ready hospitals, and warn the public. The ocean thermometer has become an early warning signal.

Whole graph →

The sea surface temperature anomaly in the Niño-3.4 region in January 2016 was 2.60°C, almost matching the 2.56°C anomaly in January 1998, when Dhaka experienced its largest cholera outbreak in 20 years and a flood that submerged more than half the city.

\lambda(t) = e^{-\nu(t-t_0)} \left[ \beta(t) \frac{I}{N} + \omega \right]
The rate at which a susceptible person becomes infected is the sum of a time-decaying trend, transmission from infected individuals with coefficient β, and a constant environmental source ω.
\beta(t) = \exp\left[ \sum_{j=1}^{6} b_j s_j(t) + \sum_{m=4}^{5} b_{Em} s_m(t) f(ENSO) \right] \frac{d\Gamma}{dt}
Seasonality is captured by six splines, the El Niño influence enters through the fourth and fifth splines via the function f(ENSO), and stochasticity is introduced through a gamma process.
f(ENSO) = A \left[ \frac{\tanh(ENSO_{Jan}(t))}{\tanh(1)} \right]
The January temperature anomaly is transformed by a hyperbolic tangent, producing a nonlinear response that saturates for large anomalies.
Concepts
La Niña ENSO