Limits...
Complex temporal climate signals drive the emergence of human water-borne disease.

Morris A, Gozlan RE, Hassani H, Andreou D, Couppié P, Guégan JF - Emerg Microbes Infect (2014)

Bottom Line: Predominantly occurring in developing parts of the world, Buruli ulcer is a severely disabling mycobacterium infection which often leads to extensive necrosis of the skin.While the exact route of transmission remains uncertain, like many tropical diseases, associations with climate have been previously observed and could help identify the causative agent's ecological niche.From this, it was possible to postulate for the first time that outbreaks of Buruli ulcer can be triggered by combinations of rainfall patterns occurring on a long (i.e., several years) and short (i.e., seasonal) temporal scale, in addition to stochastic events driven by the El Niño-Southern Oscillation that may disrupt or interact with these patterns.

View Article: PubMed Central - PubMed

Affiliation: Bournemouth University, Dorset BH12 5BB , UK ; UMR MIVEGEC, IRD-CNRS-Universités de Montpellier 1 et 2, Centre IRD de Montpellier, 34394 Montpellier cedex 5 , France.

ABSTRACT
Predominantly occurring in developing parts of the world, Buruli ulcer is a severely disabling mycobacterium infection which often leads to extensive necrosis of the skin. While the exact route of transmission remains uncertain, like many tropical diseases, associations with climate have been previously observed and could help identify the causative agent's ecological niche. In this paper, links between changes in rainfall and outbreaks of Buruli ulcer in French Guiana, an ultraperipheral European territory in the northeast of South America, were identified using a combination of statistical tests based on singular spectrum analysis, empirical mode decomposition and cross-wavelet coherence analysis. From this, it was possible to postulate for the first time that outbreaks of Buruli ulcer can be triggered by combinations of rainfall patterns occurring on a long (i.e., several years) and short (i.e., seasonal) temporal scale, in addition to stochastic events driven by the El Niño-Southern Oscillation that may disrupt or interact with these patterns. Long-term forecasting of rainfall trends further suggests the possibility of an upcoming outbreak of Buruli ulcer in French Guiana.

No MeSH data available.


Related in: MedlinePlus

Monthly time series data showing the decomposition, reconstruction and forecasting of datapoints using SSA. (A) Original rainfall time series average for 17 weather stations along the coast of French Guiana, from 1969 to 2012. (B) Periodograms of the rainfall time series identifying significant repeating patterns once per year, twice per year and three times per year. SSA extracted component corresponding to periodogram spike of (C) three times per year, 4-month component, (D) twice per year, 6-month component and (E) once per year, 12-month component. (F) The extracted rainfall trend. (G) The reconstructed rainfall time series after the removal of stochastic noise. (H) A second periodogram of the reconstructed rainfall series showing less stochastic noise around the three main repeating patterns. (I) Forecasting of the rainfall trend to 2017 using sequential SSA.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4150285&req=5

fig2: Monthly time series data showing the decomposition, reconstruction and forecasting of datapoints using SSA. (A) Original rainfall time series average for 17 weather stations along the coast of French Guiana, from 1969 to 2012. (B) Periodograms of the rainfall time series identifying significant repeating patterns once per year, twice per year and three times per year. SSA extracted component corresponding to periodogram spike of (C) three times per year, 4-month component, (D) twice per year, 6-month component and (E) once per year, 12-month component. (F) The extracted rainfall trend. (G) The reconstructed rainfall time series after the removal of stochastic noise. (H) A second periodogram of the reconstructed rainfall series showing less stochastic noise around the three main repeating patterns. (I) Forecasting of the rainfall trend to 2017 using sequential SSA.

Mentions: Periodograms of the raw rainfall time series (Figure 2A) identified seasonal components which oscillated yearly for 4-, 6- and 12-month periods (repeating patterns occurring, tri-annually, bi-annually and once per year, Figure 2B) the signals of each repeating oscillation were isolated along with the long-term trend of rainfall (Figures 2C–2F). Remaining data were considered noise and the components were reconstructed into a noise-free signal (Figure 2G), a second periodogram of the new reconstructed signal shows the repeating patterns are intact, while stochastic noise has been removed (Figure 2H). Figure 2I shows the resulting output of forecasting from sequential SSA, with rainfall beginning to decline as French Guiana enters a trough of dryer years after several years of high rainfall. The 4-month component corresponds to two rainy seasons, one long rainy season and one short, and also to the main dry season from August to December (Figure 3A). The 6-month bi-annual component corresponds to the two rainy seasons and is a reflection of the strength of these two seasons (Figure 3B), while the 12-month component shows the overall rainfall level for the year (Figure 3C). Separability of these components was confirmed with a ω-correlation matrix showing that these seasonal components did not show high levels of correlation with other components (Figure 4).


Complex temporal climate signals drive the emergence of human water-borne disease.

Morris A, Gozlan RE, Hassani H, Andreou D, Couppié P, Guégan JF - Emerg Microbes Infect (2014)

Monthly time series data showing the decomposition, reconstruction and forecasting of datapoints using SSA. (A) Original rainfall time series average for 17 weather stations along the coast of French Guiana, from 1969 to 2012. (B) Periodograms of the rainfall time series identifying significant repeating patterns once per year, twice per year and three times per year. SSA extracted component corresponding to periodogram spike of (C) three times per year, 4-month component, (D) twice per year, 6-month component and (E) once per year, 12-month component. (F) The extracted rainfall trend. (G) The reconstructed rainfall time series after the removal of stochastic noise. (H) A second periodogram of the reconstructed rainfall series showing less stochastic noise around the three main repeating patterns. (I) Forecasting of the rainfall trend to 2017 using sequential SSA.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4150285&req=5

fig2: Monthly time series data showing the decomposition, reconstruction and forecasting of datapoints using SSA. (A) Original rainfall time series average for 17 weather stations along the coast of French Guiana, from 1969 to 2012. (B) Periodograms of the rainfall time series identifying significant repeating patterns once per year, twice per year and three times per year. SSA extracted component corresponding to periodogram spike of (C) three times per year, 4-month component, (D) twice per year, 6-month component and (E) once per year, 12-month component. (F) The extracted rainfall trend. (G) The reconstructed rainfall time series after the removal of stochastic noise. (H) A second periodogram of the reconstructed rainfall series showing less stochastic noise around the three main repeating patterns. (I) Forecasting of the rainfall trend to 2017 using sequential SSA.
Mentions: Periodograms of the raw rainfall time series (Figure 2A) identified seasonal components which oscillated yearly for 4-, 6- and 12-month periods (repeating patterns occurring, tri-annually, bi-annually and once per year, Figure 2B) the signals of each repeating oscillation were isolated along with the long-term trend of rainfall (Figures 2C–2F). Remaining data were considered noise and the components were reconstructed into a noise-free signal (Figure 2G), a second periodogram of the new reconstructed signal shows the repeating patterns are intact, while stochastic noise has been removed (Figure 2H). Figure 2I shows the resulting output of forecasting from sequential SSA, with rainfall beginning to decline as French Guiana enters a trough of dryer years after several years of high rainfall. The 4-month component corresponds to two rainy seasons, one long rainy season and one short, and also to the main dry season from August to December (Figure 3A). The 6-month bi-annual component corresponds to the two rainy seasons and is a reflection of the strength of these two seasons (Figure 3B), while the 12-month component shows the overall rainfall level for the year (Figure 3C). Separability of these components was confirmed with a ω-correlation matrix showing that these seasonal components did not show high levels of correlation with other components (Figure 4).

Bottom Line: Predominantly occurring in developing parts of the world, Buruli ulcer is a severely disabling mycobacterium infection which often leads to extensive necrosis of the skin.While the exact route of transmission remains uncertain, like many tropical diseases, associations with climate have been previously observed and could help identify the causative agent's ecological niche.From this, it was possible to postulate for the first time that outbreaks of Buruli ulcer can be triggered by combinations of rainfall patterns occurring on a long (i.e., several years) and short (i.e., seasonal) temporal scale, in addition to stochastic events driven by the El Niño-Southern Oscillation that may disrupt or interact with these patterns.

View Article: PubMed Central - PubMed

Affiliation: Bournemouth University, Dorset BH12 5BB , UK ; UMR MIVEGEC, IRD-CNRS-Universités de Montpellier 1 et 2, Centre IRD de Montpellier, 34394 Montpellier cedex 5 , France.

ABSTRACT
Predominantly occurring in developing parts of the world, Buruli ulcer is a severely disabling mycobacterium infection which often leads to extensive necrosis of the skin. While the exact route of transmission remains uncertain, like many tropical diseases, associations with climate have been previously observed and could help identify the causative agent's ecological niche. In this paper, links between changes in rainfall and outbreaks of Buruli ulcer in French Guiana, an ultraperipheral European territory in the northeast of South America, were identified using a combination of statistical tests based on singular spectrum analysis, empirical mode decomposition and cross-wavelet coherence analysis. From this, it was possible to postulate for the first time that outbreaks of Buruli ulcer can be triggered by combinations of rainfall patterns occurring on a long (i.e., several years) and short (i.e., seasonal) temporal scale, in addition to stochastic events driven by the El Niño-Southern Oscillation that may disrupt or interact with these patterns. Long-term forecasting of rainfall trends further suggests the possibility of an upcoming outbreak of Buruli ulcer in French Guiana.

No MeSH data available.


Related in: MedlinePlus