Limits...
Evaluating patterns of a white-band disease (WBD) outbreak in Acropora palmata using spatial analysis: a comparison of transect and colony clustering.

Lentz JA, Blackburn JK, Curtis AJ - PLoS ONE (2011)

Bottom Line: Comparisons of DMAP results suggest the transect-level overestimated the prevalence and spatial extent of the outbreak.In contrast, more realistic prevalence estimates and spatial patterns were found by weighting each transect by the number of individual A. palmata colonies with and without WBD.Specifically, we illustrate that higher resolution sampling resulted in more realistic disease estimates.

View Article: PubMed Central - PubMed

Affiliation: Department of Oceanography and Coastal Sciences, Louisiana State University, Baton Rouge, Louisiana, United States of America. Jennifer.Lentz@gmail.com

ABSTRACT

Background: Despite being one of the first documented, there is little known of the causative agent or environmental stressors that promote white-band disease (WBD), a major disease of Caribbean Acropora palmata. Likewise, there is little known about the spatiality of outbreaks. We examined the spatial patterns of WBD during a 2004 outbreak at Buck Island Reef National Monument in the US Virgin Islands.

Methodology/principal findings: Ripley's K statistic was used to measure spatial dependence of WBD across scales. Localized clusters of WBD were identified using the DMAP spatial filtering technique. Statistics were calculated for colony- (number of A. palmata colonies with and without WBD within each transect) and transect-level (presence/absence of WBD within transects) data to evaluate differences in spatial patterns at each resolution of coral sampling. The Ripley's K plots suggest WBD does cluster within the study area, and approached statistical significance (p = 0.1) at spatial scales of 1100 m or less. Comparisons of DMAP results suggest the transect-level overestimated the prevalence and spatial extent of the outbreak. In contrast, more realistic prevalence estimates and spatial patterns were found by weighting each transect by the number of individual A. palmata colonies with and without WBD.

Conclusions: As the search for causation continues, surveillance and proper documentation of the spatial patterns may inform etiology, and at the same time assist reef managers in allocating resources to tracking the disease. Our results indicate that the spatial scale of data collected can drastically affect the calculation of prevalence and spatial distribution of WBD outbreaks. Specifically, we illustrate that higher resolution sampling resulted in more realistic disease estimates. This should assist in selecting appropriate sampling designs for future outbreak investigations. The spatial techniques used here can be used to facilitate other coral disease studies, as well as, improve reef conservation and management.

Show MeSH

Related in: MedlinePlus

The results of the Ripley's K spatial autocorrelation analysis.Normalized Ripley's K plots were used to assess the spatial distribution of white-band disease (WBD) among Acropora palmata over a distance of 2.5 m. Transect-level and colony-level versions of the K function were performed in order to compare the spatial distributions of WBD based on data analyzed at the (A) transect- and (B) colony-levels (respectively). In order to insure that the observed spatial distribution was reflecting the spatial nature of WBD, and not the spatial patterning of the underlying population, the transect and colony-level observed K values for the underlying population were subtracted from the observed Ks of WBD at the transect- and colony-levels, respectively. The resulting K values for WBD were then plotted against distance. The spatial nature of WBD was then assessed by comparing these K values for WBD (thick line) to a spatially random (Poisson) distribution (dashed line at y = 0), in which WBD values above the Poisson distribution indicates WBD was aggregated within the underlying population, while values below this line indicated WBD was more dispersed than the underlying population. The 99% confidence intervals (thin lines) generated from the observed K values for the population were used to determine the statistical significance of distribution of WBD within the underlying population of susceptible corals.
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC3139597&req=5

pone-0021830-g003: The results of the Ripley's K spatial autocorrelation analysis.Normalized Ripley's K plots were used to assess the spatial distribution of white-band disease (WBD) among Acropora palmata over a distance of 2.5 m. Transect-level and colony-level versions of the K function were performed in order to compare the spatial distributions of WBD based on data analyzed at the (A) transect- and (B) colony-levels (respectively). In order to insure that the observed spatial distribution was reflecting the spatial nature of WBD, and not the spatial patterning of the underlying population, the transect and colony-level observed K values for the underlying population were subtracted from the observed Ks of WBD at the transect- and colony-levels, respectively. The resulting K values for WBD were then plotted against distance. The spatial nature of WBD was then assessed by comparing these K values for WBD (thick line) to a spatially random (Poisson) distribution (dashed line at y = 0), in which WBD values above the Poisson distribution indicates WBD was aggregated within the underlying population, while values below this line indicated WBD was more dispersed than the underlying population. The 99% confidence intervals (thin lines) generated from the observed K values for the population were used to determine the statistical significance of distribution of WBD within the underlying population of susceptible corals.

Mentions: We used the difference function (D) to examine the spatial distribution of WBD with respect to underlying environmental heterogeneity caused by the presence of the underlying coral population. To do this we subtracted the normalized K values from the underlying population from those of the WBD corals so that we would be able to assess to what extent the spatial distributions of WBD depicted by the homogeneous analyses (see Text S1 and Figures S1–S4) were caused by the disease itself, rather than the natural background variation in the A. palmata population (Figure 3). Our resulting Disease-Population difference function was quite similar to the design of the Ripley's K function used by Jolles et al. [62] in which they set their distribution equal to that of the underlying population of susceptible corals and then plotted K-K against distance.


Evaluating patterns of a white-band disease (WBD) outbreak in Acropora palmata using spatial analysis: a comparison of transect and colony clustering.

Lentz JA, Blackburn JK, Curtis AJ - PLoS ONE (2011)

The results of the Ripley's K spatial autocorrelation analysis.Normalized Ripley's K plots were used to assess the spatial distribution of white-band disease (WBD) among Acropora palmata over a distance of 2.5 m. Transect-level and colony-level versions of the K function were performed in order to compare the spatial distributions of WBD based on data analyzed at the (A) transect- and (B) colony-levels (respectively). In order to insure that the observed spatial distribution was reflecting the spatial nature of WBD, and not the spatial patterning of the underlying population, the transect and colony-level observed K values for the underlying population were subtracted from the observed Ks of WBD at the transect- and colony-levels, respectively. The resulting K values for WBD were then plotted against distance. The spatial nature of WBD was then assessed by comparing these K values for WBD (thick line) to a spatially random (Poisson) distribution (dashed line at y = 0), in which WBD values above the Poisson distribution indicates WBD was aggregated within the underlying population, while values below this line indicated WBD was more dispersed than the underlying population. The 99% confidence intervals (thin lines) generated from the observed K values for the population were used to determine the statistical significance of distribution of WBD within the underlying population of susceptible corals.
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC3139597&req=5

pone-0021830-g003: The results of the Ripley's K spatial autocorrelation analysis.Normalized Ripley's K plots were used to assess the spatial distribution of white-band disease (WBD) among Acropora palmata over a distance of 2.5 m. Transect-level and colony-level versions of the K function were performed in order to compare the spatial distributions of WBD based on data analyzed at the (A) transect- and (B) colony-levels (respectively). In order to insure that the observed spatial distribution was reflecting the spatial nature of WBD, and not the spatial patterning of the underlying population, the transect and colony-level observed K values for the underlying population were subtracted from the observed Ks of WBD at the transect- and colony-levels, respectively. The resulting K values for WBD were then plotted against distance. The spatial nature of WBD was then assessed by comparing these K values for WBD (thick line) to a spatially random (Poisson) distribution (dashed line at y = 0), in which WBD values above the Poisson distribution indicates WBD was aggregated within the underlying population, while values below this line indicated WBD was more dispersed than the underlying population. The 99% confidence intervals (thin lines) generated from the observed K values for the population were used to determine the statistical significance of distribution of WBD within the underlying population of susceptible corals.
Mentions: We used the difference function (D) to examine the spatial distribution of WBD with respect to underlying environmental heterogeneity caused by the presence of the underlying coral population. To do this we subtracted the normalized K values from the underlying population from those of the WBD corals so that we would be able to assess to what extent the spatial distributions of WBD depicted by the homogeneous analyses (see Text S1 and Figures S1–S4) were caused by the disease itself, rather than the natural background variation in the A. palmata population (Figure 3). Our resulting Disease-Population difference function was quite similar to the design of the Ripley's K function used by Jolles et al. [62] in which they set their distribution equal to that of the underlying population of susceptible corals and then plotted K-K against distance.

Bottom Line: Comparisons of DMAP results suggest the transect-level overestimated the prevalence and spatial extent of the outbreak.In contrast, more realistic prevalence estimates and spatial patterns were found by weighting each transect by the number of individual A. palmata colonies with and without WBD.Specifically, we illustrate that higher resolution sampling resulted in more realistic disease estimates.

View Article: PubMed Central - PubMed

Affiliation: Department of Oceanography and Coastal Sciences, Louisiana State University, Baton Rouge, Louisiana, United States of America. Jennifer.Lentz@gmail.com

ABSTRACT

Background: Despite being one of the first documented, there is little known of the causative agent or environmental stressors that promote white-band disease (WBD), a major disease of Caribbean Acropora palmata. Likewise, there is little known about the spatiality of outbreaks. We examined the spatial patterns of WBD during a 2004 outbreak at Buck Island Reef National Monument in the US Virgin Islands.

Methodology/principal findings: Ripley's K statistic was used to measure spatial dependence of WBD across scales. Localized clusters of WBD were identified using the DMAP spatial filtering technique. Statistics were calculated for colony- (number of A. palmata colonies with and without WBD within each transect) and transect-level (presence/absence of WBD within transects) data to evaluate differences in spatial patterns at each resolution of coral sampling. The Ripley's K plots suggest WBD does cluster within the study area, and approached statistical significance (p = 0.1) at spatial scales of 1100 m or less. Comparisons of DMAP results suggest the transect-level overestimated the prevalence and spatial extent of the outbreak. In contrast, more realistic prevalence estimates and spatial patterns were found by weighting each transect by the number of individual A. palmata colonies with and without WBD.

Conclusions: As the search for causation continues, surveillance and proper documentation of the spatial patterns may inform etiology, and at the same time assist reef managers in allocating resources to tracking the disease. Our results indicate that the spatial scale of data collected can drastically affect the calculation of prevalence and spatial distribution of WBD outbreaks. Specifically, we illustrate that higher resolution sampling resulted in more realistic disease estimates. This should assist in selecting appropriate sampling designs for future outbreak investigations. The spatial techniques used here can be used to facilitate other coral disease studies, as well as, improve reef conservation and management.

Show MeSH
Related in: MedlinePlus