A Note on the Poisson Likelihood Ratio Test Statistic for Kulldorff's Scan Methods

Fraker, Shannon E.; Woodall, William H.; Burkom, Howard

doi:10.1080/03610920701573375

Cited by 7 publications

(2 citation statements)

References 2 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A likelihood ratio (LR) is calculated for each sub-area of the numerical domain scanned by the window. Under the Poisson assumption, the LR is computed as follows [Fraker et al, 2008]:…”

Section: Spatial-temporal Scan Statisticsmentioning

confidence: 99%

Statistical Detection of Spatio-Temporal Patterns in the Salinity Field Within an Inter-Tidal Basin

Donatelli

Duran‐Matute

Gräwe

et al. 2022

Estuaries and Coasts

View full text Add to dashboard Cite

Salinity is a key factor aff ecting biological processes and biodiversity in estuarine systems.This study investigates temporal and spatial changes in salinity at a basin-wide scale for 2005-2015 in the Dutch Wadden Sea. Scan statistics is applied to track salinity variations systematically and to detect potential clusters, i.e. estuarine regions marked by anomalous highsalinity (or low-salinity) values in a certain period (i.e., strong deviations from the expected value in a statistical sense). Clusters' statistical significance has been assessed via Monte Carlo simulations. Particular attention is devoted to event-driven spatial and temporal patterns characterized by extreme salinity values since these episodes dramatically increase stress levels on organisms living in intertidal areas. Periodic components in the modeled salinity time series are identified using wavelet analysis and eventually removed from the signal before performing scan statistics. Wavelet analysis suggests that tides are the chief agent controlling salinity fluctuations in the system at within-day time scales, whereas no dominant periodicities were detected at longer time scales. Scan statistics reveal long-lasting clusters next to the main freshwater outlets and within the areas characterized by low water exchanges. In contrast, active regions of the estuary can efficiently counteract extreme events and quickly recover their pre-perturbation conditions. Finally, by analyzing the freshwater dispersal in the system, it is found that clusters' occurrence is related to episodic events characterized by extreme conditions in the southwesterly wind and freshwater discharge. This research demonstrates that scan statistics can be used as a powerful tool for spatiotemporal analyses of marine systems and for identifying data-clustering that may be indicative of emerging environmental hazards (e.g., due to climate change).

show abstract

“…A likelihood ratio (LR) is calculated for each sub-area of the numerical domain scanned by the window. Under the Poisson assumption, the LR is computed as follows [Fraker et al, 2008]:…”

Section: Spatial-temporal Scan Statisticsmentioning

confidence: 99%

Statistical Detection of Spatio-Temporal Patterns in the Salinity Field Within an Inter-Tidal Basin

Donatelli

Duran‐Matute

Gräwe

et al. 2022

Estuaries and Coasts

View full text Add to dashboard Cite

show abstract

“…Some of what we learned was summarized in two review papers, Woodall 1 and Tsui et al 2, aimed primarily toward industrial SPC researchers and practitioners. We have studied several public health surveillance topics, including sets methods 3, temporal scan approaches 4–6, and spatiotemporal methods 7–10 [Tsui et al 2010; submitted].…”

Section: Introductionmentioning

confidence: 99%