Analyzing tiger interaction and home range shifts using a time-geographic approach.

BACKGROUND: Interaction through movement can be used as a marker to understand and model interspecific and intraspecific species dynamics, and the collective behavior of animals sharing the same space. This research leverages the time-geography framework, commonly used in human movement research, to explore the dynamic patterns of interaction between Indochinese tigers (Panthera tigris corbeti) in the western forest complex (WEFCOM) in Thailand. METHODS: We propose and assess ORTEGA, a time-geographic interaction analysis method, to trace spatio-temporal interactions patterns and home range shifts among tigers. Using unique GPS tracking data of tigers in WEFCOM collected over multiple years, concurrent and delayed interaction patterns of tigers are investigated. The outcomes are compared for intraspecific tiger interaction across different genders, relationships, and life stages. Additionally, the performance of ORTEGA is compared to a commonly used proximity-based approach. RESULTS: Among the 67 tracked tigers, 42 show concurrent interactions at shared boundaries. Further investigation of five tigers with overlapping home ranges (two adult females, a male, and two young male tigers) suggests that the mother tiger and her two young mostly stay together before their dispersal but interact less post-dispersal. The male tiger increases encounters with the mother tiger while her young shift their home ranges. On another timeline, the neighbor female tiger mostly avoids the mother tiger. Through these home range dynamics and interaction patterns, we identify four types of interaction among these tigers: following, encounter, latency, and avoidance. Compared to the proximity-based approach, ORTEGA demonstrates better detects concurrent mother-young interactions during pre-dispersal, while the proximity-based approach misses many interactions among the dyads. With larger spatial buffers and temporal windows, the proximity-based approach detects more encounters but may overestimate the duration of interaction. CONCLUSIONS: This research demonstrates the applicability and merits of ORTEGA as a time-geographic based approach to animal movement interaction analysis. We show time geography can develop valuable, data-driven insights about animal behavior and interactions. ORTEGA effectively traces frequent encounters and temporally delayed interactions between animals, without relying on specific spatial and temporal buffers. Future research should integrate contextual and behavioral information to better identify and characterize the nature of species interaction.

Early warning system for West Nile virus risk areas, California, USA.

The Dynamic Continuous-Area Space-Time (DYCAST) system is a biologically based spatiotemporal model that uses public reports of dead birds to identify areas at high risk for West Nile virus (WNV) transmission to humans. In 2005, during a statewide epidemic of WNV (880 cases), the California Department of Public Health prospectively implemented DYCAST over 32,517 km2 in California. Daily risk maps were made available online and used by local agencies to target public education campaigns, surveillance, and mosquito control. DYCAST had 80.8% sensitivity and 90.6% specificity for predicting human cases, and k analysis indicated moderate strength of chance-adjusted agreement for >4 weeks. High-risk grid cells (populations) were identified an average of 37.2 days before onset of human illness; relative risk for disease was >39× higher than for low-risk cells. Although prediction rates declined in subsequent years, results indicate DYCAST was a timely and effective early warning system during the severe 2005 epidemic.

Identifying West Nile virus risk areas: the Dynamic Continuous-Area Space-Time system.

The Dynamic Continuous-Area Space-Time (DYCAST) system was developed to identify and prospectively monitor high-risk areas for West Nile virus in New York, New York (New York City). The system is based on a geographic model that uses a localized Knox test to capture the nonrandom space-time interaction of dead birds, as an indicator of an intense West Nile virus amplification cycle, within a 1.5-mile (2.41-km) buffer area and 21-day moving window. The Knox analysis is implemented as an interpolation function to create a surface of probabilities over a grid of 1,400 cells overlaying New York City. The model's parameters were calibrated using year 2000 data and information on the vector-host transmission cycle. The DYCAST system was implemented in a geographic information system and used operationally in year 2001. It successfully identified areas of high risk for human West Nile virus infection in areas where five of seven human cases resided, at least 13 days prior to the onset of illness, and proved that it can be used as an effective tool for targeting remediation and control efforts.