Wildfire risk has become an urgent concern across the western United States, with Califor-nia in particular experiencing unprecedented fire seasons in the past two decades. According to the California Department of Forestry and Fire Protection (Cal Fire), seven of the ten largest wildfires in California history have occurred since 2017, underscoring a rapidly esca- lating crisis. The human and economic toll is immense: in addition to lives lost and homes destroyed, insurers face billions of dollars in claims each year, while local governments strug- gle to allocate limited resources for fire suppression and prevention. These events highlight the need for rigorous, data-driven models that can improve our understanding of where and when fires are most likely to occur. Los Angeles County represents a particularly compelling study area. Its unique ge- ography encompasses coastal regions, dense urban development, and mountainous terrain, creating highly heterogeneous fire regimes. The combination of Santa Ana winds, prolonged droughts, and increasing development at the wildland–urban interface (WUI) amplifies both the frequency and destructiveness of fires. High-profile events such as the 2019 Getty Fire, the 2021 Palisades Fire, and more recent incidents in 2025 have demonstrated how rapidly wildfires can ignite and spread within this region, often overwhelming response capabilities. Understanding spatial variation in fire risk within Los Angeles County is therefore crucial for both scientific inquiry and practical risk assessment. Wildfires present a rich but challenging problem for statistical modeling. Fires are in- herently spatio-temporal phenomena: they occur at specific locations and times, but their risk is influenced by both past events and environmental conditions. Unlike purely tempo- ral models, which treat events as independent over space, or purely spatial models, which ignore the dynamic evolution of risk through time, space-time statistical models allow for joint treatment of location and timing. This makes them particularly well suited to wildfire applications, where recurrence at the same site are fundamental in this process. In this thesis, I first develop exploratory analyses that provide descriptive insight into the patterns of fire occurrence, using kernel density estimates and summary statistics to highlight spatial hotspots and temporal trends. These baseline methods establish bench- marks against which more advanced models can be evaluated. Second, I apply and extend the Stoyan–Grabarnik (SG) space-time model, which explicitly incorporates covariates to estimate the conditional intensity of wildfire occurrence. A particular focus is placed on time-since-burn metrics: the time since a given location last burned, and the time since nearby locations burned. These measures capture fuel dynamics and spatial spillover, both of which are critical in understanding fire recurrence.