Fire-climate interactions in the Mediterranean climate areas of California and Southern Oregon

Wildland forest fires are, and have been, an important process shaping forest ecosystem structure and function in the fire-prone Mediterranean Climate Area (MCA) forests of the North American Pacific Coast. The mild-wet winter and warm-dry summer climate in this region predisposes many ecosystems to conditions that can burn annually. Fires are of special concern in this region because of the history of, and potential for, great loss of life and property.

Climate variation is an important driver of fire season severity and it operates over a range of temporal and spatial scales in the MCA. High interannual variability in climate is characteristic of the MCA and is linked to hemispheric and even global scale coupled ocean and atmospheric circulation patterns. This study aims at determining how the dominant modes of climatic circulation in the MCA (El Niño Southern Oscillation (ENSO), Pacific/North America Pattern (PNA), Pacific Decadal Oscillation (PDO)) are related to secondary circulation features, to surface fire weather, and to spatial and temporal variation 20th century fire extent.

A non-linear method, using a statistical clustering technique, is applied to address this question. Self-organizing maps (SOMs) provide a precipitation regionalization for the MCA, on which statistical comparisons between atmospheric circulation parameters and wildfire occurrence and severity are developed. SOMs are also used to identify wildfire regions, defined as areas of fire activity over spatially and temporally contiguous areas, based on observational fire records for the MCA (1910-present). Superposed Epoch Analysis (SEA) is then used to examine how fire anomalies and precipitation anomalies for the different regions are temporally related. The results of SEA are spatially defined wildfire areas that respond in predictable ways to precipitation on climatic time scales. On the other hand, a SOM is performed on geopotential height data (1948-present), which elicits the generalized circulation properties associated with each precipitation regime. The determined relationships between circulation and precipitation subregions are extended to wildfire regions via the previously established relationships.

This set of analyses produces quantifiable relationships among fire occurrence and severity in wildfire regions, precipitation, and the associated atmospheric circulation, with an aim towards informing future predictive models.