Causal relationships versus emergent patterns in the global controls of fire frequency

[摘要] Global controls on month-by-month fractional burnt area (2000–2005) wereinvestigated by fitting a generalised linear model (GLM) to Global FireEmissions Database (GFED) data, with 11 predictor variables representingvegetation, climate, land use and potential ignition sources. Burnt area isshown to increase with annual net primary production (NPP), number of drydays, maximum temperature, grazing-land area, grass/shrub cover and diurnaltemperature range, and to decrease with soil moisture, cropland area andpopulation density. Lightning showed an apparent (weak) negative influence,but this disappeared when pure seasonal-cycle effects were taken intoaccount. The model predicts observed geographic and seasonal patterns, aswell as the emergent relationships seen when burnt area is plotted againsteach variable separately. Unimodal relationships with mean annual temperatureand precipitation, population density and gross domestic product (GDP) arereproduced too, and are thus shown to be secondary consequences ofcorrelations between different controls (e.g. high NPP with highprecipitation; low NPP with low population density and GDP). These findingshave major implications for the design of global fire models, as severalassumptions in current models – most notably, the widely assumed dependenceof fire frequency on ignition rates – are evidently incorrect.

[学科分类] 地球化学与岩石