作者: O. Viedma , I. R. Urbieta , J. M. Moreno
DOI: 10.1038/S41598-018-36134-4
关键词:
摘要: During the last decades, wildfires have been changing in many areas across world, due to changes climate, landscapes and socioeconomic drivers. However, how role of these drivers changed over time has little explored. Here, we assessed, a spatially temporally explicit way, biophysical human-related factors on rural area west-central Spain from 1979 2008. Longitudinal Negative Binomial (NB) Zero-Inflated (ZINB) mixed models, with as interacting factor (continuous categorical), were used model number fires increasing size (≥1–10 ha, >10–100 ha, >100 ha) per 10 × 10 km cell per year, based fire statistics. We found that landscape was rather dynamic, generally became more hazardous time. Small increased spread time, medium large being stable or decreasing. NB models best for modelling small fires, while ZINB large; including categorical performed best. Best associated topography, land-use/land cover (LULC) types they underwent, well agrarian characteristics. Climate variables, forest interfaces, other variables played minor role. Wildfires initially frequent rugged conifer forests, shrublands and cells undergoing LULC nature, all sizes. As went by, lost links initial fire-prone areas, spread, lower elevation higher solar radiation, herbaceous crops, size farms. Thus, time; some decreased their explaining power, others increased. These total spatial pattern controlling limit ability predict future fires.