Identifying microclimate tree seedling refugia in post-wildfire landscapes

High-severity wildfire in arid regions has caused ecological state change, transforming previously forested areas into shrublands. This dramatically alters the microclimatic conditions, which can exceed climatic tolerance of tree seedlings, rendering likelihood returning post-wildlife landscapes to their previous relatively low. Characterizing variability across severely burned could allow for identifying locations where seedling survival is more likely. We used a combination small unmanned aircraft system imagery, satellite data and in-situ microclimate recordings, together with machine learning approach, model monthly near-ground minimum, mean max temperature as well relative humidity vapor pressure deficit area, now dominated by two different shrub species. Spatially explicit models predicted recorded (r = 0.73 0.97), projections highlight that at any given location hottest month, solar buffering capacity existing vegetation alter maximum ∼12 °C, increase ∼20% reduced 0.3mbar open areas. By harnessing these refugia, success rate reforestation efforts post-wildfire be substantially increased mitigate seedlings from climate warming local scales.