Fire, as a strong disturbance type, can exert significant impacts on biosphere, hydrosphere, geosphere, cryosphere, atmosphere and human society. It can inherently trigger both critical transitions in ecosystems and dramatic changes in landscapes, which can be detected as alternations in land cover types. However, the general changing patterns and possible influential factors of post-fire landscape change remain largely unclear on a global scale. Obtaining such knowledge is of great value in advancing the understanding of fire ecology and promoting sustainable fire management. Here, we combined the satellite observations of long-term land cover and burned areas to assess the global post-fire landscape change patterns from 2005 to 2015. The results showed that the identified areas with post-fire landscape change accounted for approximately 0.36–0.74% of the annual global burned areas during the study period and were most common in countries such as Brazil, Argentina, and the D.R. Congo. The most common landscape change types were “forest-to-agriculture” (31.93%), “forest-to-shrubland” (26.23%) and “agriculture-to-forest” (18.74%) in 2005, 2010 and 2015, respectively. In addition, the conversion between agriculture and forest as well as the shrubland and forest after fire were found to be bidirectional. After assessing 14 fire-related climatic, topographic, ecological and socioeconomic factors that could potentially influence the post-fire landscape change occurrence probability, burned area size and vegetation cover diversity were identified as the two strongest predictors, followed by aspect, fire intensity and slope. Our results provide a global overview of post-fire landscape change patterns and offer guidance for making sustainable fire management policies.

Desert expansions can cause tremendous losses to human well-being. However, the process of shifting from the non-desert state to the desert state, a representation of a system regime shift, remains unclear on the global scale. Clarifying the underpinning pattern, predictors and signals of this process is of great value in advancing understanding of both ecosystem resilience and sustainable developments. Here, we combine the climate classification map and long-term observational land cover data to assess the global desert distribution and its changes from 2000 to 2019. The identified desert areas cover approximately 7.53% of the global land in the past two decades. Only approximately 16.03% of these deserts shows expanding trends, especially in countries such as Tunisia, Tajikistan and Peru. After assessing 26 climatic, ecological and socioeconomic factors that could potentially modify desert expansion rates, vegetation cover diversity was identified as the strongest predictor in both hot and cold deserts, followed by cattle density in hot deserts and desert size in cold deserts. In addition, pronounced high fluctuation in satellite vegetation productivity and flickering between land cover states could serve as two signals for desert conversion and fast expansion, respectively. Our results provide not only a long-term overview of global desert changing patterns but also possible guidance for constraining desert expansion.