share data.rar
These data represent the research findings from our initial manuscript submitted to the journal CATENA, which includes plant information extraction, identification of key factors influencing plant distribution, as well as the simulation and future prediction of plant NDVI. Our paper is well-suited for the specific subfield of landscape ecology, as it addresses the evolutionary characteristics of early-spring ephemeral plants in the unique desert landscape and their responses to future climate change. We also believe that the study and its findings will be of interest to the readers of your journal. The highlights of the paper are as follows:
(1) This study examines the correlation between high temporal resolution remote sensing monitoring and plant phenology, establishing a distribution range identification and NDVI extraction model for early-spring ephemeral plants based on a life-cycle approach, and validating the model using species distribution point data and high spatial resolution Landsat series data.
(2) In developing the NDVI simulation model for ephemeral plants, we integrated multi-source environmental data, incorporating GIS spatial analysis and fully utilizing the factor contribution module of the RF (Random Forest) algorithm and the prediction module of the CNN (Convolutional Neural Network) model, thereby perceiving and extracting local features between ephemeral plants and climate variables.
(3) We conducted priority planning for the biodiversity conservation of the unique plant landscape of early-spring ephemeral plants in the Gurbantünggüt Desert.
The research findings of this study indicate that "it is speculated that these plants will migrate northwestward in the 2050s (to higher altitude and higher latitude areas), forming a second region in the center of the desert that is conducive to their survival and reproduction," which aligns with the inference that "due to global warming, biodiversity will migrate to higher altitudes and higher latitudes."
This study comprehensively considers the relationship between remote sensing and plant phenology in the remote sensing extraction method for the unique plant landscape of the desert (early-spring ephemeral plants). It also employs a combination model of RF and CNN to predict future plant distribution and trends. This approach can provide new ideas and suggestions for related research on plant identification and conservation globally.
