Abstract:
[Objective]Studying the relationship between species diversity of plant communities and environmental factors is of great significance for the research, evaluation, protection and comprehensive management of the diversity of mountainous ecosystem. This paper selects the Daqing Mountain Nature Reserve in Inner Mongolia as the research object, predicts the spatial distribution of plant diversity in the Daqing Mountain area, and analyzes the relationship between various environmental factors and the spatial distribution of plant diversity. [Method] The study selects the Daqing Mountain Nature Reserve in Inner Mongolia as the research subject, employs deep learning methods to construct a plant diversity index model, and validates the model's accuracy. Subsequently, it predicts the spatial distribution of plant diversity in Daqing Mountain and analyzes the variation patterns of plant diversity under different environmental factors.[Results](1) There are a total of 108 plant species in the study area, belonging to 77 genera and 31 families. The plant diversity on the shady slopes is greater than that on the sunny slopes.(2) The slope has the greatest relative contribution (0.417) to the Shannon - Wiener index (H'), Simpson dominance index (D) and Pielou evenness index (J), followed by TVDI (0.25), temperature (0.167), NDVI (0.083) and solar radiation (0.083). Temperature has the greatest relative contribution (0.382) to the Margalef richness index (R), followed by solar radiation (0.375), slope (0.088), aspect (0.084), and NDVI (0.071).(3)The predicted results of the Shannon-Wiener index (H'), Simpson dominance index (D), Pielou evenness index (J), and Margalef richness index (R) all showed strong agreement with measured values, with MSE values of 0.0811, 0.0331, 0.0265, and 0.0524, and MAE values of 0.0156, 0.0025, 0.0017, and 0.0039, respectively. Further linear regression analysis between simulated and observed values in the training set revealed that the R2 values for each diversity index reached 0.86, 0.93, 0.92, and 0.99, respectively.(4) The value range of the Shannon - Wiener index in the Daqing Mountain area is 0 - 3.87, the Simpson dominance index is 0 - 0.83, the Pielou evenness index is 0 - 0.95, and the Margalef richness index is 0 - 4.12.(5) The Shannon - Wiener index, Simpson dominance index and Pielou evenness index are linearly negatively correlated with slope, TVDI, LST and solar radiation, and linearly positively correlated with NDVI. The Margalef richness index is linearly negatively correlated with LST, solar radiation and slope, and linearly positively correlated with aspect and NDVI. Overall, plant diversity is linearly negatively correlated with LST, solar radiation and slope, and linearly positively correlated with NDVI.[Conclusion]It is feasible to predict the spatial distribution of plant diversity in mountainous landforms by using Deep Learning methods. This method can deepen the understanding of the biodiversity change patterns in the Daqing Mountain area of Inner Mongolia and contribute to the protection and restoration of plant diversity in this region.