Abstract: As a basis of ecological civilization construction, the assessment of eco-environmental quality is of great significance for adapting to climate change and reasonably developing and protecting eco-environment. Based on vegetation coverage, terrain index and population distribution, and ecological elasticity limit derived from MODIS and land use dataset from 2000 to 2020, we assessed the evolution of environmental carrying capacity of the Loess Plateau with highly fragile eco-environment. The coupling coordination degree model was used to comprehensively characterize the regional eco-environment. A standard deviation ellipse method was used to investigate the change of center of gravity and the transfer path of coordination degree. Elman Neural Network with particle swarm optimization was used to predict the future trend of ecological environment. The results showed that since 2000, the environmental carrying capacity of the Loess Plateau exhibited a spatial pattern with higher in the southeast and lower in the northwest, and the index increased slowly with a rate of 0.0625 per year. The ecological elasticity limit was consistent with land use pattern, and the index exponentially ascended with a low level. The coupled coordinate coefficient changed from slight discordance to primary coordination. The core of coupled coordination degree shifted to southeast and exhibited a spatial pattern with centripetal aggregation in the northeast-southwest direction and diffusion in northwest-southeast direction. It is predicted that the system coupling coordination degree of environmental carrying capacity and ecological elasticity limit will increase steadily in 2030, which indicates that the eco-environmental quality of the Loess Plateau will gradually become coordinated in the future, but with significant difference in the growth rate of different regions.

Key words: environmental carrying capacity, ecological elasticity limit, standard deviation ellipse, Elman Neural Network, Loess Plateau