基于优化MaxEnt模型的伊犁河谷毒害草适生区预测

打开文本图片集

AbstractThe spread of poisonous weeds poses a significant threat to the stability of grassland ecosystems， ffecting ecological balance and biodiversity.This paper used the maximum entropy（MaxEnt）model to investigate he environmental factors that affctthedistributionof typical poisonous weeds in the IliRiver Vally grassland，and predicted the potential distributionof poisonous weeds underfuture climate conditions.The results showed thatthearea underthecurve （AUC）of Aconitum leucostomum，Phlomoides umbrosa，Rheum witrockii，RumexacetosaandLigularia macrophylla all exceeded O.9，indicating high accuracy.This model had highcredibility.Aconitum leucostomum and Ligularia macrophyll were significantly affected by grazingrate.Rheum witrockii was greatly affcted bythe anual average temperature.Rumex acetosa was afected by the precipitation in the wettest season and the average daily temperature range.Phlomoides umbrosa wasmainly afected bythe precipitation inthe wettest season.The high-risk areas for potential invasionofpoisonousweeds wereconcentrated intheNalati Mountain，Nalati Grassland，Kashgar River Basin，Borokonu Mountain and the southern area of Keguqin Mountain.These poisonous weeds had strong adaptabilityto stable climateand had less invasion in their habitats.Precipitationand temperature fluctuations hada significantimpact on poisonous weeds.Under the moderate emision scenario （SSP2-4.5）， the poisonous weeds inthe li River Valley grassands generally showed a trend of distribution towards high-altitude and water abundant areas.

Keywords poisonous weed; suitable habitat; trend prediction; optimized MaxEnt model; grassand; Ili Rive Valley

毒害草是兼具毒害性和入侵性的植物群落，其扩散过程通常会引发多重负面效应。（剩余5285字）