Development Of A Landscape-based Multimetric Index To Assess The Wetland Health Status Of Poyang Lake

Wetland is located in the boundary zone between water and land,is a very important ecosystem.Landscape change caused by human activities is one of the main driving factors for the loss and degradation of wetlands.Poyang Lake wetland in China is of extraordinary importance to the regional ecology,economic and social development in the middle and lower reaches of the Yangtze River.At the same time,it has become one of the most important wetlands in the world.However,due to the longterm human disturbance and the continuous changes of landscapes(especially the increase of human-influenced land cover),Poyang Lake has experienced significant degradation,and the ecological health of its wetland has deteriorated.There is an urgent need to develop appropriate indicators to evaluate the ecosystem health status of wetlands,to clarify the natural and human factors of wetland degradation,and to provide a reference basis for the restoration and sustainable management of wetland ecological health.This paper developed a landscape-based multimetric index(LMI)to evaluate the health status of Poyang Lake wetland.Regression models were used to determine the most appropriate spatial scale for wetland health assessment.Combined with Landscape pattern index,Remote sensing metrics and Index of Biotic integrity,five core indicators were selected by correlation and discriminant analyses to evaluate the wetland health of Poyang Lake.In order to valid the feasibility of LMI in Poyang Lake wetland,the correlation analysis of the evaluation results was carried out with IBI and local interference index(LOD).Land use type transfer matrices between 1995,2005 and 2015 was used to analyze the change of land use patterns and the changes in wetland health of Poyang Lake basin.The main results were as follows:Five indicators(maximum patch index,connectivity index,Shannon diversity index,cultivated land pressure index and modified normalized difference building index)were selected as the core indicators of LMI.The multiple linear regression model showed that the landscape of the 2 km buffer zone around the sample site had the strongest explanatory power on IBIs,indicating that the landscape protection on the local scale(? 2 km)is very important for wetland protection.The LMI score showed that the overall health status of Poyang Lake wetland was “fair”.The areas near the lakeshores were mainly in a poor or very poor state,while the less accessible lake areas were in better health.The correlation analysis showed that LMI was significantly correlated with IBI and LOD(p < 0.001),which indicates that LMI has good adaptability in wetland health assessment.The area of construction land in Poyang Lake basin has increased continuously,which was mainly transformed from farmland.On the other hand,the implementation of the policy “return farmland to lake” has increased the area of Poyang Lake and natural land cover.Human activities are the main driving factors for the changes in landscape patterns of Poyang Lake basin.The LMI developed in this thesis can provide a fast and simple method for wetland ecosystem assessment in a large area,and identify the areas under great pressure of human disturbance,which will provide an important scientific basis for managers to take protective measures according to local conditions.