Study On Wetland Change Remote Sensing Diagnosis

Known as “the kidney of earth”,wetland is one of the most productive ecosystems in the world,with huge environmental functions and ecological benefits.Although the total area of global wetlands accounts for only 6.2–7.6% of the earth's land area,wetlands are the most vital ecosystems with high-level biodiversity and a wide range of ecosystem services,such as supplying freshwater and groundwater,mitigating flood and drought and conserving biodiversity.However,as the transition area between aquatic and terrestrial systems,wetlands are also the most vulnerable and endangered ecosystems.Due to the impacts of climate change and pressure from human activities,global wetlands shrank extensively during the past several decades.There is no doubt that protecting and restoring wetlands,sustainably managing wetland ecosystems and reasonably utilising wetland resources have been the main tasks for wetland management by local governments,researchers and conservation organisations.Quantitative,accurate and repeatable detection of wetlands dynamics changes is the basis and prerequisite for these tasks,and therefore is of self-evident importance.Considering that wetland ecosystems are generally inaccessibly remoted and of diverse composition,traditional field survey for dynamics changes may face several limitations,such as incompleteness.With the development of space technology,remote sensing and geographical information systems(GIS)have become important tools for wetland observation in the spatial and temporal scales,and have also been widely used to detect wetlands dynamics changes.As the appearance of environmental health remote sensing diagnosis theory,the wetland change detection based on remote sensing technology is raised to the height of methodology,which is remote sensing diagnosis of wetland change.In this paper,a remote sensing diagnostic methodolgy for wetland change was proposed under the framework of environmental health remote sensing diagnosis theory.This methodology is consisted of wetland change detection,wetland ecosystem assessment and wetland change prediction,which indicates the wetland change in the past,wetland ecosystem current statue and wetland prediction in the future.In this study,the wetland in the Erdos Relict Gull National Nature Reserve,Inner Mongolia was taken as an example.The methodology was used to diagnose wetland change.The diagnostic results were evaluated and validated by using high-resolution remote sensing images.Diagnostic results were analyzed for change driving.The main contents of this research are as follows:1.Remote sensing diagnosis methodology of wetland change is developed.Based on the theory of environmental health remote sensing diagnosis,a methodology of remote sensing diagnosis of wetland change was proposed based on the changes of wetland,the status of wetland ecosystem and the trend of wetland change.Based on the research of each aspect,the remote sensing diagnostic model of wetland change detection,the remote sensing diagnosis model of wetland ecosystem and the remote sensing diagnosis model of wetland change prediction were put forward.2.Wetland change detection method based on wetland elements is developed.Based on the researches wetland change detection using remote sensing,it is found that the researches on wetland change are mainly cosistied: change of post-classification results change detection and trendency analysis change detection.However,both two methods have some limitations,such as: post-classification change detection require a lot of wetland prior knowledge,wetland classification need lots of time and task,they may ignore change information between same classificaiton;trendency analysis change detection require input remote sensing data should be strictly radiation normalized,the change result need to be interpreted and so on.Wetland change detection method based on wetland elements combines the advantages of Spectral Mixture Analysis(SMA)and Change Vector Analysis(CVA).Firstly,the multi-temporal and multi-temporal remote sensing image data are conducted by SMA,to obtain the wetland elements'(water,soil and vegetation)fraction.Using the fraction results of the three elements in different periods as the input parameters of the change vector analysis model,the wetland change intensity and direction were obtained,and the results were validated by the survey field and high resolution remote sensing satellite image data.Comparing with the traditional change detection method,the proposed method can obtain more detailed information on wetland change(12 change types),and the change detection method avoids the complicated work such as wetland classification.3.Wetland ecosystem assessment.Based on the researches about wetland ecosystem assessments,it is found that the current assessment system of wetland ecosystems is constructed by using index system and the Analytic Hierarchy Process(AHP)method is used to evaluate the ecosystem.Based on the research literature,consultation with experts in the field and reserve managers,and field investigation,the wetland ecosystem evaluation index system was constructed from the three aspects: wetland ecosystem health,functions and value;and passed the expert demonstration.For the index weights calculation,this study chooses the Fuzzy Analytic Hierarchy Process(FAHP),which can reduce the uncertainty and fuzziness of the decision-maker comparing the traditional AHP method,and improved the weights calculation accuracy.The evaluation of wetland ecosystem health,function and value was done by calculating the index score and weights.4.Element-based wetland change prediction is developed.Based on the results of wetland change detection,the classification results wetland land cover classification results are calculated,and wetlands change are simulated and predicted based on CA-Markov model.5.Taking the wetland in Erdos Relict Gull National Nature Reserve,Inner Mongulia as an example,remote sensing diagnosis of wetland change was conducted.Based on multi-temporal remote sensing image data,meteorological data,topographic data,measured data and socio-economic data,this paper takes the wetland as an example and applies the methodology proposed in this paper.Wetland change information with 12 change types from 1977 to 2014 were obtained.The wetland ecosystem health,function and value were evaluated in 2014.Based on the wetland change detection results from 1977 to 2014,and the wetland changes in the 100 years later are predicted.Based on literature survey and field investigation,the driving forcing of wetland change were analyzed,and suggestions were provided for the protection and restoration of the wetland.The main innovations of this study are as follows:1.A methodology for remote sensing diagnosis of wetland change is proposed.2.A wetland change detection method based on wetland element is proposed.3.Assessment of wetland ecosystem based on fuzzy analytic hierarchy process(AHP)is proposed.4.A method for wetland change simulation and prediction based on the wetland elemetns was proposed.