Remote Sensing For Monitoring The Health Condition Of Artifitical Robinia Pseudoacacia Forests And Analysis Of Their Dieback Or Death In The Yellow River Delta

Artificial Robinia pseudoacacia forest in the Yellow River Delta is the largest artificial Robinia forest base of the East China Plain area, which has become the protective sand barrier of Shengli Oilfield and Dongying City, Shandong Province, which is also the attractive tourism landscape in the Yellow River Delta Nature Reserve. Since the90's last century, the phenomena of the Robinia dieback and death occurred in the Yellow River Delta. It is a great loss to the forestry development of the Yellow River Delta, and a great challenge for forestry scientists and technicians. Dynamically monitoring the change area of artificial Robinia dieback and studying its cause of dieback and death, have great significance.In this paper, the study area is the Yellow River Delta. We used GIS and RS as the main means of technique, based on the classification of the Robinia health of the Yellow River Delta, analyzed the dynamic changes of the spatial distribution pattern of1999-2007Robinia health category, and discussed the reason of the Robinia forest dieback deeply. The main work includes:(1) We customized the Robinia field survey routes and the choice of plots combined with the remote sensing methods. And we designed the selection, the sampling rules and the process of the Robinia forest plots. The Robinia health vigor indexes (live crown ratio, crown diameter, crown density and crown mortality) of the field plot were measured and calculated. The Robinia forest was assessed according to the canopy condition rating guide (CCRG), and was classified as the basis of the average of all the trees in field sampling. And then Robinia forest was classified into four classes of tree vigor:health, light dieback, moderate dieback, severe dieback or death using visual crown rating method.(2) Remote sensing image registration and Robinia forest extraction:The remote sensing images of2001,2003and2007were registered to1999image. The Robinia forest in the2007Landsat TM image was extracted by supervised classification method. A sensitivity analysis of different vegetation indices to Robinia crown vigor was carried out. NDVI, NDWI, SAVI, MSAVI vegetation indexes and TCT transformation were applied to the Robinia Landsat image, each transformation method classified Robinia into four classes using cluster analysis. The Robinia forest health classification accuracy was measured by different vegetation indexes combined with the40crown levels of field sampling data. The result showed that:Normalized difference water index (NDWI) has the highest accuracy of82.5%for classification of Robinia health.(3)We developed the research of dynamic evolution rule of the Robinia forest health classification in spatial distribution pattern. According to the classification of1999,2001,2003,2005and2007multidate Robinia images, we analyzed the previous change trends of Robinia forest health. The result showed:The healthy Robinia forest or that with light dieback in all five phases accounted for15.5%of the study area, healthy or light dieback in four phases occupied15.2%, the area of Robinia forest where the health condintion was deteriorated accounted for12.3%of the study area, the area of Robinia forest where the health condition was always poor dominated18.8%of the study area, the area of Robinia forest where the health condition had been restored accounted for5.9%.(4)According to the spatial and statistical analysis of site conditions and Robinia forest health category, we explored the factors of the Robinia forest dieback or death. The research showed that:Soil salinity, soil texture, elevation, water depth and salt content of underground water were closely related to the health condition of Robinia forest.(5) Based on the analysis of site conditions and the present Robinia health condition, we separated artificial Robinia forest into three suitable partitions in the Yellow River Delta inappropriate, more appropriate, suitable. And then corresponding recommendations for different partition types was proposed. It needs to plant Robinia, grass and wetland vegetation according to local conditions. Which can be protected from a single species, and create a better ecological environment.