Study On Extraction Methods Of Monitoring Information Of The Natural Forest Resources In Western Sichuan Based On Remote Sensing Technique

The Natural Forest Protection Project(NFPP),one of the largest investment on forest ecological construction,has been implementing for almost 20 years.The effectiveness of this project is always a focus of international and domestic social concern.As the rapid development of technologies of satellite application,the remote sensing technology is much better than the traditional ways on monitoring range,content,aging and accuracy.To support the basic data and technology for the sustainable development of the natural forest in the Western Sichuan,based on the TM images of 1989,1997 and ETM images of 2015 in the Daofu Research Area,Tibetan Autonomous Prefecture of Garze,Sichuan Province,this paper extracted the forest's information and monitored the dynamics before and after the project of forest's resources,landscape,soil and water conservation,carbon sequestration capacity and effectiveness of forest.etc..Based on the Quick Bird images in Jiajinshan Research Area,using object-oriented multi-scale segmentation method to extract the destruction information and select the risk evaluation factors,combined with the comprehensive evaluations of information value model to evaluate the risk of all kinds of damages and the landscape hazard rank division on natural forest.The main research contents and innovations of this paper are as follows:(1)Dealing with e Congnition 8.64,using object-oriented multi-scale segmentation method and screening the optimal segmentation scale and parameters for different failure types,we established the classification and recognition rule set for subalpine forest region of Western Sichuan and extracted automatically kinds of information on natural forest destruction;we extracted 48 landslide points,6.6113hm²of landslide area,19.7146km of road length,8.1295 hm²of road area,275 houses and4.5982 hm² of housing area.The total classification accuracy was more than 87%.(2)Combined with remote sensing factors,using the correlation analysis forest biomass measured and topographic and the method of linearity cure and multiple regression,we get the model respectively for arbor in subalpine forest region of Western Sichuan:Y=127.340 TM2+93.835 TM3+344.518 TM5-75.505 WVI+0.339V13-226.322 BVI+9.664(R²=0.888,P<0.05)and the model of bush:Y=-49.469TM2-141.236 TM3+0.056 BVI+16.468(R²=0.790,P<0.05).(3)We established the evaluation index system of soil erosion intensit for subalpine forest region of Western Sichuan and divided it to 4 grades and we could know from the results that the rate of the areas of severe erosion was increasing from12.38%to 14.73%as accompanied the decrease of mild erosion area before the NFPP;after the project's implementation,the intensity of severeerosion area shrank slightly and the area ratio dropped from 14.73%to 11.89%in 2015,which meant soil erosion deterioration had been effectively curbed by the NFPP.(4)By Gram-Schmidt transforming,image fusion and using the maximum likelihood method(Likelihood Maximum)for the three phases of the images classification,we extracted the location and area of different vegetation cover types and knew that the forest coverage rate was 65.05%in 1989,one year before the implementation of the NFPP,namely 1997,the rate reduced to 62.60%,the proportion had a reduction of 2.46%.NFPP which was started in recent 20 years,made the forest coverage rate back up to 65.34%in 2015.The data changes showed us that the NFPP's implementation was good for the recovery of vegetation coverage.(5)Based on the results of fusion image interpretation,we extracted the distributions of forest landscape types in three stages(1989,1997,2015)and used ENVI5.0 and Fragstats4.0 to monitor the forest landscape types changes,which showed us bush land dominated the whole experiment area,capturing 34.09%,34.27%and 33.89%,the arbor land ranked the 2^nd,capturing 30.11%,26.75%and 30.61%in the whole area.The bush landscape and the arbor landscape consist of the compound landscape,which controlled and mattered the evolution and the regulation of the regional ecological environment.Analysis on the temporal dynamics told us the number of patch,patch density and maximum patch area of bush and the arbor landscape were decreasing before the NFPP and increasing after it,the degrees of fragmentation were volatile while others'were not obvious;ED of three stages were32.7588,36.0567 and 35.2637,COHESION were 95.1919,96.5094 and 96.2046,SHDI were 0.733,0.7126,0.7523 and they behaved the characteristics of the data changes reflected to drop back.The data also told us the NFPP had promoted the evolution of landscape pattern,declined the fragmentation degree of landscape and helped the recovery of the whole landscape's connectivity,which also had promoted the material cycles and energy flows in ecosystem and also the biodiversity.(6)Based on the remote sensing biomass estimation model and data,we got the the biomass in 1989,1997 and 2015 were 2828529.7Mg,2623878.8Mg and4130166.9Mg.The contribution rate of arbor's carbon sequestration was 80.17%,77.35%and 79.16%.Statistics show that protecting the arbor has a significant influence on stabilizing the carbon sequestration capacity and effectiveness.Before and after the NFPP,the carbon sequestration capacity and efficiency showed the dynamic of first decline then up.(7)Calculations of NDVI by using red and near infrared data and dynamic changes of vegetation in the research area showed that the vegetation coverage in the research area significantly decreased in 1989?1997,the area ratio of medium coverage(40%?60%)decreased from 70.56%to 59.90%.After the NFPP,namely from 1997 to2015,the vegetation coverage in the research area was obviously picked up and so did the area ratio of medium coverage,picked up from 19.26%to 31.09%,an increase of11.83%.The data told us that the NFPP's implementation helped the recovery of vegetation coverage.(8)By using information method and combining RS with GIS,the risk of various types of damage was evaluated:the destruction of landslide was the most possibly to happen in bare land in Jiajinshan Research Area,where the slope was bigger than50°,the density of gully was bigger than 7km/km²and the distance was less than 100m away from the road.Grassland area,where the slope was within 20°and the distance was between 500m and 1000m away from the habitations,had the the destruction of grazing most clearly.The wind throw destruction was the most easily to happen in forest area,where the elevation was from 3500m to 3800m of the northwestern slope and the canopy density was less than 0.5.The region where the elevation was more than 3500m in the northen slope and tree specie was picea asperata or fir was the easiest to have the snow pressure destruction.(9)Based on information of the risk of various types of damage and Natural Breakpoint Method,we zoned the risk grade of natural forest:the total areas of severe hazardous districts was 707.9519hm²,capturing 14.49%,the total areas of mild and a very mild danger districts was 1968.0607 hm²,capturing 40.28%of the research area.the total areas of medium and hazardous districts was 2209.6133hm²,capturing45.23%.The data told us that The overall risk of natural forest destruction in Jiajinshan Research Area was in the intermediate state,as long as the protective measures were strengthened and the external environment was improved,the risk would reduce and the natural forest resources would be returned to the track of sustainable development.