Study On Forest Carbon Storage And Soil Erosion Based On Landsat Data

As an important part of terrestrial ecosystems,forest ecosystems have irreplaceable social,economic and ecological benefits.The landscape analysis method is used to analyze and evaluate the land use/cover type,carbon storage/carbon density and soil erosion in the study area,thus providing a basis for prevention and control of soil erosion and multi-functional forest management.Based on the Landsat image data of 1997,2008 and 2017,land use/cover classification and accuracy verification were carried out for Jiangle County.According to the classification results of land use/cover,the land use change is analyzed and predicted.Extracting data including visible light band,vegetation index,topographic factor,principal component,tasseled cap transformation factors and texture index.The correlation analysis between remote sensing features value and ground survey sample data is processed to obtain significant correlation features.With the selected features,a model based on BP neural network method is established to evaluate carbon density distribution in study area.According to the monthly precipitation,soil types,topography factors,vegetation coverage and land use/cover decomposition in the study area,the corresponding soil erosion factors were calculated..The soil erosion is calculated by ArcGIS10.3.The distribution soil erosion along altitude and slope is analyzed.Fragstats 4.2 is used to calculate the various landscape indexes of land use/cover,soil erosion and carbon density in the study.After analyzing,conclusions are;1)The result by Multiple Endmember Spectral Mixture Analysis are better than the constrained least squares decomposition.The vegetation-soil index derived by vegetation abundance and soil abundance can be used to enlarge the difference between different forest types.The combination of visible light band,vegetation abundance,vegetation-soil index and NDVI are used in the classification of different forest types for 2007 and 2017 in forest area.The overall accuracy of the first step in 1997,2007 and 2017 are 90.40%,91.90%and 88.01%,respectively,with the kappa indexes of 0.85,0.88 and 0.83,respectively.The overall accuracy for the forest area in 2007 and 2017 are 85.17%and 86.62%,with the kappa indexes of 0.80 and 0.82.According to the classification results,the water area first decreases and then increases.The area of'build-up and forests continues to increase.The area of cultivated land and bare land continucs to decrease.The change of bare land is the most severe,mainly affected by urban expansion and forestry production and management.Moreover,the distribution of bare land was in certain randomness.Secondly,the cultivated land or build-up were mainly represented by the reduction of cultivated land area and the increase of build-up area,and the build-up area mainly comes from cultivated land.The forest area has the largest change area and the smallest change rate,which is mainly converted to cultivated land.The kappa value of the land use/cover results between actual results and predicted results for 2017 was 0.81,indicating the suitability of the CA-Markov model.Therefore,the land use/cover of the study area in 2027 is predicted.2)According to the Pearson correlation analysis,the reciprocal of the band 2 and band 6,and the Mean texture of the band 2 and band 5 under window 5X5 were selected as the input variables of the neural network model.Finally,a neural network model with a coefficient of determination of 0.66,a root mean square error of 5.81 t/ha and a relative root mean square error of 0.72 is establishd.According to the model,the carbon density distribution of forest land in the study area was obtained.The average carbon density is 95.58 t/ha,the standard deviation is 56.47 t/ha,and the total carbon stock of forest land is 1.92×107 t.According to the statistics of different forest types,the average carbon density of different types is ranked as mixed forest>Masson pine forest>broad-leaved forest>Chinese fir forest,with values of 103.76 t/ha,98.19 t/ha,95.43 t/ha and 90.90 t/ha,respectively.The total carbon storage of Pinus,Chinese fir,broad-leaved forest and mixed forest were 7.07×107t,6.14x1 07 t,4.26×107 t and 1.72× 107 t,respectively.3)In 2007 and 2017,the soil erosion area of Jiangle County was 991.02 km2 and 1001.12 km2,respectively.The average soil erosion modulus was 508.96 t/(km2·a)and 522.82 t/(km2·a).The total amount of soil erosion was 1,146,200 t/a,1,175,700 t/a;the intensity of soil erosion is mainly micro-and light.The average soil erosion modulus increases first and then decreases with the increasing of altitude while it increases with the increasing of slope.The slight erosion is mainly happened on the slope of 0°-15“.While the mild erosion intensity is mainly happened on the slope of 15° or above.The proportion of erosion in the altitude 200-800m reached 80.56%and 82.48%,respectively,while the proportion of erosion in slope 1 5°-25°reached 72.68%and 72.52%,respectively.It indicates that the area of soil erosion in the study area is mainly happened at an altitude of 200-800m and the slope of 15°-25°.The main reason is the greate impact of human activities in the region,the management measures such as clear cutting have certain influence on it.Therefore,the area should be considered more when developing measures.4)During the period of 1997-2017,the land use/cover of the study area has undergone dramatic changes,the degree of landscape fragmentation has decreased,while the shape tends to be regularized and the connectivity between the landscapes has increased.The main reason is the development process,especially the construction land.The massive annexation of unutilized landscapes reduces fragmentation and increases the concentration of landscapes.The forest carbon density patch area is small,the types are diverse,the distribution is balanced,and the patches are generally fragmented,which is greatly affected by human activities.The Level 4(66.21 t/ha-122.69 t/ha)carbon density level centered on the average carbon density represents the actual carbon density landscape distribution of most of the forest land,indicating that the quality of the forest land in Jiangle County is better and has high connectivity.The number of soil erosion patches is large,while the average area of patches is small,indicating that there is no large-area erosion of the same type in the study area,but it has good connectivity and should be prevented from being connected.The number of soil erosion patches increased first and then decreased with the increase of carbon density level,and the distribution of erosion patches was relatively dispersed.The forest soil erosion was affected by its carbon density distribution.Erosion is mainly happened in Leve13,Leve14 and Leve15.The number of patches and patch area are small for moderate erosion,the serious erosion and the extreme erosion area are mainly within one pixel.In general,human activities play an important role on land use/cover,soil erosion and forest carbon density.The main driving forces are the implementation of urbanization expansion,forest management,and Grain for Green Project.