Research On The Remote Sensing Inversion And Spatial Distribution Of Soil Organic Carbon In Forest

Forest ecosystems occupy the leading status in regional carbon cycle and carbon balance. Effectivemonitoring of SOC in forest land provides important information for soil quality assessment, sustainableforest management and regional climate change. Considering the measured data, remote sensingmonitoring can expand study scope from point to surface. And it also provides technician support tospace monitoring of SOC. Therefore, based on104sample points, the remote sensing data were used inthis paper. To model the soil basal respiration, the Van't Hoff model was used to combine with CASAmodel. Finally, the estimation of SOC had been realized.The main conclusions were as follows:(1) After spectrum enhancement processing, the threshold segmentation method was used in order todivide the image into two spectrum patches which named A and B. For better accuracy, we madesubmodels respectively. In this way we could reduce the impact of mountainous terrain shadow and thenonlinear characteristics in vegetation index.(2) The traditional CASA model was improved by separating NDVI and SR. The Van't Hoff model wasused to combine with NPP. Then we estimated the distribution of soil basal respiration based on NDVIand SR respectively. Divided measured sample point into two submodels by different spectrum patches.At the same time, to keep the estmiation of SOC well matched with the measured SOC, we constructedtwo submodels asOC₁=092×A_ndvi^1.021and OC₂=A_sr^0.982+3.944px+7.907The researchshowed that, soil basal respiration base on NDVI was better suited for independent variable insubmode-one, the best R²came from power exponent curve which reached0.585(P<0.05)with82.80%average relative accuracy. In the meantime, soil basal respiration base on SR was better suited forindependent variable in submode-two. The best R²(0.643) came from nonlinear model by introduceingthe quantification slope factor, with84.47%average relative accuracy.(3) Under different terrain and biological conditions, we analysed the spatial distributing rules of SOCin forest. The study figured that: spatial heterogeneity can obviously be found in woodland soil organic carbon. In general, the SOC increased significantly with altitude,gradient and stand ages.The meanvalue of SOC was higher in south-facing slope than in north-facing slope.The mean value of SOC wasin the sequence of broad-leaved forest> Pinus Massoniana> Chinese Fir. SOC in natural forest washigher than planted forest. The above conclusions were basically the same as the results of manydomestic reviews.