Remote sensing inversion of soil organic carbon in Jining city based on decision tree classification

摘要

This study used Landsat8 remote sensing data and the measured data of soil organic carbon, taking the carbon content of the surface 0-20 cm of soil in the study area as the research object, through multiple linear regression analysis of SPSS, building the prediction model: soc = 24.476 + 51.503 × b_1-49.824 × b_4-11.297 × b_6-21.778 × b_7, R~2 = 0.747, P < 0.01. Then, based on the Band Math and Decision Tree of ENVI, the soil organic carbon content and distribution in the whole study area were obtained. The results showed as follows: (1) The average soil organic carbon content in the study area was about 12.45 g/kg, which was in the middle even lower level. And the organic carbon content in some areas was close to 0 g/kg, the highest content was 22.5 ± 4.5g/kg. (2) The organic carbon content was mostly concentrated in 12-18 g/kg, which area of about 6165.13 km~2, accounting for the total land area of 59.99%. It was the most widely distributed.