Study On The Change Of Soil Organic Carbon And The Prediction Model Of Total Organic Carbon Based On Hyperspectral After Biochar Addition

Soil organic carbon(SOC)is one of the most important factors affecting soil fertility.Due to its highly aromatic carbon structure and developed pore structure,biochar can be used as a soil improver to increase soil organic carbon content and improve soil physical structure.In recent years,biochar has become a research hotspot in the field of soil science and agricultural environment.The rapid and accurate measurement of soil organic carbon content is an important basis for the study of biochar agronomy and soil environmental effects.This study is based on two long-term positioning experiments by Zhejiang Academy of Agricultural Sciences: "Effects of Biochar Application on Soil Organic Carbon in Paddy Field"(Experiment 1)and "Effects of Different Fertilization Treatments on Soil Quality and Crop Productivity in Paddy Field"(Experiment 2).To explore the effect of biochar on the content of soil organic carbon.The soil spectra of "Effects of Biochar Application Amount on Paddy Soil Organic Carbon" and "Effects of Different Fertilization Treatments on Soil Organic Carbon" were measured through indoor experiments.The prediction models of soil total organic carbon(TOC)after biochar application were established by comparing different spectral pretreatment methods,sample selection methods,three models and four characteristic band selection methods,and the prediction effects of these models were compared and analyzed.The main results are as follows: The main research results show that:(1)The experimental results of the effects of different biochar application rates on soil organic carbon showed that,the TOC content was significantly increased in rape with medium and high application rates,but the growth trend in rice season was significantly less than that in rape season,and there was no significant difference between different application rates.Biochar significantly increased the content of soil dissolved organic carbon(DOC),among which the soil DOC content was the highest under the application rate,which increased by 111.48% compared with the control.Biochar significantly increased the content of soil readily oxidized organic carbon(ROOC)in rape season,and there was no significant difference between middle and high application rates.The maximum content of soil microbial biomass carbon microbial biomass carbon(MBC)in rape season and rice season appeared under medium application rate.When the application amount was large,biochar had a significant effect on soil Light Fraction Organic Carbon(LFOC),and the proportion of LFOC in rape season was significantly higher than that in rice season.By comparing different pretreatment methods,modeling sample selection methods,modeling methods,and feature band extraction algorithms,an optimal model of soil TOC for biochar application was established,indicating that a pre-treatment method combined with SG smoothing and multiplication scatter correction(MSC),Kennard-Stone(KS)sample selection method,Genetic Algorithm-Support vector machine(GA-SVM)model established by Successive Projections Algorithm(SPA)characteristic band algorithm,which was suitable for fast prediction of TOC.(2)The experimental results of the effects of different fertilization treatments on soil organic carbon showed that,biochar significantly increased soil TOC content in rape season,and the maximum TOC content appeared in biochar treatment,which increased by 36.21% compared with the control.The TOC content of the treatment with biochar was the highest,which was 36.21% higher than that of the control.In rape season,straw returning and biochar treatment significantly increased soil DOC content,and soil DOC content reached the maximum under biochar treatment.There were significant differences between biochar treatment,control and biochar reduction fertilization treatment,indicating that reduced the effect of biochar on soil DOC.However,there was no significant difference between rice season and rape season,but there was no significant difference between straw returning,biochar and biochar reduction fertilization.In rice season,there was a significant difference between biochar reduced fertilization and biochar treatment,indicating that reduced fertilization may reduce the effect of biochar on soil ROOC.The MBC content of straw returning treatment reached the maximum in rape season,which increased by 23.62% compared with fertilization control and only 7.30% in rice season.There was no significant difference between biochar reduction fertilization and biochar treatment in the two seasons.In the rape season,rape season straw returning and biochar treatment significantly increased the proportion of soil LFOC.By comparing different pretreatment methods,modeling sample selection methods,modeling methods,and feature band extraction algorithms,an optimal model of soil TOC for different fertilization treatments was established,indicating that the SG convolution smoothing pretreatment method and the sample set partitioning based on joint x-y distances(SPXY)sample selection method,as well as the Partial Least Squares(PLS)model established by the SPA characteristic band algorithm was suitable for the rapid prediction of TOC.