Effect of straw management practices on spectral based measurement of soil organic carbon

Simple item page
dc.contributor.advisorBector, Vishal
dc.contributor.authorJoshi, Raghav
dc.date.accessioned2023-06-25T04:52:09Z
dc.date.available2023-06-25T04:52:09Z
dc.date.issued2022
dc.description.abstractSoil organic carbon (SOC) is an important soil health parameter, which helps in determining additional inputs to be added in soil. Various straw management practices improve soil organic carbon of fields. Determination of SOC in laboratory by chemical analysis is time, labor and resource consuming. VIS-NIR spectroscopy offers simple, accurate, quick and effective methods for measurement of soil organic carbon. Therefore, in the present study, spectral signatures were captured from soil samples of different straw management fields using spectroradiometer, with the aim of determining soil organic carbon. Straw managed fields, where same practices were followed from the last three years, used in the study were residue incorporated fields, conventional tillage fields and mulched fields. Partial least squares regression (PLSR), random forest regression (RF) and cubist regression model (CRM) were developed for estimation of soil organic carbon using spectral signatures. Predictive performances of the developed models were evaluated using different indices like coefficient of determination (R2), root mean square error (RMSE), mean absolute percentage error (MAPE) and ratio of performance to deviation (RPD). Spectral signatures for dried fields at depths 0-5 cm, 5-10 cm and 10-15 cm were then measured. The field data were fed into the developed models for prediction of SOC, and compared with the laboratory SOC values. Results showed that residue incorporated fields had highest mean SOC of 0.65 %, whereas mulched field had mean SOC of 0.61 % and conventional tillage field had mean SOC of 0.55 %. While checking the model performances in validation, CRM obtained highest linear R2 (0.76) and RPD values (2.02); with PLSR obtaining R2 as 0.74 and RPD as 1.98; and RF obtaining R2 as 0.63 and RPD as 1.52. For field data, highest accuracy was attained for residue incorporated fields at 5-10 cm depth when PLSR model was used for prediction.en_US
dc.identifier.citationJoshi, Raghav (2022). Effect of straw management practices on spectral based measurement of soil organic carbon (Unpublished M.Tech. thesis). Punjab Agricultural University, Ludhiana, Punjab, India.en_US
dc.identifier.urihttps://krishikosh.egranth.ac.in/handle/1/5810197706
dc.keywordsSoil organic carbon, straw management practices, VIS-NIR spectroscopy, PLSR, RF, cubist, R2, RMSE, RPD, MAPEen_US
dc.language.isoEnglishen_US
dc.pages83en_US
dc.publisherPunjab Agricultural University, Ludhianaen_US
dc.research.problemEffect of straw management practices on spectral based measurement of soil organic carbonen_US
dc.subFarm, Machinery and Poweren_US
dc.themeEffect of straw management practices on spectral based measurement of soil organic carbonen_US
dc.these.typeM.Tech.en_US
dc.titleEffect of straw management practices on spectral based measurement of soil organic carbonen_US
dc.typeThesisen_US
Files
Original bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
Raghav thesis after viva 22.12.2022.pdf
Size:
2.63 MB
Format:
Adobe Portable Document Format
Description:
M.Tech.
Download
License bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
1.71 KB
Format:
Item-specific license agreed upon to submission
Description:
Download
Collections
Thesis