| Author | Nazarmavloev,Farrukh |
| Note | A thesis submitted in partial fulfillment of the requirements for the
degree of Master of Science in
Remote Sensing and Geographic Information Systems, School of Engineering and Technology |
| Publisher | Asian Institute of Technology |
| Abstract | Soil degradation monitoring and environmental management by using the up-to-date and
accurate information of soil properties is very important. The conventional methods of soil
analysis which are time consuming and expansive can be replaced by the remote sensing
technologies for more rapid and efficient soil mapping. This study armed to evaluate space
born image spectroscopy capacity for mapping soil organic carbon in the loess area with low
to medium organic carbon content. The loess hills of Faizabad watershed in the western part of
Tajikistan was selected as study area, where the soil erosion is considered the fastest and most
wide spread soil degradation process and reduction of soil organic carbon has been identified
as one of the key degradation processes. Soil samples (n=30) collected from the field where
analyzed for organic carbon content by conventional method in the laboratory. The spectral
reflectance of the samples was measured in the lab and in the field using ASD Field
spectrometer (350–2500 nm). Hyperion image was acquired, preprocessed and
atmospherically corrected using FLAASH model of ENVI. The total 160 bands have been
selected for the analysis. The spectra of the samples site (n=23) were extracted from the
Hyperion image (400–2500 nm). Partial Least Squares (PLS) regression with cross validation
was used to develop calibrations models. The coefficients of determination in calibration (R2)
and standard errors in cross validation (SECV) were 0.78 and 0.14 for laboratory reflectance
spectra (n=30). For filed spectra (n=30), PLS regression result of R2 of 0.71 and SECV 0.16.
and R2 of 0.67 and SECV of 0.18 of Hyperion image spectra (n=23) respectively. The SOC
map was generated by applying the result of the PLS model to the Hyperion image. The result
showed that predictions of SOC using the Hyperion spectra were less accurate than those of
the lab and field spectra. However, for rapid and cost-effective approach, Hyperion
demonstrated the good potential to predict soil organic carbon and digital soil mapping in the
study area. |
| Year | 2011 |
| Type | Thesis |
| School | School of Engineering and Technology (SET) |
| Department | Department of Civil and Infrastucture Engineering (DCIE) |
| Academic Program/FoS | Remote Sensing (RS) |
| Chairperson(s) | Honda. Kiyoshi ; |
| Examination Committee(s) | Tripathi, Nitin K. ;Nagai, Masahiko ;Wolfgramm, Bettina ; |
| Scholarship Donor(s) | University of Central Asia, Kyrgyz Republic; |