Volume 14, Issue 27 (8-2023)
J Watershed Manage Res 2023, 14(27): 38-51 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Danesh M, Bahrami H, Emadi S M. (2023). Applicability of Proximal Sensing Technology in the study of Silt in the soils of Mazandaran Province. J Watershed Manage Res. 14(27), 38-51. doi:10.61186/jwmr.14.27.38
URL: http://jwmr.sanru.ac.ir/article-1-1181-en.html
URL: http://jwmr.sanru.ac.ir/article-1-1181-en.html
1- Sari Agricultural Sciences and Natural Resources University
2- Department of Soil Science and Engineering, Tarbiat Modares University, Tehran
2- Department of Soil Science and Engineering, Tarbiat Modares University, Tehran
Abstract: (2195 Views)
Extended Abstract
Introduction and Objective: Silt is one the most important constituents of soil texture that directly influence the soil erosion process and should take into account in many projects of soil erosion management and conservation. The study of this fraction using the traditional and prevalent lab methods, especially on large scales, is time-consuming, laborious and costly. Today, this can be done in a quick and cost-effective method applying new high-techs such as the spectroscopy technology. The present work intends to investigate the spectral behaviours of the soil silt fraction using the reflectance spectroscopy technology in Mazandaran province.
Material and Methods: Accordingly, 128 soil samples were collected from 20 cm of soil surface using the SRS method and auxiliary info-layers like as geology, pedology, landuse and road map of Mazandaran province. First, the sample set was sub-divided into two subsets: calibration and validation. Spectral signatures and domains specific to the silt components were detected and specified utilizing the PLSR and Cross-Validation techniques, as well, the hyperspectral pre-processing methods such as averaging, smoothing and 1st derivative algorithms based on the Savitzky-Golay Algorithm were done.
Results: Modeling process was done based on the PLS technique to investigate the spectral signatures and behaviours of silt constituents. The final model with 4 latent factors (LFs) was calibrated with these specs: Rc: 0.55, RMSEc: 8.31 %, RPDc: 1.20 and RPIQc: 1.71 and was eventually selected as the best model for studying the soil silt of Mazandaran province. Results showed the model potentiality in prediction of soil silt of the study area, as well, the most influential spectral domains and ranges were detected and recognized. The correlation coefficients of silt contents with the influential spectral ranges and wavebands were also defined as follows, UV-390 nm: 0.27, Vis-680 nm: 0.31, NIR-970 to 990 nm: 0.32, SWIR- 1400 to 1410 nm wavebands: 0.34, 1910-1930 nm: 0.38, 2200-2210 nm: 0.39, 2340-2350 nm: 0.41 and finally, for 2430-2460 wavebands calculated as 0.43. The obtained spectral wavebands with the highest correlation coefficients (R(CCmax)) indicate the high impact as the independent predictor variables in the processes of soil silt modeling of Mazandaran province. Finally, the capability of the proximal sensing of diffuse reflectance spectroscopy technology (VNIR-PS) was demonstrated in the study of silt contents of Mazandaran province.
Conclusion: In this approach, the spectral ranges and bands affected by the silt components were defined, in addition to the predictive modeling processes. That can be used as a basis for studying silt contents at large scales applying the upscaling operation via airborne/satellite hyperspectral data. Also, it indicates the importance of soil reflectance spectroscopy technology as a fundament for detecting and recognizing the useful and effective spectral wavelengths as well as creating the optimized model for the utilization by remotely sensed satellite data. Moreover, the use of data with higher coefficient of variation and greater amplitude is highly recommended to improve and boost the model preciseness so that, the PLS algorithm can process better.
Introduction and Objective: Silt is one the most important constituents of soil texture that directly influence the soil erosion process and should take into account in many projects of soil erosion management and conservation. The study of this fraction using the traditional and prevalent lab methods, especially on large scales, is time-consuming, laborious and costly. Today, this can be done in a quick and cost-effective method applying new high-techs such as the spectroscopy technology. The present work intends to investigate the spectral behaviours of the soil silt fraction using the reflectance spectroscopy technology in Mazandaran province.
Material and Methods: Accordingly, 128 soil samples were collected from 20 cm of soil surface using the SRS method and auxiliary info-layers like as geology, pedology, landuse and road map of Mazandaran province. First, the sample set was sub-divided into two subsets: calibration and validation. Spectral signatures and domains specific to the silt components were detected and specified utilizing the PLSR and Cross-Validation techniques, as well, the hyperspectral pre-processing methods such as averaging, smoothing and 1st derivative algorithms based on the Savitzky-Golay Algorithm were done.
Results: Modeling process was done based on the PLS technique to investigate the spectral signatures and behaviours of silt constituents. The final model with 4 latent factors (LFs) was calibrated with these specs: Rc: 0.55, RMSEc: 8.31 %, RPDc: 1.20 and RPIQc: 1.71 and was eventually selected as the best model for studying the soil silt of Mazandaran province. Results showed the model potentiality in prediction of soil silt of the study area, as well, the most influential spectral domains and ranges were detected and recognized. The correlation coefficients of silt contents with the influential spectral ranges and wavebands were also defined as follows, UV-390 nm: 0.27, Vis-680 nm: 0.31, NIR-970 to 990 nm: 0.32, SWIR- 1400 to 1410 nm wavebands: 0.34, 1910-1930 nm: 0.38, 2200-2210 nm: 0.39, 2340-2350 nm: 0.41 and finally, for 2430-2460 wavebands calculated as 0.43. The obtained spectral wavebands with the highest correlation coefficients (R(CCmax)) indicate the high impact as the independent predictor variables in the processes of soil silt modeling of Mazandaran province. Finally, the capability of the proximal sensing of diffuse reflectance spectroscopy technology (VNIR-PS) was demonstrated in the study of silt contents of Mazandaran province.
Conclusion: In this approach, the spectral ranges and bands affected by the silt components were defined, in addition to the predictive modeling processes. That can be used as a basis for studying silt contents at large scales applying the upscaling operation via airborne/satellite hyperspectral data. Also, it indicates the importance of soil reflectance spectroscopy technology as a fundament for detecting and recognizing the useful and effective spectral wavelengths as well as creating the optimized model for the utilization by remotely sensed satellite data. Moreover, the use of data with higher coefficient of variation and greater amplitude is highly recommended to improve and boost the model preciseness so that, the PLS algorithm can process better.
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
