Document Type : Research Paper
Authors
1 MSc student, Faculty of Agriculture, Tarbiat Modares University, Iran
2 Professor, Faculty of Agriculture, Tarbiat Modares University, Iran
3 Assistant Professor, Soil Conservation and Watershed Management Research Institute, Iran
Abstract
Soil Organic Matter (SOM) is one of the most vital components in soil solid phase and plays an important role in soil aggregation. SOM considerably improves soil physical conditions during long time. Moreover, knowing the SOM content of arid and semi-arid areas could significantly affect the soil management in such SOM-lacking regions. The objective of this research was to compare four common statistical models with the purpose of acheiving the best model for estimating the SOM in Ivanekey and test the obtained model in another semi-arid area of Urmia plain. Two Hyperion images were collected for Ivanekey and Urmia plain which were used to identify the relationship between spectral bands and SOM by applying SWR, MinR, PLSR and PCA models. Models were calibrated by Ivanekey Hyperion images and verified by nine Ivanekey and 23 Urmia random samples. Results indicated that MinR and PCA models do not have reasonable accuracy, but SWR and PLSR can provide the highest accuracy with the same RMSE value of 0.18, and R2 of 0.69 and 0.66, respectively. It is also demonstrated that there is a close relation between SOM and Red and Short Wave Infra Red spectral regions.
Keywords
)