Header

UZH-Logo

Maintenance Infos

Evaluation of the capability of SPOT5-HRG data for predicting tree density in the northern Zagros forests


Pir Bavaghar, M; Ghahramani, L; Fatehi, P (2011). Evaluation of the capability of SPOT5-HRG data for predicting tree density in the northern Zagros forests. Iranian Journal of Forest and Poplar Research, 19(2):242-253.

Abstract

Quantitative attributes of forest stands are valuable data that are very important for the evaluation of forest resources. Regarding to unique structure of Zagros forests, we tried to predict tree density using SPOT5-HRG satellite data in this study. A systematic random grid consisting of 319 circle plots (0.1 ha) were used to collect field data. Spectral values related to field plots were extracted from original and the artificial bands composed of vegetation indices and principle component analysis. Ancillary data such as slope, aspect and elevation were also used. Multiple regression and stepwise method were used to predict tree density from 4 original spectral bands and 16 artificial bands as independent variables. Ancillary data didn' t improve the results. For considering geographic aspects effects, the study also was done for different aspects, separately. In the general model, predictive variables were PCAC2 (the 2nd component of PCA) and B2 (Red band) with the adjusted coefficient of determination of 0.26%. In the suggested models for the northern, southern, eastern and western forests, independent variables are PCAC2, Ratio, PCAC2, AVI, B1 and PVI, AVI, B3, with the adjusted coefficient of determination of 31%, 34%, 19% and 42%, respectively. The Results of model validation tests showed that all of the presented equations had a reliable validation and are useful for this area, however, for better estimation of tree density, we should find the other approaches.

Abstract

Quantitative attributes of forest stands are valuable data that are very important for the evaluation of forest resources. Regarding to unique structure of Zagros forests, we tried to predict tree density using SPOT5-HRG satellite data in this study. A systematic random grid consisting of 319 circle plots (0.1 ha) were used to collect field data. Spectral values related to field plots were extracted from original and the artificial bands composed of vegetation indices and principle component analysis. Ancillary data such as slope, aspect and elevation were also used. Multiple regression and stepwise method were used to predict tree density from 4 original spectral bands and 16 artificial bands as independent variables. Ancillary data didn' t improve the results. For considering geographic aspects effects, the study also was done for different aspects, separately. In the general model, predictive variables were PCAC2 (the 2nd component of PCA) and B2 (Red band) with the adjusted coefficient of determination of 0.26%. In the suggested models for the northern, southern, eastern and western forests, independent variables are PCAC2, Ratio, PCAC2, AVI, B1 and PVI, AVI, B3, with the adjusted coefficient of determination of 31%, 34%, 19% and 42%, respectively. The Results of model validation tests showed that all of the presented equations had a reliable validation and are useful for this area, however, for better estimation of tree density, we should find the other approaches.

Statistics

Downloads

3 downloads since deposited on 09 Apr 2013
0 downloads since 12 months
Detailed statistics

Additional indexing

Item Type:Journal Article, refereed, original work
Communities & Collections:07 Faculty of Science > Institute of Geography
Dewey Decimal Classification:910 Geography & travel
Language:Persian
Date:2011
Deposited On:09 Apr 2013 06:59
Last Modified:05 Apr 2016 16:44
Publisher:Research Institute of Forests and Rangelands
ISSN:1735-0883
Related URLs:http://www.sid.ir/en/ViewPaper.asp?ID=219279&vRadif=6&vWriter=PIR%20BAVAGHAR%20M.,GHAHRAMANI%20L.,FATEHI%20P.&vJournal=IRANIAN+JOURNAL+OF+FOREST+AND+POPLAR+RESEARCH&vDate=SUMMER%202011&vVolume=19&vNo=2%20(44)&vStart=242&vEnd=253

Download

Preview Icon on Download
Content: Published Version
Filetype: PDF - Registered users only
Size: 987kB