browser icon
You are using an insecure version of your web browser. Please update your browser!
Using an outdated browser makes your computer unsafe. For a safer, faster, more enjoyable user experience, please update your browser today or try a newer browser.

Applicability of kriging to predict spatial distribution of carbon stocks of Acacia mangium plantations

Posted by on March 30, 2010

Tiryana, T., Tatsuhara, S., and Shiraishi, N.  2009.  Applicability of kriging to predict spatial distribution of carbon stocks of Acacia mangium plantations.  Journal of Forest Planning 14(1): 17‒26. Available at: http://ci.nii.ac.jp/naid/110009357740 or here: http://tinyurl.com/6q8o9ea

Abstract

The prediction of carbon stocks is essential to evaluate the environmental benefits of forests. Although carbon stocks can be predicted by combining sample plots from several or all forest stands, this approach cannot directly provide spatial information on the distribution of the carbon stocks. We investigated the applicability of kriging as an alternative method to spatially predict the carbon stocks of an Acacia mangium plantation in Indonesia. The carbon stock data obtained from 247 sample plots and variogram analysis were used to characterize and model the spatial autocorrelation of the carbon stocks. The spatial prediction of carbon stocks was carried out using ordinary kriging (OK), universal kriging (UK), and co-kriging (CK) methods. UK used the geographical coordinates as a secondary variable, whereas CK used stand age. The prediction accuracy of the kriging methods was assessed using cross-validation. The carbon stocks were spatially autocorrelated up to a distance of 3600 m and tended to be more similar along the direction of 135. The kriging methods produced similar prediction maps of the spatial distribution of the carbon stocks. There were no significant differences among the kriging predictions and their standard deviations. Cross-validation, however, confirmed that CK performed better than OK and UK. The predicted carbon stocks varied from 7.25 to 37.95 tC/ha, with a standard deviation ranging from 9.44 to 12.21 tC/ha. Thus, kriging can be used as an alternative method to predict the spatial distribution of carbon stocks or other forest attributes.

Keywords: carbon stocks, geostatistics, kriging, spatial autocorrelation, variogram

Comments are closed.