Selection of key terrain attributes for SOC model

Publikation: KonferencebidragKonferenceabstrakt til konferenceForskning

  • Mogens Humlekrog Greve
  • Kabindra Adhikari, Department of Soil Science, University of Wisconsin-Madison, Madison, WI, USA
  • Menaka Chellasamy, Danmark
  • Zhixing Guo, Danmark
As an important component of the global carbon pool, soil organic carbon (SOC) plays an important role in the global carbon cycle. SOC pool is the basic information to carry out global warming research, and needs to sustainable use of land resources. Digital terrain attributes are often use of predictors in Digital soil mapping of SOC. But there are no rules only few empirical guidelines on which digital terrain attributes to use. The aim of this paper was to select and the evaluate 21 digital terrain attributes and use the best for mapping. A typical 7500 km2 region located in Denmark was selected, total 2,514,820 data mining models were constructed by 71 differences grid from 12m to 2304m and 22 attributes, 21 attributes derived by DTM and the original elevation. Relative importance and usage of each attributes in every model were calculated. Comprehensive impact rates of each attribute which was processed by weighting, grouping, summation and normalization was taken as the important indicator in SOC model. The results showed that there use differences in the SOC model using different terrain attributes. Relative Slope Position(RSP), Channel Altitude(Chnl_alti),and Standard High (standh) are the first three key terrain attributes in 5-attributes-model in all resolutions, the rest 2 of 5 attributes are Normal High (NormalH) and Valley Depth (Vall_depth) at the resolution finer than 40m, and Elevation and Channel Base (Chnl_base) coarser than 40m. The models at pixels size at 88m, 92.8m which were nearest to 90m of the Globalsoilmap project and 30.4m nearest to 30m TM satellite image, were validated, the prediction accuracy reached extremely significant level.
Antal sider1
StatusUdgivet - 2015
BegivenhedThe 6th Global Workshop on Digital Soil Mapping - Nanjing, Kina
Varighed: 11 nov. 201414 nov. 2014


KonferenceThe 6th Global Workshop on Digital Soil Mapping

Se relationer på Aarhus Universitet Citationsformater

ID: 96315546