SCIENTIA GEOGRAPHICA SINICA ›› 2014, Vol. 34 ›› Issue (12): 1511-1517.doi: 10.13249/j.cnki.sgs.2014.012.1511

• Orginal Article • Previous Articles     Next Articles

Characteristic Scales and the Multi-scale Correlation Between DEM and NDVI Factors: A Case Study in the Mongolian Plateau

Yun-feng HU1(), Zhi-ying XU1,2   

  1. 1. Institute of Geographical Sciences and Natural Resources Research,Chinese Acadamy of Sciences, Beijing100101,China
    2. University of Chinese Academy of Sciences, Beijing 100049,China
  • Received:2013-07-11 Revised:2013-11-05 Online:2014-12-10 Published:2014-12-10


As one of multi-scale analysis methods, wavelet analysis can effectively solve multi-scale problems in geo-sciences research. In order to effectively analyze the spatial distribution pattern and identify the multi-scale features of DEM and NDVI, four sampling lines were set up along the longitudinal and latitudinal transects (43°N, 47°N, 100°E, and 110°E). And then supported by wavelet transformation using db6 as wavelet base, the wavelet variances and multi-scale relationships of DEM and NDVI were calculated to identify the characteristic scale and analyze the coupling relationship between DEM and NDVI. The results showed: 1) In Mongolian Plateau, the spatial variations of DEM in northern and western parts were stronger than that in the southern and eastern parts. Along the latitudinal and longitudinal transects, there existed one characteristic scale around 20 km. 2) The spatial variations of NDVI was similar to that of DEM in the whole Mongolia Plateau. While two characteristic scales of NDVI distribution were identified, with a small scale of 2-4 km and a larger scale of 20 km. 3) As to the DEM and NDVI factor, there existed a coupling relationship in the 20 km scale. As a great change of the topography(DEM) had taken place in the specific scales, a large change of vegetation (NDVI)then occurred correspondingly. Generally, the terrain factors tended to affect the spatial distribution of NDVI in macroscopic scale, instead of microscopic scale. This article finally summarized that the wavelet analysis method had the abilities indeed to recognize the characteristic scale and to quantitatively analyze multi-scale correlation among different factors. At the same time, the authors also pointed out that the method had some incapacities for the two-dimension spatial data analyses, including continuous scale analysis, the multi-scale correlation analysis in a specific position and two-dimensional analysis.

Key words: wavelet, scale, DEM, NDVI, Mongolia Plateau

CLC Number: 

  • Q149