Preprints
https://doi.org/10.5194/gmd-2018-274
https://doi.org/10.5194/gmd-2018-274

Submitted as: development and technical paper 30 Nov 2018

Submitted as: development and technical paper | 30 Nov 2018

Review status: this preprint has been withdrawn by the authors.

SBDM v1.0: A scaling-based discretization method for the Geographical Detector Model

Xiaoyu Meng1,2, Xin Gao1, Shengyu Li1, Wenjing Huang3, and Jiaqiang Lei1 Xiaoyu Meng et al.
  • 1State Key Laboratory of Desert and Oasis Ecology, Xinjiang Institute of Ecology and Geography, Chinese Academy of Sciences, 818 South Beijing Road, Urumqi 830011, Xinjiang, China
  • 2University of Chinese Academy of Science, Beijing 100049, China
  • 3College of Resources, Environment and Tourism, The Capital Normal University, Beijing 100048, China

Abstract. Geographical Detector Model (GDM) can be used to assess the affinity between potential environmental factors and the response variables. If environmental factors entered are continuous, the first step for application of GDM is to discretize the continuous variable into category strata with an appropriate discretization method. Many one-dimensional discretization methods have been arbitrarily applied to GDM but failed to obtain the optimal strata of environmental factors, resulting in an inaccurate model output. In this paper, we present the Scaling-Based Discretization Method (SBDM) as a novel discretization method that can be used to obtain the optimal strata for GDM. The SBDM takes the power of determinant as a criterion function through upscaling and downscaling processes to obtain the optimal discretization. The software was tested with two case studies: (1) The distance to river was discretized with SBDM to reveal the effect of rivers on the sand cover ratio in the Maowusu (Mu Us) Sandy Land, northern China. The SBDM obtained more accurate information for the influence of rivers on the sand cover ratio than the results from Priori Knowledge discretization method. (2) Seven environmental factors were discretized using SBDM to detect potential associations between these factors and NDVI spatial pattern in Xinjiang, north-western China. Then we compared the q values from SBDM with the values from four commonly used one-dimensional discretization methods, demonstrating that for all considered factors, SBDM gets a larger q value than other methods. Collectively, SBDM offers a new way for data discretization that accurately reveals the relationship between controlling factors and response variables.

This preprint has been withdrawn.

Xiaoyu Meng et al.

Interactive discussion

Status: closed
Status: closed
AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment
Printer-friendly Version - Printer-friendly version Supplement - Supplement

Interactive discussion

Status: closed
Status: closed
AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment
Printer-friendly Version - Printer-friendly version Supplement - Supplement

Xiaoyu Meng et al.

Model code and software

SBDM X. Meng https://doi.org/10.5281/zenodo.1475892

Xiaoyu Meng et al.

Viewed

Total article views: 793 (including HTML, PDF, and XML)
HTML PDF XML Total Supplement BibTeX EndNote
507 250 36 793 43 44 41
  • HTML: 507
  • PDF: 250
  • XML: 36
  • Total: 793
  • Supplement: 43
  • BibTeX: 44
  • EndNote: 41
Views and downloads (calculated since 30 Nov 2018)
Cumulative views and downloads (calculated since 30 Nov 2018)

Viewed (geographical distribution)

Total article views: 670 (including HTML, PDF, and XML) Thereof 667 with geography defined and 3 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 26 Oct 2021
Download

This preprint has been withdrawn.