Comparison of Modeling Grassland Degradation with and without Considering Localized Spatial Associations in Vegetation Changing Patterns

Abstract

Abstract:Grassland ecosystems worldwide are confronted with degradation. It is of great importance to understand long-term trajectory patterns of grassland vegetation by advanced analytical models. This study proposes a new approach called a binary logistic regression model with neighborhood interactions, or BLR-NIs, which is based on binary logistic regression (BLR), but fully considers the spatio-temporally localized spatial associations or characterization of neighborhood interactions (NIs) in the patterns of grassland vegetation. The BLR-NIs model was applied to a modeled vegetation degradation of grasslands in the Xilin river basin, Inner Mongolia, China. Residual trend analysis on the normalized difference vegetation index (RESTREND-NDVI), which excluded the climatic impact on vegetation dynamics, was adopted as a preprocessing step to derive three human-induced trajectory patterns (vegetation degradation, vegetation recovery, and no significant change in vegetation) during two consecutive periods, T 1 (2000)(2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008) and T 2 (2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015). Human activities, including livestock grazing intensity and transportation accessibility measured by road network density, were included as explanatory variables for vegetation degradation, which was defined for locations if vegetation recovery or no significant change in vegetation in T 1 and vegetation degradation in T 2 were observed. Our work compared the results of BLR-NIs and the traditional BLR model that did not consider NIs. The study showed that: (1) both grazing intensity and road density had a positive correlation to vegetation degradation based on the traditional BLR model; (2) only road density was found to positively correlate to vegetation degradation by the BLR-NIs model; NIs appeared to be critical factors to predict vegetation degradation; and (3) including NIs in the BLR model improved the model performance substantially. The study provided evidence for the importance of including localized spatial associations between the trajectory patterns for mapping vegetation degradation, which has practical implications for designing management policies to counterpart grassland degradation in arid and semi-arid areas.

Using Shapley additive explanations to interpret extreme gradient boosting predictions of grassland degradation in Xilingol, China

“…However, the SHAP values indicate that when sheep density decreases, the probability of grassland degradation increases. Overgrazing has been the dominant driver for grassland degradation on the Mongolian plateau before, which has changed the grassland ecosystem significantly towards lower grass coverage (Nkonya et al, 2016;Wang et al, 2017). However, there is recent evidence that this causal relationship has changed.…”

Assessment of Human-Related Driving Forces for Reduced Carbon Uptake Using Neighborhood Analysis and Geographically Weighted Regression: A Case Study in the Grassland of Inner Mongolia, China

“…RCU was assessed for each year during the time window. The explanatory variables grazing and population were compiled from yearbooks, and the road network was extracted from Open Street Map (OSM) [26]. All resultant raster layers were resampled to 1 km × 1 km if they were not at that resolution.…”

Profiling Human-Induced Vegetation Change in the Horqin Sandy Land of China Using Time Series Datasets

“…In step 2, candidates that could not simultaneously satisfy the following two conditions lost their qualifications: (1) there must be a significant correlation of ∑ NDV I trajectories between x and the candidates; (2) the differential of mean value of the ∑ NDV I trajectory between x and candidates must be below 3δ. The 3δ method has been used to detect land cover change in some studies [14,56], and helped determine homogeneous spatial neighborhoods in this study. All of the candidates selected by this step constituted the homogeneous spatial neighborhood of x.…”

“…The complicated interaction between human and vegetation is critical for almost all of the research studies that have been done about earth system models [12]. Discriminating human-induced vegetation change has always been a challenging issue [1,13,14], even considering the long-term perspective of palaeoecology [15,16].…”