Using principal component analysis and annual seasonal trend analysis to assess karst rocky desertification in southwestern China

Abstract

Increasing exploitation of karst resources is causing severe environmental degradation because of the fragility and vulnerability of karst areas. By integrating principal component analysis (PCA) with annual seasonal trend analysis (ASTA), this study assessed karst rocky desertification (KRD) within a spatial context. We first produced fractional vegetation cover (FVC) data from a moderate-resolution imaging spectroradiometer normalized difference vegetation index using a dimidiate pixel model. Then, we generated three main components of the annual FVC data using PCA. Subsequently, we generated the slope image of the annual seasonal trends of FVC using median trend analysis. Finally, we combined the three PCA components and annual seasonal trends of FVC with the incidence of KRD for each type of carbonate rock to classify KRD into one of four categories based on K-means cluster analysis: high, moderate, low, and none. The results of accuracy assessments indicated that this combination approach produced greater accuracy and more reasonable KRD mapping than the average FVC based on the vegetation coverage standard. The KRD map for 2010 indicated that the total area of KRD was 78.76 × 10 3 km 2 , which constitutes about 4.06% of the eight southwest provinces of China. The largest KRD areas were found in Yunnan province. The combined PCA and ASTA approach was demonstrated to be an easily implemented, robust, and flexible method for the mapping and assessment of KRD, which can be used to enhance regional KRD management schemes or to address assessment of other environmental issues.

A novel‐optimal monitoring model of rocky desertification based on feature space models with typical surface parameters derived from LANDSAT_8 OLI

“…Most previous studies on the detection of the spatial distribution of rocky desertification have used single indices, a comprehensive index, or image classification methods (Zhang et al, 2017;Zhang et al, 2021). However, the process of rocky desertification is influenced by various types of factors and their interactions, including vegetation, geology, geomorphology, climate change, and human activities (Guo, Zang, Yang, Huang, et al, 2020).…”

“…Among all the vegetation indices, the NDVI had better applicability to indicate the vegetation coverage and growth. The MSAVI had the worst monitoring precision of rocky desertification with 77.35%, which could be attributed to the fact that this index could better reflect the vegetation information in zones with sparse vegetation and more exposed bare soil Zhang et al, 2017;Zong et al, 2014). However, in the severe karst area, the bare rock was one of the dominant surface compositions, which would lead to worse monitoring results of vegetation conditions.…”

Analysis of Landsat-8 OLI Imagery for Estimating Exposed Bedrock Fractions in Typical Karst Regions of Southwest China Using a Karst Bare-Rock Index

“…In future studies, GEE will be a very useful tool to effectively mask the clouds/shadows on satellite images and derive long time series of annual/seasonal composite images. Finally, due to the fact that the high-quality sampling data are highly critical for the model establishment [60,61], steep slopes and rugged landscapes lead to very difficult access to the collection of sufficient ground truth training and validation data for modeling [20,62,63]. Therefore, how to develop a more optimized estimation model using limited training data remains a challenge in the next study.…”

“…The karst ecosystem in southwest China is famous for its high degree of landscape heterogeneity and broken surface [19]. Consequently, land covers are often mixtures of several types in karst regions [20], making it quite difficult to accurately extract the main land-surface symptoms of KRD. Vegetation indices and spectral mixture analysis (SMA) have been widely used to extract land cover fractions applied on a subpixel scale by remote sensing [21][22][23].…”

A stratified environmental reference system for better understanding of the relationship between remote sensing observations and ground monitoring of karst rocky desertification

“…The normalized difference vegetation index (NDVI), a key product from MODIS or other remote sensing data, has a high positive correlation with vegetation coverage, suggesting an effective basis for KRD mapping (Li & Wu, 2015; Zhang et al, 2017). One case study showed that as KRD intensity decreases, NDVI increases (Chen et al, 2014; Guo et al, 2021).…”