Optimization of Rocky Desertification Classification Model Based on Vegetation Type and Seasonal Characteristic

Abstract

Building a high-precision, stable, and universal automatic extraction model of the rocky desertification information is the premise for exploring the spatiotemporal evolution of rocky desertification. Taking Guizhou province as the research area and based on MODIS and continuous forest inventory data in China, we used a machine learning algorithm to build a rocky desertification model with bedrock exposure rate, temperature difference, humidity, and other characteristic factors and considered improving the model accuracy from the spatial and temporal dimensions. The results showed the following: (1) The supervised classification method was used to build a rocky desertification model, and the logical model, RF model, and SVM model were constructed separately. The accuracies of the models were 73.8%, 78.2%, and 80.6%, respectively, and the kappa coefficients were 0.61, 0.672, and 0.707, respectively. SVM performed the best. (2) Vegetation types and vegetation seasonal phases are closely related to rocky desertification. After combining them, the model accuracy and kappa coefficient improved to 91.1% and 0.861. (3) The spatial distribution characteristics of rocky desertification in Guizhou are obvious, showing a pattern of being heavy in the west, light in the east, heavy in the south, and light in the north. Rocky desertification has continuously increased from 2001 to 2019. In conclusion, combining the vertical spatial structure of vegetation and the differences in seasonal phase is an effective method to improve the modeling accuracy of rocky desertification, and the SVM model has the highest rocky desertification classification accuracy. The research results provide data support for exploring the spatiotemporal evolution pattern of rocky desertification in Guizhou.

Dynamic Monitoring of Desertification in Ningdong Based on Landsat Images and Machine Learning

“…It is not easily affected by factors such as noise variables, feature correlations, etc. RF and SVM algorithms are commonly used in remote sensing monitoring research, and their good performance has been verified in many studies [25,61,93,94]. They also showed a good recognition effect in this study.…”

Spatiotemporal Evolution Analysis and Future Scenario Prediction of Rocky Desertification in a Subtropical Karst Region

“…Thanks to the support of sufficient sample data, the method of [51], in this paper, obtained remote sensing inversion results of rocky desertification level in 2020 bearing similarity to the real results, providing a reliable data basis for CA-Markov prediction of future prospects. Although there were errors within a certain allowable range, the results can still be considered reliable.…”

“…Using the MODIS data set and National Forest Continuous Inventory data (NFCI) on the Google Earth Engine (GEE) platform, and referring to previous research methods, we constructed rocky desertification maps for different periods [48][49][50]. In the National Forest Continuous Inventory data (NFCI) of Guizhou, rocky desertification is divided into five categories, coded as 00, 10, 21, 22, and 23-24, representing NRD, PRD, LRD, MRD, and SRD, respectively [51]. We obtained rocky desertification level maps for Guizhou Prov- The state of rocky desertification in Guizhou presents a distribution pattern of "heavy in the west, light in the east, heavy in the south, and light in the north".…”

“…First, the rocky desertification level (NRD, PRD, LRD, MRD, SRD) map was obtained from the previous research [51], where the multiple line model (ML), random forest model (RF), and support vector machine model (SVM) were compared to the constructed rocky desertification classification model. The overall accuracies (OAs) of these models were 73.8%, 78.2%, and 80.6%, respectively.…”

Desertificação: uma visão geral dos processos e conceitos, fundamentados em aplicação de índices orbitais através do sensoriamento remoto

“…Em diversos estudos foram utilizados indicadores como albedo, NDVI e SAVI para se identificar o grau de desertificação em que as regiões afetadas se encontravam, mas devido ao grau de processos de desertificação, diferentes texturas de solo (Wei et al, 2018), torna-se inviabilizando muitas vezes a utilização desses indicadores. Dessa forma o indicador utilizado para identificar a degradação do solo em áreas desertificadas é o TGSI (Topsoil Grain Size Index) (Gebru et al, 2021;Qian et al, 2021). Lamchin et al (2017) descobriram que as maiores correlações estavam entre TGSI e albedo em todos os níveis de desertificação.…”