Imagen de Portada

Assessing the Accuracy and Consistency of Cropland Datasets and Their Influencing Factors on the Tibetan Plateau

Guardado en:
Publicado en:Remote Sensing vol. 17, no. 11 (2025), p. 1866
Autor principal: Zhang Fuyao
Otros Autores: Wang, Xue, Liangjie, Xin, Li, Xiubin
Publicado:
MDPI AG
Materias:
Tibetan Plateau
China
Vegetation
Food security
Accuracy
Agricultural land
Agricultural production
Datasets
Elevation
Topography
Remote sensing
Remote monitoring
Land use
Machine learning
Land cover
Climate change
Land surveys
Precipitation
Low altitude
Climatic conditions
Vegetation index
Temperature
Cloud computing
Classification
Regions
Impact analysis
Agricultural management
Crops
Plastic pollution
Acceso en línea:Citation/Abstract
Full Text + Graphics
Full Text - PDF
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
Resumen:With advancements in cloud computing and machine learning algorithms, an increasing number of cropland datasets have been developed, including the China land-cover dataset (CLCD) and GlobeLand30 (GLC). The unique climatic conditions of the Tibetan Plateau (TP) introduce significant differences and uncertainties to these datasets. Here, we used a quantitative and visual integrated assessment approach to assess the accuracy and spatial consistency of five cropland datasets around 2020 in the TP, namely the CLCD, GLC30, land-use remote sensing monitoring dataset in China (CNLUCC), Global Land Analysis and Discovery (GLAD), and global land-cover product with a fine classification system (GLC_FCS). We analyzed the impact of terrain, climate, population, and vegetation indices on cropland spatial consistency using structural equation modeling (SEM). In this study, the GLAD cropland area had the highest fit with the national land survey (R2 = 0.88). County-level analysis revealed that the CLCD and GLC_FCS underestimated cropland areas in high-elevation counties, whereas the GLC and CNLUCC tended to overestimate cropland areas on the TP. Considering overall accuracy, GLC and GLAD performed the best with scores of 0.76 and 0.75, respectively. In contrast, CLCD (0.640), GLC_FCS (0.640), and CNLUCC (0.620) exhibited poor overall accuracy. This study highlights the significantly low spatial consistency of croplands on the TP, with only 10.60% consistency in high and complete agreement. The results showed substantial differences in spatial accuracy among zones, with relatively higher consistency observed in low-altitude zones and notably poorer accuracy in zones with sparse or fragmented cropland. The SEM results indicated that elevation and slope directly influenced cropland consistency, whereas temperature and precipitation indirectly affected cropland consistency by influencing vegetation indices. This study provides a valuable reference for implementing cropland datasets and future cropland mapping studies on the TP region.
ISSN:2072-4292
DOI:10.3390/rs17111866
Fuente:Advanced Technologies & Aerospace Database