Collaborative validation of GlobeLand30: Methodology and practices

Abstract

30-m Global Land Cover (GLC) data products permit the detection of land cover changes at the scale of most human land activities, and are therefore used as fundamental information for sustainable development, environmental change studies, and many other societal benefit areas. In the past few years, increasing efforts have been devoted to the accuracy assessment of GlobeLand30 and other finer-resolution GLC data products. However, most of them were conducted either within a limited percentage of map sheets selected from a global scale or in some individual countries (areas), and there are still many areas where the uncertainty of 30-m resolution GLC data products remains to be validated and documented. In order to promote a comprehensive and collaborative validation of 30-m GLC data products, the GEO Global Land Cover Community Activity had organized a project from 2015 to 2017, to examine and explore its major problems, including the lack of international agreed validation guidelines and on-line tools for facilitating collaborative validation activities. With the joint effort of experts and users from 30 GEO member countries or participating organizations, a technical specification for 30m GLC validation was developed based on the findings and experiences. An on-line validation tool, GLCVal, was developed by integrating land cover validation procedures with the service computing technologies. About 20 countries (regions) have completed the accuracy assessment of GlobeLand30 for their territories with the guidance of the technical specification and the support of GLCVal.

Human expansion into Asian highlands in the 21st Century and its effects

“…Globeland30 datasets cover 10 land cover types: cultivated land, artificial surface, forest, grassland, shrubland, wetland, water, tundra, bare land, glaciers and permanent snow. The accuracy of Globeland30 was assessed using the landscape shape index-based sampling model and more than 230 thousand samples, and the overall accuracy and kappa coefficient, respectively, were 85.72% and 0.82 for the 2020 version and about 83.50% and 0.78 for the previous versions 45 , 46 , which meets the accuracy requirement of land cover change analysis 47 . Consequently, the Globeland30 products have been extensively used for a variety of applications 48 , 49 .…”

“…According to the classification system of GlobeLand30, cultivated land denotes the land used for planting crops, including paddy fields, irrigated dry land, rain-fed dry land, vegetable land, pasture land, greenhouse land and land mainly for planting crops with fruit trees and other economic trees and economic crops, such as tea and coffee gardens 46 . Artificial surface is impervious surface formed by artificial construction activities, including various residential settlements, such as towns, industrial and mining areas and transportation facilities, but excluding contiguous green spaces and water bodies within construction lands 46 . Therefore, based on the above definitions, the expansions of cultivated lands and artificial surfaces cover almost all intensive human activities that may occur in the highlands.…”

Consistency Assessments of the Land Cover Products on the Tibetan Plateau

“…The confusion matrix, based on land cover types, is still a widely used metric to assess product accuracy [42], [43]. It includes the overall accuracy (OA), producer accuracy (PA), user's accuracy (UA) and kappa coefficient to assess the accuracy of remote images [44], [45].…”

Human Expansion-Induced Biodiversity Crisis over Asia from 2000 to 2020

“…Globeland30 datasets were developed by combining several multispectral data sources, including Landsat Thematic Mapper, Enhanced Thematic Mapper Plus, Chinese HJ-1 images, and Gaofen multispectral images (2020 version), and they have been extensively used for various areas [70] . The classification was performed using a split-and-merge strategy, Pixel–Object–Knowledge approach was applied to classify each land cover type, and a knowledge-based interactive verification procedure was applied to improve the mapping accuracy [ 71 , 72 ]. The accuracy of the Globeland30 dataset was assessed using the landscape shape index-based sampling model and more than 230,000 samples, and the overall accuracy and kappa coefficient, respectively, were 85.7% and 0.82 for the 2020 version and about 83.5% and 0.78 for the previous versions [ 72 ], which meet the accuracy requirements of land cover change analysis, especially for large-scale applications [ 46 , 47 ].…”