GLC30    Papers    Download    Online Tagging      简体中文  
1 2 3
BackGround           Product           Technology           Science

Information regarding land cover and its change over time is essential for a variety of societal needs, ranging from natural resources management, environmental studies, urban planning to sustainable development. Remote sensing has long been recognized as an effective tool for broad-scale land cover mapping. As a result, a number of land cover datasets at a global scale have been developed with resolution ranging from 300 m to 1 km, using coarse resolution satellite imagery such as AVHRR, MODIS and MERIS. Although these GLC data products have been widely used, their quality is far from satisfaction for many applications. Various researchers have highlighted the shortfalls of these datasets, e.g. the considerably low accuracies and low-level agreement amongst themselves. Consequently, the demand for new GLC products with improved spatial resolution and accuracy has been increasingly recognized by the remote sensing community e.g. the Group on Earth Observations (GEO) and the International Society for Photogrammetry and Remote Sensing (ISPRS).

With the long-term archive and free availability of Landsat and Landsat-like image data, the development of GLC data products at 30 m resolution has become feasible. Such products have been considered as a superior option for the next generation of GLC maps, since most significant human activities on the land system can be captured at this scale. During past two decades, the extraction of land cover information from Landsat-like imagery has been intensively studied, and a variety of automated and semi-automated methods/algorithms have been developed. These have been applied to a number of national and regional land cover mapping projects using Landsat imagery. For instance, a set of 30 m land cover data with 13 different classes was produced by MDA (MDA, 2014), which covers the USA and a large proportion of Africa and Asia.

Land cover mapping with 30 m resolution at a global scale is much more complex than that at national or regional scale due to a number of factors, including the availability of good-quality imagery covering the land surface of the entire earth (about 150 million km2) and the complex spectral and textual characterization of global landscapes. This makes the development of reliable 30 m GLC data products a very difficult task, as it requires a substantive level of technical innovation, as well as human and financial resources. This could be the reason why so far only global datasets with limited classes at this resolution have been reported. Of course, forest data is very important for various applications, however a 30 m GLC product with a more comprehensive set of land cover types is desirable for wider studies.

In 2010, China launched a GLC mapping project, and finally produced the 30 m GLC data product (GlobeLand30) with 10 classes for years 2000 and 2010 within a four year period.


GlobeLand30 refers to the land cover of the earth between latitude 80N to 80S. Presently, our website has provided the browsing and downloading services for GlobeLand30-2010 (GlobeLand30-2010) in baseline year of 2010. The brief information of this data product is as follows.

1.Images and Auxiliary Data

The images utilized for GlobeLand30 classification are multispectral images with 30 meters, including the TM5 and ETM + of America Land Resources Satellite (Landsat) and the multispectral images of China Environmental Disaster Alleviation Satellite (HJ-1). Besides multispectral images, plenty of Auxiliary data are also used in the process of data production such as sample collection and classification, etc. They mainly contains: the existing land cover data (global and regional), MODIS NDVI, global geographic information, global DEM, thematic data (global mangrove forest, wetland and glacier, etc.) and online resources (Google Earth, Bing Map, OpenStreetMap and Map World) and so on.

2. Classification Scheme

The classification system includes 10 land cover types, namely cultivated land, forest, grassland, shrubland, wetland,water bodies, tundra, artificial surfaces, bareland, permanent snow and ice. Please classification scheme is as follows:

1) Cultivated Land. Lands used for agriculture, horticulture and gardens, including paddy fields, irrigated and dry farmland, vegetation and fruit gardens, etc.

2) Forest. Lands covered with trees, with vegetation cover over 30%, including deciduous and coniferous forests, and sparse woodland with cover 10 - 30%, etc.

3) Grassland. Lands covered by natural grass with cover over 10%, etc.

4) Shrubland. Lands covered with shrubs with cover over 30%, including deciduous and evergreen shrubs, and desert steppe with cover over 10%, etc.

5) Water bodies. Water bodies in the land area, including river, lake, reservoir, fish pond, etc.

6) Wetland. Lands covered with wetland plants and water bodies, including inland marsh, lake marsh, river floodplain wetland, forest/shrub wetland, peat bogs, mangrove and salt marsh, etc.

7) Tundra. Lands covered by lichen, moss, hardy perennial herb and shrubs in the polar regions, including shrub tundra, herbaceous tundra, wet tundra and barren tundra, etc.

8) Artificial surfaces. Lands modified by human activities, including all kinds of habitation, industrial and mining area, transportation facilities, and interior urban green zones and water bodies, etc.

9) Bareland. Lands with vegetation cover lower than 10%, including desert, sandy fields, Gobi, bare rocks, saline and alkaline lands, etc.

10) Permanent snow and ice. Lands covered by permanent snow, glacier and icecap.

3. Data Composition and Format

GlobeLand30 data adopts raster data format for storage, with the non-destructive GeoTIFF compression format and the 256 indexed color pattern of the 8 Bit. The data consist of 5 parts, namely classification result file, coordinate information file, map setting file of classification image, metadata file and illustrative file.

4. Data Organization

GlobeLand30 data adopts WGS84 coordinate system, UTM projection, 6-degree zoning and the reference ellipsoid is WGS 84 ellipsoid. According to different latitude situations, 2 methods are adopted to organize the data tiles. Within the area of 60°N and 60°S, the data tile is implemented according to the size of 5° (latitude) * 6° (longitude); within the area of 60° to 80° degrees north and south of the equator, the data tile is implemented according to the size of 5° (latitude) * 12° (longitude) and the projection is conducted according to the central meridian of 6°zone with an odd number.

There are 853 data tiles in the whole world in total and the covering situation of specific picture is shown as Figure 1.

Figure 1 Data tile of GlobeLand30-2010

5. Accuracy Assessment

Nine types and over 150,000 test samples are evaluated in terms of accuracy assessment. The overall accuracy of GlobeLand30-2010 reaches 80.33%. The Kappa indicator is 0.75.

6. Contact Information

If you need a further understanding of the detailed information of the product, please refer to Product Manual ( If you have any question and suggestion, please contact Doctor Chen Lijun and his contact details are as follows:

Lijun Chen Ph.D

National Geomatics Center of China (NGCC)

28 Lianhuachi West Road, Beijing, 100830, China

Tel: +86-10-63880216

Cellphone: +86-13641011589



Producing global land cover with 30m resolution is a huge remote sensing project of great challenges, which encountered a good deal of technical problems to be solved. For example, how to realize the effective cover of global images within a limited time frame? How to establish and refine the classification workflow? How to enhance the operating efficiency? How to control the classification quality? And so forth.

In order to solve the above problems, our researchers have made a lot of effort to the researches on image processing, reference materials and knowledge integration, hybrid classification and the workflow for data quality control. A series of methods and algorithms have been approached to support the effectively generation of 30m global land cover data (GlobeLand30).

1.Geometric and radiometric restoration of classification images

The generation of global land cover needs to obtain global multispectral remote sensing images of 30m resolution in 2000 and 2010. However, the applicable image resources are limited at present, which requires to comprehensively utilize the multispectral remote sensing data of Landsat TM, ETM7 (SLC-off) and HJ-1A/B, etc. with certain deficiencies.

In order to guarantee these remote sensing images can be applied to the production of GlobeLand30, some methods in aspects of the strict geometric registration, missing data interpolation have been developed. These methods can effectively enhance the precision of geometric correction for satellite image with wide field angle, and have been widely applied in cloud decontamination, band interpolation of Landsat ETM + SLC-off data and many other processes. Finally, the effective global over of remote sensing image was achieved by using Landsat TM5, ETM + SLC - off and HJ-1A/B.

2.Hierarchic image classification method

Since the land cover situations in different regions are inconsistent with each other, and we cannot ensure all the images collected are of the same season, it is impossible to use a certain classification method to complete the production of global land cover data.

To achieve high-precision global land cover data, the POK classification technology was proposed by the comprehensive utilization of pixel classification, object-oriented classification and knowledge rules processing. By adopting the hierarchical classification strategy of POK, we can not only make an effective use of the advantages of all sorts of classification methods and knowledge rules of land cover, but also achieve the effective matching of classification technology process with classification images and auxiliary data, thus effectively solving the difficulty in extracting land cover data brought by the diversity of global land cover types and complexity of spectral textures. On this basis, a system supporting the production of GLC30 was established and utilized in our project.

3. A rule-based workflow for data quality control

The results generated by classification algorithms are often accompanied with low accuracy and fail to meet the requirement of classification indicator. Interactive operations are needed to reduce the classification error, so as to obtain reliable quality of land cover classification. Due to the complicated global situations, a large volume of data, and numerous personnel involved, it is of great importance to formulate unified and effective verification rules, and to develop an interactive platform.

In order to guarantee the verification efficiency, a rule-based workflow was developed to control the data quality. Firstly, we classified all kinds of knowledge into nature-based land cover knowledge, culture-based land cover knowledge and temporal-constraint land cover knowledge. On this basis, verification rules such as spatial coincidence, relationship coordination, time continuity etc. could be extracted and defined. Finally, a web-based verification system was developed by publishing all the data and functions required in the form of web service, which can not only solve the control the verification process but also enhance the working efficiency.

4. The integration of multisource and heterogeneous reference materials

The classification and verification processes involved a lot of multisource images and heterogeneous web service, such as remote sensing images, existing classification results, sample data, and on-line map services etc. These data and services are inconsistent with each other in coordinate reference, classification standard, data format and service interfaces. Thus a production environment with uniform data standard and compatible service interface needs to be established. After an in-deep research on web service technology, a web-based land cover data service platform has been developed by using data transformation and service integration methods. Finally, the sharing and integration of multisource and heterogeneous reference materials has been realized, thus providing the production and verification of land cover data with a unified data environment.

For more details, please see the reference [General Technologies for Global 30-meter Land Cover Remote Sensing Mapping]


Our earth is now supporting more than 7 billion people for their survival and development, which requires the forest, ocean and wet land system to maintain their balance and stability. As GlobeLand30 has the highest resolution in all the global land cover productions, we can utilize it to realize the spatial distribution of land cover and its change more accurately. It not only provides reliable support of essential data for the global change research and the development of earth system model, etc., but also has vital significance for revealing the change in global ecology, environment and resource brought by human beings, deeply analyzing the conflict between human and the earth, as well as scientifically making the global sustainable development. This can be reflected in four aspects as follows:

1.Global change

To cope with global changes, it is necessary to fully master the holding quantity of various resources contributing to climate warming, as well as its spatial distribution information. Based on GlobeLand30, the area, spatial distribution and changes of the land coverage type closed related to the health of the earth can be described exactly, and both positive and negative influences that human activities have brought on global environment, ecology, and so on can be evaluated. For example, the GlobeLand30 can be used to take an accurate evaluation on global carbon pool source, carbon sink source, emission zone of greenhouse gas, etc.; what's more, it can also be applied to evaluate the effect of each country in performing the Convention on Tackling Climate Change.

2.Sustainable development

Maintaining global sustainable development is the central challenge for human society. We can obtain the information of global land cover from the GlobeLand30 and combine it with the data of population, society and economy to analyze and assess the natural resource endowment of each country and the key factors which have led to the conflict between human and the earth, so as to provide decision reference frame for the global food security, desertification control, wetland conservation and biodiversity conservation.

3.Serve for global scientific management

All kinds of different information are needed for global scientific management, among which the land coverage is very important. Land coverage status quo of different periods obtained from the GlobeLand30 data is capable of presenting the variation trend and variable quantity of various land coverage, which could offer decision reference for adjusting the global policy of governance and assessing governance effects in an accurate way.

4.Earth system modeling/simulation

Scientific acknowledge of the earth has always been a striving direction in the field of geoscience around the world. All kinds of earth models need the data of land cover as the basic input parameter prior to conducting model simulation and prediction. On account of higher resolution and accuracy, GlobeLand30 can be widely used in the research work with different measurements (global model, zone model, etc.) or different models (atmospherically model, hydrological model, land model, etc.) in a bid to reduce the uncertainty of various model systems.