Land cover is the physical material at the surface of the Earth. As the cause and result of global environmental change, land cover change (LCC) influences the global energy balance and biogeochemical cycles. Continuous and dynamic monitoring of global LC is urgently needed. Effective monitoring and comprehensive analysis of LCC at the global scale are rare. With the latest version of GLASS (Global Land Surface Satellite) CDRs (climate data records) from 1982 to 2015, we built the first record of 34-year-long annual dynamics of global land cover (GLASS-GLC) at 5 km resolution using the Google Earth Engine (GEE) platform. Compared to earlier global land cover (LC) products, GLASS-GLC is characterized by high consistency, more detail, and longer temporal coverage. The average overall accuracy for the 34 years each with seven classes, including cropland, forest, grassland, shrubland, tundra, barren land, and snow/ice, is 82.81 % based on 2431 test sample units. We implemented a systematic uncertainty analysis and carried out a comprehensive spatiotemporal pattern analysis. Significant changes at various scales were found, including barren land loss and cropland gain in the tropics, forest gain in the Northern Hemisphere, and grassland loss in Asia. A global quantitative analysis of human factors showed that the average human impact level in areas with significant LCC was about 25.49 %. The anthropogenic influence has a strong correlation with the noticeable vegetation gain, especially for forest. Based on GLASS-GLC, we can conduct long-term LCC analysis, improve our understanding of global environmental change, and mitigate its negative impact. GLASS-GLC will be further applied in Earth system modeling to facilitate research on global carbon and water cycling, vegetation dynamics, and climate change. The GLASS-GLC data set presented in this article is available at https://doi.org/10.1594/PANGAEA.913496 (Liu et al., 2020).

Urbanization is a substantial contributor to anthropogenic environmental change, and often occurs at a rapid pace that demands frequent and accurate monitoring. Time series of satellite imagery collected at fine spatial resolution using stable spectral bands over decades are most desirable for this purpose. In practice, however, temporal spectral variance arising from variations in atmospheric conditions, sensor calibration, cloud cover, and other factors complicates extraction of consistent information on changes in urban land cover. Moreover, the construction and application of effective training samples is time-consuming, especially at continental and global scales. Here, we propose a new framework for satellite-based mapping of urban areas based on transfer learning and deep learning techniques. We apply this method to Landsat observations collected during 1984–2016 and extract annual records of urban areas in four cities in the temperate zone (Beijing, New York, Melbourne, and Munich). The method is trained using observations of Beijing collected in 1999, and then used to map urban areas in all target cities for the entire 1984–2016 period. The method addresses two central challenges in long term detection of urban change: temporal spectral variance and a scarcity of training samples. First, we use a recurrent neural network to minimize seasonal urban spectral variance. Second, we introduce an automated transfer strategy to maximize information gain from limited training samples when applied to new target cities in similar climate zones. Compared with other state-of-the-art methods, our method achieved comparable or even better accuracy: the average change detection accuracy during 1984–2016 is 89% for Beijing, 94% for New York, 93% for Melbourne, and 89% for Munich, and the overall accuracy of single-year urban maps is approximately 96 ± 3% among the four target cities. The results demonstrate the practical potential and suitability of the proposed framework. The method is a promising tool for detecting urban change in massive remote sensing data sets with limited training data.

This study examines the suitability of 30 m Landsat Thematic Mapper (TM), 250 m time-series Moderate Resolution Imaging Spectrometer (MODIS) Enhanced Vegetation Index (EVI) and other auxiliary datasets for mapping forest extent in China at 30 m resolution circa 2010. We calculated numerous spectral features, EVI time series, and topographical features that are helpful for forest/non-forest distinction. In this research, extensive efforts have been made in developing training samples over difficult to map or complex regions. Scene by scene quality checking was done on the initial forest extent results and low quality results were refined until satisfactory. Based on the forest extent mask, we classified the forested area into 6 types (evergreen/deciduous broadleaf, evergreen/deciduous needleleaf, mixed forests, and bamboos). Accuracy assessment of our forest/non-forest classification using 2195 test sample units independent of the training sample indicates that the producer’s accuracy (PA) and user’s accuracy (UA) are 92.0% and 95.7%, respectively. According to this map, the total forested area in China was 164.90 million ha (Mha) circa 2010. It is close to the forest area of 7th National Forest Resource Inventory with the same definition of forest. The overall accuracy for the more detailed forest type classification is 72.7%.

Substantial progress has been made in the field of large-area land cover mapping as the spatial resolution of remotely sensed data increases. However, a significant amount of human power is still required to label images for training and testing purposes, especially in high-resolution (e.g., 3-m) land cover mapping. In this research, we propose a solution that can produce 3-m resolution land cover maps on a national scale without human efforts being involved. First, using the public 10-m resolution land cover maps as an imperfect training dataset, we propose a deep learning based approach that can effectively transfer the existing knowledge. Then, we improve the efficiency of our method through a network pruning process for national-scale land cover mapping. Our proposed method can take the state-of-the-art 10-m resolution land cover maps (with an accuracy of 81.24% for China) as the training data, enable a transferred learning process that can produce 3-m resolution land cover maps, and further improve the overall accuracy (OA) to 86.34% for China. We present detailed results obtained over three mega cities in China, to demonstrate the effectiveness of our proposed approach for 3-m resolution large-area land cover mapping.

Seasonal dynamic land cover maps could provide useful information to ecosystem, water-resource and climate modelers. However, they are rarely mapped more frequent than annually. Here, we propose an approach to map dynamic land cover types with frequently available satellite data. Landsat 8 data acquired from nine dates over Beijing within a one-year period were used to map seasonal land cover dynamics. A two-step procedure was performed for training sample collection to get better results. Sample sets were interpreted for each acquisition date of Landsat 8 image. We used the random forest classifier to realize the mapping. Nine sets of experiments were designed to incorporate different input features and use of spatial temporal information into the dynamic land cover classification. Land cover maps obtained with single-date data in the optical spectral region were used as benchmarks. Texture, NDVI and thermal infrared bands were added as new features for improvements. A Markov random field (MRF) model was applied to maintain the spatio-temporal consistency. Classifications with all features from all images were performed, and an MRF model was also applied to the results estimated with all features. The best overall accuracies achieved for each date ranged from 75.31% to 85.61%.