The scale and severity of forest disturbances across the globe are increasing due to climate change and human activities. Remote sensing analysis using time series data is a powerful approach for detecting large-scale forest disturbances and describing detailed forest dynamics. Various large-scale forest disturbance detection algorithms have been proposed, but most of them are only suitable for detecting high-magnitude forest disturbances (e.g., fire, harvest). Conversely, more continuous, subtle, and gradual lower-magnitude forest disturbances (e.g., thinning, pests, and diseases) have been subject to less focus. Deep learning (DL) can distinguish subtle differences in information within time series data, offering new opportunities to capture forest disturbances in a complete and detailed way. This study proposes an approach for analyzing forest dynamics across large areas and long time periods by combining DL time series classification and prior knowledge constraint. The approach consists of two stages: (1) an improved self-attention model used for time series classification to identify sequences with forest disturbance characteristics; (2) developed skip-disturbance recovery index (S-DRI) characterizing the temporal context, using prior knowledge constraint to identify forest disturbance years in time series with disturbance characteristics. In this study, the year of forest disturbances in five study areas located in the United States, Canada, and Poland from 2001 to 2020 was mapped. A total of 3082 manually interpreted test data with different disturbance causal agents (such as fire, harvest, conversion, hurricane, and pests) were sampled from five research areas for validation. Our approach was also evaluated against two forest disturbance benchmark datasets derived from LandTrendr and the Global Forest Change (GFC) dataset. The results demonstrate that our approach achieved an overall accuracy of 87.8%, surpassing the accuracy of LandTrendr (84.6%) and the Global Forest Change dataset (81.4%). Furthermore, our approach demonstrated lower omission rates (ranging from 10.0% to 67.4%) in detecting subtle to severe causal agents of forest disturbance, in comparison to LandTrendr (with a range of 18.0% to 81.6%) and GFC (with a range of 15.0% to 88.8%). This study, which involved mapping large-scale and long-term forest disturbance in multiple regions, revealed that our approach can be applied to new areas without a requirement for complex parameter adjustments. These results demonstrate the potential of our approach in generating comprehensive and detailed forest disturbance data, thus providing a new and effective method in this domain.

Dense time series of remote sensing images with high spatio-temporal resolution are critical for monitoring land surface dynamics in heterogeneous landscapes. Spatio-temporal fusion is an effective solution to obtaining such time series images. Many spatio-temporal fusion methods have been developed for producing high spatial resolution images at frequent intervals by blending fine spatial images and coarse spatial resolution images. Previous studies have revealed that the accuracy of fused images depends not only on the fusion algorithm, but also on the input image pairs being used. However, the impact of input images dates on the fusion accuracy for time series with different temporal variation patterns remains unknown. In this paper, the impact of input image pairs on the fusion accuracy for monotonic linear change (MLC), monotonic non-linear change (MNLC), and non-monotonic change (NMC) time periods were evaluated, respectively, and the optimal selection strategies of input image dates for different situations were proposed. The 16-day composited NDVI time series (i.e., Collection 6 MODIS NDVI product) were used to present the temporal variation patterns of land surfaces in the study areas. To obtain sufficient observation dates to evaluate the impact of input image pairs on the spatio-temporal fusion accuracy, we utilized the Harmonized Landsat-8 Sentinel-2 (HLS) data. The ESTARFM was selected as the spatio-temporal fusion method for this study. The results show that the impact of input image date on the accuracy of spatio-temporal fusion varies with the temporal variation patterns of the time periods being fused. For the MLC period, the fusion accuracy at the prediction date (PD) is linearly correlated to the time interval between the change date (CD) of the input image and the PD, but the impact of the input image date on the fusion accuracy at the PD is not very significant. For the MNLC period, the fusion accuracy at the PD is non-linearly correlated to the time interval between the CD and the PD, the impact of the time interval between the CD and the PD on the fusion accuracy is more significant for the MNLC than for the MLC periods. Given the similar change of time intervals between the CD and the PD, the increments of R² of fusion result for the MNLC is over ten times larger than those for the MLC. For the NMC period, a shorter time interval between the CD and the PD does not lead to higher fusion accuracies. On the contrary, it may lower the fusion accuracy. This study suggests that temporal variation patterns of the data must be taken into account when selecting optimal dates of input images in the fusion model.

Dynamic Habitat Indices (DHIs) are crucial for understanding species richness patterns and provide a powerful tool for large-scale biodiversity conservation research. DHIs summarize three key aspects of vegetation productivity: (a) cumulative annual productivity, (b) minimum productivity, and (c) intra-annual variability. While DHIs have demonstrated potential for predicting richness patterns globally and in tropical regions, their predictive ability across climate zones and using multiple sources of species richness data remains untested. We assess the feasibility of using DHIs to predict the richness of resident birds in China and explore approaches to improve model performance. We used (a) IUCN range maps of terrestrial resident birds in China, and (b) species distribution models (SDMs) to delineate bird richness patterns, classifying species into six habitat guilds: forest, shrubland, grassland, cropland, wetland, and all resident birds. We linked DHIs to richness in each guild and quantified their predictive power using three modeling approaches: linear (GLM) and non-linear models (GAM, Random Forest). We also recorded the relative importance of each DHI component in the models. Our results show that DHIs best predicted cropland bird richness (SDM-based richness, adjusted R² = 0.73), while for IUCN-based guilds, adjusted R² ranged from 0.68 to 0.71. The Random Forest model achieved the highest performance and interpretability. Among DHI components, cumulative DHI consistently played the most dominant role in predicting richness from both SDM and IUCN sources. DHIs effectively capture the link between energy availability and resident bird richness in China, demonstrating considerable potential for biodiversity assessment and conservation planning.

Towers often have complex structures and higher reflectances than the surrounding surfaces. To identify the uncertainty of the tower in tower-based albedo observations, a combined approach involving simulation and field observations was applied to three newly constructed towers in the Saihanba region. The simulation results demonstrate that tower presence influenced the albedo measurements, with the impact primarily controlled by tower reflectance rather than sensor height. When the tower-body albedo markedly differs from the background, systematic correction is necessary. Field validation indicates that for the towers in the Saihanba region, the contrast in albedo between the tower and the surrounding background is small, and both the simulated and observed albedos agree well with the ground measurements, with differences below 0.0075. By identifying when and how tower structures distort albedo observations, this study enhances confidence in ground-based validation and provides practical guidance for future tower design and material selection.

The global distribution and disturbance information of forest have strong impact on the change of Earth’s ecosystems. In the 1990s, the Eurasian continent forest cover an area of 182 million ha, accounting about 33.2% of the Eurasian continent land area. However, we lack a complete mapping of high-resolution forest disturbances in Eurasia. Remote sensing can regularly obtain forest cover data across expansive range. Therefore, a complete set of Landsat time-series-based forest disturbance detection method is constructed in this paper to map a 30-meter forest disturbance detection distribution map of Eurasian continent.In the construction of Landsat time series(LTS) data, the Landsat TM, ETM +, and OLI images of forest growth season were selected and synthesized into inter-annual time series over 35 years from 1986 to 2020. And the appropriate indices, NBR and NDVI, were selected as the input data for time series analysis. In time series analysis, the adaptive threshold of model learning is effectively applied in the process of extracting potential disturbance points, and the rich temporal information of LTS is fully mined to optimize and filter the disturbances.The LTS images and forest disturbance based on adaptive threshold model are used to map three decades of forest disturbances, including the characteristics of the disturbance, spatiotemporal distribution and disturbance frequency across Eurasian continent. The derived disturbance year maps revealed that the disturbed forest area is 237 million ha and 12.8% of Eurasia’s forest area. In order to validate the accuracy of the map, 10066 interpreted Landsat pixels, including 3932 disturbed samples and 6134 undisturbed samples, are selected as reference data. The overall accuracy of the disturbance map is 86.6%, with a commission error of 13.4% and an omission error of 9.4%. The results indicated that the LTS and adaptive threshold model can effectively support the mapping of forest disturbance in Eurasian continent.