查看更多>>摘要：Tidal wetlands, the global hotspots of biodiversity and carbon stocks, are currently experiencing widespread modifications in their composition due to human disturbances and changing climate. Accurate mapping of tidal wetland composition is crucial and urgently required for the conservation and management of coastal ecosystem, as well as for maximizing their associated services. However, remote sensing of tidal wetlands is still challenge due to periodic tidal fluctuations, frequent cloud cover, and similar spectral characteristics with terrestrial land cover types. Previous approaches to mapping the tidal wetlands have been restricted to small study regions or have focused on an individual tidal wetland type, thus limiting their ability to consistently monitor the composition of tidal wetlands over large geographic extents. To address the above issues, we proposed a novel algorithm on Google Earth Engine, called Multi-class Tidal Wetland Mapping by integrating Tide-level and Phenological features (MTWM-TP), to simultaneously map mangroves, salt marshes and tidal flats for specifying large-scale tidal wetland composition. The MTWM-TP algorithm firstly generates several noise-free composite images with different tide levels and phenological stages and then concatenates them into a random forest classifier for further classification. The usage of tide-level and phenological features eliminates inland landscapes and help to distinguish deciduous salt marshes and evergreen mangroves, leading to a statistically significant improvement in accuracy. We applied the algorithm to 10,274 Sentinel-2 images of East Asia and derived a 10-m resolution multi-class tidal wetland map with an overall accuracy of 97.02% at a sub-continental scale. We found that tidal wetlands occupied 1,308,241 ha of areas in East Asia in 2020, of which 89.12% were tidal flats, 9.39% were salt marshes, and only 1.49% were mangroves. This spatially explicit map of tidal wetland composition will provide valuable guidance for coastal biodiversity protection and blue carbon restoration. In addition, the proposed MTWM-TP algorithm can serve as a reliable means for monitoring sub -continental-or larger-scale tidal wetland composition more broadly.
查看更多>>摘要：Spatially continuous estimates of forest canopy height at national to global scales are critical for quantifying forest carbon storage, understanding forest ecosystem processes, and developing forest management and restoration policies to mitigate global climate change. Spaceborne light detection and ranging (lidar) platforms, especially the Global Ecosystem Dynamics Investigation (GEDI) and Ice, Cloud, and land Elevation Satellite-2 (ICESat-2) Advanced Topographic Laser Altimeter System (ATLAS), can measure forest canopy height in discrete footprints globally. Their coverage provides a promising data source for national to global-scale forest canopy height estimates. However, previous studies usually used a regression-based approach to develop spatially continuous forest canopy height distribution through the aid of optical images, which cannot take full advantage of the dense spaceborne lidar footprints and may still suffer from the saturation effect of optical images. In this study, we developed a novel neural network guided interpolation (NNGI) method to map forest canopy height by fusing GEDI, ICESat-2 ATLAS, and Sentinel-2 images. To evaluate the performance of the proposed NNGI method, we generated a 30-m forest canopy height product of China for the year 2019. More than 140 km2 drone-lidar data were collected across the country to train and validate the NNGI method. The results showed that the average forest canopy height of China is 15.90 m with a standard deviation of 5.77 m. We evaluated the interpolated forest canopy height product of China by over 1,100,000 GEDI validation footprints (R2 = 0.55, RMSE = 5.32 m), about 33 km2 drone-lidar validation data (R2 = 0.58, RMSE = 4.93 m), and over 59,000 field plot measurements (R2 = 0.60, RMSE = 4.88 m). Benefiting from the interpolation-based mapping strategy, the resulting product had almost no saturation effect in areas with tall forest canopies. The high mapping accuracy demonstrates the feasibility of the proposed NNGI method for monitoring spatially continuous forest canopy height at national to global scales by integrating multi-platform spaceborne lidar data and optical images, enabling opportunities to provide more accurate quantification of terrestrial carbon storage and better understanding of forest ecosystem processes.
查看更多>>摘要：Recently, the frequency of natural and environmental disasters has increased significantly, causing constant changes on the Earth's surface. Synthetic Aperture Radar (SAR) data have been proved to be useful for operational change monitoring tasks. The multiscale framework presented in this paper aims at detecting and analyzing changes using SAR image time series composed of large-size images. Spatio-temporal changes are initially detected at the subimage scale analysis stage to determine regions and image acquisition dates related to the change occurrence. Detailed changes are then identified at the pixel scale analysis stage between selected acquisitions at each recognized region. This framework was used for flood monitoring over a large area along the central coast of Vietnam (from Thua Thien Hue province to Quang Nam province). We exploited a Sentinel-1 image time series acquired during two rainy seasons and typhoon seasons in the Western Pacific (from September to December of the two years 2017 and 2018). The proposed framework detected flooded areas with a high overall accuracy of 90.4% and could analyze different types of changes that occurred in this time series, i.e. dirac, periodic, chaotic changes, and temporal stability.
查看更多>>摘要：Snowpack on sea ice can adjust changes in sea ice conditions and plays a vital role in the Earth's climate system. Snow depth, an important parameter of snowpack, is a necessary variable for retrieving sea ice thicknesses based on satellite altimeter data. Here, regression analysis (RA) is used to determine the best gradient ratio (GR) combination of brightness temperatures for estimating snow depths, and the RA model is proposed. Based on the RA model, one additional deep learning model is built, namely, the 5-variable long short-term memory (5VLSTM) model (or the RA-5VLSTM model). Meanwhile, an additional neural network model is built for comparisons, namely, the 5-variable genetic wavelet neural network (5VGWNN) model (or the RA-5VGWNN model). Using Operation IceBridge (OIB) and ice mass balance buoy (IMB) data, these three models, plus three existing algorithms, are compared to assess their performances in estimating snow depth. The results show that the RA-5VLSTM model performed pretty well among the six algorithms, with an RMSE of 7.16 cm. The RA5VLSTM model, a robust approach, was less influenced by the uncertainty in the input data. From January to April during 2012-2020, the average monthly snow depth in the Beaufort Sea and the Chukchi Sea mainly showed a downward trend, while an upward trend was observed in the Central Arctic in most months. Variations in snow depth in the Central Arctic were mainly affected by the autumn 2-m air temperature (T2m) and the sea surface temperature (SST). Variations in snow depth in the Chukchi Sea were mainly affected by the autumn sea ice velocity.
查看更多>>摘要：Sinkhole activity in Florida is a major hazard for people and property. Its increasing frequency is often related to an accelerated use of ground-water and land resources in the region. In this work, we use a combined approach of radar interferometry and spatial clustering analysis over three selected sites in West-Central Florida to identify localized deformation that may be caused by sinkhole activity. The region of West-Central Florida is a densely active sinkhole region, where sinkholes tend to be small and land cover is mixed resulting in variable interferometric coherence that complicates Interferometric Synthetic Aperture Radar (InSAR) surveys. In this work, we present a combined methodology implementing InSAR and a Density-Based Spatial Clustering Analysis (DBSCAN) algorithm to detect unknown sinkhole activity and to alert to possible precursors of sinkhole collapse. The data used for the study consist of acquisitions from three TerraSAR-X frames covering time spans of similar to 1.7 and 2.5 years with spatial resolutions ranging from 25 cm up to 1 m. We applied the Persistent Scatterer Interferometry (PSI) technique using the Stanford Method for Persistent Scatterers (StaMPS) and confirmed the observed deformation signals by also processing the data using the SAR PROcessing tool (SARPROZ). Results show several areas of localized subsidence, from which the cluster with highest rates for each site was selected for detailed inspection. Locations of selected clusters were found in buildings with sizes ranging from 300 m(2) to nearly 2000 m(2), with subsidence trends ranging from -3 to -6 mm/yr. Results were compared with in-situ observations such as ground penetrating radar (GPR), electrical resistivity tomography (ERT) surveys, visual structural inspection and public county archive documents to help as ground truthing; subsiding locations were found to be related to sinkhole presence or development.
查看更多>>摘要：Floods are causing massive losses of crops and agricultural infrastructures in many regions across the globe. During the 2018/2019 agricultural year, heavy rains from Cyclone Idai caused flooding in Central Mozambique and had the greatest impact on Sofala Province. The main objectives of this study are to map the flooding durations, evaluate how long crops survived the floods, and analyse the dynamics of the affected crops and their recovery following various flooding durations using multi-source satellite data. Our results indicate that Otsu method-based flooding mapping provides reliable flood extents and durations with an overall accuracy higher than 90%, which facilitates the assessment of how long crops can survive floods and their recovery progress. Croplands in both Buzi and Tica administrative units were the most severely impacted among all the regions in Sofala Province, with the largest flooded cropland extent at 23,101.1 ha in Buzi on 20 March 2019 and the most prolonged flooding duration of more than 42 days in Tica and Mafambisse. Major summer crops, including maize and rice, could survive when the fields were inundated for up to 12 days, while all crops died when the flooding duration was longer than 24 days. The recovery of surviving crops to pre-flooding status took a much longer time, from approximately 20 days to as long as one month after flooding. The findings presented herein can assist decision making in developing countries or remote regions for flood monitoring, mitigation and damage assessment.
查看更多>>摘要：Digital elevation models (DEMs) contain some of the most important data for providing terrain information and supporting environmental analyses. However, the applications of DEMs are significantly limited by data voids, which are commonly found in regions with rugged terrain. We propose a novel deep learning-based strategy called a topographic knowledge-constrained conditional generative adversarial network (TKCGAN) to fill data voids in DEMs. Shuttle Radar Topography Mission (SRTM) data with spatial resolutions of 3 and 1 arc-seconds are used in experiments to demonstrate the applicability of the TKCGAN. Qualitative topographic knowledge of valleys and ridges is transformed into new loss functions that can be applied in deep learning-based algorithms and constrain the training process. The results show that the TKCGAN outperforms other common methods in filling voids and improves the elevation and surface slope accuracy of the reconstruction results. The performance of the TKCGAN is stable in the test areas and reduces the error in the regions with medium and high surface slopes. Furthermore, the analysis of profiles indicates that the TKCGAN achieves better performance according to a visual inspection and quantitative comparison. In addition, the proposed strategy can be applied to DEMs with different resolutions. This work is an endeavour to transform topographic knowledge into computer-processable rules and benefits future research related to terrain reconstruction and modelling.
查看更多>>摘要：Water-leaving albedo (alpha w), defined as the ratio of water-leaving irradiance to downwelling irradiance just above the surface, is a major component of ocean surface albedo (alpha) but has long been ignored or underrepresented. A semi-analytical scheme based on inherent optical properties (IOPs), termed IOPs-alpha w, is proposed in this study to estimate spectral alpha w(lambda) from ocean color measurements. Evaluations with numerical simulations of radiative transfer show that IOPs-alpha w outperforms the conventional scheme based on chlorophyll-a (Chl) concentration. The median absolute percentage difference (MAPD) of derived alpha w(lambda) from IOPs-alpha w is generally less than 3% in the blue-green spectral domain, in comparison to MAPD of over 40% for estimated alpha w(lambda) from the Chl-based scheme. IOPs-alpha w is later implemented to monthly composite data of the Visible Infrared Imaging Radiometer Suite (VIIRS), where reasonable spatial distributions and seasonal patterns of alpha w(lambda) are obtained. In particular, broadband alpha w in the visible domain, termed alpha w_VIS, obtained via IOPs-alpha w is over 50% higher than the previous estimation by the Chl-based scheme in most oceanic waters. Furthermore, this study concludes that alpha w_VIS could contribute up to 20% to alpha in oceanic waters under low solar-zenith angles. Thus, we suggest that neither the spatial variability of alpha w_VIS nor the contribution of alpha w_VIS to alpha shall be neglected, and it is necessary to incorporate IOPs-alpha w into current parameterizations of alpha in coupled ocean-atmosphere and climate models.
查看更多>>摘要：Monitoring agricultural systems becomes increasingly important in the context of global challenges like climate change, biodiversity loss, population growth, and the rising demand for agricultural products. High-resolution, national-scale maps of agricultural land are needed to develop strategies for future sustainable agriculture. However, the characterization of agricultural land cover over large areas and for multiple years remains challenging due to the locally diverse and temporally variable characteristics of cultivated land. We here propose a workflow for generating national agricultural land cover maps on a yearly basis that accounts for varying environmental conditions. We tested the approach by mapping 24 agricultural land cover classes in Germany for the three years 2017, 2018, and 2019, in which the meteorological conditions strongly differed. We used a random forest classifier and dense time series data from Sentinel-2 and Landsat 8 in combination with monthly Sentinel-1 composites and environmental data and evaluated the relative importance of optical, radar, and environmental data. Our results show high overall accuracy and plausible class accuracies for the most dominant crop types across different years despite the strong inter-annual meteorological variability and the presence of drought and nondrought years. The maps show high spatial consistency and good delineation of field parcels. Combining optical, SAR, and environmental data increased overall accuracies by 6% to 10% compared to single sensor approaches, in which optical data outperformed SAR. Overall accuracy ranged between 78% and 80%, and the mapped areas aligned well with agricultural statistics at the regional and national level. Based on the multi-year dataset we mapped major crop sequences of cereals and leaf crops. Most crop sequences were dominated by winter cereals followed by summer cereals. Monocultures of summer cereals were mainly revealed in the Northwest of Germany. We showcased that high spatial and thematic detail in combination with annual mapping will stimulate research on crop cycles and studies to assess the impact of environmental policies on management decisions. Our results demonstrate the capabilities of integrated optical time series and SAR data in combination with variables describing local and seasonal environmental conditions for annual large-area crop type mapping.
查看更多>>摘要：Spartina alterniflora is an aggressive invasive plant spreading along the coastal China, spanning a latitudinal range of 20 degrees N-39 degrees N, and its invasion resulted in dramatic decline in both native plant diversity and ecosystem functioning. Phenology of S. alterniflora saltmarshes is a critical feature to elucidate the invasion dynamics over geographical regions but has not been well understood yet. In this study, we examined the variation of S. alterniflora saltmarsh phenology across coastal China during 2018-2020 by using time series Landsat 7/8 and Sentinel-2 images. Combined Landsat 7/8 and Sentinel-2 images provided more good-quality observations in a year, which could facilitate phenological retrieval. We applied and assessed three widely used phenology retrieval methods (i.e., NDVI-based pixel-specific statistical threshold, NDVI-based double logistic mathematical equation, and LSWI-based biological threshold) for retrieving the start and end of season (SOS and EOS) as well as the length of season (LOS) of S. alterniflora saltmarshes. The SOS and EOS dates derived from three phenology retrieval methods showed similar patterns in latitudinal phenology variation: SOS became later and LOS became shorter as latitude increased, and the latitudinal trend of EOS was not as large as that of SOS. This study shows the potential of Landsat 7/8 and Sentinel-2 to quantify land surface phenology of S. alterniflora saltmarshes, which not only enhances our understanding of the spatial-temporal dynamics of coastal saltmarshes in China but also improves the management of this plant invader that threatens native saltmarshes in the world.
NSTL
Elsevier