首页期刊导航|Remote Sensing of Environment
期刊信息/Journal information
Remote Sensing of Environment
Elsevier
Remote Sensing of Environment

Elsevier

0034-4257

Remote Sensing of Environment/Journal Remote Sensing of EnvironmentSCIISTPEI
正式出版
收录年代

    Using radiant intensity to characterize the anisotropy of satellite-derived city light at night

    Li, XiShang, XiaoyuZhang, QinglingLi, Deren...
    18页
    查看更多>>摘要：Recent studies have described the anisotropy of ALAN based on statistical analysis of a quadratic model-based relationship between radiances derived from the VIIRS Day/Night Band sensor onboard Suomi-NPP and viewing zenith angle (VZA) (Li et al., 2019). Contrary to conventional wisdom, the satellite-observed radiance of ALAN always decreases at first and then increase with VZA, especially for high building areas. This leads to a socalled 'cold-spot' effect (i.e. radiance reaches local minimum in a specific VZA) in the VZA-radiance relationship, one that has not been understood using existing remote sensing methods. In this paper, we propose using radiant intensity - a measure of light power in a specific direction- to characterize the anisotropy of satellite-observed artificial lights at night (ALAN). Accordingly, we propose a new study design based on analysis across fourteen global cities. A linear regression model describing the relationship between VZA and radiant intensity resulted in an averaged regression R-2 between 0.26 and 0.73 for the cities, suggesting a decay of radiant intensity with increased VZA. We then introduced a cosine-corrected linear model to describe the VZA-radiance relationship, which is mathematically transformed from the linear VZA-intensity model. Our results suggest that the 'cold spot' effect in the VZA-radiance relationship is consistent with the revealed phenomenon that the radiant intensity decays with increased VZA. We also constructed indexes related to urban morphology derived from LiDAR data, including the Blocking Index (BI), the Standard Deviation of Building Height (SDBI) and Average Building Height (ABH). These variables were all moderately or strongly correlated (Pearson correlation coefficient < -0.45) to the Change Index (CI), which describes the anisotropic pattern of ALAN in the VZA-intensity relationship, in the four U.S. cities where LiDAR data is accessible. We conclude that radiant intensity is a suitable physical index to characterize the directional distribution of ALAN that can help understand the anisotropy of ALAN.
      原文链接:
    • NSTL
    • Elsevier

    Soil moisture retrieval over croplands using dual-pol L-band GRD SAR data

    Bhogapurapu, NarayanaraoDey, SubhadipMandal, DipankarBhattacharya, Avik...
    26页
    查看更多>>摘要：Vegetation cover significantly influences the hydrometeorological processes of land surfaces. The heterogeneity of vegetation cover makes these processes more complex and impacts the interaction between water held in the soil matrix and vegetation cover. The backscatter measured by Synthetic Aperture Radar (SAR) is sensitive to target dielectric and morphological properties. In addition, SAR acquisitions are weather-independent, providing a major advantage over optical imaging during periods of cloud cover. Most often, vegetation properties are measured using vegetation indices, including simple ratios of backscatter intensities. However, this approach can miss the scattered wave purity from vegetation targets. The scattered wave characteristic is an essential parameter as it is susceptible to the complex random structure of vegetation. In this study, we propose a novel dual-polarimetric SAR vegetation descriptor based on the co-pol purity component of the wave. We use this descriptor within a semi-empirical vegetation model to estimate soil moisture. It is noteworthy that this proposed method to retrieve soil moisture uses only the dual-polarimetric Ground Range Detected (GRD) SAR product, i.e., only backscatter intensities. Therefore, the proposed method has the potential for operational scale monitoring applications at a global scale. This study validated different crop types (viz. canola, oats, forage, maize, soybeans, sugar beet, wheat, winter wheat, winter barley) over two test sites using airborne SAR data. The Root Mean Square Error (RMSE) values for soil moisture are within the range of 4.3% to 7.7% with the Pearson correlation coefficient r greater than 0.6. This novel descriptor will facilitate the operational use of dual-polarimetric GRD SAR data for vegetation monitoring and soil moisture estimation. The code to generate DpRVI(c) in Google Earth Engine is available at: https://github.com/Narayana-Rao/dua l_pol_descriptors.
      原文链接:
    • NSTL
    • Elsevier

    Two-phase forest inventory using very-high-resolution laser scanning

    Persson, Henrik J.Olofsson, KennethHolmgren, Johan
    14页
    查看更多>>摘要：In this study, we compared a two-phase laser-scanning-based forest inventory of stands versus a traditional field inventory using sample plots. The two approaches were used to estimate stem volume (VOL), Lorey's mean height (HL), Lorey's stem diameter (DL), and VOL per tree species in a study area in Sweden. The estimates were compared at the stand level with the harvested reference values obtained using a forest harvester. In the first phase, a helicopter acquired airborne laser scanning (ALS) data with >500 points/m2 along 50-m wide strips across the stands. These strips intersected systematic plots in phase two, where terrestrial laser scanning (TLS) was used to model DL for individual trees. In total, phase two included 99 plots across 10 boreal forest stands in Sweden (lat 62.9 degrees N, long 16.9 degrees E). The single trees were segmented in both the ALS and TLS data and linked to each other. The very-high-resolution ALS data enabled us to directly measure tree heights and also classify tree species using a convolutional neural network. Stem volume was predicted from the predicted DBH and the estimated height, using national models, and aggregated at the stand level. The study demonstrates a workflow to derive forest variables and stand-level statistics that has potential to replace many manual field inventories thanks to its time efficiency and improved accuracy. To evaluate the inventories, we estimated bias, RMSE, and precision, expressed as standard error. The laser-scanning-based inventory provided estimates with an accuracy considerably higher than the field inventory. The RMSE was 17 m3/ha (7.24%), 0.9 m (5.63%), and 16 mm (5.99%) for VOL, HL, and DL respectively. The tree species classification was generally successful and improved the three species-specific VOL estimates by 9% to 74%, compared to field estimates. In conclusion, the demonstrated laser-scanning-based inventory shows potential to replace some future forest inventories, thanks to the increased accuracy demonstrated empirically in the Swedish forest study area.
      原文链接:
    • NSTL
    • Elsevier

    Uncertainty of city-based urban heat island intensity across 1112 global cities: Background reference and cloud coverage

    Li, KangningChen, YunhaoGao, Shengjun
    19页
    查看更多>>摘要：As an urban heat island (UHI) is closely associated with a wide variety of environmental issues, it is necessary to monitor its spatiotemporal variations in an accurate and timely manner. However, the accuracy of surface urban heat island intensity (SUHII) estimation is greatly challenged by three issues. Namely, less attention has been given to city-based global studies, LSUHII (SUHII difference) is induced when various methods are used to the determine background reference, and uncertainty arises in the methods for coping with missing land surface temperature (LST). Thus, this paper quantitatively evaluated these different methods and their impacts on spatiotemporal SUHII variations across 1112 global cities. There are two major findings: (1) the modified equal area-rural (MEA-R) method can overcome the limitations of the other methods. The fixed buffer (FB) method is limited by the difficulty of buffer selection and the inapplicability of a single fixed buffer for various cities, particularly in large-scale studies. The simplified urban-extent (SUE) method is challenged by cluster discrepancies and data obsolescence issues, and the modified equal area-suburban (MEA-S) method underestimates SUHII resulting from strong human activities in suburban areas. The daytime LSUHII induced by different methods for the reference definition reaches 0.62 K, accounting for 42% of the average SUHII. LSUHII is 0.34 K (23%) at night. LSUHII with evident spatiotemporal variations indicates that the different methods used to define the background reference introduce a considerable uncertainty to SUHII research. (2) After comparing methods for coping with missing LST, the first spatial and after temporal aggregation with threshold removal (FSAT-T) method is demonstrated to provided improved robustness to data missing over other methods. Daytime LSUHII is over 0.1 K in 62% of cities, and nighttime LSUHII is over 0.1 K in 45% of cities. Strong latitudinal and seasonal variations of LSUHII induced by the different methods used to cope with missing data introduce uncertainty and incomparability to SUHII research. Moreover, various image quality control (QC) methods and the effects from these methods on SUHII are discussed. The accuracy of QC methods changes with the error distribution of the image quality. Further investigations are required to determine the optimal QC method. The questions of whether different background definition affects the atmospheric UHI were also discussed. In this paper, a promising strategy for estimating city-based SUHII is proposed and a global-scale database based on the Google Earth Engine (GEE) is established to support further research on the variations and controls of SUHI.
      原文链接:
    • NSTL
    • Elsevier

    Quantifying tropical forest structure through terrestrial and UAV laser scanning fusion in Australian rainforests

    Terryn, LouiseCalders, KimBartholomeus, HarmBartolo, Renee E....
    15页
    查看更多>>摘要：Accurately quantifying tree and forest structure is important for monitoring and understanding terrestrial ecosystem functioning in a changing climate. The emergence of laser scanning, such as Terrestrial Laser Scanning (TLS) and Unoccupied Aerial Vehicle Laser Scanning (UAV-LS), has advanced accurate and detailed forest structural measurements. TLS generally provides very accurate measurements on the plot-scale (a few ha), whereas UAV-LS provides comparable measurements on the landscape-scale ( 10 ha). Despite the pivotal role dense tropical forests play in our climate, the strengths and limitations of TLS and UAV-LS to accurately measure structural metrics in these forests remain largely unexplored. Here, we propose to combine TLS and UAV-LS data from dense tropical forest plots to analyse how this fusion can further advance 3D structural mapping of structurally complex forests. We compared stand (vertical point distribution profiles) and tree level metrics from TLS, UAV-LS as well as their fused point cloud. The tree level metrics included the diameter at breast height (DBH), tree height (H), crown projection area (CPA), and crown volume (CV). Furthermore, we evaluated the impact of point density and number of returns for UAV-LS data acquisition. DBH measurements from TLS and UAV-LS were compared to census data. The TLS and UAV-LS based H, CPA and CV measurements were compared to those obtained from the fused point cloud. Our results for two tropical rainforest plots in Australia demonstrate that TLS can measure H, CPA and CV with an accuracy (RMSE) of 0.30 m (H-average =27.32 m), 3.06 m(2) (CPAaverage =66.74 m(2)), and 29.63 m(3) (CVaverage =318.81 m(3)) respectively. UAV-LS measures H, CPA and CV with an accuracy (RMSE) of <0.40 m, <5.50 m(2), and <30.33 m(3) respectively. However, in dense tropical forests single flight UAV-LS is unable to sample the tree stems sufficiently for DBH measurement due to a limited penetration of the canopy. TLS can determine DBH with an accuracy (RMSE) of 5.04 cm, (DBHaverage =45.08 cm), whereas UAV-LS can not. We show that in dense tropical forests stand-alone TLS is able to measure macroscopic structural tree metrics on plot-scale. We also show that UAV-LS can be used to quickly measure H, CPA, and CV of canopy trees on the landscape-scale with comparable accuracy to TLS. Hence, the fusion of TLS and UAV-LS, which can be time consuming and expensive, is not required for these purposes. However, TLS and UAV-LS fusion opens up new avenues to improve stand-alone UAV-LS structural measurements at the landscape scale by applying TLS as a local calibration tool.
      原文链接:
    • NSTL
    • Elsevier

    Thermally derived evapotranspiration from the Surface Temperature Initiated Closure (STIC) model improves cropland GPP estimates under dry conditions

    Bai, YunBhattarai, NishanMallick, KaniskaHu, Tian...
    19页
    查看更多>>摘要：Satellite-based gross primary productivity (GPP) monitoring in croplands is challenging due to our limited ability to empirically constrain photosynthetic capacity and associated parameters. Here, we investigated if integrating land surface temperature (T-R)-based evapotranspiration (ET) or latent heat flux (lambda E) into a Remote sensing-driven approach to Coupling Ecosystem Evapotranspiration and Photosynthesis (RCEEP) model, which is modified from the underlying water use efficiency method, can better characterize GPP under dry conditions as compared to the light use efficiency (LUE)-based models. We developed the new GPP model, termed STIC-RCEEP, by combining an ET model, called STIC (Surface Temperature Initiated Closure), and RCEEP. We compared the performance of STIC-RCEEP against four conventional LUE-based models (Vegetation Photosynthesis Model, MOD17, STIC-MOD17, STIC-LUE), using tower-based daily GPP data from different cropland flux sites across the globe. An evaluation of the five GPP models, all optimized using available 170 site-years data from the 22 flux sites, revealed the relatively better performance of the STIC-RCEEP (R-2 = 0.78 and RMSE = 2.5 gC m(-2) d(-1)) with respect to the other four LUE models. Error analysis revealed substantially less error of STICRCEEP particularly under the dry conditions and RMSE of STIC-RCEEP was 9%-14% less than the other models. This finding highlights the enhanced ability of thermal sensors to capture the water stress signals in ET and GPP, which is also evidenced by the substantially better performance of STIC than another widely-regarded nonthermal ET model (the Priestly Taylor - Jet Propulsion Laboratory model), under dry conditions. The improved GPP estimates from STIC-RCEEP under water-stressed environment opens avenues for further research and applications using existing and new TR-based sensors in coupled crop water use-productivity modeling.
      原文链接:
    • NSTL
    • Elsevier

    Attention mechanism-based generative adversarial networks for cloud removal in Landsat images

    Plaza, Antonio J.Xu, MengDeng, FurongJia, Sen...
    15页
    查看更多>>摘要：The existence of clouds affects the quality of optical remote sensing images. Cloud removal is an important preprocessing procedure to effectively improve the utilization of optical remote sensing images. Thin clouds partly obscure the land surfaces beneath them, making it possible to correct the cloudy scenes according to the available information. In this research, we introduce the attention mechanism-based generative adversarial networks for cloud removal (AMGAN-CR) method for Landsat images. First, attention maps of the input cloudy images are generated to extract the cloud distributions and features through an attentive recurrent network. Second, clouds are removed by an attentive residual network under the guidance of the attention maps. Finally, the generated feature maps are fed to a reconstruction network to restore the final cloud-free images. The networks are trained by cloudy and cloud-free Landsat image pairs, and the cloudy images are tested to validate the effectiveness of AMGAN-CR. Both simulated and real cloud experimental results show that the proposed method is more outstanding than the other five state-of-the-art traditional and deep learning methods in removing cloud.
      原文链接:
    • NSTL
    • Elsevier

    Time series analysis for global land cover change monitoring: A comparison across sensors

    Xu, LiliHerold, MartinTsendbazar, Nandin-ErdeneMasiliunas, Dainius...
    14页
    查看更多>>摘要：Comparing the performance of different satellite sensors in global land cover change (LCC) monitoring is necessary to assess their potential and limitations for more accurate and operational LCC estimations. This paper aims to examine and compare the performance in LCC monitoring using three satellite sensors: PROBA-V, Landsat 8 OLI, and Sentinel-2 MSI. We utilized a unique set of global reference data containing four years of records (2015-2018) at 29,263 land cover change/no-change 100 x 100-m sites. The LCC monitoring was conducted using the BFAST(s)-Random Forest (BRF) change detection framework involving 15 global timeseries vegetation indices and three BFAST models. Due to the different spectral characteristics and data availability of the sensors, we designed 30 comparison scenarios to extensively evaluate their performance. The overall results were: 1) for global general LCC monitoring, Landsat 8 OLI slightly outperformed Sentinel-2, and PROBA-V performed the worst. The performance among the three sensors differed consistently despite different data availability and spectral observation regions. Sentinel-2 was more competitive with Landsat 8 when the red-edge 1 band was included; 2) Landsat 8 was more accurate in forest, herbaceous vegetation, and water monitoring. Sentinel-2 performed particularly well in wetland monitoring. In addition, we further observed: 3) missing data in time series decreased the accuracy in all sensors, but had little influence on the relative performance across sensors; 4) combining sensors would not necessarily improve the accuracy because the complementary effects enhanced the accuracy only when there was a large amount of data missing for all sensors; 5) the BRF framework maintained the performance gap among sensors, but obtained a higher and more balanced accuracy overall when compared with using BFAST methods alone, by involving ensemble learning with an embedded sample-balancing strategy; 6) among the random forest variables, the 'magnitude' proved to be the most important contributor, and the NDVI had the most consistently good performance across sensors when compared against other vegetation indices. All sensors using BRF still had some errors in change detection, with a tendency to underestimate the global LCC. A potential reason for this is the complexity of the diverse change/no-change characteristics at the global extent and the fact that smaller, more subtle LCCs might not be well detected. These limitations could be addressed by taking advantage of ensemble learning approaches with a combination of multiple independent region/thematic-adapted LCC monitoring models and using the original Sentinel-2 (10 m) and Landsat 8 (30 m) in the future.
      原文链接:
    • NSTL
    • Elsevier

    Monitoring glacial lake outburst flood susceptibility using Sentinel-1 SAR data, Google Earth Engine, and persistent scatterer interferometry

    Wangchuk, SonamBolch, TobiasRobson, Benjamin Aubrey
    18页
    查看更多>>摘要：Continuous monitoring of glacial lakes, their parent glaciers and their surroundings is crucial because possible outbursts of these lakes pose a serious hazard to downstream areas. Ongoing climate change increases the risk of this hazard globally due to recession of glaciers leading to formation and expansion of glacial lakes, and permafrost degradation which impacts the stability of glaciers, slopes and moraines. Here, we demonstrate the capability of our approach for monitoring lake outburst susceptibility using time-series of Sentinel-1 Synthetic Aperture Radar (S-1 SAR) data. We selected Lunana in the Bhutanese Himalayas as an example region as it is highly susceptible to glacial lake outburst floods and suitable baseline data were available. We used Google Earth Engine (GEE) to calculate average radar backscatter intensity (ARBI) of glaciers, lakes, basins, and moraines. To determine the periodicity of the highest and the lowest radar backscatter intensity, we denoised the ARBI data using a Fast Fourier Transform and autocorrelated using a Pearson correlation function. Additionally, we determined glacier melt area, basin melt area, lake area, open water area, and lake ice area using radar backscatter intensity data. The Persistent Scatterer Interferometry (PSI) technique was used to investigate the stability of moraines and slopes around glacial lakes. The PSI results were qualitatively validated by comparison with high-resolution digital elevation model differencing results. Our approach showed that glaciers and basins in the region underwent seasonal and periodic changes in their radar backscatter intensity related to changes in ice and snow melt. Lakes also showed seasonal changes in their radar backscatter intensity related to the variation of lake ice and open water area, but the radar backscatter intensity change was not periodic. We could also infer lake area change using a time-series radar backscatter intensity data such as the rapid expansion of Bechung Tsho. The PSI analysis showed that all the terminal moraines were stable except Drukchung Tsho. Its terminal moraine showed subsidence at the rate of -5.18 mm/yr. Sidewalls of lakes were also stable with the exception of Lugge Tsho at site 4. Due to the free availability of S-1 SAR data, the efficiency of processing a large amount of imagery within GEE, and the PSI technique, we were able to understand the outburst susceptibility of glacial lakes in the region at great detail. The regular acquisition of S-1 SAR data enables continuous monitoring of glacial lakes. A similar approach and concept can be transferred to any geographic region on earth that shares similar challenges in glacial lake monitoring.
      原文链接:
    • NSTL
    • Elsevier

    Shifts in structural diversity of Amazonian forest edges detected using terrestrial laser scanning

    Maeda, Eduardo EijiNunes, Matheus HenriqueCalders, Kimde Moura, Yhasmin Mendes...
    11页
    查看更多>>摘要：Forest edges are an increasingly common feature of Amazonian landscapes due to human-induced forest frag-mentation. Substantial evidence shows that edge effects cause profound changes in forest biodiversity and productivity. However, the broader impacts of edge effects on ecosystem functioning remain unclear. Assessing the three-dimensional arrangement of forest elements has the potential to unveil structural traits that are scalable and closely linked to important functional characteristics of the forest. Using over 600 high-resolution terrestrial laser scanning measurements, we present a detailed assessment of forest structural metrics linked to ecosystem processes such as energy harvesting and light use efficiency. Our results show a persistent change in forest structural characteristics along the edges of forest fragments, which resulted in a significantly lower structural diversity, in comparison with the interior of the forest fragments. These structural changes could be observed up to 35 m from the forest edges and are likely to reflect even deeper impacts on other ecosystem variables such as microclimate and biodiversity. Traits related to vertical plant material allocation were more affected than traits related to canopy height. We demonstrate a divergent response from the forest understory (higher vegetation density close to the edge) and the upper canopy (lower vegetation density close to the edge), indicating that assessing forest disturbances using vertically integrated metrics, such as total plant area index, can lead to an erroneous interpretation of no change. Our results demonstrate the strong potential of terrestrial laser scanning for benchmarking broader-scale (e.g. airborne and space-borne) remote sensing assessments of forest distur-bances, as well as to provide a more robust interpretation of biophysical changes detected at coarser resolutions.
      原文链接:
    • NSTL
    • Elsevier