Science of Remote Sensing最新文献

Cover crops are planted to reduce soil erosion, increase soil fertility, and improve watershed management. In the Delmarva Peninsula of the eastern United States, winter cover crops are essential for reducing nutrient and sediment losses from farmland. Cost-share programs have been created to incentivize cover crops to achieve conservation objectives. This program required that cover crops be planted and terminated within a specified time window. Usually, farmers report cover crop termination dates for each enrolled field (∼28,000 per year), and conservation district staff confirm the report with field visits within two weeks of termination. This verification process is labor-intensive and time-consuming and became restricted in 2020–2021 due to the COVID-19 pandemic. This study used Harmonized Landsat and Sentinel-2 (HLS, version 2.0) time-series data and the within-season termination (WIST) algorithm to detect cover crop termination dates over Maryland and the Delmarva Peninsula. The estimated remote sensing termination dates were compared to roadside surveys and to farmer-reported termination dates from the Maryland Department of Agriculture database for the 2020–2021 cover crop season. The results show that the WIST algorithm using HLS detected 94% of terminations (statuses) for the enrolled fields (n = 28,190). Among the detected terminations, about 49%, 72%, 84%, and 90% of remote sensing detected termination dates were within one, two, three, and four weeks of agreement to farmer-reported dates, respectively. A real-time simulation showed that the termination dates could be detected one week after termination operation using routinely available HLS data, and termination dates detected after mid-May are more reliable than those from early spring when the Normalized Difference Vegetation Index (NDVI) was low. We conclude that HLS imagery and the WIST algorithm provide a fast and consistent approach for generating near-real-time cover crop termination maps over large areas, which can be used to support cost-share program verification.

Tropical peatlands are estimated to hold carbon stocks of 70 Pg C or more as partly decomposed organic matter, or peat. Peat may accumulate over thousands of years into gently mounded deposits called peat domes with a relief of several meters over distances of kilometers. The mounded shapes of tropical peat domes account for much of the carbon storage in these landscapes, but their subtle topographic relief is difficult to measure. As many of the world's tropical peatlands are remote and inaccessible, spaceborne laser altimetry data from missions such as NASA's Global Ecosystem Dynamics Investigation (GEDI) on the International Space Station (ISS) and the Advanced Topographic Laser Altimeter System (ATLAS) instrument on the Ice, Cloud and land Elevation Satellite-2 (ICESat-2) observatory could help to describe these deposits. We evaluate retrieval of ground elevations derived from GEDI waveform data, as well as single-photon data from ATLAS, with reference to an airborne lidar dataset covering an area of over 300 km² in the Belait District of Brunei Darussalam on the island of Borneo. Spatial filtering of GEDI L2A version 2, algorithm 1 quality data reduced mean absolute deviations from airborne-lidar-derived ground elevations from 8.35 m to 1.83 m, root-mean-squared error from 15.98 m to 1.97 m, and unbiased root-mean-squared error from 13.62 m to 0.72 m. Similarly, spatial filtering of ATLAS ATL08 version 3 ground photons from strong beams at night reduced mean absolute deviations from 1.51 m to 0.64 m, root-mean-squared error from 3.85 m to 0.77 m, and unbiased root-mean-squared error from 3.54 m to 0.44 m. We conclude that despite sparse ground retrievals, these spaceborne platforms can provide useful data for tropical peatland surface altimetry if postprocessed with a spatial filter.

This study describes the generation and comprehensive validation of 30 m Landsat-based annual percent tree cover and forest cover loss products for the conterminous United States (CONUS). The products define (i) forest status with respect to three thematic classes: stable forest, stable non-forest, forest cover loss, (ii) percent tree cover (PTC, 0–100%), (iii) percent tree cover decrease (ΔPTC), and (iv) the Landsat acquisition dates bounding mapped forest cover loss occurrence. Forest was defined, based on the U.S. federal government forested land definition, as 30 m pixels with mapped PTC >10%. Annual products were derived using temporally overlapping 9-year periods (mapping within each central 5-year period) of USGS Landsat Analysis Ready Data (ARD) with reconciliation of the results between periods. The products for 2013 are presented and were validated rigorously by comparison with 1910 30 m independent reference data interpreted from bi-temporal <1 m resolution aerial imagery selected using a Stage 3 CONUS stratified random sampling design. The stable forest, stable non-forest, and forest cover loss results were validated using standard accuracy metrics derived from the confusion matrix. The overall accuracy was high (0.92), and class-specific user's accuracy (UA) and producer's accuracy (PA) metrics were also high for the stable forest (UA = 0.94, PA = 0.84) and stable non-forest (UA = 0.90, PA = 0.97) classes. The forest cover loss class had similarly high UA (0.89) but significantly lower PA (0.61) indicating non-negligible omission errors. All standard errors were <5%. The total area of stable forest over CONUS for year 2013 was estimated as 3,049,380 ± 114,392 km² and the total area of forest cover loss was estimated as 31,382 ± 4751 km², with 95% confidence interval. The PTC and ΔPTC products were validated by linear regression with the reference data, indicating good PTC precision reflected by a high coefficient of determination (R² = 0.79), and accuracy with a regression slope close to unity (0.86) and small intercept (3.48). The regression between mapped ΔPTC and the reference data had a high coefficient of determination (R² = 0.74) but a regression slope further away from unity (0.78) and small intercept (1.68) consistent with the forest cover loss omission errors revealed by the confusion matrix. State-level comparison of the stable forest mapped area with forest land area statistics published by the U.S. federal government for the 48 CONUS states indicated reasonable correspondence (R² = 0.97) but with a 1.15 regression line slope indicating relative over estimation of the mapped stable forest area, likely related to forest land reporting differences.

Tropical deforestation is a recent phenomenon that started in the second part of the twentieth century. One may argue that the Brazilian state of Maranhão is an excellent case study for ex-amining deforestation trends and the effects of environmental policies. A man-made line sepa-rates Maranhão into two sections. Due to the administrative divide between the Legal Amazon Maranhão (LM) and the Cerrado Maranhão (CM), one may hypothesize about differences in deforestation between the two regions. This research employs a nonlinear modelling approach based on Generalized Additive Models (GAMs) with a quasi-Poisson distribution and a logarith-mic function to detect deforestation patterns in these areas. Deforestation is linked to the year and a variety of climatic variables. These covariates differ substantially across seasons (rainy and dry) and regions. During times of above-average precipitation, including in the dry and wet seasons, deforestation occurred in the LM area. However, in the non-enforced region, this regime was not followed. According to the statistics, deforestation decreased in the LM region when precipitation levels were below average.

Intertidal mudflats are highly productive ecosystems where elevation gradients and complex microtopography drive the growth of benthic microalgae (microphytobenthos; MPB) that form the basis of estuarine foodwebs and are crucial for nutrient cycling, shoreline stabilization, and the persistence of marine and coastal species. Mudflat ecosystems are threatened by human activity and natural stressors and thus need to be mapped and monitored. Unoccupied aerial vehicle (UAV) technologies and digital aerial photogrammetry (DAP) have been successfully implemented to study mudflat environments. However, standardizing the UAV flight parameters needed for optimal DAP performance on mudflats remains outstanding. Here, we systematically determined the optimal flight parameters for collection and photogrammetric processing of UAV-acquired data on mudflats by (1) testing across-track overlap (50, 70, and 90%) and flight elevation (73 m and 110 m) parameters, assessing the accuracy of DAP results against reference data from a mobile laser scanner (MLS), and (2) comparing semi-variograms of digital surface models (DSMs) from two UAV flight elevations. We found that all combinations of UAV flight parameters yielded accurate DAP products; flight elevation had a marginal effect on image alignment and had no effect on accuracy, while across-track overlap had no effect on image alignment of DSM of difference (DoD) values. All UAV and MLS point clouds were aligned with and accuracy of < 0.016 m and absolute values of mean DoDs were all sub-millimeter, ranging from 0.0001 ± 0.0322 to 0.0083 ± 0.0270 m. We conclude that conducting UAV surveys at 110 m elevation with 50% across-track image overlap is sufficient for high-accuracy DAP in mudflats. Finally, we tested the utility of such fine-scale topographic data for ecological applications by comparing elevation and topographic position indices (TPI) of DAP-derived DSMs to MPB abundance, measured as chlorophyll a (chl-a), calculated from UAV-acquired NDVI data. We found that elevation and TPI account for 1.6–17% of the variation in chl-a concentration, and that these relationships depend on distance from shore and mudflat morphology. Our findings contribute to standardizing the application of UAV technologies in mudflats and demonstrate the potential of UAV-acquired data for modeling the relationship between microtopography and MPB on ecologically important mudflats.

Machine learning algorithms are a frequently used crop classification method and have been applied to identify the distribution of various crops over regional and national scales. Previous studies have underscored that the number of training samples strongly influences the classification accuracy of machine learning algorithms, resulting in extensive training sample collection efforts. This study, taking winter wheat as an example, challenges the above principle by selecting training samples with the time-weighted dynamic time warping (TWDTW) method and finds that the classification accuracy of machine learning algorithms highly relies on the representativeness and proportion of training samples rather than the quantity. With the increase of the representativeness of training samples, i.e. more comprehensively reflected the characteristics of winter wheat, the classification accuracy is continually improved. The best classification accuracy is further achieved when selecting the training samples of winter wheat and non-winter wheat according to the ratio of their statistical areas. On the contrary, only a slight difference was found in overall accuracy (91.26% and 90.74%), producer’s accuracy (86.33% and 86.65%) and user’s accuracy (97.37% and 96.01%) when using 1,000 and 10,000 training samples. Overall, this study demonstrates that the characteristics of training samples have a great impact on the classification accuracy of machine learning algorithms, and the training samples generated by TWDTW method are reliable for crop mapping.

The groundwater abstraction and injection cycle in coastal aquifer systems can locally change the piezometric head in aquifer system, leading to differential settlement on the ground that may compromise infrastructure safety. Furthermore, long-term, extensive groundwater extraction may cause significant damage to water resources. Ironically, Florida, a state known for its abundant water resources, has been experiencing major water supply issues in some areas that began to intensify with rapid population growth over the last five decades. As the demand for drinking water in Florida continues to rise, local authorities have turned to using brackish and saline water sources. As of 2022, more than 80% of the desalination plants in the United States are concentrated in the coastal areas of central and south Florida. Using satellite radar interferometry, we have investigated the spatiotemporal evolution of surface subsidence driven by groundwater pumping for brackish-water reverse osmosis (BWRO) desalination facilities in the City of Cape Coral, Florida. We employed Persistent Scatterer Interferometry (PSI) to process all available Sentinel 1A and 1B scenes over the region along two ascending orbits. The deformation time-series obtained from independent SAR data sets are compared spatiotemporally with the groundwater level that provides feed water to the BWRO facilities. The deformation pattern shows one main lobe of subsidence with rates of up to 25 mm/year centred around the operating wells in the north BWRO wellfield that we interpret as human-induced compaction. The spatial correlation between the subsiding area and the active production wells argues in favour of surface deformation induced by the BWRO operations. Based on the InSAR-derived displacement field and well data, we propose a model to explain the spatial heterogeneity of the subsidence process. The ground deformation is reproduced by an elastic model mimicking the reservoir compaction using planar negative closing dislocations. Modelling of the subsidence shows ∼ 0.67 Mm³ yr⁻¹ vol loss due to compaction of the aquifer. The subsidence deformation was also used to compute the cumulative drainage area of the producing wells.

Global measurement of plant structural traits like canopy height, canopy cover and Plant Area Volume Density (PAVD) profiles form a key input for many emerging fields in ecology and meteorology. Here, we test the ability of an ICESat-2 simulator, based on the GEDI simulator presented by Hancock et al. (2019) and pre-launch ICESat-2 simulations by Neuenschwander and Magruder (2016), to replicate measurements of plant structural traits retrieved from ICESat-2 observations and, through this, explore the sensitivity of ICESat-2 to plant structural traits not currently in the ATL08 product. The simulator takes Airborne Laser Scanning (ALS) data, produces a pseudo-waveform and then samples individual photons to replicate real ICESat-2 measurements. Because the simulator assumes that the ICESat-2 photon-cloud distribution is proportional to the ALS vertical profile, which has been shown to be sensitive to canopy cover and structure, an accurate ICESat-2 simulator would indicate that ICESat-2 is sensitive to plant structural traits. ALS data are used to re-classify real ICESat-2, removing any classification error from the ATL08 product in order to allow a direct comparison of the returned photon profiles, from which the key simulation parameters - pure vegetation and pure ground photon rates - are calculated. ICESat-2 tracks that intersect ALS measurements from a range of sites and forest types are identified and simulated, allowing for one-to-one comparison of simulated and observed ICESat-2 photon-profiles and plant structural trait measurements. The canopy height, canopy cover, Relative Height metrics and PAVD profiles calculated from simulated and observed ICESat-2 photons are similar, with the simulator having an average canopy height bias of less than 50 cm and canopy cover bias less than 1.5% relative to the observed ICESat-2 data for sites where canopy:ground reflectance ratio is well constrained, indicating that ICESat-2 is sensitive to stand-level plant structural traits. Noise and differences between ground and canopy reflectances are found to be two key influences on the accuracy of ICESat-2 simulations and so plant structural trait measurement. This research suggests that, with global mapping of ground and canopy reflectances and correctly classified photons in the ATL08 product, it is possible to derive stand-level plant structural trait measurements from ICESat-2.