Cloud-free Observations Research Articles

Mapping corn and soybean phenometrics at field scales over the United States Corn Belt by fusing time series of Landsat 8 and Sentinel-2 data with VIIRS data

Accurate and timely crop phenology is essential for efficient crop management and precise modelling of crop yield and productivity. It has been commonly mapped from moderate or coarse resolution (≥500 m) satellite data at a national or global scale, which generally represents the mixture of multiple crops or between crops and natural vegetation. Crop phenological progress at field scales (≤30 m) is critical for monitoring the growth of specie-specific crop types, but it is very challenging at large scales because of limited cloud-free observations in the satellite time series. To map corn and soybean phenometrics at a field scale (30 m) over entire Corn Belt in the United States, this study for the first time develops a new approach that fuses high temporal resolution Visible Infrared Imaging Radiometer Suite (VIIRS) observations (500 m) and high spatial resolution Harmonized Landsat and Sentinel-2 (HLS) data (30 m). The fused HLS-VIIRS time series (daily 30 m) is applied to map crop phenometrics of greenup onset, mid-greenup phase, maturity onset, senescence onset, mid-senescence phase, and dormancy onset. These six phenometrics are correlated to crop progress (CP) of seven corn and six soybean growth stages from planting to harvest that are observed by the National Agricultural Statistics Service (NASS) of the United States Department of Agriculture (USDA) and are further validated using PhenoCam near-surface time series imagery. The result indicates that the HLS-VIIRS crop phenometrics are capable of tracking the CP very well. Although their biophysical definitions of CP stages and sample coverage were not the same in our comparisons, the HLS-VIIRS crop phenometrics and NASS CP at a state level are closely comparable with a difference<4 days for corn and soybean emergence, 1 day for soybean dropping leaves, and 6–10 days for other corn and soybean growth stages. Their significant correlations (R2 ranging from 0.81 to 0.99, p < 0.01) suggest that the HLS-VIIRS crop phenometrics are robust proxies for predicting NASS CP with a high accuracy. Moreover, the HLS-VIIRS phenometrics aligns closely with PhenoCam phenological timings at various crop growth stages with an absolute average difference from 3 days to 5.3 days and bias from 1.5 to 3.5 days. The result demonstrates that the proposed approach could be implemented for monitoring field scale crop progress at a national or global scale.

Cloudy Image Arithmetic: A Cloudy Scene Synthesis Paradigm With an Application to Deep-Learning-Based Thin Cloud Removal

We present a cloudy scene synthesis paradigm that produces cloudy images for arbitrary optical cloud-free observations. The synthesis paradigm consists of two fundamental operations, i.e., 1) cloud self-subtraction and 2) cloud addition-to-scene. Cloud self-subtraction extracts cloud ingredient images from cloudy images of weak texture regions (typically sea areas). The cloud ingredient images exhibit clouds in more realistic forms than simulated clouds. Cloud addition-to-scene incorporates the cloud ingredient images into arbitrary cloud-free land images, synthesizing cloudy scenes. It provides a means of constructing data pairs of cloud-free scene images and cloudy scene images, which are highly needed but considerably insufficient in the remote sensing literature. We refer to the overall paradigm consisting of the two fundamental operations as cloudy image arithmetic. We explore the use of the cloudy image arithmetic for the purpose of thin cloud removal. To this end, we develop a multi scale generative adversarial net (MSGAN) that removes thin clouds from cloudy scenes. We use the cloudy image arithmetic to construct a comprehensive training dataset for the MSGAN. Experimental evaluations validate that the cloudy image arithmetic synthesizes good cloudy scenes and the MSGAN with aid of the cloudy image arithmetic gives effective results in thin cloud removal.

Mapping grassland mowing events across Germany based on combined Sentinel-2 and Landsat 8 time series

Spatially explicit knowledge on grassland extent and management is critical to understand and monitor the impact of grassland use intensity on ecosystem services and biodiversity. While regional studies allow detailed insights into land use and ecosystem service interactions, information on a national scale can aid biodiversity assessments. However, for most European countries this information is not yet widely available. We used an analysis-ready-data cube that contains dense time series of co-registered Sentinel-2 and Landsat 8 data, covering the extent of Germany. We propose an algorithm that detects mowing events in the time series based on residuals from an assumed undisturbed phenology, as an indicator of grassland use intensity. A self-adaptive ruleset enabled to account for regional variations in land surface phenology and non-stationary time series on a pixel-basis. We mapped mowing events for the years from 2017 to 2020 for permanent grassland areas in Germany. The results were validated on a pixel level in four of the main natural regions in Germany based on reported mowing events for a total of 92 (2018) and 78 (2019) grassland parcels. Results for 2020 were evaluated with combined time series of Landsat, Sentinel-2 and PlanetScope data. The mean absolute percentage error between detected and reported mowing events was on average 40% (2018), 36% (2019) and 35% (2020). Mowing events were on average detected 11 days (2018), 7 days (2019) and 6 days (2020) after the reported mowing. Performance measures varied between the different regions of Germany, and lower accuracies were found in areas that are revisited less frequently by Sentinel-2. Thus, we assessed the influence of data availability and found that the detection of mowing events was less influenced by data availability when at least 16 cloud-free observations were available in the grassland season. Still, the distribution of available observations throughout the season appeared to be critical. On a national scale our results revealed overall higher shares of less intensively mown grasslands and smaller shares of highly intensively managed grasslands. Hotspots of the latter were identified in the alpine foreland in Southern Germany as well as in the lowlands in the Northwest of Germany. While these patterns were stable throughout the years, the results revealed a tendency to lower management intensity in the extremely dry year 2018. Our results emphasize the ability of the approach to map the intensity of grassland management throughout large areas despite variations in data availability and environmental conditions.

Aerosol properties and aerosol–radiation interactions in clear-sky conditions over Germany

Abstract. The clear-sky radiative effect of aerosol–radiation interactions is of relevance for our understanding of the climate system. The influence of aerosol on the surface energy budget is of high interest for the renewable energy sector. In this study, the radiative effect is investigated in particular with respect to seasonal and regional variations for the region of Germany and the year 2015 at the surface and top of atmosphere using two complementary approaches. First, an ensemble of clear-sky models which explicitly consider aerosols is utilized to retrieve the aerosol optical depth and the surface direct radiative effect of aerosols by means of a clear-sky fitting technique. For this, short-wave broadband irradiance measurements in the absence of clouds are used as a basis. A clear-sky detection algorithm is used to identify cloud-free observations. Considered are measurements of the short-wave broadband global and diffuse horizontal irradiance with shaded and unshaded pyranometers at 25 stations across Germany within the observational network of the German Weather Service (DWD). The clear-sky models used are the Modified MAC model (MMAC), the Meteorological Radiation Model (MRM) v6.1, the Meteorological–Statistical solar radiation model (METSTAT), the European Solar Radiation Atlas (ESRA), Heliosat-1, the Center for Environment and Man solar radiation model (CEM), and the simplified Solis model. The definition of aerosol and atmospheric characteristics of the models are examined in detail for their suitability for this approach. Second, the radiative effect is estimated using explicit radiative transfer simulations with inputs on the meteorological state of the atmosphere, trace gases and aerosol from the Copernicus Atmosphere Monitoring Service (CAMS) reanalysis. The aerosol optical properties (aerosol optical depth, Ångström exponent, single scattering albedo and asymmetry parameter) are first evaluated with AERONET direct sun and inversion products. The largest inconsistency is found for the aerosol absorption, which is overestimated by about 0.03 or about 30 % by the CAMS reanalysis. Compared to the DWD observational network, the simulated global, direct and diffuse irradiances show reasonable agreement within the measurement uncertainty. The radiative kernel method is used to estimate the resulting uncertainty and bias of the simulated direct radiative effect. The uncertainty is estimated to −1.5 ± 7.7 and 0.6 ± 3.5 W m−2 at the surface and top of atmosphere, respectively, while the annual-mean biases at the surface, top of atmosphere and total atmosphere are −10.6, −6.5 and 4.1 W m−2, respectively. The retrieval of the aerosol radiative effect with the clear-sky models shows a high level of agreement with the radiative transfer simulations, with an RMSE of 5.8 W m−2 and a correlation of 0.75. The annual mean of the REari at the surface for the 25 DWD stations shows a value of −12.8 ± 5 W m−2 as the average over the clear-sky models, compared to −11 W m−2 from the radiative transfer simulations. Since all models assume a fixed aerosol characterization, the annual cycle of the aerosol radiation effect cannot be reproduced. Out of this set of clear-sky models, the largest level of agreement is shown by the ESRA and MRM v6.1 models.

Evaluating a spatiotemporal shape-matching model for the generation of synthetic high spatiotemporal resolution time series of multiple satellite data

Time series of high spatiotemporal resolution satellite data are required to monitor land surface biophysical properties and their seasonal and inter-annual dynamics at sub-field to field scales. To generate such time series, various algorithms have been developed to fuse infrequent cloud-free Landsat observations with daily Moderate Resolution Imaging Spectroradiometer (MODIS) observations. An early and widely used Spatial and Temporal Adaptive Reflectance Fusion Model (STARFM) was developed based on the assumptions that observations from the two sensors are consistent in terms of spatial aggregation and temporal variation. These assumptions are not always valid in the real world, especially in a complex heterogeneous region. In this study, we investigated a Spatiotemporal Shape-Matching Model (SSMM) to generate synthetic time series of high spatiotemporal resolution satellite data. The SSMM, which is conceptually different from the image pair-based STARFM and STARFM-like approaches, makes full use of all spatiotemporally matched fine and coarse resolution data in an entire time series to establish a temporally uniformed fusion model for a given fine resolution pixel. The SSMM assumes that the temporal shapes of both the fine and coarse resolution time series are similar, but their magnitudes and phenological phases could differ largely even for the same vegetation type. This study assessed the capability of the SSMM to generate high spatiotemporal resolution time series of two-band Enhanced Vegetation Index (EVI2). Specifically, we generated the synthetic 30 m time series using the SSMM and STARFM algorithms in the northeastern United States based on 500 m Visible Infrared Imaging Radiometer Suite (VIIRS) data and 30 m Landsat 8 and Sentinel-2 observations. We then evaluated the SSMM-derived 30 m time series across 15 land cover types and various degrees of heterogeneity. The result indicates that the SSMM is able to effectively generate synthetic time series in all different land cover types, which has advantages over the STARFM approach. Although the SSMM performance is degraded in heterogeneous regions, it can explain 82–91% of variations in 30 m EVI2 time series and produce constant root mean square error (0.053–0.056) across various levels of heterogeneity. Moreover, the SSMM can explain 87%-93% and 69%-87% of variation in 30 m EVI2 time series, respectively, with the models established using more than 10 pairs and 4–10 pairs of fine and coarse resolution observations. This suggests that the SSMM is also capable of generating high spatiotemporal resolution time series using historical Landsat time series with limited cloud-free observations, which are critical for studying vegetation dynamics and monitoring crop conditions.

A practical approach to reconstruct high-quality Landsat NDVI time-series data by gap filling and the Savitzky–Golay filter

Normalized Difference Vegetation Index (NDVI) data derived from Landsat satellites are important resources for vegetation monitoring. However, Landsat NDVI time-series data are usually temporally discontinuous owing to the nominal 16-day revisit cycle, frequent cloud contamination, and other factors. Although several methods have been proposed to reconstruct continuous Landsat NDVI time-series data, some challenges remain in the existing reconstruction methods. In this study, we developed a simple but effective Gap Filling and Savitzky–Golay filtering method (referred to as “GF-SG”) to reconstruct high-quality Landsat NDVI time-series data. This new method first generates a synthesized NDVI time series by filling missing values in the original Landsat NDVI time-series data by integrating the MODIS NDVI time-series data and cloud-free Landsat observations. Then, a weighted Savitzky-Golay filter was designed to remove the residual noise in the synthesized time series. Compared with three previous typical methods (IFSDAF, STAIR, and Fill-and-Fit) in two challenging areas (the Coleambally irrigated area in Australia and the Taian cultivated area in China) with heterogeneous parcels and complex NDVI profiles, we found that GF-SG performed the best with three obvious improvements. First, GF-SG improved the reconstruction of long-term continuous missing values in Landsat NDVI time series, whereas the other methods were less reliable for reconstructing these long data gaps. Second, the performance of GF-SG was less affected by the residual noise caused by cloud detection errors in the Landsat image, which is due to the incorporation of the weighted SG filter in the new method. Third, GF-SG was simple and could be implemented on the computing platform Google Earth Engine (GEE), which is particularly important for the practical application of the new method at a large spatial scale. The GEE code is freely available at https://code.earthengine.google.com/3a883c9e84ad119045bcb88e4de77b47?noload=true. We expect that this practical approach can further popularize the use of Landsat NDVI time-series data in ecological, geographical, and environmental research.

Cloud Cover throughout All the Paddy Rice Fields in Guangdong, China: Impacts on Sentinel 2 MSI and Landsat 8 OLI Optical Observations

Cloud cover hinders the effective use of vegetation indices from optical satellite-acquired imagery in cloudy agricultural production areas, such as Guangdong, a subtropical province in southern China which supports two-season rice production. The number of cloud-free observations for the earth-orbiting optical satellite sensors must be determined to verify how much their observations are affected by clouds. This study determines the quantified wide-ranging impact of clouds on optical satellite observations by mapping the annual total observations (ATOs), annual cloud-free observations (ACFOs), monthly cloud-free observations (MCFOs) maps, and acquisition probability (AP) of ACFOs for the Sentinel 2 (2017–2019) and Landsat 8 (2014–2019) for all the paddy rice fields in Guangdong province (APRFG), China. The ATOs of Landsat 8 showed relatively stable observations compared to the Sentinel 2, and the per-field ACFOs of Sentinel 2 and Landsat 8 were unevenly distributed. The MCFOs varied on a monthly basis, but in general, the MCFOs were greater between August and December than between January and July. Additionally, the AP of usable ACFOs with 52.1% (Landsat 8) and 47.7% (Sentinel 2) indicated that these two satellite sensors provided markedly restricted observation capability for rice in the study area. Our findings are particularly important and useful in the tropics and subtropics, and the analysis has described cloud cover frequency and pervasiveness throughout different portions of the rice growing season, providing insight into how rice monitoring activities by using Sentinel 2 and Landsat 8 imagery in Guangdong would be impacted by cloud cover.

A global analysis of the temporal availability of PlanetScope high spatial resolution multi-spectral imagery

The PlanetScope CubeSat constellation is providing unprecedented global coverage, visible to near infrared, atmospherically corrected, 3 m imagery. The revisit interval between successive overpasses varies in space and time in a complex manner because of a variety of factors and particularly because of the different sensor orbits. The temporal availability of PlanetScope imagery is quantified in this study considering all of the publicly available images acquired globally for a 12 month period from December 1st 2019 to November 30th 2020. A total of 175.8 million images were acquired by the constellation that was composed of between 100 and 133 unique PlanetScope sensors each month and three sensor generations. The local morning overpass times of the three sensor generations were quantified and the most frequently occurring times were 10:16, 10:29, and 10:03 (to the nearest minute) with 90% of the images acquired with a range of morning overpass times of 2 h and 13 min, 1 h and 30 min, and 1 h and 50 min, for PlanetScope-0, PlanetScope-1, and PlanetScope-2, respectively. Maps, histograms and summary statistics of the total number of observations and revisit intervals are derived with respect to a global grid of 4.7 million land points spaced 5.6 km apart in the equal area sinusoidal projection. The annual and monthly number of PlanetScope observations and average revisit intervals did not vary in a geographically uniform manner. This is due to several factors including the different PlanetScope orbits, seasonal high latitude darkness at the time of sensor overpass, and because of the changing number of sensors on orbit as PlanetScope sensors were decommissioned and later generations became operational over the 12 month study period. In addition, the images in each frame of sensed data are not made available if they cannot be geolocated due to cloud and/or featureless or unstructured terrain precluding ground control matching. The PlanetScope constellation provided higher temporal resolution than provided by sensors such as Landsat-8 or Sentinel-2 although 9% of the global land grid locations, predominantly in the interior of Greenland and non-coastal Antarctica, had no observations. Considering the 12 months of global observations, the median average revisit interval was only 30.3 h, and 9.6%, 71.8%, and 88.4% of the land points had average revisit intervals <24 h, <36 h, and < 48 h, respectively. Globally, the median minimum revisit interval was 25 s and the median maximum revisit interval was 9.15 days; 95.4% of the land grid points had a minimum revisit <180 s, and 89.1% had a maximum revisit <480 h (20 days). The PlanetScope images are labelled as “standard” or as “test” quality based on solar geometry, saturated pixel, and geolocation accuracy criteria. The median annual proportion of observations labelled as “standard” at each land grid point over the 12 months was 78.14%. A global cloud analysis was undertaken to quantify the probability of there being at least one and at least two cloud-free PlanetScope observations within 5, 7 and 10 day consecutive periods. Lower probabilities occurred in cloudy regions and where there were fewer observations. The global mean average probability of there being at least one cloud-free observation over the 12 study months was 0.84, 0.88 and 0.92 for the 5, 7 and 10 day periods respectively. The global mean average probability of there being at least two cloud-free observations was 0.65, 0.76 and 0.84 for the 5, 7 and 10 day periods respectively. The probabilities varied seasonally and the northern hemisphere winter (December–February) and spring (March–May) had lower and higher global mean average seasonal probabilities, respectively, than those derived over the 12 months. The high temporal global coverage provided by the PlanetScope constellation will benefit new applications in particular those concerned with assessment of rapidly changing phenomena and assessment of phenomena that cannot be resolved at moderate and coarse resolution.

SENTINEL-2 SURFACE REFLECTANCE PRODUCTS GENERATED BY CNES AND DLR: METHODS, VALIDATION AND APPLICATIONS

Abstract. To allow for a robust and automatic exploitation of Sentinel-2 data, Analysis Ready Data (ARD) products are requested by most users. The processors of ARD products take care of the common burdens necessary for most applications, that include precise orthorectification, cloud detection and atmospheric correction steps, as well as the generation of periodic syntheses of cloud free surface reflectances. The French Theia land data center, and the German Earth Observation Center (EOC) started delivering Sentinel-2 surface reflectance products to users in 2016 in France and 2019 in Germany respectively. Both centers produce and distribute these data sets in near real time, over large regions requested by French users such as Western Europe, Maghreb, Sahel, Madagascar… Theia’s and EOC products include an instantaneous surface reflectance product (Level-2A), and a monthly cloud free synthesis of surface reflectance (Level-3A). This article shortly describes the methods used to generate the Level-2A products with the MAJA processor, and the Level-3A products with theWASP processor. The MAJA processor is based on multi-temporal methods, that use the slow variation of surface reflectance to detect clouds and estimate aerosol depth, while WASP, thanks to the quality of MAJA cloud mask, calculates a weighted average of all the cloud free observations over 45 days, every month. The article also provides validation results for Level-2A and Level-3A products, resulting from comparison with in-situ data and with other methods. A last section gives first insights from the monitoring of user uptake of the distributed products.

Impacts of MOPITT cloud detection revisions on observation frequency and mapping of highly polluted scenes

Measurements made by the MOPITT (“Measurements of Pollution in the Troposphere”) instrument on the NASA Terra polar-orbiting platform enable the retrieval of tropospheric carbon monoxide (CO) concentrations. As determined by the Terra orbit and MOPITT swath width, the frequency of MOPITT observations at a specific location, or measurement sampling frequency, is typically about once every three to four days. However, because the MOPITT retrieval algorithm is only applicable to clear-sky scenes, MOPITT retrieval sampling frequency strongly depends on regional cloudiness and can be much smaller than the measurement sampling frequency. Moreover, highly polluted scenes, characterized by high aerosol optical depths, can be confused with cloudy scenes and thus be discarded unnecessarily by the MOPITT cloud detection algorithm. Herein are described revisions to this algorithm which substantially increase retrieval sampling over land in varying pollution conditions. The performance of the revised cloud detection algorithm is evaluated through validation, case studies, and continental-scale maps of retrieval sampling frequency. Presented case studies illustrate (1) why the current operational MOPITT cloud detection algorithm excludes extended areas of potentially valuable cloud-free MOPITT observations and (2) how, for the same scenes, improved retrieval coverage benefits analyses of regional CO variability. Maps of retrieval sampling frequency for South America and Asia exhibit well-defined improvements, especially in regions with poor sampling frequency in the current product.

Annual Time Series of Global VIIRS Nighttime Lights Derived from Monthly Averages: 2012 to 2019

A consistently processed annual global nighttime lights time series (2012–2019) was produced using monthly cloud-free radiance averages made from low light imaging day/night band (DNB) data collected by the NASA/NOAA Visible Infrared Imaging Radiometer Suite (VIIRS). The processing steps are modified from the original methods developed to produce annual nighttime lights products from nightly data. Only two years of VIIRS nighttime lights (VNL) were produced with the V.1 methods: 2015 and 2016. Here we report on methods used to produce a V.2 VNL time series from the monthly averages with filtering to remove extraneous features such as biomass burning, aurora, and background. In this case, outlier removal is achieved with a twelve-month median, which discards high and low radiance outliers, thus isolating the background to a narrow range of radiances under 1 nW/cm2/sr. Background areas with no detectable lighting are further isolated using a statistical measure of texture, 3 × 3 data range (DR). The DR threshold for zeroing out background rises as the number of cloud-free observations falls. The V.2 method extends the temporal leverage in the noise filtering by developing the DR threshold from a multiyear maximum DR and a multiyear percent cloud-free grid. Additional noise filtering is achieved by zeroing out grid cells that have low average radiances (&lt;0.6 nW/cm2/sr) and detection in only one or two years out of eight. The spatial extent and average radiance levels are compared for the V.1 and V.2 2015 VNL. For the vast majority of grid cells, the average radiances are nearly the same in the two products. However, the V.2 product has more areas of dim lighting detected. The key advantages of the V.2 time series include consistent processing and threshold levels across all years, thus optimizing the set for change detection analyses.

New generation geostationary satellite observations support seasonality in greenness of the Amazon evergreen forests

Assessing the seasonal patterns of the Amazon rainforests has been difficult because of the paucity of ground observations and persistent cloud cover over these forests obscuring optical remote sensing observations. Here, we use data from a new generation of geostationary satellites that carry the Advanced Baseline Imager (ABI) to study the Amazon canopy. ABI is similar to the widely used polar orbiting sensor, the Moderate Resolution Imaging Spectroradiometer (MODIS), but provides observations every 10–15 min. Our analysis of NDVI data collected over the Amazon during 2018–19 shows that ABI provides 21–35 times more cloud-free observations in a month than MODIS. The analyses show statistically significant changes in seasonality over 85% of Amazon forest pixels, an area about three times greater than previously reported using MODIS data. Though additional work is needed in converting the observed changes in seasonality into meaningful changes in canopy dynamics, our results highlight the potential of the new generation geostationary satellites to help us better understand tropical ecosystems, which has been a challenge with only polar orbiting satellites.

Prototyping of LAI and FPAR Retrievals From GOES-16 Advanced Baseline Imager Data Using Global Optimizing Algorithm

The latest Geostationary (GEO) Operational Environmental Satellite-16 (GOES-16) equipped with Advanced Baseline Imager (ABI) has comparable spectral and spatial resolution as low earth orbiting (LEO) sensors [i.e., the Moderate Resolution Imaging Spectroradiometer (MODIS)], but with up-to-the-minute image acquisition capability. This enables greater opportunities to generate two essential climate variables-Leaf area index (LAI) and the fraction of photosynthetically active radiation (FPAR) absorbed by vegetation with more cloud-free observations and at much higher frequency. The improved GEO LAI/FPAR products will increase the capacity for monitoring highly dynamic ecosystems in a timely manner. However, the radiative transfer (RT)-based MODIS operational algorithm cannot be directly applied to GOES-16 ABI data due to different sensor characteristics. Fortunately, it has been shown theoretically and practically, that the RT-based algorithm can be transplanted to any other optical sensors by optimizing the sensor-specific parameters-the single scattering albedo (SSA) and relative stabilized precision (RSP). We built the RT-based ABI-specific lookup tables (LUTs) using a global optimizing algorithm (SCE-UA) that can quickly find the optimal solution. SCE-UA optimizes the SSAs and RSPs in the LUTs by minimizing the difference between ABI and MODIS retrievals and maximizing the main algorithm execution rate. Our efforts indicate that these strategies of parametric optimization is able to decrease the discrepancy between the ABI and MODIS LAI/FPAR products. Comprehensive evaluations were conducted to evaluate ABI retrievals. These indirect inter-comparisons suggest a spatiotemporal consistency between ABI and the benchmark MODIS products, while direct validation with field measurements increases confidence in their accuracy. The proposed approach is applicable to any other optical sensors for LAI/FPAR estimation, especially, GEO sensors (i.e., Himawari-8, Geo-KOMPSAT-2A, FengYun-4 etc.).

Mapping Crop Phenology in Near Real-Time Using Satellite Remote Sensing: Challenges and Opportunities

Crop phenology is critical for agricultural management, crop yield estimation, and agroecosystem assessment. Traditionally, crop growth stages are observed from the ground, which is time-consuming and lacks spatial variability. Remote sensing Vegetation Index (VI) time series has been used to map land surface phenology (LSP) and relate to crop growth stages mostly after the growing season. In recent years, high temporal and spatial resolution remote sensing data have allowed near-real-time mapping of crop phenology within the growing season. This paper summarizes two classes of near-real-time mapping methods, i.e., curve-based and trend-based approaches. The curve-based approaches combine the time series VIs and crop growth stages from historical years with the current observations to estimate crop growth stages. The curve-based approaches are capable of a short-term prediction. The trend-based approaches detect upward or downward trends from time series and confirm the trends using the increasing or decreasing momentum and VI thresholds. The trend-based approaches only use current observations. Both curve-based and trend-based approaches are promising in mapping crop growth stages timely. Nevertheless, mapping crop phenology near real-time is challenging since remote sensing observations are not always sensitive to crop growth stages. The accuracy of crop phenology detection depends on the frequency and availability of cloud-free observations within the growing season. Recent satellite datasets such as the harmonized Landsat and Sentinel-2 (HLS) are promising for mapping crop phenology within the season over large areas. Operational applications in the near future are feasible.

Characterizing spring phenology of temperate broadleaf forests using Landsat and Sentinel-2 time series

Vegetation phenology has a great impact on land-atmosphere interactions like carbon cycling, albedo, and water and energy exchanges. To understand and predict these critical land-atmosphere feedbacks, it is crucial to measure and quantify phenological responses to climate variability, and ultimately climate change. Coarse-resolution sensors such as MODIS and AVHRR have been useful to study vegetation phenology from regional to global scales. These sensors are, however, not capable of discerning phenological variation at moderate spatial scales. By offering increased observation density and higher spatial resolution, the combination of Landsat and Sentinel-2 time series might provide the opportunity to overcome this limitation.In this study, we analyzed the potential of combined Sentinel-2 and Landsat time series for estimating start of season (SOS) of broadleaf forests across Germany for the year 2018. We tested two common statistical modeling approaches (logistic and generalized additive models using thin plate splines) and the two most commonly used vegetation indices, the Normalized Difference Vegetation Index (NDVI) and the Enhanced Vegetation Index (EVI).We found strong agreement between SOS estimates from logistic and spline models (rEVI = 0.86; rNDVI = 0.65), whereas agreement was higher for EVI than for NDVI (RMSDEVI = 3.07, RMSDNDVI = 5.26 days). The choice of vegetation index thus had a higher impact on the results than the fitting method. The EVI-based SOS also showed higher correlation with ground observations compared to NDVI (rEVI = 0.51, rNDVI = 0.42). Data density played an important role in estimating land surface phenology. Models combining Sentinel-2A/B, with an average cloud-free observation frequency of 12 days, were largely consistent with the combined Landsat and Sentinel-2 models, suggesting that Sentinel-2A/B may be sufficient to capture SOS for most areas in Germany in 2018. However, in non-overlapping swath areas and mountain areas, observation frequency was significantly lower, underlining the need to combine Landsat and Sentinel-2 for consistent SOS estimates over large areas. Our study demonstrates that estimating SOS of temperate broadleaf forests at medium spatial resolution has become feasible with combined Landsat and Sentinel-2 time series.

Consistency evaluation of landsat-7 and landsat-8 for improved monitoring of colored dissolved organic matter in complex water

A series of Landsat satellites has been widely used to monitor water quality in coastal and inland complex water, but a solo Landsat satellite may not provide enough information because of its long revisit time (16-day) and frequent cloud cover, and hence sometimes it is necessary to fuse or combine the images from two or more satellites, for example, Landsat-7 and Landsat-8. However, the consistency of the two satellites has not been fully evaluated when they are used to estimate the colored dissolved organic matter (CDOM), an important water quality indicator. In this study, we used 210 Landsat-7 images and 185 neighboring Landsat-8 images from April 8, 2013 to December 26, 2018 in Sacramento River and Northern San Francisco Bay, USA, and extracted 225,265 paired samples to make the evaluation. A new Landsat-7 CDOM remote sensing model was established using these samples. The results show that the two satellites have good consistency for estimating low-level CDOM (R2 = 0.79, RMSE = 0.05 m–1, and MRE = 11.8%). This study shows that by combining Landsat-7 and Landsat-8 data, the available cloud-free observation frequency can be doubled compared with the frequency of using solo Landsat-8, and hence provide more detailed information for monitoring quality of complex water.

Estimating tree canopy cover using harmonic regression coefficients derived from multitemporal Landsat data

The goal of this study was to evaluate whether harmonic regression coefficients derived using all available cloud-free observations in a given Landsat pixel for a three-year period can be used to estimate tree canopy cover (TCC), and whether models developed using harmonic regression coefficients as predictor variables are better than models developed using median composite predictor variables, the previous operational standard for the National Land Cover Database (NLCD). The two study areas in the conterminous USA were as follows: West (Oregon), bounded by Landsat Worldwide Reference System 2 (WRS-2) paths/rows 43/30, 44/30, and 45/30; and South (Georgia/South Carolina), bounded by WRS-2 paths/rows 16/37, 17/37, and 18/37. Plot-specific tree canopy cover (the response variable) was collected by experienced interpreters using a dot grid overlaid on 1 m spatial resolution National Agricultural Imagery Program (NAIP) images at two different times per region, circa 2010 and circa 2014. Random forest model comparisons (using 500 independent model runs for each comparison) revealed the following (1) harmonic regression coefficients (one harmonic) are better predictors for every time/region of TCC than median composite focal means and standard deviations (across times/regions, mean increase in pseudo R2 of 6.7% and mean decrease in RMSE of 1.7% TCC) and (2) harmonic regression coefficients (one harmonic, from NDVI, SWIR1, and SWIR2), when added to the full suite of median composite and terrain variables used for the NLCD 2011 product, improve the quality of TCC models for every time/region (mean increase in pseudo R2 of 3.6% and mean decrease in RMSE of 1.0% TCC). The harmonic regression NDVI constant was always one of the top four most important predictors across times/regions, and is more correlated with TCC than the NDVI median composite focal mean. Eigen analysis revealed that there is little to no additional information in the full suite of predictor variables (47 bands) when compared to the harmonic regression coefficients alone (using NDVI, SWIR1, and SWIR2; 9 bands), a finding echoed by both model fit statistics and the resulting maps. We conclude that harmonic regression coefficients derived from Landsat (or, by extension, other comparable earth resource satellite data) can be used to map TCC, either alone or in combination with other TCC-related variables.

THE SPATIAL AND TEMPORAL AVAILABILITY DIFFERENCES OF CLOUD-FREE LANDSAT IMAGES OVER THREE GORGES RESERVOIR AREA

Abstract. Availability analysis of cloud-free optical remote sensing data is a prerequisite for remote sensing applications. In this study, spatio-temporal availability differences of cloud-free Landsat TM, ETM+, and OLI sensors images over Three Gorges Reservoir Area (TGRA) were analyzed from 1986 to 2019 based on the Google Earth Engine (GEE). The results show that: 1) in Summer, especially in August, the probabilities of obtaining Landsat images with no more than 30% cloud cover (CC) is higher. 2) the northeast of TGRA has higher probability of acquiring cloudless images than the southwest. 3) In TGRA, annual monitoring which require at least one cloud-free observation in a year largely unaffected by CC, but when considering seasonal monitoring, cloud contaminate will become a limitation, and monthly monitoring in this area is basically not feasible even if the three sensors data are combined. The results of this paper will provide important references for the research of using optical data in this area, and although the research area is relatively small, the analysis method and the program developed in this paper have no restrictions on the area.

User needs for future Landsat missions

Landsat satellites have been operating since 1972, providing the longest continuous observation record of the Earth's land surface. Over the past half century, the Landsat user community has grown exponentially, encompassing more diverse and evolving scientific research and operational uses. Understanding current and future user needs is crucial to informing the design of Landsat missions beyond Landsat 9. The U.S. Geological Survey (USGS) initiated a user needs collection process to document needs from U.S. Federal civil subject matter experts who rely on moderate-resolution land imaging data across a diverse range of scientific research and application domains. In total, 379 moderate-resolution land imaging user needs were collected through structured interviews. The findings indicate that, at present, users need continuity in Landsat capabilities with free and open data access. Improvements to future Landsat systems should include 10 m spatial resolution and at least weekly cloud-free observation frequency. Spectral enhancements should include the addition of red edge bands, and multiple, narrower visible, near infrared, shortwave infrared, and thermal infrared bands. Ideally, a variety of applications need continuous, full-spectrum coverage in 10 nm-wide bands spanning the visible to shortwave infrared (VSWIR) region (400–2500 nm) and 5 to 8 multispectral thermal infrared bands. Non-Federal (state, local, commercial, academic, and international) sources found similar results, but a more comprehensive comparison across these communities through a broader survey may provide additional insights. USGS-collected moderate-resolution land imaging user needs are an input to the Landsat 10 Architecture Study to develop and assess feasible Landsat 10 mission architectures.

Assessing the radiative impacts of an extreme desert dust outbreak and the potential improvements on short-term weather forecasts: The case of February 2015

On 1st and 2nd February 2015, the central and eastern Mediterranean, as well as parts of the Black Sea, were affected by an extreme desert dust outbreak. This exceptional dust event took place under the prevalence of a strong southwesterly airflow over the region, with maximum wind speeds up to 70 (36 ms−1) and 35 knots (18 ms−1) at 700 hPa and at mean sea-level pressure, respectively. Based on MODIS-Terra/Aqua aerosol optical depth (AOD) retrievals, the intensity of the transported dust loads maximized over the central Mediterranean (4.2) while very high AODs (up to 2.5) were recorded over Greece and the Black Sea. In addition, according to in-situ measurements obtained at Finokalia (Crete, southern Greece), the maximum PM10 concentrations reached up to 758 μgr m−3 (maximum levels for 2015). Through the implementation of the WRF-Chem model, the clear-sky direct radiative effects (DREs) have been computed for the SW, LW and NET (SW + LW) radiation, at the top of the atmosphere (TOA), within the atmosphere (ATM) as well as for the downwelling (SURF) and the absorbed (NETSURF) radiation at the ground. According to our simulations, during daytime, the instantaneous NET DREs reach down to −278.1 Wm−2 (Solar Zenith Angle (SZA) = 53°) and − 252.2 Wm−2 for SURF and NETSURF, respectively, indicating a strong surface cooling, while the corresponding values for ATM can be as large as 147.3 Wm−2 (SZA = 50°) revealing a strong atmospheric warming. At TOA, the computed DREs vary from −120.1 Wm−2 (planetary cooling; SZA = 51°) to 59.4 Wm−2 (planetary warming; SZA = 64°) being negative and positive over dark and bright surfaces, respectively. During nighttime, reverse effects of lower magnitude are found at all levels of the Earth-Atmosphere system. At a regional scale, throughout the forecast period, the clear-sky NET DREs range from −16.60 Wm−2 to −0.76 Wm−2, from −46.18 Wm−2 to 16.11 Wm−2, from −32.31 Wm−2 to 13.70 Wm−2 and from −14.63 Wm−2 to 20.40 Wm−2 for TOA, SURF, NETSURF and ATM, respectively. As a response to the perturbation of the surface radiation budget, the 2 m air temperature decreases/increases by up to 1.5 °C (5.3 °C) during daytime/nighttime. Moreover, at noon, the low-level dust layer (up to 3 km above sea level), cools (heats) the lowest troposphere by up to 0.5 °C (1.4 °C) over land (sea). At night, the temperature of the atmospheric layers where the dust aerosols are confined decreases (by up to −1.8 °C), whereas a warming effect (by up to 2.8 °C) is evident in the air masses beneath the dust layer. Through the inclusion of dust-radiation interactions in the numerical simulations, the model's predictive skill in terms of reproducing the downwelling SW radiation at the ground is improved. This is justified via the comparison of the RADON (dust-radiation interactions are activated) and RADOFF (dust-radiation interactions are deactivated) outputs against cloud-free observations derived by 31 NOAAN stations. More specifically, from RADOFF to RADON, the bias reduces from 157.63 to 26.02 Wm−2 and the slope decreases from 1.25 to 0.98. On the contrary, the evaluation analysis of the temperature at 2 m does not reveal a remarkably better performance of the RADON run, particularly during sunlight hours, reflecting thus the predominance of first order model errors over the expected improvements attributed to dust-radiation interactions.