Surface Reflectance Data Research Articles

High Spatial Resolution Leaf Area Index Estimation for Woodland in Saihanba Forestry Center, China

Owing to advancements in satellite remote sensing technology, the acquisition of global land surface parameters, notably, the leaf area index (LAI), has become increasingly accessible. The Sentinel-2 (S2) satellite plays an important role in the monitoring of ecological environments and resource management. The prevalent use of the 20 m spatial resolution band in S2-based inversion models imposes significant limitations on the applicability of S2 data in applications requiring finer spatial resolution. Furthermore, although a substantial body of research on LAI retrieval using S2 data concentrates on agricultural landscapes, studies dedicated to forest ecosystems, although increasing, remain relatively less prevalent. This study aims to establish a viable methodology for retrieving 10 m resolution LAI data in forested regions. The empirical model of the soil adjusted vegetation index (SAVI), the backpack neural network based on simulated annealing (SA-BP) algorithm, and the variational heteroscedastic Gaussian process regression (VHGPR) model are established in this experiment based on the LAI data measured and the corresponding 10 m spatial resolution S2 satellite surface reflectance data in the Saihanba Forestry Center (SFC). The LAI retrieval performance of the three models is then validated using field data, and the error sources of the best performing VHGPR models (R2 of 0.8696 and RMSE of 0.5078) are further analyzed. Moreover, the VHGPR model stands out for its capacity to quantify the uncertainty in LAI estimation, presenting a notable advantage in assessing the significance of input data, eliminating redundant bands, and being well suited for uncertainty estimation. This feature is particularly valuable in generating accurate LAI products, especially in regions characterized by diverse forest compositions.

Cross-comparison of Landsat-8 and Landsat-9 data: a three-level approach based on underfly images

ABSTRACT The recently launched Landsat-9 has an important mission of working together with Landsat-8 to reduce the revisit period of Landsat Earth observations to eight days. This requires the data of Landsat-9 to be highly consistent with that of Landsat-8 to avoid bias caused by data inconsistency when the two satellites are simultaneously used. Therefore, this study evaluated the consistency of the surface reflectance (SR) and land surface temperature (LST) data between Landsat-8 and Landsat-9 based on five test sites from different parts of the world using synchronized underfly image pairs of both satellites. Previous cross-comparisons have demonstrated high consistency between the spectral bands of Landsat-8 and Landsat-9, with differences of around 1%. However, it is unclear whether this low deviation will be amplified in subsequent multiband calculations. It is also necessary to determine whether the difference is consistent across different land cover types. Therefore, this study used a three-level cross-comparison approach to specifically examine these concerns. Besides the commonly used band-by-band comparison, which served as the first-level comparison in this study, this approach included a second-level comparison based on the calculations of several indicators and a third-level comparison based on a composite index calculated from the indicators obtained in the second-level comparison. This three-level approach will examine whether the difference found in the first-level per-band comparison would change after the subsequent calculations in the second- and third-level comparisons. The Remote Sensing based Ecological Index (RSEI) was used for this approach because it is a composite index integrating four indicators. The results of this three-level comparison show that the first-level per-band comparison exhibited high consistency between the two satellites’ SR data, with an average absolute percent change (PC) of 1.88% and an average R 2 of 0.957 across six bands in the five test sites. This deviation increased to 2.21% in the third-level composite index-based comparison, with R 2 decreasing to 0.956. This indicates that after complex calculations, the deviation between the bands of the two satellites was amplified to some extent. However, when analyzing specific land cover types, notable differences emerged between the two satellites for the water category, with an average absolute PC ranging from 18% to 35% and an R 2 of lower than 0.6. Additionally, there were also nearly 5% differences for the built-up land category, with an average R 2 value of lower than 0.7. The comparison of LST data between both satellites also reveals that the Landsat-9 LST is on average 0.24°C lower than Landsat-8 LST across the five test areas but can be 0.58°C lower in built-up land-dominated areas and 0.42°C higher in desert environments. Overall, the SR and LST data between Landsat-8 and Landsat-9 are consistent. However, their performance varies depending on different land cover types. Caution is needed particularly for water-related research when utilizing both satellites simultaneously. Significant discrepancies may also arise in the areas characterized by deserts and built-up lands.

Impact of large kernel size on yield prediction: a case study of corn yield prediction with SEDLA in the U.S. Corn Belt

Predicting agricultural yields is imperative for effective planning to sustain the growing global population. Traditionally, regression-based, simulation-based, and hybrid methods were employed for yield prediction. In recent times, there has been a notable shift towards the adoption of Machine Learning (ML) methods, with Deep Learning (DL), particularly Convolutional Neural Networks (CNNs) and Long-Short Term Memory (LSTM) networks, emerging as popular choices for their enhanced predictive accuracy. This research introduces a cost-effective DL architecture tailored for corn yield prediction, considering computational efficiency in processing time, data size, and NN architecture complexity. The proposed architecture, named SEDLA (Simple and Efficient Deep Learning Architecture), leverages the spatial and temporal learning capabilities of CNNs and LSTMs, respectively, with a unique emphasis on exploring the impact of kernel size in CNNs. Simultaneously, the study aims to exclusively employ satellite and yield data, strategically minimizing input variables to enhance the model’s simplicity and efficiency. Notably, the study demonstrates that employing larger kernel sizes in CNNs, especially when processing histogram-based Surface Reflectance (SR) and Land Surface Temperature (LST) data from Moderate Resolution Imaging Spectroradiometer (MODIS), allows for a reduction in the number of hidden layers. The efficacy of the architecture was evaluated through extensive testing on corn yield prediction across 13 states in the United States (U.S.) Corn Belt at county-level. The experimental results showcase the superiority of the proposed architecture, achieving a Mean Absolute Percentage Error (MAPE) of 6.71 and Root Mean Square Error (RMSE) of 14.34, utilizing a single-layer CNN with a 15 × 15 kernel in conjunction with LSTM. These outcomes surpass existing benchmarks in the literature, affirming the efficacy and potential of the suggested DL framework for accurate and efficient crop yield predictions.

Assessing Forest Species Diversity in Ghana’s Tropical Forest Using PlanetScope Data

This study utilized a remotely sensed dataset with a high spatial resolution of 3 m to predict species diversity in the Bobiri Forest Reserve (BFR), a moist semi-deciduous tropical forest in Ghana. We conducted a field campaign of tree species measurements to achieve this objective for species diversity estimation. Thirty-five field plots of 50 m × 20 m were established, and the most dominant tree species within the forest were identified. Other measurements, such as diameter at breast height (DBH ≥ 5 cm), tree height, and each plot’s GPS coordinates, were recorded. The following species diversity indices were estimated from the field measurements: Shannon–Wiener (H′), Simpson diversity index (D2), species richness (S), and species evenness (J′). The PlanetScope surface reflectance data at 3 m spatial resolution was acquired and preprocessed for species diversity prediction. The spectral/pixel information of all bands, except the coastal band, was extracted for further processing. Vegetation indices (VIs) (NDVI—normalized difference vegetation index, EVI—enhanced vegetation index, SRI—simple ratio index, SAVI—soil adjusted vegetation index, and NDRE—normalized difference red edge index) were also calculated from the spectral bands and their pixel value extracted. A correlation analysis was then performed between the spectral bands and VIs with the species diversity index. The results showed that spectral bands 6 (red) and 2 (blue) significantly correlated with the two main species diversity indices (S and H′) due to their influence on vegetation properties, such as canopy biomass and leaf chlorophyll content. Furthermore, we conducted a stepwise regression analysis to investigate the most important spectral bands to consider when estimating species diversity from the PlanetScope satellite data. Like the correlation results, bands 6 (red) and 2 (blue) were the most important bands to be considered for predicting species diversity. The model equations from the stepwise regression were used to predict tree species diversity. Overall, the study’s findings emphasize the relevance of remotely sensed data in assessing the ecological condition of protected areas, a tool for decision-making in biodiversity conservation.

National-scale remotely sensed lake trophic state from 1984 through 2020

Lake trophic state is a key ecosystem property that integrates a lake’s physical, chemical, and biological processes. Despite the importance of trophic state as a gauge of lake water quality, standardized and machine-readable observations are uncommon. Remote sensing presents an opportunity to detect and analyze lake trophic state with reproducible, robust methods across time and space. We used Landsat surface reflectance data to create the first compendium of annual lake trophic state for 55,662 lakes of at least 10 ha in area throughout the contiguous United States from 1984 through 2020. The dataset was constructed with FAIR data principles (Findable, Accessible, Interoperable, and Reproducible) in mind, where data are publicly available, relational keys from parent datasets are retained, and all data wrangling and modeling routines are scripted for future reuse. Together, this resource offers critical data to address basic and applied research questions about lake water quality at a suite of spatial and temporal scales.

Global 500 m seamless dataset (2000–2022) of land surface reflectance generated from MODIS products

Abstract. The Moderate Resolution Imaging Spectroradiometer (MODIS) is widely utilized for retrieving land surface reflectance to reflect plant conditions, detect ecosystem phenology, monitor forest fires, and constrain terrestrial energy budgets. However, the state-of-the-art MODIS surface reflectance products suffer from temporal and spatial gaps due to atmospheric conditions (e.g. clouds and aerosols), limiting their use in ecological, agricultural, and environmental studies. Therefore, there is a need for reconstructing spatiotemporally seamless (i.e. gap-filled) surface reflectance data from MODIS products, which is difficult due to the intrinsic inconsistency of observations resulting from various sun/view geometry and the prolonged missing values resulting from polar night or heavy cloud coverage, especially in monsoon season. We built a framework for generating the global 500 m daily seamless data cubes (SDC500) based on MODIS surface reflectance dataset, which contains the generation of a land-cover-based a priori database, bidirectional reflectance distribution function (BRDF) correction, outlier detection, gap filling, and smoothing. The first global spatiotemporally seamless land surface reflectance at 500 m resolution was produced, covering the period from 2000 to 2022. Preliminary evaluation of the dataset at 12 sites worldwide with different land cover demonstrated its robust performance. The quantitative assessment shows that the SDC500 gap-filling results have a root-mean-square error (RMSE) of 0.0496 and a mean absolute error (MAE) of 0.0430. The SDC500 BRDF correction results showed an RMSE of 0.056 and a bias of −0.0085 when compared with MODIS nadir BRDF-adjusted reflectance (NBAR) products, indicating the acceptable accuracy of both products. From a temporal perspective, the SDC500 eliminates abnormal fluctuations while retaining the useful localized feature of rapid disturbances. From a spatial perspective, the SDC500 shows satisfactory spatial continuity. In conclusion, the SDC500 is a well-processed global daily surface reflectance product, which can serve as the fundamental input for large-scale ecological, agricultural, and environmental applications and quantitative remote sensing studies. The SDC500 is available at http://data.starcloud.pcl.ac.cn/resource/27 (Liang et al., 2023b) or https://doi.org/10.12436/SDC500.27.20230701 (Liang et al., 2023a).

青藏高原色林错2000–2021年湖冰数据集

Lake ice is a sensitive indicator of climate change. Research on the response mechanism between lake ice and climate offer valuable insights into the mechanisms driving global climate change and enhance our ability to predict climate patterns. Using the Landsat surface albedo products, we delineated the boundaries of Lake Selin Co from 2000 to 2021; and by leveraging the MODIS Terra surface reflectance data with a 250 m resolution and daily revisit time, we generated a dataset of lake ice distribution in Lake Selin Co from 2000 to 2021. We selected the results of the freezing period and the complete freezing period respectively, compared them with the visual results to perform accuracy verification and calculate the confusion matrix. The overall accuracy of the user reached 94%, and the Kappa coefficient exceeded 0.89. Continuous monitoring data for lake ice phenology in Selin Co are of great significance for tracking climate change, which help fill gaps in measured data and provide valuable support for related research.

Continuous Time-Series Land Cover Maps for South Korea using Google Earth Engine

In this study, we utilized Google Earth Engine to construct, for the first time in South Korea, a long-term (1986-2021) continuous time series of land cover maps with a spatial resolution of 30 meters. Derived from the surface reflectance data of the Landsat satellite series, a total of 44 input variables were generated, including various spectral bands and indices related to land cover. For accuracy verification of the maps, 4,824 reference data were established using areas where land cover remained unchanged, identified by comparing the most recent (2018) and historical (1988) land cover maps from the Ministry of Environment. The Random Forest model was employed to classify seven land cover types (settlements, cropland, forest land, grassland, wetlands, bare land, and water bodies), with an overall accuracy of 0.97 and a Macro F1-score of 0.91, indicating a generally high performance of the model. However, considering the annual variability, potentially due to unidentified or untraceable errors, a composite land cover map dataset, integrated in five-year intervals, was suggested to ensure the generation of stable data.

Retrieving BRDFs from UAV-based radiometers for fiducial reference measurements: caveats and recommendations

Surface Bidirectional reflectance distribution function (BRDF) is a key intrinsic geophysical variable depending only on the characteristics of the observed medium. It is therefore the most suitable measurand to support the definition of fiducial reference measurements (FRM). Field acquisition of surface reflectance data relies on substantial assumptions and simplifications, often without accounting for their impact. For example, the BRDF is a theoretical concept and can never be measured in the field. In contrast, the hemispherical conical reflectance factor (HCRF), which is the measurand obtained during field campaigns, is impacted by all scene elements and is not intrinsic to the surface. This study analyses the impact of four parameters (atmospheric scattering, measurement device field of view cropping, acquisition duration, non-Lambertian reference panels) on HCRF estimation. Simulations are performed on a 3D vegetation scene, using the new radiative transfer model Eradiate. It is found that among the aforementioned parameters, atmospheric scattering alone leads to a relative root-mean-square error (RRMSE) of more than 10% between HCRF and reference Bidirectional reflectance factor (BRF).

GF-1 WFV Surface Reflectance Quality Evaluation in Countries along “the Belt and Road”

The GaoFen-1 wide field of view (GF-1 WFV) has produced level 1 digital number data globally; however, most applications have focused on China, and data quality outside China has not been validated. This study presents a preliminary assessment of the 2020 GF-1 WFV surface reflectance data for Nepal, Azerbaijan, Kenya, and Sri Lanka along “the Belt and Road” route using Sentinel-2 Multi-Spectral Instrument (MSI), Landsat-8 Operational Land Image (OLI), and Moderate Resolution Imaging Spectroradiometer (MODIS) data. A method for obtaining the GF-1 WFV surface reflectance data was also proposed, with steps including atmospheric correction, cross-radiation calibration, and bidirectional reflectance distribution function correction. The results showed that WFV surface reflectance data was not significantly different from MSI, OLI, and MODIS surface reflectance data. In the visible and near-infrared bands, for most landcover types, the bias was less than 0.02, and the precision and root mean square error were less than 0.04. When the landcover types were forest and water, the MSI, OLI, and MODIS surface reflectance data were higher than that of WFV in the near-infrared band. The results of this study provide a basis for assessing the global application potential of GF-1 WFV.

Improved global 250 m 8-day NDVI and EVI products from 2000–2021 using the LSTM model

Satellite vegetation index (VI) products, such as normalized difference vegetation index (NDVI) and enhanced vegetation index (EVI), have been widely used. However, they are severely contaminated by clouds and other factors and provide false signals of the surface vegetation conditions. In this study, the new global seamless 250 m, eight-day NDVI and EVI products from 2000–2021 were developed from Moderate Resolution Imaging Spectroradiometer (MODIS) surface reflectance data using a long short-term memory (LSTM) neural network method. High-quality globally representative time series VI samples were constructed to train the model using a combination of the Savitzky-Golay filter (SG), Global LAnd Surface Satellite (GLASS) leaf area index (LAI) fitting and upper envelope methods. To evaluate the proposed method and the 250 m VI products, the MODIS VI product (MOD13Q1) was used for the inter-comparisons using four widely used VI reconstruction methods. Assuming that the MODIS VI data of high quality represents the true values, the root mean square error (RMSE) for NDVI and EVI generated by the LSTM model are 0.0734 and 0.0509, respectively.

The Search for Topographic Correlations within the Reiner Gamma Swirl

Lunar swirls have been traditionally considered to be unaffected by topographic changes. Yet, correlations between areas of high albedo and lower elevations are observed in regions of the Mare Ingenii swirl. Here, we apply similar techniques used at Mare Ingenii to determine if correlations between swirl units and topography also exist within Reiner Gamma. We generate topography using the techniques of stereophotoclinometry to Lunar Reconnaissance Orbiter Camera Narrow Angle Camera images to derive meter- to submeter-scale topography. We choose a 50 km2 study region with a 2.6 m ground sample distance (GSD), and within this region there is a 1 km2 subregion with a 0.8 m GSD. We use surface reflectance data at multiple viewing geometries to classify these regions into different swirl units using machine learning techniques. Statistical analyses of the data show mean height variations between on- and off-swirl of ∼4 m, with on-swirl at a lower elevation. It is not clear how this scale of elevation difference influences the formation of swirl units, but it supports postulations of dust migration and magnetic sorting contributing to their formation.

A dataset of JLS-5-20M black soil image in Northeast China

This is a dataset of the surface reflectance data covering some black soil areas in the three northeastern provinces of China, based on the L1 data of the JL1GP satellites’ multispectral imager with a spatial resolution of 5-20 meters. Specifically, we adopted the on-orbit absolute calibration technology based on multiple stable sites to ensure that the on-orbit absolute radiometric calibration accuracy of the JL1GP satellites’ multispectral imagers is no less than 7%. Then, we selected the L1 data of JL1GP satellite multispectral imager that imaged in the three provinces of Northeast China with clear weather and no clouds are as the input image of atmospheric correction process. Finally, the L1-level image data are atmospherically corrected and orthorectified by the atmospheric correction algorithm based on the atmospheric radiative transfer model, to generate the reflectance product dataset covering some black soil areas in the three northeastern provinces of China with an absolute accuracy of no less than 7%. This dataset takes advantage of the domestic JL1GP satellites’ advantages of high spatial resolution, a large number of bands, and stable imaging quality, hence it can provide reliable support for quantitative remote sensing applications in the black soil areas of the three northeastern provinces in China.

Mapping Grassland Based on Bio-Climate Probability and Intra-Annual Time-Series Abundance Data of Vegetation Habitats

Accurate inventories of grasslands are important for studies of greenhouse gas (GHG) dynamics, as grasslands store about one-third of the global terrestrial carbon stocks. This paper develops a framework for large-area grassland mapping based on the probability of grassland occurrence and the interactive pathways of fractional vegetation and soil-related endmember nexuses. In this study, grassland occurrence probability maps were produced based on data on bio-climate factors obtained from MODIS/Terra Land Surface Temperature (MOD11A2), MODIS/Terra Vegetation Indices (MOD13A3), and Tropical Rainfall Measuring Mission (TRMM 3B43) using the random forests (RF) method. Time series of 8-day fractional vegetation-related endmembers (green vegetation, non-photosynthetic vegetation, sand land, saline land, and dark surfaces) were generated using linear spectral mixture analysis (LSMA) based on MODIS/Terra Surface Reflectance data (MOD09A1). Time-series endmember fraction maps and grassland occurrence probabilities were employed to map grassland distribution using an RF model. This approach improved the accuracy by 5% compared to using endmember fractions alone. Additionally, based on the grassland occurrence probability maps, we identified extensive ecologically sensitive regions, encompassing 1.54 (104 km2) of desert-to-steppe (D-S) and 2.34 (104 km2) of steppe-to-meadow (S-M) transition regions. Among these, the D-S area is located near the threshold of 310 mm/yr in precipitation, an annual temperature of 10.16 °C, and a surface comprehensive drought index (TVPDI) of 0.59. The S-M area is situated close to the line of 437 mm/yr in precipitation, an annual temperature of 5.49 °C, and a TVPDI of 0.83.

Generating daily 100 m resolution land surface temperature estimates continentally using an unbiased spatiotemporal fusion approach

Fine spatial resolution (i.e., ≤ 100 m) land surface temperature (LST) data are crucial to study heterogeneous landscapes (e.g., agricultural and urban). Some well-known spatiotemporal fusion methods like the Spatial and Temporal Adaptive Reflectance Fusion Model (STARFM) and the Enhanced STARFM (ESTARFM), which were originally developed to fuse surface reflectance data, may not be suitable for direct application in LST studies due to the high sub-diurnal dynamics of LST. Furthermore, the effectiveness of spatiotemporal fusion methods for LST data has not been thoroughly evaluated in previous studies that only focused on relatively small spatiotemporal extents. To address these limitations, we proposed a variant of ESTARFM, referred to as the unbiased ESTARFM (ubESTARFM), specifically designed to accommodate the high temporal dynamics of LST to generate fine-resolution LST estimates. We evaluated ubESTARFM and ESTARFM against in-situ LST and the ECOsystem Spaceborne Thermal Radiometer Experiment on Space Station (ECOSTRESS) LST across 12 regions throughout Australia, encompassing various land covers and environments. Independent validation showed that ubESTARFM had a bias of 2.55 K, unbiased root mean squared error (ubRMSE) of 2.57 K, and Pearson correlation coefficient (R) of 0.95 against the in-situ LST over 11,290 observations at the 12 sites, all of which were considerably better than those calculated for ESTARFM, being a bias of 4.73 K, ubRMSE of 3.80 K and R of 0.92. When compared to ECOSTRESS data, ubESTARFM LST had a bias of −1.69 K, ubRMSE of 2.00 K, and R of 0.70 over 43 near clear-sky scenes, while ESTARFM LST had a bias of 1.79 K, ubRMSE of 2.68 K, and R of 0.59. Overall, our results demonstrated that ubESTARFM can avoid systematic bias accumulation, substantially reduce uncertainty deviation, and maintain a good level of correlation with validation datasets when compared to ESTARFM. A further assessment underscored the potential of ubESTARFM for application using LST data acquired from geostationary platforms (e.g., Himawari-8), with a mean ubRMSE (R) of 2.22 K (0.97) against in-situ LST over 1327 observations at 3 sites from southeast Australia at the overpass time of MODIS/Terra. This promising method leverages reliable numeric values from coarse-resolution LST while borrowing spatial heterogeneity from fine-resolution LST and has the potential to be coupled with energy balance and/or radiative transfer models thus enabling better farm and/or regional-scale water management strategies to be implemented. Furthermore, both the input and generated LST data, encompassing a comprehensive spatial extent over diverse land covers and climatic conditions, are publicly available for benchmarking future algorithmic refinements.

Land Use and Land Cover (LULC) Assessment within the Batanes Protected Landscapes and Seascapes

Declared protected areas have ecologically important landscapes that must be conserved and protected. Status of protected areas could be monitored through land use and land cover (LULC) assessments. LULC offers baseline data for integrated land use planning and improvement of existing policies are therefore necessary to be conducted. This study was conducted to monitor the existing LULC of six islands within the Batanes Protected Landscapes and Seascapes (BPLS) through a machine learning (ML)-based random forest (RF) classifier using multi-sourced data such as Landsat imageries’ surface reflectance (SR), Landsat-derived land surface temperature (LST), and global ecosystem dynamic investigation (GEDI)-derived height (Ht) metrics and to determine the effects of the LST and Ht metrics to LULC classification. Four layer stacked images with different features were analyzed – including SR, SR-LST, SR-Ht, and SR-LST-Ht. The result of the LULC classification showed an accuracy based on Macro F1-score and Kappa (K) of 0.81 and 0.83, 0.83and 0.86, 0.86 and 0.89, and 0.93 and 0.94, for SR, SR-LST, SR-Ht, and SR-LST-Ht, respectively. When compared to the existing global-scale LULC, this study has higher accuracy than the GLAD and ESRI products, which have Macro F1-scores and K-values of 0.73 and 0.71, and 0.59 and 0.64, respectively. To conclude, the inclusion of LST and Ht information in addition to SR data in LULC classification can improve the accuracy by up to 12% and 11% based on Macro F1-score and K,respectively. The result of this study can serve as a reference for achieving improved and reliable LULC information that is necessary for monitoring fluctuations of the global earth’s resources and comprehensive LULC planning. In addition, the technique used in this study can serve as a reference in generating reliable LULC information that can aid in the sustainable implementation of policies, rules, and regulations intended for declared protected areas like BPLS.

Enhanced crop classification through integrated optical and SAR data: a deep learning approach for multi-source image fusion

ABSTRACT Agricultural crop mapping has advanced over the last decades due to improved approaches and the increased availability of image datasets at various spatial and temporal resolutions. Considering the spatial and temporal dynamics of different crops during a growing season, multi-temporal classification frameworks are well-suited for mapping crops at large scales. Addressing the challenges posed by imbalanced class distribution, our approach combines the strengths of different deep learning models in an ensemble learning framework, enabling more accurate and robust classification by capitalizing on their complementary capabilities. This research aims to enhance the crop classification of maize, soybean, and wheat in Bei’an County, Northeast China, by developing a novel deep learning architecture that combines a three-dimensional convolutional neural network (3D-CNN) with a variant of convolutional recurrent neural networks (ConvRNN). The proposed method integrates multi-temporal Sentinel-1 polarimetric features with Sentinel-2 surface reflectance data for multi-source fusion and achieves an overall accuracy of 91.7%, a Kappa coefficient of 85.7%, and F1 scores of 93.7%, 92.2%, and 90.9% for maize, soybean, and wheat, respectively. Our proposed model is also compared with alternative data augmentation techniques, maintaining the highest mean F1 score (87.7%). The best performer was weakly supervised with ten per cent of ground truth data collected in Bei’an in 2017 and used to produce an annual crop map for measuring the model’s generalizability. The model learning reliability of the proposed method is interpreted through the visualization of model soft outputs and saliency maps.

Remote sensing-enabled machine learning for river water quality modeling under multidimensional uncertainty

Two fundamental problems have inhibited progress in the simulation of river water quality under climate (and other) uncertainty: 1) insufficient data, and 2) the inability of existing models to account for the complexity of factors (e.g., hydro-climatic, basin characteristics, land use features) affecting river water quality. To address these concerns this study presents a technique for augmenting limited ground-based observations of water quality variables with remote-sensed surface reflectance data by leveraging a machine learning model capable of accommodating the multidimensionality of water quality influences. Total Suspended Solids (TSS) can serve as a surrogate for chemical and biological pollutants of concern in surface water bodies. Historically, TSS data collection in the United States has been limited to the location of water treatment plants where state or federal agencies conduct regularly-scheduled water sampling. Mathematical models relating riverine TSS concentration to the explanatory factors have therefore been limited and the relationships between climate extremes and water contamination events have not been effectively diagnosed. This paper presents a method to identify these issues by utilizing a Long Short-Term Memory Network (LSTM) model trained on Moderate Resolution Imaging Spectroradiometer (MODIS) satellite reflectance data, which is calibrated to TSS data collected by the Ohio River Valley Water Sanitation Commission (ORSANCO). The methodology developed enables a thorough empirical analysis and data-driven algorithms able to account for spatial variability within the watershed and provide effective water quality prediction under uncertainty.

MODIS data assessment of the relationship between vegetation and land surface temperature for central vegetation zone of Adamawa State, Nigeria

Global concern over climate change has necessitated the need for the study of the impact of the ever-increasing land surface temperature on the greenness of land covers. Grassland phenological events are dependent on land surface temperature. This work is aimed at assessing the spatio-temporal trend of the relationship between vegetation cover and the day and night-times land surface temperatures in the Northern Guinea Savanna vegetation zone of Adamawa state, Nigeria. Twenty years (2000 to 2019) data series of Moderate Resolution Imaging Spectroradiometer (MODIS) surface reflectance data (MOD09A1), the day-time land surface temperature data (MOD11A), and the night-time land surface temperature (MYOD11A1) were used to spatially compute the Normalized Difference Vegetation Index (NDVI) and day-time and night-time Land Surface Temperatures (LST) for each of the respective years of the study period. The day- and night-time LST were correlated with the NDVI using simple linear regression analysis. The temporal analysis of the results indicate nonlinear changes in both the NDVI, day-time LST, and night-time LST. The results also indicate a negative correlation between LST and NDVI in both cases. Regression analyses of the day-time LST as a function of NDVI revealed a vegetation cover reduction of 0.0063 per Kelvin change in LST (R2 = 0.3026, p > 0005, RMSE = 8.1042) and a night-time vegetation cover reduction of 0.0103 per Kelvin rise in LST (R2 = 0.3697, p > 0.05, RMSE = 3.7511). The relatively low coefficient of determination and the p values implies a weak relationship between the two variables. The parameters however indicate that the night-time LST is a better predictor of the vegetation greenness of the area. The results revealed that LST is a weak determinant of plant phenological events with respect to the study area.

Calibration of the SNPP and NOAA 20 VIIRS sensors for continuity of the MODIS climate data records

Accurate long-term sensor calibration and periodic re-processing to ensure consistency and continuity of atmospheric, land and ocean geophysical retrievals from space within the mission period and across different missions is a major requirement of climate data records. In this work, we applied the Multi-Angle Implementation of Atmospheric Correction (MAIAC)-based vicarious calibration technique over Libya-4 desert site to perform calibration analysis of Visible Infrared Imaging Radiometer Suite (VIIRS) on Suomi National Polar-orbiting Partnership (SNPP) and NOAA-20 satellites. For both VIIRS sensors we characterized residual linear calibration trends and cross-calibrated both sensors to MODerate resolution Imaging Spectroradiometer (MODIS) Aqua regarded as a calibration standard. The relative spectral response (RSR) differences were accounted for using the German Aerospace Center (DLR) Earth Sensing Imaging Spectrometer (DESIS) hyperspectral surface reflectance data. Our results agree with independent vicarious calibration results of both the MODIS/VIIRS Characterization Support Team as well as the CERES Imager and Geostationary Calibration Group within estimated uncertainty of 1–2%. Analysis of MAIAC geophysical products with the new calibration shows a high level of agreement of MAIAC aerosol, surface reflectance and NDVI records between MODIS and VIIRS. Excluding high aerosol optical depth (AOD), all three sensors agree in AOD with mean difference (MD) less than 0.01 and residual mean squared difference rmsd ∼ 0.04. Spectral geometrically normalized surface reflectance agrees within rmsd of 0.003–0.005 in the visible and 0.01–0.012 at longer wavelengths. The residual surface reflectance differences are fully explained by differences in spectral filter functions. Finally, difference in NDVI is characterized by rmsd ∼ 0.02 and MD less than 0.003 for NDVI based on VIIRS imagery bands I1/I2 and less than 0.01 for NDVI based on VIIRS radiometric bands M5/M7. In practical sense, these numbers indicate consistency and continuity in MAIAC records ensuring the smooth transition from MODIS to VIIRS.