Sentinel-2 Satellite Data Research Articles

DEVELOPMENT OF AUTOMATED SATELLITE DATA DOWNLOADING AND PROCESSING PIPELINE ON AWS CLOUD FOR NEAR-REAL-TIME AGRICULTURE APPLICATIONS

Abstract. Remote sensing satellites allow users to acquire detailed information about the Earth's surface on a temporal basis. Widen time-series analysis at a large geographical scale involves a huge amount (in Terabytes) of satellite data downloading and processing operations. Such processes need good computational power, large storage, and sophisticated tools. Maintaining such infrastructure can cost heavily to the research/commercial enterprises. To overcome such issues, Amazon Web Service (AWS) offers a sophisticated cloud computing environment. We developed an in-house automated satellite data downloading and processing (ADDPro) pipeline on the AWS platform. The ADDPro pipeline employed Sentinel-2 satellite data to offer current and relative vegetation health information of the agriculture region on a temporal basis at the pan-India scale. Image compositing and multi-sensor data fusion technique have been incorporated into the ADDPro pipeline to produce cloud-free raster (GeoTIFF) outputs. ADDPro pipeline also facilitates lossless raster data compression, which reduces AWS data transfer costs between regions. Data compression also aids in reducing raster publishing time on GeoServer. Operationally, AWS allows users to download only the bands required to generate a certain index (e.g. NDVI) rather than the entire Sentinel-2 data package. The entire ADDPro pipeline is extremely cost-effective, efficient, and scalable.

Mapping land-use intensity of grasslands in Germany with machine learning and Sentinel-2 time series

Information on grassland land-use intensity (LUI) is crucial for understanding trends and dynamics in biodiversity, ecosystem functioning, earth system science and environmental monitoring. LUI is a major driver for numerous environmental processes and indicators, such as primary production, nitrogen deposition and resilience to climate extremes. However, large extent, high resolution data on grassland LUI is rare. New satellite generations, such as Copernicus Sentinel-2, enable a spatially comprehensive detection of the mainly subtle changes induced by land-use intensification by their fine spatial and temporal resolution. We developed a methodology quantifying key parameters of grassland LUI such as grazing intensity, mowing frequency and fertiliser application across Germany using Convolutional Neural Networks (CNN) on Sentinel-2 satellite data with 20 m × 20 m spatial resolution. Subsequently, these land-use components were used to calculate a continuous LUI index. Predictions of LUI and its components were validated using comprehensive in situ grassland management data. A feature contribution analysis using Shapley values substantiates the applicability of the methodology by revealing a high relevance of springtime satellite observations and spectral bands related to vegetation health and structure. We achieved an overall classification accuracy of up to 66% for grazing intensity, 68% for mowing, 85% for fertilisation and an r2 of 0.82 for subsequently depicting LUI. We evaluated the methodology's robustness with a spatial 3-fold cross-validation by training and predicting on geographically distinctly separated regions. Spatial transferability was assessed by delineating the models' area of applicability. The presented methodology enables a high resolution, large extent mapping of land-use intensity of grasslands.

Synergic use of neural networks model and remote sensing algorithms to estimate water clarity indicators in Khanpur reservoir, Pakistan

Freshwater reservoirs are limited and facing issues of over-exploitation, climate change effects and poor maintenance which have serious consequences for water quality. Developing countries face the challenge of collecting in situ information on ecological status and water quality of these reservoirs due to constraints of cost, time and infrastructure. In this study, a practical method of retrieval of two water clarity indicators, total suspended matter and secchi disk depth, using Sentinel-2 satellite data is adopted for preliminary assessment of water quality and trophic conditions in Khanpur reservoir, Pakistan. The study explores the synergy of utilizing two independent models, i.e., case 2 regional coast color analytical neural network model and semiempirical remote sensing algorithms to understand the spatiotemporal dynamics of water clarity patterns in the dammed reservoir, in the absence of ground measurements. The drinking water quality and trophic state of the reservoir water is determined based purely on satellite measurements. Out of the five months studied, the reservoir water has high turbidity and poor eutrophic status in three months. The results from both computational models are compared, which exhibit a high degree of statistical agreement. The study demonstrates the effective utilization of relevant analytical and semiempirical methods on satellite data to map water clarity indicators and understand their dynamics in both space and time. This solution is particularly useful for regions where routine ground sampling and observation of environmental variables are absent.

Analysing the change in water quality parameters along river Ganga at Varanasi, Mirzapur and Ghazipur using Sentinel-2 and Landsat-8 satellite data during pre-lockdown, lockdown and post-lockdown associated with COVID-19

The proper functioning of the river ecosystem has been symbolised by healthy aquatic life. The river Ganga has shown signs of rejuvenation due to lockdown. In this study, an attempt has been made to analyse the change in river water quality using Sentinel-2 and Landsat-8 imageries. The quantitative analysis has been performed for temperature and normalised difference turbidity index (NDTI). The qualitative analysis has been performed for pH, dissolved oxygen (DO) and total suspended solids (TSSs). Ghazipur, Varanasi and Mirzapur stretches have been selected for this study. In the Ghazipur stretch, the river temperature decreased by 7.14% in May 2020 (lockdown period) as compared to May 2019 (1 year before lockdown). Similarly, in the Varanasi stretch, this decrease has been by 8.62%, and in the Mirzapur stretch, this decrease has been by 12.06% in May 2020 compared to May 2019. For the same period, NDTI in the Ghazipur, Varanasi and Mirzapur stretch has been decreased by 0.22, 0.26 and 0.24, respectively. The pH and DO of the river increased, and TSS decreased for the considered time period. The lockdown during the second wave of the coronavirus disease 2019 was not helpful for river rejuvenation. This study elicited how the behaviour of the parameters changed during the lockdown.Research highlights River Ganga becomes much cleaner in the lockdown period (May 2020) compared to the pre-lockdown time.In the Mirzapur stretch, the temperature decreased most in May 2020 as compared to May 2019.In the Varanasi stretch, there is a maximum variation in the NDTI value in May 2020 in comparison with that of May 2019.The most significant task will be to maintain river conditions during post-lockdown similar to that prevailed during lockdown.In the second wave COVID-19 lockdown the river again became polluted like the pre-COVID times.

Development of automated marine floating plastic detection system using Sentinel-2 imagery and machine learning models

The increasing level of marine plastic pollution poses severe threats to the marine ecosystem and biodiversity. Open remote sensing data and advanced machine learning (ML) algorithms could be a cost-effective solution for identifying large plastic patches across the scale. The potential application of such resources in detecting and discriminating marine floating plastics (MFP) are not fully explored. Therefore, the present study attempted to explore the full functionality of open Sentinel satellite data and ML models for detecting and classifying the MFP in Mytilene (Greece), Limassol (Cyprus), Skala Loutron, Greece, Calabria (Italy), and Beirut (Lebanon). Two ML models, i.e. Support Vector Machine (SVM) and Random Forest (RF), were utilized to perform the classification analysis. In-situ plastic location data was collected from the control experiments conducted in Mytilene, Greece (in 2018 and 2019), Skala Loutron, Greece (2021), and Limassol, Cyprus (2018), and the same was considered for training the models. The accuracy and performances of the trained models were further tested on unseen new data collected from Calabria, Italy and Beirut, Lebanon. Both remote sensing bands and spectral indices were used for developing the ML models. A spectral signature profile for marine plastic was created for discriminating the floating plastic from other marine debris. A newly developed index, kernel Normalized Difference Vegetation Index (kNDVI), was incorporated into the modelling to examine its contribution to model performances. Both SVM and RF were performed well in five models and test case combinations. Among the two ML models, the highest performance was measured for the RF. The inclusion of kNDVI was found effective and increased the model performances, reflected by high balanced accuracy measured for model 2 (~89% to ~100% for SVM and ~92% to ~98% for RF). An automated floating plastic detection system was developed and tested in Calabria and Beirut using the best-performed model. The trained model had detected the floating plastic for both sites with ~80%–90%% accuracy. Among the six predictors, the Floating Debris Index (FDI) was the most important variable for detecting marine floating plastic. These findings collectively suggest that high-resolution remote sensing imagery and the automated ML models can be an effective alternative for the cost-effective detection of MFP. Future research will be directed toward collecting quality training data to develop robust automated models and prepare a spectral library for different plastic objects for discriminating plastic from other marine floating debris and advancing the marine plastic pollution research by taking full advantage of open-source data and technologies.

The Identification of Irrigated Crop Types Using Support Vector Machine, Random Forest and Maximum Likelihood Classification Methods with Sentinel-2 Data in 2018: Tashkent Province, Uzbekistan

Accurately mapping land use and land cover including agricultural use and the state of crops at various stages is important to address specific agro-ecological challenges, to implement sustainable agricultural practices, and monitor crops periodically. This study aims to provide a timely and accurate main irrigated crop types mapping at 10m resolution for Tashkent province based on multi-temporal Sentinel-2 data acquired for the growing season in 2018. This paper shows the potential use of multitemporal Sentinel-2 satellite data to derive an up-to-date irrigated crop types classification map of the study area. As single-date satellite imagery does not allow proper cropland classification, multitemporal and high-resolution Sentinel-2 data was used to capture small cropland fields and specific crop types for the vegetation period (April to October 2018). NDVI monthly profiles of crop types as well as additional 10 m resolution bands 2 and 3 were used as input data to perform and assess three classification algorithms: Support Vector Machine (SVM), Random Forest (RF), and Maximum Likelihood Classification (MLC). Accuracy assessment results showed that SVM showed the highest Overall Accuracy (OA) and Kappa Accuracy (KA). KA of classified images for SVM were 0.90 and 0.89 for the RF algorithm. Both performed well with close values. But MLC showed a lower result of KA 0.60. The paper also compares the area of derived irrigated cropland area with data from the State Committee for Statistics of Uzbekistan for selected crop types. Values for the crops "cotton" and "wheat" derived by SVM and RF methods show a high correlation with the provided statistical data. Based on the results, the SVM classification method is recommended for further mapping and monitoring of irrigated crop types in the region when Sentinel-2 data is used.

Estimation of land subsidence hazard using interferometry of satellite radar images

Increased exploitation of underground resources is one of the main reasons of subsidence occurrence. Subsidence causes infrastructure damage and ultimately leads to increased risk to society and the economy. In some areas of Iran recently, buildings, runways, bridges, tunnels, streets, railways and roads have been severely damaged. In this study, the amount of land subsidence was analysed in the western region of Kerman City using interferometry of radar images of Sentinel-1 satellite data from 2014 to 2020 in the Sentinel Application Platform software. The results show that the subsidence rate in the area changed between 3.3 and 13.2 cm, confirming a significant increase. Field studies indicate that increased exploitation of groundwater resources, compaction due to ground drainage and imposition of heavy loads can all cause subsidence, which leads huge damage to the infrastructure.

Sentinel-1A for monitoring land subsidence of coastal city of Pakistan using Persistent Scatterers In-SAR technique

Karachi is located in the southern part of Pakistan along the Arabian Sea coast. Relevant institutions are concerned about the possibility of ground subsidence in the city, contributing to the comparative sea-level rise. So yet, no direct measurement of the subsidence rate and its relation to city submergence danger has been made. SAR (Synthetic Aperture Radar) interferometry is a powerful method for obtaining millimeter-accurate surface displacement measurements. The Sentinel-1 satellite data provide extensive geographical coverage, regular acquisitions, and open access. This research used the persistent scatterer interferometry synthetic aperture radar (PS-InSAR) technology with Sentinel-1 SAR images to monitor ground subsidence in Karachi, Pakistan. The SARPROZ software was used to analyze a series of Sentinel-1A images taken from November 2019 to December 2020 along ascending and descending orbit paths to assess land subsidence in Karachi. The cumulative deformation in Line of Sight (LOS) ranged from − 68.91 to 76.06 mm/year, whereas the vertical deformation in LOS ranged from − 67.66 to 74.68 mm/year. The data reveal a considerable rise in subsidence from 2019 to 2020. The general pattern of subsidence indicated very high values in the city center, whereas locations outside the city center saw minimal subsidence. Overall, the proposed technique effectively maps, identifies, and monitors land areas susceptible to subsidence. This will allow for more efficient planning, construction of surface infrastructure, and control of subsidence-induced risks.

An object-based sparse representation model for spatiotemporal image fusion

Many algorithms have been proposed for spatiotemporal image fusion on simulated data, yet only a few deal with spectral changes in real satellite images. An innovative spatiotemporal sparse representation (STSR) image fusion approach is introduced in this study to generate global dense high spatial and temporal resolution images from real satellite images. It aimed to minimize the data gap, especially when fine spatial resolution images are unavailable for a specific period. The proposed approach uses a set of real coarse- and fine-spatial resolution satellite images acquired simultaneously and another coarse image acquired at a different time to predict the corresponding unknown fine image. During the fusion process, pixels located between object classes with different spectral responses are more vulnerable to spectral distortion. Therefore, firstly, a rule-based fuzzy classification algorithm is used in STSR to classify input data and extract accurate edge candidates. Then, an object-based estimation of physical constraints and brightness shift between input data is utilized to construct the proposed sparse representation (SR) model that can deal with real input satellite images. Initial rules to adjust spatial covariance and equalize spectral response of object classes between input images are introduced as prior information to the model, followed by an optimization step to improve the STSR approach. The proposed method is applied to real fine Sentinel-2 and coarse Landsat-8 satellite data. The results showed that introducing objects in the fusion process improved spatial detail, especially over the edge candidates, and eliminated spectral distortion by preserving the spectral continuity of extracted objects. Experiments revealed the promising performance of the proposed object-based STSR image fusion approach based on its quantitative results, where it preserved almost 96.9% and 93.8% of the spectral detail over the smooth and urban areas, respectively.

Dynamic detection of offshore wind turbines by spatial machine learning from spaceborne synthetic aperture radar imagery

In order to change the energy supply structure of traditional thermal power generation with excessive carbon emissions, more and more offshore wind turbines are being installed and put into operation worldwide. However, the dynamic monitoring of offshore wind turbines is still an emerging problem to be solved at a global scale. This paper proposes an effective approach for dynamic detection of offshore wind turbines by machine learning from spaceborne synthetic aperture radar (SAR) Sentinel-1 satellite data. First, the radar images are preprocessed to reduce the speckle noise and ocean wave clutter. Specifically, the time series radar data are cumulatively averaged according to the monthly interval to eliminate the influence of moving temporary objects such as ships; then, the speckle noise is suppressed using refined lee filter; the imagery affected by ocean wave clutters are also mitigated by using a constant false alarm rate (CFAR) technique. Next, representative data are carefully selected as the labeled training data set; and the random forest (RF) model is trained on the Google Earth Engine (GEE) cloud computing platform. Validation demonstrates an overall accuracy of 99.99%, a producer accuracy of 100%, and an user accuracy of 94.12%. Then, mathematical morphology-based time series spatial data differentiation is proposed for monitoring the change of wind turbine. Finally, the dynamic change of the offshore wind farms in the Yellow Sea of China and the North Sea of Europe are monitored by using the proposed approach from 2015 to 2021. Cross-comparisons with ground truth data show an accuracy of 93.67% for dynamic detection of offshore wind turbines. From a large number of experimental results, it is shown that the proposed approach has the advantages of large scale monitoring and high precision, and can be used for global dynamic detection of offshore wind turbines.

Impacts of Soil and Water Conservation Practice on Soil Moisture in Debre Mewi and Sholit Watersheds, Abbay Basin, Ethiopia

Soil and water conservation (SWC) practices have been widely implemented to reduce surface runoff in the Debre Mewi watershed. However, studies on the issue have disproportionately focused on the lost or preserved soils, expressed in tons per hectare, while the impacts on the lost or preserved moisture were inadequately addressed. This study aimed to investigate the impacts of soil and water conservation practice on soil moisture in the Debre Mewi and Sholit watersheds, Abbay basin, Ethiopia. We compared soil moisture between the treated (Debre Mewi) and the untreated (Sholit) watersheds with SWCs, based on Sentinel-1A data and the field-measured soil moisture, Leaf Area Index (LAI), and water cloud model (WCM). Field-measurement was based on satellite-synchronized 63 soil moisture samples, systematically collected from the two treatment slope positions, two treatment positions, and two depths. We employed ANOVA to compare samples and discern patterns along space and time. The result indicated that the LAI, a predictor of crop yield, was higher in the SWC treated watershed, demonstrating the potential of conserving moisture for boosting crop production. In addition, the results reveal that the higher soil moisture was recorded on the grasslands of the treated watershed at a depth of 15–30 cm, while the lowest was from croplands and eucalyptus trees at 0–15 cm depth. A higher correlation was observed between the measured and estimated soil moisture across three stages of crop development. The soil moisture estimation using WCM from the Sentinel-1 satellite data gives promising results with good correlation (R2 = 0.69, 0.43 and 0.75, RMSE = 0.16, 2.24 and 0.02, and in Sholit (0.7539, 0.933, and 0.3673 and the RMSEs are 0.17%, 0.02%, and 1.02%) for different dates: August, September, and November 2020, respectively. We conclude that in the face of climate change-induced rainfall variability in tropical countries, predicted to elongate the dry spell during the cropping season, the accurate measurement of soil moistures with the mix of satellite and in-situ data could support rain-fed agriculture planning and assist in fine-tuning the climate adaptation measures at the local and regional scales.

Using Sentinel-1 and GRACE satellite data to monitor the hydrological variations within the Tulare Basin, California

Subsidence induced by groundwater depletion is a grave problem in many regions around the world, leading to a permanent loss of groundwater storage within an aquifer and even producing structural damage at the Earth’s surface. California’s Tulare Basin is no exception, experiencing about a meter of subsidence between 2015 and 2020. However, understanding the relationship between changes in groundwater volumes and ground deformation has proven difficult. We employ surface displacement measurements from Interferometric Synthetic Aperture Radar (InSAR) and gravimetric estimates of terrestrial water storage from the Gravity Recovery and Climate Experiment (GRACE) satellite pair to characterize the hydrological dynamics within the Tulare basin. The removal of the long-term aquifer compaction from the InSAR time series reveals coherent short-term variations that correlate with hydrological features. For example, in the winter of 2018–2019 uplift is observed at the confluence of several rivers and streams that drain into the southeastern edge of the basin. These observations, combined with estimates of mass changes obtained from the orbiting GRACE satellites, form the basis for imaging the monthly spatial variations in water volumes. This approach facilitates the quick and effective synthesis of InSAR and gravimetric datasets and will aid efforts to improve our understanding and management of groundwater resources around the world.

Satellite data reveal differential responses of Swiss forests to unprecedented 2018 drought.

Extreme events such as the summer drought of 2018 in Central Europe are projected to occur more frequently in the future and may cause major damages including increased tree mortality and negative impacts on forest ecosystem services. Here, we quantify the response of >1 million forest pixels of 10 × 10 m across Switzerland to the 2018 drought in terms of resistance, recovery, and resilience. We used the Normalized Difference Water Index (NDWI) derived from Sentinel‐2 satellite data as a proxy for canopy water content and analyzed its relative change. We calculated NDWI change between the 2017 pre‐drought and 2018 drought years (indicating resistance), 2018 and the 2019 post‐drought (indicating recovery), and between 2017–2019 (indicating resilience). Analyzing the data from this large natural experiment, we found that for 4.3% of the Swiss forest the NDWI declined between 2017 and 2018, indicating areas with low resistance of the forest canopy to drought effects. While roughly 50% of this area recovered, in 2.7% of the forested area NDWI continued to decline from 2018 to 2019, suggesting prolonged negative effects or delayed damage. We found differential forest responses to drought associated with site topographic characteristics and forest stand characteristics, and to a lesser extent with climatic conditions and interactions between these drivers. Low drought resistance and high recovery were most prominent at forest edges, but also on south‐facing slopes and lower elevations. Tree functional type was the most important driver of drought resilience, with most of the damage in stands with high conifer abundance. Our results demonstrate the suitability of satellite‐based quantification of drought‐induced forest damage at high spatial resolution across large areas. Such information is important to predict how local site characteristics may impact forest vulnerability to future extreme events and help in the search for appropriate adaptation strategies.

Remote sensing-based biomass estimation of dry deciduous tropical forest using machine learning and ensemble analysis

Forests play a vital role in maintaining the global carbon balance. However, globally, forest ecosystems are increasingly threatened by climate change and deforestation in recent years. Monitoring forests, specifically forest biomass is essential for tracking changes in carbon stocks and the global carbon cycle. However, developing countries lack the capacity to actively monitor forest carbon stocks, which ultimately adds uncertainties in estimating country specific contribution to the global carbon emissions. In India, authorities use field-based measurements to estimate biomass, which becomes unfeasible to implement at finer scales due to higher costs. To address this, the present study proposed a framework to monitor above-ground biomass (AGB) at finer scales using open-source satellite data. The framework integrated four machine learning (ML) techniques with field surveys and satellite data to provide continuous spatial estimates of AGB at finer resolution. The application of this framework is exemplified as a case study for a dry deciduous tropical forest in India. The results revealed that for wet season Sentinel-2 satellite data, the Random Forest (adjusted R2 = 0.91) and Artificial Neural Network (adjusted R2 = 0.77) ML models were better-suited for estimating AGB in the study area. For dry season satellite data, all the ML models failed to estimate AGB adequately (adjusted R2 between −0.05 – 0.43). Ensemble analysis of ML predictions not only made the results more reliable, but also quantified spatial uncertainty in the predictions as a metric to identify its robustness.

Flood inundation maps using Sentinel-2: a case study in Berdan Plain

Abstract This study assessed the performance of remote sensing and geographic information system techniques for analysing flood events in agricultural areas and estimating flooded areas in agricultural lands. The Berdan Plain in the southern part of Turkey is below sea level and prone to flood events. The flooded agricultural areas were estimated by using Sentinel-2 satellite data for the Berdan Irrigation Scheme from December 2019-January 2020. The flooded area in the old Aynaz marsh was estimated as 486.79 hectares on 9 January 2020, while the total flooded area in the Berdan Irrigation Project was 4,515.58 hectares. Flood volume was calculated as 2.61 hm3 and the highest water level was estimated to be −0.76 metres in the old Aynaz marsh.

Soil heterogeneity: A comparative assessment of soils from two different AESR, southern India

A comparative study was carried out to assess the nature and characteristics of soils from two agro-ecological sub regions (AESU and AESR) from two states of South India. Using sentinel-2 satellite data, google imagery and top sheet, a detailed soil survey was taken up in AESU and AESR. Major soils from upland as well as from lowlands were compared to assess the soils heterogeneity. Results showed that there was a high heterogeneity between properties of soils AESU and AESR. Variability existed in climatic conditions (tropical humid region in AESU and semiarid condition in AESR), physiographic (valley plain to gently, very gently, moderate, steep and very steep slopes in AESU and levelled to gently and very gently slopes in AESR ), geology (charnockite and granite gneiss in AESU whereas granite gneiss in AESR), land use (rubber, paddy, coconut in AESU and paddy, mango, pigeon pea, groundnut in AESR), soil classification (Ultisols and Alfisols in AESU whereas Alfisols and Inceptisols in AESR) and land capability (soil and water conservation measures must in AESU and slope was manageable with agronomic management practices in AESR). Soils varied in texture (clay, sandy clay and clay loam in AESU whereas sandy clay loam to sandy loam in AESR), consistency and depth of soils (shallow to very deep in AESU and shallow to deep in AESR). Clay content (32.9 to 63.5%) was dominant fraction in AESU whereas sand (52.1 to 72.1%) was in AESR. Soil acidity (strongly acid to very strongly acid) was a major problem (both hydrogen and aluminium saturation) in AESU whereas slightly acidic to neutral and slightly alkaline in AESR. Organic carbon content was high (up to 25.9 gkg-1) in AESU whereas it was up to 10.0 gkg-1 in AESR. Base saturation was <35% in AESU and >75% up to 100% in AESR. Calcium carbonate equivalent is >5% in AESR and not a problem in AESU. CEC was low (<12 cmol (p+) kg–1) in AESU and good (>12 cmol (p +) kg–1) in AESR.

Soil heterogeneity: A comparative assessment of soils from two different AESR, southern India

A comparative study was carried out to assess the nature and characteristics of soils from two agro-ecological sub regions (AESU and AESR) from two states of South India. Using sentinel-2 satellite data, google imagery and top sheet, a detailed soil survey was taken up in AESU and AESR. Major soils from upland as well as from lowlands were compared to assess the soils heterogeneity. Results showed that there was a high heterogeneity between properties of soils AESU and AESR. Variability existed in climatic conditions (tropical humid region in AESU and semiarid condition in AESR), physiographic (valley plain to gently, very gently, moderate, steep and very steep slopes in AESU and levelled to gently and very gently slopes in AESR ), geology (charnockite and granite gneiss in AESU whereas granite gneiss in AESR), land use (rubber, paddy, coconut in AESU and paddy, mango, pigeon pea, groundnut in AESR), soil classification (Ultisols and Alfisols in AESU whereas Alfisols and Inceptisols in AESR) and land capability (soil and water conservation measures must in AESU and slope was manageable with agronomic management practices in AESR). Soils varied in texture (clay, sandy clay and clay loam in AESU whereas sandy clay loam to sandy loam in AESR), consistency and depth of soils (shallow to very deep in AESU and shallow to deep in AESR). Clay content (32.9 to 63.5%) was dominant fraction in AESU whereas sand (52.1 to 72.1%) was in AESR. Soil acidity (strongly acid to very strongly acid) was a major problem (both hydrogen and aluminium saturation) in AESU whereas slightly acidic to neutral and slightly alkaline in AESR. Organic carbon content was high (up to 25.9 gkg-1) in AESU whereas it was up to 10.0 gkg-1 in AESR. Base saturation was <35% in AESU and >75% up to 100% in AESR. Calcium carbonate equivalent is >5% in AESR and not a problem in AESU. CEC was low (<12 cmol (p+) kg–1) in AESU and good (>12 cmol (p +) kg–1) in AESR.

Primary productivity estimation of forest based on in-situ biophysical parameters and sentinel satellite data using vegetation photosynthesis model in an eastern Indian tropical dry deciduous forest

Primary productivity estimation of forest based on in-situ biophysical parameters and sentinel satellite data using vegetation photosynthesis model in an eastern Indian tropical dry deciduous forest

Designing a European-Wide Crop Type Mapping Approach Based on Machine Learning Algorithms Using LUCAS Field Survey and Sentinel-2 Data

One of the most challenging aspects of obtaining detailed and accurate land-use and land-cover (LULC) maps is the availability of representative field data for training and validation. In this manuscript, we evaluate the use of the Eurostat Land Use and Coverage Area frame Survey (LUCAS) 2018 data to generate a detailed LULC map with 19 crop type classes and two broad categories for woodland and shrubland, and grassland. The field data were used in combination with Copernicus Sentinel-2 (S2) satellite data covering Europe. First, spatially and temporally consistent S2 image composites of (1) spectral reflectances, (2) a selection of spectral indices, and (3) several bio-geophysical indicators were created for the year 2018. From the large number of features, the most important were selected for classification using two machine-learning algorithms (support vector machine and random forest). Results indicated that the 19 crop type classes and the two broad categories could be classified with an overall accuracy (OA) of 77.6%, using independent data for validation. Our analysis of three methods to select optimum training data showed that by selecting the most spectrally different pixels for training data, the best OA could be achieved, and this already using only 11% of the total training data. Comparing our results to a similar study using Sentinel-1 (S1) data indicated that S2 can achieve slightly better results, although the spatial coverage was slightly reduced due to gaps in S2 data. Further analysis is ongoing to leverage synergies between optical and microwave data.

High-resolution agriculture soil property maps from digital soil mapping methods, Czech Republic

Detailed maps of soil properties are essential for soil protection planning and management; however, creating very high-resolution maps at the national level with sufficient accuracy is a challenging task and remains unavailable for most countries. For the Czech Republic, very high resolution (20 m·pixel−1) soil property maps (soil organic carbon—SOC, texture, pH, bulk density, soil depth) were created using digital soil mapping methods, combined with a wide database of soil legacy and current samples. The latest approaches were employed for predictive mapping: a quantile random forest model with the determination of prediction intervals, a mosaic of bare soils from Sentinel-2 satellite data, a Gaussian pyramid of terrain attributes, and a buffer distance map. These variables were found to be among the most important in the resulting models. The properties were mapped with an RMSE accuracy of 0.43% SOC, 5.56–11.14% for texture fractions, 0.70 pH, 0.13 g·cm−3 bulk density, and 20.03 cm for soil depth, thus providing detailed data on soil cover. Greater levels of inaccuracy were found in areas with extreme values, for which further investigation is necessary either through more detailed sampling based on active learning, or adapted methods for enhanced predictive ability.