Given the rapid global urbanization and climate change, understanding the influence of urban form on urban vegetation phenology is of great importance. However, previous studies rarely focused on how the unique local climates created by urban structures impact urban vegetation phenology. Here, we investigated these effects across 934 cities in the Northern Hemisphere spanning various climate regions, using fine-resolution Local Climate Zones (LCZs) data and the VNP22Q2 phenology product. By employing a Linear Mixed-Effects Model (LMM) and implementing a strict 50% purity pixel filtering strategy to ensure the rigor of our results, our findings indicate that intense urban forms, characterized by building surface fraction and height, extend the length of the growing season (LOS) by advancing the start and delaying the end of growing seasons, despite those large-scale variations of vegetation phenology are predominantly explained by background climate. On average, the LOS for urban areas exceeds that of surrounding natural environments by 13.07 days across all studied cities. Furthermore, the LMM-based quantification reveals that the height of urban structures significantly impacts vegetation phenology more than building density, with taller types extending the LOS by 38.47 days, whereas denser types extend it by only 15.00 days. These insights reveal that urbanization contributes substantially to prolonging the urban vegetation phenological season. Highlighting the influence of urban form on vegetation phenology underscores its critical role in addressing immediate climate challenges and achieving urban sustainability. This study provides essential insights for urban planning and policy to foster ecological balance in rapidly developing urban areas.

Rising precipitation under climate change can increase groundwater recharge in many regions and accelerate the mobilization of legacy nitrate stored in the vadose zone. We couple a machine learning emulator of a global hydrological model with the nitrate time bomb (NTB) model to quantify nitrate migration from 1958 to 2100. Nitrate migration velocity increases across all climate zones except the tropics, with the most pronounced gains in the cold (+0.20/+0.25 m year-1 under SSP2-4.5/SSP5-8.5) and temperate (+0.17/+0.15 m year-1) zones. Groundwater table shallowing further shortens transport distance; in the arid zone, climate change advances nitrate peak arrival by about 6/8 years (locally >20 years). Among grid cells where nitrate peaks reach groundwater before 2100, more than 60% show shortened NTB countdowns under both scenarios. After the first global nitrate accumulation peak, depth-projected leaching shows renewed increase by 2019, suggesting a second phase of nitrate loading to the vadose zone. Integrating legacy nitrate mass with climate-adjusted transport identifies the North China Plain, South Asia, and Western Europe as NTB hotspots. These results show that climate forcing can accelerate the release of long buried agricultural pollutants, underscoring the urgency of proactive groundwater protection.

Soil structure governs water storage, aeration, root growth, trafficability, nutrient transport, and carbon stabilization, but agricultural monitoring still relies largely on sparse, destructive, and labor-intensive measurements. Digital imaging, spectroscopy, geophysics, in situ sensors, proximal sensing, and remote sensing now provide multiscale observations, yet converting them into reliable structural information remains limited by sparse ground truth, scale mismatch, domain shift, and inconsistent validation. This structured narrative review synthesizes how artificial intelligence (AI) transforms these heterogeneous observations into useful soil structure information for agricultural monitoring and management. We organize the literature around a sensor-to-decision workflow: observation, representation, scaling, discovery, and decision-making. At the pore scale, deep learning improves image enhancement, segmentation, and reconstruction of pore and aggregate architecture from imaging technologies. At pedon and field scales, machine learning links spectra, geophysical signals, mobile proximal sensors, and in situ time series to structural proxies and subsurface states. At landscape scales, AI enables data fusion, downscaling, and transfer learning from UAV, satellite, LiDAR, and mobile sensing platforms. We further examine how physics-guided learning, uncertainty quantification, explainable AI, surrogate modeling, and soil digital twins can improve transferability and support management of transport processes, mechanical resilience, and biological carbon dynamics. By linking sensing modality, AI method, validation design, and management applications, this review provides practical guidance for selecting deployable soil structure monitoring workflows for compaction avoidance, irrigation management, trafficability assessment, carbon stabilization, and resilience monitoring.

Aboveground biomass (AGB) is a critical indicator for assessing crop growth status and productivity, yet accurately linking fine-scale ground measurements with coarse-resolution satellite imagery remains challenging. Here, we propose an integrated ground-UAV-satellite framework that combines high-resolution UAV observations with an optimized systematic sampling-Global Moran's I (SS-GMI) procedure and a simple allometric growth model. Multi-variety sugar beet cultivated across heterogeneous habitats was used as a case study. Results indicate that a power-law model effectively captures the allometric relationships between AGB, plant height, and the Dreg vegetation index in sugar beet, achieving high accuracy and strong transferability. Incorporating phenological information from Biologische Bundesanstalt, Bundessortenamt und CHemische Industrie (BBCH) codes and a thermal index further enhanced model robustness across independent habitat trials, yielding coefficients of determination (R2) of 0.80 and 0.83. The SS-GMI sampling procedure integrates systematic sampling with Global Moran's I to reduce spatial autocorrelation while ensuring uniform spatial coverage, thereby enabling the acquisition of representative and spatially independent samples from UAV-derived AGB maps. These samples were used to develop satellite-based AGB estimation models for PlanetScope and Sentinel2A imagery, achieving R2 values of 0.83 and 0.73, respectively. This study provides a practical and scalable framework for field-to-satellite AGB upscaling, offering new insights for the scale conversion of multi-source data in agricultural remote sensing.

Understanding evapotranspiration (ET) partitioning into soil evaporation (E) and plant transpiration (T) is crucial for improving agricultural water use efficiency in water-scarce regions. The isotope mass balance (IMB) method and AquaCrop model are two widely used approaches for ET partitioning, yet their comparative performance across different crop growth stages remains poorly characterized. This study systematically compared these two methods using two consecutive years (2012-2013) of field isotopic observations in a summer maize field on the North China Plain, a core maize production area facing severe agricultural water scarcity. Stable isotope analysis showed that the local meteoric water line (LMWL) had a slope lower than the global meteoric water line. The 0-5 cm surface soil water evaporation lines had slopes of 5.84 (2012) and 8.06 (2013), confirming significant evaporative enrichment in the topsoil. Plant water isotopic composition closely resembled that of 40-100 cm deep soil water, indicating limited root uptake from the surface layer. IMB-estimated transpiration ratio (T/ET) exhibited distinct phenological patterns, increasing from 37 to 44% at jointing to a peak of 94-96% at filling, then declining to 84-85% at maturity. The two methods agreed well during filling to maturity (differences of 2-10%), but compared with the IMB method, AquaCrop substantially underestimated T/ET at jointing (0.9% vs. 43.8% in 2013) due to its canopy-cover-based transpiration algorithm. These findings identify the filling stage as the critical water demand period, providing a quantitative reference for precision irrigation management under similar climate and soil conditions.

Low mobility and bioavailability of phosphorus (P) constrain fertilizer use efficiency in the black soil region of Northeast China (NEC). Conservation tillage (CST) has been shown to improve soil structure and nutrient bioavailability and therefore has the potential to increase P use efficiency. In this study, we evaluated the effects of CST and conventional tillage (CT) on P use efficiency-quantified through both fertilizer-based partial factor productivity (PFP-P) and uptake-based utilization efficiency (PUtE)-maize dry matter, and P content, root vertical distribution, and soil physicochemical properties based on field experiments at three representative sites in the black soil region of NEC. Compared to CT, CST significantly increased PFP-P (+15.3 %) and PUtE (+25.3 %), promoted grain dry matter accumulation, and enhanced P translocation from vegetative organs to grains. CST decreased the stratification ratio (0-10 cm / 10-50 cm) of total P (TP) and Olsen-P by 24.9 % and 49.5 %, respectively (except at the XL site), thereby alleviating surface P enrichment. CST also increased average root biomass by 28.5 % and significantly enhanced the root distribution ratio (20-30 cm / 0-50 cm) (p < 0.05), suggesting an improved soil-root configuration for accessing subsoil P. Redundancy analysis showed that variations in P efficiency were primarily associated with Olsen-P stratification, root distribution patterns (20-30 cm), soil penetration resistance (20-30 cm), and SOM (0-10 cm), which together explained over 86.0 % of the variance. Overall, CST improved maize P use efficiency by synchronizing root-soil-P interactions, providing scientific insight and practical strategies for sustainable agriculture in the black soil region of NEC.

Pyrogenic carbon is widely regarded as a component of soil organic carbon, yet how small amounts of vertically redistributed pyrogenic carbon influence organo mineral interfaces in subsoils remains unresolved. Here we combine a ten year field ageing experiment along a 200 cm calcareous soil profile in China with electrochemical assays, spectroscopy and nanoscale microscopy to link pyrogenic carbon electron transfer capacity to mineral association. Redistributed pyrogenic carbon retains electron accepting and donating capacities in subsoils despite low concentrations and exhibits faster electron transfer kinetics. This retention reflects oxidative surface transformation, enriching quinone and phenolic redox moieties and mineral complexing oxygen groups. Nanoscale observations show oxidised subsoil pyrogenic carbon surfaces with organo mineral coatings associated with iron and calcium bearing phases, consistent with a coupled redox and sorptive interface rather than passive presence alone. These findings suggest that field-aged pyrogenic carbon contributes to redox coupled mineral stabilisation in calcareous subsoils. In calcareous alkaline soils, pyrogenic carbon acquires and retains persistent redox-coupled mineral stabilisation and faster electron transfer kinetics that remains expressed after vertical redistribution, based on a ten-year long field experiment in China.

Soil carbon sequestration should mitigate climate change, yet the dynamics of deep soil carbon are poorly known. Here, we conducted a 12-year field experiment in a Fluvic Cambisol under a wheat–maize rotation system, comparing straw return with straw removal, and quantified carbon stocks across a 0–200 cm soil profile. We traced carbon sources using both natural 13C abundance, and 13C-labeled glucose in a 60-day laboratory incubation experiment. Results show that straw return increased soil inorganic carbon stocks by 57.6 t ha−1 in the 100–200 cm layers. Pedogenic carbon increased by 96.6

Background and aimsBrace roots are crucial for stalk anchorage and plant growth, yet the effects of soil compaction and moisture conditions on their development remain unclear. This study examined the interactive effects of soil compaction and moisture on maize brace root traits and explored the associated hormonal regulatory mechanisms.MethodsA completely randomized design was employed with two factors: soil bulk density (1.2 g cm-3, uncompact, B1.2; and 1.5 g cm-3, compacted, B1.5) and soil water content (75% and 100% field capacity, W75 and W100). Brace root traits, hormone content in nodal tissue, and soil properties were analyzed.ResultsHigher soil water content increased brace root node number (BRNN), spread width (BRSW), and diameter (BRD), whereas soil compaction increased brace root angle (BRA). Nodal 1-aminocyclopropane-1-carboxylic acid (ACC) was positively correlated with BRNN, BRSW, and BRD, suggesting a positive regulatory role of ethylene in brace root development. Cytokinins (CKs) were negatively associated with brace root development, whereas indole-3-acetic acid (IAA) had minimal effect. The higher nodal ACC observed under wetter conditions may be linked to changes in soil gas diffusivity.ConclusionThese findings indicate that compacted and wet soils enhance brace root development via ethylene regulation, offering a potential strategy to manage soil physical conditions for improved lodging resistance and nutrient acquisition in maize.

On-the-go proximal soil sensing systems mounted on vehicles can rapidly acquire high-density soil information for precision agriculture. However, raw measurements are often affected by factors such as warm-up drift, attitude changes, abrupt turns, and microtopography, degrading data quality and the reliability of subsequent soil property maps. In this study, a two-stage data filtering framework was developed to improve on-the-go apparent electrical conductivity (ECa) measured at four depths (0.54, 1.03, 1.55, and 3.18 m) using a DUALEM-21S electromagnetic induction sensor. Stage 1 applied physically interpretable, rule-based screening to remove gross outliers. Stage 2 encompassed density-based spatial clustering of applications with noise (DBSCAN) in a time-space median-deviation feature space to detect residual outliers, with parameters derived automatically using kernel density estimation. In a 41-ha field dataset, 39% of raw records were removed as erroneous, primarily due to noise from metal pipelines and headland turns. The mean leave-one-out cross-validation RMSE for ordinary kriging interpolation maps, averaged across the four depths, decreased by 67%. Soil texture prediction was also improved using filtered data. Specifically, for a partial least squares regression (PLSR) model predicting surface sand content, R2 increased from 0.62 to 0.76, and RMSE decreased from 6.90% to 5.52%. These results demonstrate that the proposed framework enhanced the spatial consistency of on-the-go ECa data and the accuracy of derived soil maps. The framework provides an objective, portable, and computationally efficient protocol for field-scale implementation that can be generalized to filter most high-density proximal soil-sensing data collected during mobile sensor operations.

Soil compaction is a prevalent physical constraint that adversely affects root development. Brace roots are critical for providing structural support and facilitating nutrient and water uptake of maize, but the responses of brace roots to soil compaction are understudied. In this study, we examined maize brace root development under three soil compaction levels created by different wheeling intensities: no wheeling passes as a control, five, and ten wheeling passes of compaction. Increased wheeling intensity significantly increased soil bulk density, shear strength, and penetration resistance while decreasing gas diffusivity. These changes promoted brace root development, increasing node number, root number, diameter, angle, and lateral root branching density. Soil compaction and the associated changes in soil physical properties significantly affected hormone content at maize nodes, with increased 1-aminocyclopropane-1-carboxylic acid and abscisic acid that were positively correlated with brace root development. Meanwhile, compaction reduced cytokinin content, which was negatively correlated with brace root development. These results demonstrate that maize brace roots exhibit adaptive morphological responses to soil compaction, which was mediated by changes in nodal hormone content. This study provides insights into hormone-mediated maize root plasticity under compaction stress, offering potential strategies for improving crop resilience. Further studies are need to explore the mechanisms that soil compactions influence the hormone of maize.

Accurate pre-harvest prediction of sugar beet yield is vital for effective agricultural management and decisionmaking. However, traditional methods are constrained by reliance on empirical knowledge, time-consuming processes, resource intensiveness, and spatial-temporal variability in prediction accuracy. This study presented a plot-level approach that leverages UAV technology and recurrent neural networks to provide field yield predictions within the same growing season, addressing a significant gap in previous research that often focuses on regional scale predictions relied on multi-year history datasets. End-of-season yield and quality parameters were forecasted using UAV-derived time series data and meteorological factors collected at three critical growth stages, providing a timely and practical tool for farm management. Two years of data covering 185 sugar beet varieties were used to train a developed stacked Long Short-Term Memory (LSTM) model, which was compared with traditional machine learning approaches. Incorporating fresh weight estimates of aboveground and root biomass as predictive factors significantly enhanced prediction accuracy. Optimal performance in prediction was observed when utilizing data from all three growth periods, with R2 values of 0.761 (rRMSE = 7.1 %) for sugar content, 0.531 (rRMSE = 22.5 %) for root yield, and 0.478 (rRMSE = 23.4 %) for sugar yield. Furthermore, combining data from the first two growth periods shows promising results for making the predictions earlier. Key predictive features identified through the Permutation Importance (PIMP) method provided insights into the main factors influencing yield. These findings underscore the potential of using UAV time-series data and recurrent neural networks for accurate pre-harvest yield prediction at the field scale, supporting timely and precise agricultural decisions. (c) 2025 The Authors. Publishing services by Elsevier B.V. on behalf of KeAi Communications Co., Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

No-tillage with straw return is an effective strategy for improving soil quality and fostering earthworm community. Straw return could increase soil microbial resource limitations, which strongly change the processes of microbial metabolism and subsequently nutrient cycling in agroecosystems. However, the soil microbial resource limitations in response to increased earthworm community remain unclear. A microcosm experiment was conducted to quantify how earthworms regulate the effect of straw addition on microbial resource limitations and subsequent impact on microbial carbon use efficiency (CUE). Results showed that straw return enhanced microbial nitrogen (N) limitation, but this enhancement was mitigated by earthworms, primarily due to increased nutrients availability. Earthworms with straw addition significantly enhanced microbial CUE, primarily attributed to the increased nutrients availability and reduced microbial resource limitations. Overall, our results revealed that earthworms decreased straw-induced microbial N limitation and increased microbial CUE, emphasizing the importance of earthworms in balancing soil microbial resource limitations and C sequestration in the no-tillage agricultural system.

Groundwater serves as a vital water resource for agricultural irrigation and domestic use in farmland areas. Its chemical composition is jointly influenced by agricultural fertilization, land use practices, and natural geological processes. However, research on the controlling factors and spatial distribution characteristics of groundwater hydrochemistry in agricultural regions remains insufficient. In this study, 56 groundwater samples were collected from the central-eastern plain of Henan Province, China. A combination of hierarchical cluster analysis, ionic ratio methods, principal component analysis, and kriging interpolation was employed to investigate the hydrochemical characteristics, spatial patterns, and primary controlling factors of regional groundwater. The results indicate that the first group of samples is characterized by high total dissolved solids (TDS), elevated Na+ and Cl− concentrations, predominantly controlled by evaporation and concentration processes. The second group exhibits high pH and low Ca2+ concentrations, mainly influenced by silicate weathering, with reverse cation exchange acting as a secondary controlling process. The third group is characterized by elevated concentrations of Ca2+ and NO3−, primarily controlled by carbonate weathering and agricultural activities. The western part of the study area serves as the main groundwater recharge zone and has the highest NO3− and Ca2+ concentrations. In the central area, most ion concentrations are relatively high, forming a distinct gradient with surrounding regions. Meanwhile, the eastern area displays elevated concentrations of HCO3−, TDS, Na+, and Cl−, highlighting pronounced spatial heterogeneity. Overall, the hydrochemical composition of groundwater in the study area is shaped by both natural processes and anthropogenic activities, exhibiting significant spatial heterogeneity. Notably, the spatial variation of NO3− concentrations is substantial, indicating that certain localities have already been affected by agricultural non-point source pollution.

Leaf shape is of great significance in plant phenotype research. Landmarks method is a widely used morphometric approach, which can comprehensively describe the morphological differences among leaves. However, the selection of landmarks is time-consuming and laborious. An automatic landmarking algorithm is proposed here. Based on conformal mapping, the leaf outline can be transformed into a monotonically increasing function curve, referred to as the 'fingerprint function'. The Dynamic Time Warping (DTW) algorithm was introduced to match landmarks between different leaves. Two leaf datasets were used to validate the algorithm separately in different species and developmental stages. Dataset1 is a public dataset which covers 26 different types of leaves. The average positional difference between automatic and manual landmarks for dataset1 was only 2.95%. Dataset2 consists of cotton leaves collected in the field at various growth stages, and the positional difference for this dataset was all below 5%. These results validate that our algorithm is applicable to a wide range of leaf types and capable of identifying and locating novel features that emerge during leaf growth. The automatic landmarking algorithm can simulate manual landmarking to a great extent. It provides a new approach for automated acquisition of plant leaf shape homology tailored to the research needs of botanists.

Maize root lodging causes yield and grain quality reduction. We hypothesized that conservation tillage (CST) could increase root lodging resistance compared to conventional tillage (CVT) by facilitating root development. In this study, we compared maize root pushing resistance (RPR), a proxy for root lodging, in paired CST and CVT fields at 14 field sites and evaluated the relationship between RPR, soil physical properties and the maize plant traits. We found CST significantly increase maize RPR by 33.0 % in CVT system. Soil bulk density (BD), penetration resistance (PR) and shear strength (SS) of the topsoil (0-20 cm) was also significantly higher in CST. Brace root traits, including diameter (BRD), whorl number (BRWN) and angle (BRA), and stalk width were significantly increased following CST relative to CVT. Correlation analysis showed the variation in RPR can be attributed to maize stalk width and brace root phenotypes. A positive correlation was also found between soil strength and brace root traits. These findings indicate that improved soil properties are key factors for stimulating maize brace root development, and increasing maize root lodging resistance in CST fields. These results shed new light on the optimizing tillage practice to minimize maize root lodging.

Study region: Haihe Basin (HB), North China. Study focus: Studying the impact of extreme precipitation on watershed hydrological factors plays a crucial role in water resource management, climate adaptation, and disaster resilience. An improved Soil and Water Assessment Tool (SWAT) was employed to assess the impact of extreme precipitation indices (EPIs) on temporal and spatial variations in hydrological factors in the HB, China. Five EPIs were identified in this study, including R10 (moderate rain), R20 (heavy rain), R50 (torrential rain), R95p (95th percentile of precipitation), and R99p (99th percentile of precipitation). New hydrological insights for the region: The EPIs with the greatest contribution rates to precipitation, water yield, and percolation in the historical period were R20 (32.1 %), R50 (14.3 %), and R20 (29.0 %), respectively, for the entire basin. During the historical period, there were more occurrences of extreme precipitation events in the plain area compared to the mountainous area. In the plain area, rainfall was beneficial for replenishing groundwater when daily precipitation exceeded 50 mm. Over the entire future period (2041-2100), R50 contributed the greatest water yield (18.4 %) and percolation (36.3 %) in the HB. Furthermore, the number of days with rainfall from 20 to 50 mm d- 1 and those exceeding 50 mm d- 1 increased in the future period relative to the historical period. The results of this study provide a reference for understanding the spatiotemporal distribution pattern of extreme precipitation in the HB and for relevant departments to formulate response strategies.

Accurate regional mapping of soil organic carbon (SOC) in croplands is essential for assessing soil carbon sequestration potential. However, accurate SOC mapping of cropland at a regional scale is challenging due to numerous natural and anthropogenic management factors. The impact of covered crop residue remains undervalued when mapping surface SOC, despite the significant impact of crop residue coverage (CRC) on SOC. In particular, the agricultural management practice of returning crop residues to the soil significantly alters the spatio temporal patterns of SOC in northeast China. Given these issues, we used the Shapley Additive exPlanations (SHAP) approach to interpret the influence of natural and anthropogenic factors on SOC estimation using the random forest model. Our results show the high SHAP values of air temperature, CRC, and clay content due to their significant influence on SOC estimation. Interestingly, our analysis showed a significant increase in SHAP values when the CRC reached 0.30, which refers to the CRC threshold of conservation tillage. Furthermore, our results revealed that integrating crop residue coverage significantly improved the accuracy of SOC mapping as the Lin Concordance Correlation Coefficient (LCCC) increased from 0.75 to 0.83 and the root mean squared error (RMSE) decreased from 6.70 g kg−1 to 5.60 g kg−1. This study provides actionable insights for optimizing CRC management practices for SOC sequestration in Northeast China.

Dealing with heterogeneity in leaf canopies when calculating light interception per species in a mixed canopy is a challenge. Goudriaan developed a computationally simple, though conceptually sophisticated, model for light interception in strip canopies, which can be reasonably represented as 'blocks', such as vineyards and crop rows. This model is widely used, but there is no independent verification of the model. Hence, we developed a comparison of light interception calculations with Goudriaan's model and with detailed spatially explicit three-dimensional functional-structural plant models (FSPM) of maize in which plant architecture can be represented explicitly. Two models were developed, one with small randomly oriented leaves in blocks, similar to Goudriaan's assumption, which we refer to as the intermediate model (IM), and another with a realistic representation of individual plants with stems and leaves having shape, orientation and so on, referred as FSPM. In IM and FSPM, light interception was calculated using ray tracing. In Goudriaan's model, the light extinction coefficient (k), including both its daily and seasonal average values, was generated using the FSPM. Correspondence between the three models was excellent in terms of light capture for different levels of crop height, leaf area and uniformity, with the difference less than 3.3 %. The results are strong support for the use of Goudriaan's summary model for calculating light interception in strip canopies.

Soil analysis using near-infrared spectroscopy has shown great potential to be an alternative to traditional laboratory analysis, and there is continuously increasing interest in building large-scale soil spectral libraries (SSLs). However, due to issues such as high non-linearity in soil spectral data and complexity in soil spatial variation, the establishment of robust prediction models for soil spectral libraries remains a challenge. This study aimed to investigate the performance of deep learning algorithms, including long short-term memory (LSTM) and LSTM–convolutional neural networks (LSTM–CNN) integrated models, to predict the soil organic matter (SOM) of a provincial-scale SSL, and compare it to the normally used local weighted regression (LWR) model. The Hebei soil spectral library (HSSL) contains 425 topsoil samples (0–20 cm), of which every 3 soil samples were collected from dry land, irrigated land, and paddy fields, respectively, in different counties of Hebei Province, China. The results show that the accuracy of the validation dataset rank as follows: LSTM–CNN (R2p = 0.96, RMSEp = 1.66 g/kg) > LSTM (R2p = 0.83, RMSEp = 3.42 g/kg) > LWR (R2p = 0.82, RMSEp = 3.79 g/kg). The LSTM–CNN model performed the best, mainly due to its comprehensive ability to effectively extract spatial and temporal features. Meanwhile, the LSTM model achieved higher accuracy than the LWR model, owing to its built-in memory unit and its advantage of faster feature band extraction. Thus, it was suggested to use deep learning algorithms for SOM predictions in SSLs. However, their performance on larger-scale SSLs such as continental/global SSLs still needs to be further investigated.