Search

With the continuous maturity of hyperspectral remote sensing imaging technology, it has been widely adopted by scholars to improve the performance of feature classification. However, due to the challenges in acquiring hyperspectral images and producing training samples, the limited training sample is a common problem that researchers often face. Furthermore, efficient algorithms are necessary to excavate the spatial and spectral information from these images, and then, make full use of this information with limited training samples. To solve this problem, a novel two-branch deep learning network model is proposed for extracting hyperspectral remote sensing features in this paper. In this model, one branch focuses on extracting spectral features using multi-scale convolution and a normalization-based attention module, while the other branch captures spatial features through small-scale dilation convolution and Euclidean Similarity Attention. Subsequently, pooling and layering techniques are employed to further extract abstract features after feature fusion. In the experiments conducted on two public datasets, namely, IP and UP, as well as our own labeled dataset, namely, YRE, the proposed DMAN achieves the best classification results, with overall accuracies of 96.74%, 97.4%, and 98.08%, respectively. Compared to the sub-optimal state-of-the-art methods, the overall accuracies are improved by 1.05, 0.42, and 0.51 percentage points, respectively. The advantage of this network structure is particularly evident in unbalanced sample environments. Additionally, we introduce a new strategy based on the RPNet, which utilizes a small number of principal components for feature classification after dimensionality reduction. The results demonstrate its effectiveness in uncovering compressed feature information, with an overall accuracy improvement of 0.68 percentage points. Consequently, our model helps mitigate the impact of data scarcity on model performance, thereby contributing positively to the advancement of hyperspectral remote sensing technology in practical applications.

Accurately estimating forest aboveground biomass (AGB) is imperative for comprehending carbon cycling, calculating carbon budgets, and formulating sustainable forest management plans. Currently, random forest (RF) and other machine learning models are widely used to estimate forest AGB, as they can effectively handle nonlinear relationships. However, by constructing a global model using all the samples collected from a study area, these models fail to account for the spatial heterogeneity in the AGB and cannot correct the prediction biases, thereby constraining the estimation accuracy. To overcome these limitations, we proposed a novel approach termed geographical random forest and empirical Bayesian kriging (GRFEBK). This hybrid model combines the localized modeling capability of geographical random forest (GRF) with the bias correction strength of empirical Bayesian kriging (EBK). GRF adapts RF to account for the spatial heterogeneity of the AGB, while EBK utilizes the spatial autocorrelation of residuals to correct the prediction deviations. This study was conducted in Hainan Island, utilizing spectral bands, vegetation indices, tasseled cap components derived from Landsat-8 imagery, backscattering coefficients from ALOS-2 synthetic aperture radar, topographic features, and the forest canopy height as the explanatory variables. A total of 195 forest aboveground biomass (AGB) samples were collected for modeling and assessing the predictive accuracy. The results demonstrate that, among the tested models, including GRFEBK, RF, support vector machine (SVM), k-nearest neighbor (KNN), geographically weighted regression (GWR), GRF, and EBK, GRFEBK attains the highest R2 (0.78) and the lowest RMSE (36.04 Mg/ha) and RRMSE (22.87%), significantly outperforming the conventional models and using GRF or EBK alone. These results demonstrate that by accounting for local non-stationarity in AGB and correcting prediction biases, GRFEBK achieves significantly higher accuracy than conventional RF and other models. While the results are promising, the computational cost of GRFEBK and its performance under varying geographical conditions warrant further investigation at larger scales to assess its broader applicability. Nevertheless, GRFEBK provides an innovative and more reliable approach for accurate forest AGB estimation with great potential to support global forest resource monitoring.

The Yellow River Delta wetlands in China belong to the coastal wetland ecosystem, which is one of the youngest and most characteristic wetlands in the world. The Yellow River Delta wetlands are constantly changed by inland sediment and the influence of waves and storm surges, so the accurate classification of the coastal wetlands in the Yellow River Delta is of great significance for the rational utilization, development and protection of wetland resources. In this study, the Yellow River Delta sentinel-2 multispectral data were processed by super-resolution synthesis, and the feature bands were optimized. The optimal feature-band combination scheme was screened using the OIF algorithm. A deep learning model attention mechanism ResNet based on feature optimization with attention mechanism integration into the ResNet network is proposed. Compared with the classical machine learning model, the AM_ResNet model can effectively improve the classification accuracy of the wetlands in the Yellow River Delta. The overall accuracy was 94.61% with a Kappa of 0.93, and they were improved by about 6.99% and 0.1, respectively, compared with the best-performing Random Forest Classification in machine learning. The results show that the method can effectively improve the classification accuracy of the wetlands in the Yellow River Delta.

Soil salinization has seriously affected agricultural production and ecological balance in the Yellow River Delta region. Rapid and accurate monitoring of soil salinity has become an urgent need. Traditional machine learning models tend to fall into local optimal values during the learning process, which reduces their accuracy. This paper introduces Circle map to enhance the crayfish optimization algorithm (COA), which is then integrated with the regularized extreme learning machine (RELM) model, aiming to improve the accuracy of soil salinity content (SSC) inversion in the Yellow River Delta region. We employed Landsat5 TM remote sensing images and measured salinity data to develop spectral indices, such as the band index, salinity index, vegetation index, and comprehensive index, selecting the optimal modeling variable group through Pearson correlation analysis and variable projection importance analysis. The back propagation neural network (BPNN), RELM, and improved crayfish optimization algorithm–regularized extreme learning machine (ICOA-RELM) models were constructed using measured data and selected variable groups for SSC inversion. The results indicate that the ICOA-RELM model enhances the R2 value by an average of about 0.1 compared to other models, particularly those using groups of variables filtered by variable projection importance analysis as input variables, which showed the best inversion effect (test set R2 value of 0.75, MAE of 0.198, RMSE of 0.249). The SSC inversion results indicate a higher salinization degree in the coastal regions of the Yellow River Delta and a lower degree in the inland areas, with moderate saline soil and severe saline soil comprising 48.69% of the total area. These results are consistent with the actual sampling results, which verify the practicability of the model. This paper’s methods and findings introduce an innovative and practical tool for monitoring and managing salinized soils in the Yellow River Delta, offering significant theoretical and practical benefits.

Soil moisture (SM) is a critical variable affecting ecosystem carbon and water cycles and their feedback to climate change. In this study, we proposed a convolutional neural network (CNN) model embedded with a residual block and attention module, named SMNet, to spatially downscale the European Space Agency (ESA) Climate Change Initiative (CCI) SM product. In the SMNet model, a lightweight Convolutional Block Attention Module (CBAM) dual-attention mechanism was integrated to comprehensively extract the spatial and channel information from the high-resolution input remote sensing products, the reanalysis meteorological dataset, and the topographic data. The model was employed to downscale the ESA CCI SM from its original spatial resolution of 25 km to 1 km in California, USA, in the annual growing season (1 May to 30 September) from 2003 to 2021. The original ESA CCI SM data and in situ SM measurements (0–5 cm depth) from the International Soil Moisture Network were used to validate the model’s performance. The results show that compared with the original ESA CCI SM data, the downscaled SM data have comparable accuracy with a mean correlation (R) and root mean square error (RMSE) of 0.82 and 0.052 m3/m3, respectively. Moreover, the model generates reasonable spatiotemporal SM patterns with higher accuracy in the western region and relatively lower accuracy in the eastern Nevada mountainous area. In situ site validation results in the SCAN, the SNOTEL network, and the USCRN reveal that the R and RMSE are 0.62, 0.63, and 0.77, and 0.077 m3/m3, 0.093 m3/m3, and 0.078 m3/m3, respectively. The results are slightly lower than the validation results from the original ESA CCI SM data. Overall, the validation results suggest that the SMNet downscaling model proposed in this study has satisfactory performance in handling the task of soil moisture downscaling. The downscaled SM model not only preserves a high level of spatial consistency with the original ESA CCI SM model but also offers more intricate spatial variations in SM depending on the spatial resolution of model input data.

Accurately predicting actual evapotranspiration (ETa) at the regional scale is crucial for efficient water resource allocation and management. While previous studies mainly focused on predicting site-scale ETa, in-depth studies on regional-scale ETa are relatively scarce. This study aims to address this issue by proposing a MulSA-ConvLSTM model, which combines the multi-headed self-attention module with the Pyramidally Attended Feature Extraction (PAFE) method. By extracting feature information and spatial dependencies in various dimensions and scales, the model utilizes remote sensing data from ERA5-Land and TerraClimate to attain regional-scale ETa prediction in Shandong, China. The MulSA-ConvLSTM model enhances the efficiency of capturing the trend of ETa successfully, and the prediction results are more accurate than those of the other contrast models. The Pearson’s correlation coefficient between observed and predicted values reaches 0.908. The study has demonstrated that MulSA-ConvLSTM yields superior performance in forecasting various ETa scenarios and is more responsive to climatic changes than other contrast models. By using a convolutional network feature extraction method, the PAFE method extracts global features via various convolutional kernels. The customized MulSAM module allows the model to concentrate on data from distinct subspaces, focusing on feature changes in multiple directions. The block-based training method is employed for the large-scale regional ETa prediction, proving to be effective in mitigating the constraints posed by limited hardware resources. This research provides a novel and effective method for accurately predicting regional-scale ETa.

Vegetation plays a vital role in the global carbon cycle, a function of particular significance in regulating carbon dioxide fluxes within tropical ecosystems. Therefore, it is crucial to enhance the precision of carbon dioxide flux estimates for tropical vegetation and to explore the determinants influencing carbon sequestration. In this study, Landsat series images and Sentinel-2 Multispectral Instrument satellite data were used to invert vegetation biophysical parameters, thereby improving the timeliness and resolution of state variables from the boreal ecosystem productivity simulator (BEPS). The BEPS model at a 30 m resolution was developed to accurately capture tropical vegetation carbon dioxide fluxes across Hainan Island (HN) over the preceding two decades. The impacts of climate variations and anthropogenic activities on the carbon dioxide fluxes of tropical vegetation were further quantified using quantile regression models and a land-use transfer matrix. Results indicate significant increases in both net primary productivity (NPP) and net ecosystem productivity (NEP) in HN during the period 2000–2020, by 5.81 and 4.29 g C/m2 year, respectively. Spatial trends in vegetation carbon dioxide fluxes exhibited a consistent decline from inland regions to coastal zones. Anthropogenic activities were the dominant factor in the reduced stability of coastal NPP, while the post-2005 vegetation restoration promoted the southward expansion of high NPP (>1200 g C/m2) in the central part of HN. NPP in this tropical island was more sensitive to temperature than to precipitation, with a 1 °C temperature increase resulting in 4.1 g C/m2 reduction in dry-season NPP compared to wet-season NPP. Upgrades of cropland quality and grassland restoration have improved NPP yields, and land use transfers have resulted in a 0.301 Tg C net increase in NPP. This study provides new insight into the improvement of the carbon dioxide flux model at a finer scale for tropical vegetation and highlights ecological construction as an adaptation strategy to enhance the carbon sinks of tropical vegetation under negative climate change conditions.

Accurately mapping the forest canopy height is vital for conserving forest ecosystems. Employing the forest height measured by satellite light detection and ranging (LiDAR) systems as ground samples to establish forest canopy height extrapolation (FCHE) models presents promising opportunities for mapping large-scale wall-to-wall forest canopy height. However, despite the potential to provide more samples and alleviate the stripe effect by synergistically using the data from two existing LiDAR datasets, Global Ecosystem Dynamics Investigation (GEDI) and Ice, Cloud, and land Elevation Satellite-2 (ICESat-2), the fundamental differences in their operating principles create measurement biases, and thus, there are few studies combining them for research. Furthermore, previous studies have typically employed existing regression algorithms as FCHE models to predict forest canopy height, without customizing a model that achieves optimal performance based on the current samples. These shortcomings constrain the accuracy of predicting forest canopy height using satellite LiDAR data. To surmount these difficulties, we proposed a genetic programming (GP) guided method for mapping forest canopy height by combining the GEDI and ICESat-2 LiDAR data with Sentinel-1/2, terrain, and climate data. In this method, GP autonomously constructs the fusion model of the GEDI and ICESat-2 datasets (hereafter GIF model) and the optimal FCHE model based on the explanatory variables for the specific study area. The outcomes demonstrate that the fusion of the GEDI and ICESat-2 data shows high consistency (R2 = 0.85, RMSE = 2.2m, pRMSE = 11.24%). The synergistic use of the GEDI and ICESat-2 data, coupled with the optimization of the FCHE model, substantially improves the precision of forest canopy height predictions, and finally achieves R2, RMSE, and pRMSE of 0.64, 3.38m, and 16.08%, respectively. In summary, our research presents a reliable approach to accurately estimate forest canopy height using remote sensing data by addressing measurement biases between the GEDI and ICESat-2 data and overcoming the limitations of traditional FCHE models.

Partitioning evapotranspiration (ET) into vegetation transpiration (T) and soil evaporation (E) is challenging, but it is key to improving the understanding of plant water use and changes in terrestrial ecosystems. Considering that the transpiration of vegetation at night is minimal and can be negligible, we established a machine learning model (i.e., extreme gradient boosting algorithm (XGBoost)) for soil evaporation estimation based on night-time evapotranspiration observation data from eddy covariance towers, remote sensing data, and meteorological reanalysis data. Daytime T was consequently calculated as the difference between the total evapotranspiration and predicted daytime soil evaporation. The soil evaporation estimation model was validated based on the remaining night-time ET data (i.e., model test dataset), the non-growing season ET data of the natural ecosystem, and ET data during the fallow periods of croplands. The validation results showed that XGBoost had a better performance in E estimation, with the average overall accuracy of NSE 0.657, R 0.806, and RMSE 11.344 W/m2. The average annual T/ET of the examined ten ecosystems was 0.50 ± 0.08, with the highest value in deciduous broadleaf forests (0.68 ± 0.11), followed by mixed forests (0.61 ± 0.04), and the lowest in croplands (0.40 ± 0.08). We further examined the impact of the leaf area index (LAI) and vapor pressure deficit (VPD) on the variation in T/ET. Overall, at the interannual scale, LAI contributed 28% to the T/ET variation, while VPD had a small (5%) influence. On a seasonal scale, LAI also exerted a stronger impact (1~90%) on T/ET compared to VPD (1~77%). Our study suggests that the XGBoost machine learning model has good performance in ET partitioning, and this method is mainly data-driven without prior knowledge, which may provide a simple and valuable method in global ET partitioning and T/ET estimation.

Remote sensing techniques for shoreline extraction are crucial for monitoring changes in erosion rates, surface hydrology, and ecosystem structure. In recent years, Convolutional neural networks (CNNs) have developed as a cutting-edge deep learning technique that has been extensively used in shoreline extraction from remote sensing images, owing to their exceptional feature extraction capabilities. They are progressively replacing traditional methods in this field. However, most CNN models only focus on the features in local receptive fields, and overlook the consideration of global contextual information, which will hamper the model’s ability to perform a precise segmentation of boundaries and small objects, consequently leading to unsatisfactory segmentation results. To solve this problem, we propose a parallel semantic segmentation network (TCU-Net) combining CNN and Transformer, to extract shorelines from multispectral remote sensing images, and improve the extraction accuracy. Firstly, TCU-Net imports the Pyramid Vision Transformer V2 (PVT V2) network and ResNet, which serve as backbones for the Transformer branch and CNN branch, respectively, forming a parallel dual-encoder structure for the extraction of both global and local features. Furthermore, a feature interaction module is designed to achieve information exchange, and complementary advantages of features, between the two branches. Secondly, for the decoder part, we propose a cross-scale multi-source feature fusion module to replace the original UNet decoder block, to aggregate multi-scale semantic features more effectively. In addition, a sea–land segmentation dataset covering the Yellow Sea region (GF Dataset) is constructed through the processing of three scenes from Gaofen-6 remote sensing images. We perform a comprehensive experiment with the GF dataset to compare the proposed method with mainstream semantic segmentation models, and the results demonstrate that TCU-Net outperforms the competing models in all three evaluation indices: the PA (pixel accuracy), F1-score, and MIoU (mean intersection over union), while requiring significantly fewer parameters and computational resources compared to other models. These results indicate that the TCU-Net model proposed in this article can extract the shoreline from remote sensing images more effectively, with a shorter time, and lower computational overhead.

Accurate prediction of future chlorophyll-a (Chl-a) concentrations is of great importance for effective management and early warning of marine ecological systems. However, previous studies primarily focused on chlorophyll-a inversion and reconstruction, while methods for predicting Chl-a concentrations remain limited. To address this issue, we adopted four deep learning approaches, including Convolutional LSTM Network (ConvLSTM), Convolutional Neural Network-Long Short-Term Memory (CNN-LSTM), Eidetic 3D LSTM (E3D-LSTM), and Self-Attention ConvLSTM (SA-ConvLSTM) models, to predict Chl-a over the Yellow Sea and Bohai Sea (YBS) in China. Furthermore, 14 environmental variables obtained from the remote sensing data of Moderate-resolution Imaging Spectroradiometer (MODIS) and ECMWF Reanalysis v5 (ERA5) were utilized to predict the Chl-a concentrations in the study area. The results showed that all four models performed satisfactorily in predicting Chl-a concentrations in the YBS, with SA-ConvLSTM exhibiting a closer approximation to true values. Furthermore, we analyzed the impact of the Self-Attention Memory Module (SAM) on the prediction results. Compared to the ConvLSTM model, the SA-ConvLSTM model integrated with the SAM module better captured subtle large-scale variations within the study area. The SA-ConvLSTM model exhibited the highest prediction accuracy, and the one-month Pearson correlation coefficient reached 0.887. Our study provides an available approach for anticipating Chl-a concentrations over a large area of sea.

Sea fog is a weather hazard along the coast and over the ocean that seriously threatens maritime activities. In the deep learning approach, it is difficult for convolutional neural networks (CNNs) to fully consider global context information in sea fog research due to their own limitations, and the recognition of sea fog edges is relatively vague. To solve the above problems, this paper puts forward an ECA-TransUnet model for daytime sea fog recognition, which consists of a combination of a CNN and a transformer. By designing a two-branch feed-forward network (FFN) module and introducing an efficient channel attention (ECA) module, the model can effectively take into account long-range pixel interactions and feature channel information to capture the global contextual information of sea fog data. Meanwhile, to solve the problem of insufficient existing sea fog detection datasets, we investigated sea fog events occurring in the Yellow Sea and Bohai Sea and their territorial waters, extracted remote sensing images from Moderate Resolution Imaging Spectroradiometer (MODIS) data at corresponding times, and combined data from the Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observation (CALIPSO), cloud and sea fog texture features, and waveband feature information to produce a manually annotated sea fog dataset. Our experiments showed that the proposed model achieves 94.5% accuracy and an 85.8% F1 score. Compared with the existing models relying only on CNNs such as UNet, FCN8s, and DeeplabV3+, it achieves state-of-the-art performance in sea fog recognition.

Understanding variations in the temporal and spatial distribution of vegetation phenology is essential for adapting to and mitigating future climate change and urbanization. However, there have been limited vegetation phenology studies within small-scale areas such as urban environments over the past decades. Therefore, the present study focuses on Jinan city, Shandong Province, China as the study area and employs a more refined local climate zone (LCZ) approach to investigate spatial and temporal variations in vegetation phenology. The three phenological indicators used in this study from 2007 to 2018, namely, the start of growing season (SOS), the end of growing season (EOS), and the length of growing season (LOS), were provided by MODIS satellite data. The SOS, EOS, and LOS were superimposed on the LCZ and urban–rural gradient to analyze the changes in vegetation phenology, and the applicability of these two analysis methods in the study of urban vegetation phenology was compared by the honest significant difference test. We found that the SOS, EOS, and LOS of vegetation in the study area generally showed an advance, delay, and extension trend, respectively. The means of the SOS and EOS along different LCZ types varied noticeably more than those along urban–rural gradients. In 2016, 77.5%, 80.0%, and 75.8% of LCZ pairs indicated statistically significant differences for SOS, EOS, and LOS, respectively. This study provides a new perspective for the study of urban vegetation phenology which can help in management of urban-scale environments, identification of areas rich in biodiversity, and conservation and restoration of biodiversity in urban areas.

Accurately estimating aboveground biomass (AGB) is essential for assessing the ecological functions of coastal wetlands, and AGB of coastal wetlands is affected by Land use/land cover (LULC) types of conversion. To address this issue, in the current study, we used the Boreal Ecosystem Productivity Simulator (BEPS) model to simulate the AGB of the Yellow River Delta during 2000–2015. Based on the LULC types transform, we analyzed the spatiotemporal dynamics of the AGB simulation results and their relationship with the human-nature driving process. At the same time, combined with the actual situation of LULC transformation in the Yellow River Delta, a new driving process (Replace) is introduced. The results show that from 2000 to 2015, 755 km2 of natural wetlands in the Yellow River Delta were converted into constructed wetlands, and AGB increased by 386,121 Mg. Both single and multiple driving processes contributed to the decrease in AGB, with 72.6% of the increase in AGB associated with single artificial (such as Restore) or natural (such as Accretion) driving processes and 27.4% of the increase in AGB associated with multiple driving processes. Naturally driven processes bring much more AGB gain than loss, and human-driven processes bring the largest AGB gain. LULC conversion brought on by anthropogenic and natural driving processes has a large impact on AGB in coastal wetlands, and exploring this impact has a significant role in planning coastal wetland land use and protecting blue carbon ecosystems.

The rapid intensification of drought, commonly known as flash drought, has recently drawn widespread attention from researchers. However, how the characteristics and drivers, as well as the ecological impacts of rapidly intensified droughts, differ from those of slowly intensified ones still remains unclear over the globe. To this end, we defined three types of droughts based on the root zone soil moisture (RZSM) decline rates, flash droughts, general droughts, and creep droughts, and then implemented a comparative analysis between them across the globe and the 26 Intergovernmental Panel on Climate Change Special Report on Extremes (IPCC-SREX) regions. The ensemble of RZSM from multiple reanalysis datasets was used to reduce the uncertainties. According to the frequency analysis, our findings suggest that flash droughts contributed to the majority of drought events during 1980–2019, indicating the prevalence of rapid transition from an energy-limited to a water-limited condition in most of the regions. The comparative results of vegetation responses show that flash droughts are more likely to happen in the growing season, leading to faster but relatively minor vegetation deterioration compared to the slowly intensified ones. By analyzing the precipitation and temperature anomalies in the month of drought onset, we found the role of temperature (precipitation) on drought intensification can be generalized as the warmer (drier) the climate is or the faster the drought intensifies, but the main driving forces vary by region. Unlike temperature dominating in midwestern Eurasia and northern high latitudes, precipitation plays a prominent role in the monsoon regions. However, the temperature is expected to be the decisive driver in the warming future, given its monotonically increased contribution over the past four decades.

Estimating gross primary productivity (GPP) is important for simulating the subsequent carbon cycle elements and assessing the capacity of terrestrial ecosystems to support the sustainable development of human society. Light use efficiency (LUE) models were widely used to estimate GPP due to their concise model structures. However, quantifying LUEmax (maximum light use efficiency) and representing the responses of photosynthesis to environmental factors are still subject to large uncertainties, which lead to substantial errors in GPP simulations. In this study, we developed a hybrid model based on machine learning and a LUE model for GPP estimates. This hybrid model was built by targeting LUE with a machine learning approach, namely multi-layer perceptron (MLP), and then, estimating GPP within a LUE model framework with the MLP-based LUE and other required inputs. We trained the hybrid LUE (H-LUE) model and then, compared it against two conventional LUE models, the vegetation photosynthesis model (VPM) and vegetation photosynthesis and respiration model (VPRM), regarding GPP estimation, using tower-based daily-scale observations from 180 flux sites that cover nine different plant function types (PFTs). The results revealed better performance (R2 = 0.86 and RMSE = 1.79 gC m−2 d−1 on the test dataset) of the H-LUE model compared to the VPM and VPRM. Evaluations of the three models under four different extreme conditions consistently revealed better performance of the H-LUE model, indicating greater adaptability of the model to varied environments in the context of climate change. Furthermore, we also found that the H-LUE model can reasonably represent the responses of the LUE to meteorological variables. Our study revealed the reliable and robust performance of the developed hybrid LUE when simulating GPP across global biomes, providing references for developing better hybrid GPP models.

Drought is an extremely dangerous natural hazard that causes water crises, crop yield reduction, and ecosystem fires. Researchers have developed many drought indices based on ground-based climate data and various remote sensing data. Ground-based drought indices are more accurate but limited in coverage; while the remote sensing drought indices cover larger areas but have poor accuracy. Applying data-driven models to fuse multi-source remote sensing data for reproducing composite drought index may help fill this gap and better monitor drought in terms of spatial resolution. Machine learning methods can effectively analyze the hierarchical and non-linear relationships between the independent and dependent variables, resulting in better performance compared with traditional linear regression models. In this study, seven drought impact factors from the Moderate Resolution Imaging Spectroradiometer (MODIS) satellite sensor, Global Precipitation Measurement Mission (GPM), and Global Land Data Assimilation System (GLDAS) were used to reproduce the standard precipitation evapotranspiration index (SPEI) for Shandong province, China, from 2002 to 2020. Three machine learning methods, namely bias-corrected random forest (BRF), extreme gradient boosting (XGBoost), and support vector machines (SVM) were applied as regression models. Then, the best model was used to construct the spatial distribution of SPEI. The results show that the BRF outperforms XGBoost and SVM in SPEI estimation. The BRF model can effectively monitor drought conditions in areas without ground observation data. The BRF model provides comprehensive drought information by producing a spatial distribution of SPEI, which provides reliability for the BRF model to be applied in drought monitoring.

Accurate extraction of crop cultivated area and spatial distribution is essential for food security. Crop classification methods based on machine learning and deep learning and remotely sensed time-series data are widely utilized to detect crop planting area. However, few studies assess the effectiveness of machine learning and deep learning algorithm integrated time-series satellite data for identifying multiple crop type classification over a large-scale region. Hence, this study aims to evaluate the effectiveness of machine learning and deep learning models in crop classification and provide a framework for large-scale multiple crop type classification based on time-series of satellite data. The time-series of the Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), and NaE (combined NDVI and EVI) were adopted as input features, and four widely used machine learning models, including Support Vector Machine (SVM), Random Forest (RF), K-Nearest Neighbor (KNN), and their integration (Stacking) were compared to examine the performance of multiple crop types (i.e., wheat, corn, wheat-corn, early rice, and early rice-late rice) classification in the North China Plain. The performance of two types of deep learning classifiers: the One-dimensional Convolutional Neural Network (Conv1D) and Long Short-Term Memory Networks (LSTM), were also tested. The results showed that the NaE feature performed best among three input features, and the Stacking model produced the highest accuracy (77.12%) compared to other algorithms.

Droughts are ranked among the most devastating agricultural disasters that occur naturally in the world. East Africa is the most vulnerable and drought-prone region worldwide. In this study, four drought indices were used as input variables for drought assessment from 1982 to 2015. This work applied the SMDM algorithm to the integrated approach of OLR and Hurst exponent. The Detrended Fluctuation Analysis (DFA) and Ordinary Least Square (OLR) were merged to compute the trend and persistence (Hurst exponent) of the drought indices. Result indicates that the OLR at time scale 1, 6, and 12 shows a similar distribution with positive (negative) trends scattered in the Northwest (Northeast and Southern) parts of the study area which differs with the OLR aggregated at a 3-month time scale. The percentage pixel distribution for OLR-1, OLR-3, OLR-6, and OLR-12 is 18.2 (81.8), 72.5 (27.5), 32.9 (67.1), and 36.9 (63.1) for increasing (decreasing) trends respectively. Additionally, results indicate that DFA-1 is highly persistent with few random pixels scattered around Ethiopia, South Sudan and Tanzania, with percentage pixels as 88.7, 11.3 and 0.1 representing h > 0.5, h = 0.5, and h < 0.5, respectively. DFA-6 shows high (low) pixels representing h > 0.5 (h > 1), respectively. Meanwhile, for DFA-3 and DFA-12, the distribution shows persistence and a random walk, respectively. Drought conditions may eventually persist, reverse or vary drastically in an unpredictable manner depending on the driving forces. Overall, the drought risk map at 1-, 3-, and 6-month aggregates has shown severe degradation in Southern Kenya and Tanzania while noticeable improvements are seen in western Ethiopia and South Sudan.

Droughts are one of the world’s most destructive natural disasters. In large regions of Africa, droughts can have strong environmental and socioeconomic impacts. Understanding the mechanism that drives drought and predicting its variability is important for enhancing early warning and disaster risk management. Taking North and West Africa as the study area, this study adopted multi-source data and various statistical analysis methods, such as the joint probability density function (JPDF), to study the meteorological drought and return years across a long term (1982–2018). The standardized precipitation index (SPI) was used to evaluate the large-scale spatiotemporal drought characteristics at 1–12-month timescales. The intensity, severity, and duration of drought in the study area were evaluated using SPI–12. At the same time, the JPDF was used to determine the return year and identify the intensity, duration, and severity of drought. The Mann-Kendall method was used to test the trend of SPI and annual precipitation at 1–12-month timescales. The pattern of drought occurrence and its correlation with climate factors were analyzed. The results showed that the drought magnitude (DM) of the study area was the highest in 2008–2010, 2000–2003, and 1984–1987, with the values of 5.361, 2.792, and 2.187, respectively, and the drought lasting for three years in each of the three periods. At the same time, the lowest DM was found in 1997–1998, 1993–1994, and 1991–1992, with DM values of 0.113, 0.658, and 0.727, respectively, with a duration of one year each time. It was confirmed that the probability of return to drought was higher when the duration of drought was shorter, with short droughts occurring more regularly, but not all severe droughts hit after longer time intervals. Beyond this, we discovered a direct connection between drought and the North Atlantic Oscillation Index (NAOI) over Morocco, Algeria, and the sub-Saharan countries, and some slight indications that drought is linked with the Southern Oscillation Index (SOI) over Guinea, Ghana, Sierra Leone, Mali, Cote d’Ivoire, Burkina Faso, Niger, and Nigeria.

Grasslands are one of the most important terrestrial ecosystems on the planet and have significant economic and ecological value. Accurate and rapid discrimination of grassland communities is critical to the conservation and utilization of grassland resources. Previous studies that explored grassland communities were mainly based on field surveys or airborne hyperspectral and high-resolution imagery. Limited by workload and cost, these methods are typically suitable for small areas. Spaceborne mid-resolution RS images (e.g., Sentinel, Landsat) have been widely used for large-scale vegetation observations owing to their large swath width. However, there still keep challenges in accurately distinguishing between different grassland communities using these images because of the strong spectral similarity of different communities and the suboptimal performance of models used for classification. To address this issue, this paper proposed a superpixel-based grassland community classification method using Genetic Programming (GP)-optimized classification model with Sentinel-2 multispectral bands, their derived vegetation indices (VIs) and textural features, and Sentinel-1 Synthetic Aperture Radar (SAR) bands and the derived textural features. The proposed method was evaluated in the Siziwang grassland of China. Our results showed that the addition of VIs and textures, as well as the use of GP-optimized classification models, can significantly contribute to distinguishing grassland communities, and the proposed approach classified the seven communities in Siziwang grassland with an overall accuracy of 84.21% and a kappa coefficient of 0.81. We concluded that the classification method proposed in this paper is capable of distinguishing grassland communities with high accuracy at a regional scale.

Accurate estimates of evapotranspiration (ET) over croplands on a regional scale can provide useful information for agricultural management. The hybrid ET model that combines the physical framework, namely the Penman-Monteith equation and machine learning (ML) algorithms, have proven to be effective in ET estimates. However, few studies compared the performances in estimating ET between multiple hybrid model versions using different ML algorithms. In this study, we constructed six different hybrid ET models based on six classical ML algorithms, namely the K nearest neighbor algorithm, random forest, support vector machine, extreme gradient boosting algorithm, artificial neural network (ANN) and long short-term memory (LSTM), using observed data of 17 eddy covariance flux sites of cropland over the globe. Each hybrid model was assessed to estimate ET with ten different input data combinations. In each hybrid model, the ML algorithm was used to model the stomatal conductance (Gs), and then ET was estimated using the Penman-Monteith equation, along with the ML-based Gs. The results showed that all hybrid models can reasonably reproduce ET of cropland with the models using two or more remote sensing (RS) factors. The results also showed that although including RS factors can remarkably contribute to improving ET estimates, hybrid models except for LSTM using three or more RS factors were only marginally better than those using two RS factors. We also evidenced that the ANN-based model exhibits the optimal performance among all ML-based models in modeling daily ET, as indicated by the lower root-mean-square error (RMSE, 18.67–21.23 W m−2) and higher correlations coefficient (r, 0.90–0.94). ANN are more suitable for modeling Gs as compared to other ML algorithms under investigation, being able to provide methodological support for accurate estimation of cropland ET on a regional scale.

Estimating yield potential (Yp) and quantifying the contribution of suboptimum field managements to the yield gap (Yg) of crops are important for improving crop yield effectively. However, achieving this goal on a regional scale remains difficult because of challenges in collecting field management information. In this study, we retrieved crop management information (i.e., emerging stage information and a surrogate of sowing date (SDT)) from a remote sensing (RS) vegetation index time series. Then, we developed a new approach to quantify maize Yp, total Yg, and the suboptimum SDT-induced Yg (Yg0) using a process-based RS-driven crop yield model for maize (PRYM–Maize), which was developed in our previous study. PRYM–Maize and the newly developed method were used over the North China Plain (NCP) to estimate Ya, Yp, Yg, and Yg0 of summer maize. Results showed that PRYM–Maize outputs reasonable estimates for maize yield over the NCP, with correlations and root mean standard deviation of 0.49 ± 0.24 and 0.88 ± 0.14 t hm−2, respectively, for modeled annual maize yields versus the reference value for each year over the period 2010 to 2015 on a city level. Yp estimated using our new method can reasonably capture the spatial variations in site-level estimates from crop growth models in previous literature. The mean annual regional Yp of 2010–2015 was estimated to be 11.99 t hm−2, and a Yg value of 5.4 t hm−2 was found between Yp and Ya on a regional scale. An estimated 29–42% of regional Yg in each year (2010–2015) was induced by suboptimum SDT. Results also show that not all Yg0 was persistent over time. Future studies using high spatial-resolution RS images to disaggregate Yg0 into persistent and non-persistent components on a small scale are required to increase maize yield over the NCP.

Urbanization is an increasing phenomenon around the world, causing many adverse effects in urban areas. Urban heat island is are of the most well-known phenomena. In the present study, surface urban heat islands (SUHI) were studied for seven megacities of the South Asian countries from 2000–2019. The urban thermal environment and relationship between land surface temperature (LST), land use landcover (LULC) and vegetation were examined. The connection was explored with remote-sensing indices such as urban thermal field variance (UTFVI), surface urban heat island intensity (SUHII) and normal difference vegetation index (NDVI). LULC maps are classified using a CART machine learning classifier, and an accuracy table was generated. The LULC change matrix shows that the vegetated areas of all the cities decreased with an increase in the urban areas during the 20 years. The average LST in the rural areas is increasing compared to the urban core, and the difference is in the range of 1–2 (°C). The SUHII linear trend is increasing in Delhi, Karachi, Kathmandu, and Thimphu, while decreasing in Colombo, Dhaka, and Kabul from 2000–2019. UTFVI has shown the poor ecological conditions in all urban buffers due to high LST and urban infrastructures. In addition, a strong negative correlation between LST and NDVI can be seen in a range of −0.1 to −0.6.

As climate change intensifies, surface vegetation productivity and carbon exchange between terrestrial ecosystems and the atmosphere are significantly affected by the variation of climatic factors. Due to the sensitivity of grasslands to these climatic factors, it is crucial to understand the response of vegetation greenness, or carbon exchange within grasslands, to environment factor dynamics. In this study, we used solar-induced chlorophyll fluorescence (SIF), precipitation (P), vapor pressure deficit (VPD), evaporative stress (ES), and root zone soil moisture (RSM) derived from remote sensing, reanalysis, and assimilation datasets to explore the response of vegetation greenness within Eurasian Steppe to climatic factors. Our results indicated deseasonlization based on the Seasonal-Trend decomposition using Loess (STL) method, which was an effective means to remove the seasonality disturbances that affect the qualification of the relationship between SIF and the four climatic factors. The response of SIF had a time lag effect on these climatic factors, and the longer the response period, the greater the impact on the correlation of SIF with P, VPD, ES, and RSM. We also found, among the four factors, that the response of SIF to ES was the timeliest. The findings of this study emphasized the impact of the seasonality and time lag effect on the dynamic response between variables, and provided references to the attribution and monitoring of vegetation greenness and ecosystem productivity.

Drought has devastating impacts on agriculture and other ecosystems, and its occurrence is expected to increase in the future. However, its spatiotemporal impacts on net primary productivity (NPP) in Mongolia have remained uncertain. Hence, this paper focuses on the impact of drought on NPP in Mongolia. The drought events in Mongolia during 2003–2018 were identified using the Moderate Resolution Imaging Spectroradiometer (MODIS) normalized difference vegetation index (NDVI). The Boreal Ecosystem Productivity Simulator (BEPS)-derived NPP was computed to assess changes in NPP during the 16 years, and the impacts of drought on the NPP of Mongolian terrestrial ecosystems was quantitatively analyzed. The results showed a slightly increasing trend of the growing season NPP during 2003–2018. However, a decreasing trend of NPP was observed during the six major drought events. A total of 60.55–87.75% of land in the entire country experienced drought, leading to a 75% drop in NPP. More specifically, NPP decline was prominent in severe drought areas than in mild and moderate drought areas. Moreover, this study revealed that drought had mostly affected the sparse vegetation NPP. In contrast, forest and shrubland were the least affected vegetation types.

Drought, a climate-related disaster impacting a variety of sectors, poses challenges for millions of people in South Asia. Accurate and complete drought information with a proper monitoring system is very important in revealing the complex nature of drought and its associated factors. In this regard, deep learning is a very promising approach for delineating the non-linear characteristics of drought factors. Therefore, this study aims to monitor drought by employing a deep learning approach with remote sensing data over South Asia from 2001–2016. We considered the precipitation, vegetation, and soil factors for the deep forwarded neural network (DFNN) as model input parameters. The study evaluated agricultural drought using the soil moisture deficit index (SMDI) as a response variable during three crop phenology stages. For a better comparison of deep learning model performance, we adopted two machine learning models, distributed random forest (DRF) and gradient boosting machine (GBM). Results show that the DFNN model outperformed the other two models for SMDI prediction. Furthermore, the results indicated that DFNN captured the drought pattern with high spatial variability across three penology stages. Additionally, the DFNN model showed good stability with its cross-validated data in the training phase, and the estimated SMDI had high correlation coefficient R2 ranges from 0.57~0.90, 0.52~0.94, and 0.49~0.82 during the start of the season (SOS), length of the season (LOS), and end of the season (EOS) respectively. The comparison between inter-annual variability of estimated SMDI and in-situ SPEI (standardized precipitation evapotranspiration index) showed that the estimated SMDI was almost similar to in-situ SPEI. The DFNN model provides comprehensive drought information by producing a consistent spatial distribution of SMDI which establishes the applicability of the DFNN model for drought monitoring.

Satellite-borne multispectral data are suitable for regional-scale grassland community classification owing to comprehensive coverage. However, the spectral similarity of different communities makes it challenging to distinguish them based on a single multispectral data. To address this issue, we proposed a support vector machine (SVM)–based method integrating multispectral data, two-band enhanced vegetation index (EVI2) time-series, and phenological features extracted from Chinese GaoFen (GF)-1/6 satellite with (16 m) spatial and (2 d) temporal resolution. To obtain cloud-free images, the Enhanced Spatial and Temporal Adaptive Reflectance Fusion Model (ESTARFM) algorithm was employed in this study. By using the algorithm on the coarse cloudless images at the same or similar time as the fine images with cloud cover, the cloudless fine images were obtained, and the cloudless EVI2 time-series and phenological features were generated. The developed method was applied to identify grassland communities in Ordos, China. The results show that the Caragana pumila Pojark, Caragana davazamcii Sanchir and Salix schwerinii E. L. Wolf grassland, the Potaninia mongolica Maxim, Ammopiptanthus mongolicus S. H. Cheng and Tetraena mongolica Maxim grassland, the Caryopteris mongholica Bunge and Artemisia ordosica Krasch grassland, the Calligonum mongolicum Turcz grassland, and the Stipa breviflora Griseb and Stipa bungeana Trin grassland are distinguished with an overall accuracy of 87.25%. The results highlight that, compared to multispectral data only, the addition of EVI2 time-series and phenological features improves the classification accuracy by 9.63% and 14.7%, respectively, and even by 27.36% when these two features are combined together, and indicate the advantage of the fine images in this study, compared to 500 m moderate-resolution imaging spectroradiometer (MODIS) data, which are commonly used for grassland classification at regional scale, while using 16 m GF data suggests a 23.96% increase in classification accuracy with the same extracted features. This study indicates that the proposed method is suitable for regional-scale grassland community classification.

Studying the significant impacts of drought on vegetation is crucial to understand its dynamics and interrelationships with precipitation, soil moisture, and temperature. In North and West Africa regions, the effects of drought on vegetation have not been clearly stated. Therefore, the present study aims to bring out the drought fluctuations within various types of Land Cover (LC) (Grasslands, Croplands, Savannas, and Forest) in North and West Africa regions. The drought characteristics were evaluated by analyzing the monthly Self-Calibrating Palmer Drought Severity Index (scPDSI) in different timescale from 2002 to 2018. Then, the frequency of droughts was examined over the same period. The results have revealed two groups of years (dry years and normal years), based on drought intensity. The selected years were used to compare the shifting between vegetation and desert. The Vegetation Condition Index (VCI), the Temperature Condition Index (TCI), the Precipitation Condition Index (PCI), and the Soil Moisture Condition Index (SMCI) were also used to investigate the spatiotemporal variation of drought and to determine which LC class was more vulnerable to drought risk. Our results revealed that Grasslands and Croplands in the West region, and Grasslands, Croplands, and Savannas in the North region are more sensitive to drought. A higher correlation was observed among the Normalized Difference Vegetation Index (NDVI), Land Surface Temperature (LST), Tropical Rainfall Measuring Mission (TRMM), and Soil Moisture (SM). Our findings suggested that NDVI, TRMM, and SM are more suitable for monitoring drought over the study area and have a reliable accuracy (R2 > 0.70) concerning drought prediction. The outcomes of the current research could, explicitly, contribute progressively towards improving specific drought mitigation strategies and disaster risk reduction at regional and national levels.

The Himalayan region, a major source of fresh water, is recognized as a water tower of the world. Many perennial rivers originate from Nepal Himalaya, located in the central part of the Himalayan region. Snowmelt water is essential freshwater for living, whereas it poses flood disaster potential, which is a major challenge for sustainable development. Climate change also largely affects snowmelt hydrology. Therefore, river discharge measurement requires crucial attention in the face of climate change, particularly in the Himalayan region. The snowmelt runoff model (SRM) is a frequently used method to measure river discharge in snow-fed mountain river basins. This study attempts to investigate snowmelt contribution in the overall discharge of the Budhi Gandaki River Basin (BGRB) using satellite remote sensing data products through the application of the SRM model. The model outputs were validated based on station measured river discharge data. The results show that SRM performed well in the study basin with a coefficient of determination (R2) >0.880. Moreover, this study found that the moderate resolution imaging spectroradiometer (MODIS) snow cover data and European Centre for Medium-Range Weather Forecasts (ECMWF) meteorological datasets are highly applicable to the SRM in the Himalayan region. The study also shows that snow days have slightly decreased in the last three years, hence snowmelt contribution in overall discharge has decreased slightly in the study area. Finally, this study concludes that MOD10A2 and ECMWF precipitation and two-meter temperature products are highly applicable to measure snowmelt and associated discharge through SRM in the BGRB. Moreover, it also helps with proper freshwater planning, efficient use of winter water flow, and mitigating and preventive measures for the flood disaster.

Accurate and timely information on soil salinity is crucial for vegetation growth and agricultural productivity in coastal regions. This study investigates the potential of using Wifi POGO, an in situ electromagnetic sensor, for soil salinity assessment over saline coastal regions in eastern China. The sensor readings, soil moisture, and temperature-corrected apparent electrical conductivity (ECa) were used to generate models for EC1:5 (a surrogate for soil salinity) estimation. Two salty areas with distinct soil textures, sandy loam (Shuntai) and clay (Dongxin), were selected. This study revealed that the difference between soil salinity and the in situ measured soil ECa (i.e., EC1:5-ECa) had a strong curvilinear relationship with soil moisture. Such a relationship allows for the direct estimation of soil salinity from soil ECa with the aid of soil moisture information. Both ECa and soil moisture can be measured in situ using a Wifi POGO, a low-cost ground-based soil sensor. By using the leave-one-out cross-validation (LOOCV), the achieved root mean square error (RMSE) and relative RMSE (RRMSE) in EC1:5 estimation were 0.0109 S/m and 19.24% respectively in Shuntai, and 0.0157 S/m and 16.05%, in Dongxin. This new method offers a simple, cost-effective and reliable tool for assessing soil salinity in dynamic coastal regions.

Aerosol optical depth (AOD) is a critical variable in estimating aerosol concentration in the atmosphere, evaluating severity of atmospheric pollution, and studying their impact on climate. With the assistance of the 6S radiative transfer model, we simulated apparent reflectancein relation to AOD in each Moderate Resolution Imaging Spectroradiometer (MODIS) waveband in this study. The closeness of the relationship was used to identify the most and least sensitive MODIS wavebands. These two bands were then used to construct three aerosol indices (difference, ratio, and normalized difference) for estimating AOD quickly and effectively. The three indices were correlated, respectively, with in situ measured AOD at the Aerosol Robotic NETwork (AERONET) Lake Taihu, Beijing, and Xianghe stations. It is found that apparent reflectance of the blue waveband (band 3) is the most sensitive to AOD while the mid-infrared wavelength (band 7) is the least sensitive. The difference aerosol index is the most accurate in indicating aerosol-induced atmospheric pollution with a correlation coefficient of 0.585, 0.860, 0.685, and 0.333 at the Lake Taihu station, 0.721, 0.839, 0.795, and 0.629 at the Beijing station, and 0.778, 0.782, 0.837, and 0.643 at the Xianghe station in spring, summer, autumn and winter, respectively. It is concluded that the newly proposed difference aerosol index can be used effectively to study the level of aerosol-induced air pollution from MODIS satellite imagery with relative ease.