Search

Chromophoric dissolved organic matter (CDOM) in aquatic environments is an important component of the biogeochemical cycle and carbon cycle. The aim of this study is to investigate the long‐term changes in CDOM in shallow and eutrophic Chaohu Lake, as well as its relationship with climate, environment and social factors. Using long time series Landsat image data and machine learning technology, the spatiotemporal evolution of Chaohu CDOM since 1987 was reconstructed. A total of 180 samples were collected, which were divided into three parts based on regional and hydrological characteristics. The results show that the water quality in different regions were significantly different, and TN may be the key factor driving the change of CDOM in Chaohu Lake. Machine learning algorithms including random forest (RF), support vector regression (SVR), neural network (NN), multimodal deep learning (MDL) model and Extreme Gradient Boosting (XGBoost) were used, among which XGBoost model performed best (R2 = 0.955, mean absolute error [MAE] = 0.024 mg/L, root mean square error [RMSE] = 0.036 mg/L, bias = 1.005) and was used for CDOM spatiotemporal variation retrieval. The change of CDOM was seasonal, highest in August (0.67 m−1) and lowest in December (0.48 m−1), and the western lake is the main source of CDOM. Annual variability of the CDOM indicates that it began to decline after the completion of water pollution control in 2000. Temperature changes were closely related to CDOM (P < 0.01) and agricultural non‐point source pollution plays an important role in Chaohu Lake. This study will provide feasible methods and scientific basis for the long‐term remote sensing supervision of CDOM. [ABSTRACT FROM AUTHOR]

It is well known that the dry/wet boundaries of land surface temperature fractional vegetation coverage (LSTfc) trapezoid framework vary linearly with vegetation coverage (fc). In this study, the theoretical end-members algorithm is modified to continuously estimate the dry/wet end-members under varying vegetation conditions, causing the theoretical dry/wet boundaries to become non-linear. The findings revealed that the non-linear dry/wet boundaries were generally below the conventional linear dry/wet boundaries. Furthermore, the non-linear boundary scheme adopted herein provided better performance in estimating the latent heat flux (LE) and vegetation latent heat flux fraction (LEv/LE) compared to the linear boundary scheme. The parametric schemes of aerodynamic and thermodynamic roughness length and the aerodynamic resistance were the major drivers that result in dry/wet boundaries characteristics being highly non-linear. This study enhanced the physical process description in the LST-fc trapezoid framework and improved the prediction accuracy of regional LE and its components. [ABSTRACT FROM AUTHOR]

• A generalized flood mapping model named EMF was proposed for urban land-use types. • The optimal features were pinpointed through the random forest feature ranking method. • Social sensing combined with remote sensing data enhanced EMF modelʼs performance. • EMF mapping suggests that farmland and urban intersections are more prone to flooding. Flood susceptibility mapping is crucial for urban disaster management. However, the heterogeneity of urban land use and the complexity of terrain pose challenges to the accuracy and generalizability of flood models. In this paper, we propose a framework named the EMF model for urban flood mapping. Specifically, the social sensing and remote sensing were used for flooding information collection. The XGBoost, Support Vector Classifier (SVC), Multilayer Perceptron (MLP), and Multimodal Deep Learning (MDL), were used as the predictive models, and their results were ensembled using a Random Forest model to produce the final outcome. Results show that the EMF model outperforms the standalone models, with an accuracy of 0.942 on the training set and 0.940 on the testing set. The accuracy of the five models is ranked as EMF > MDL > XGBoost > SVC > MLP. The flood map indicates that flooding has a greater impact on the outskirts and suburbs of the city compared to its central urban areas. Farmland is the most affected land type, making up 54.8 % of the flooded area. Overall, the proposed framework enables rapid and accurate identification of flood-prone areas, providing technical support for managers in formulating effective flood prevention strategies. [ABSTRACT FROM AUTHOR]

• A deep learning approach is proposed to remotely assess CTP in the Great Lakes. • The model reliably estimates CTP from MODIS and outperforms 7 other algorithms. • The near-surface CTP of the Great Lakes decreased significantly from 2002 to 2022. • Reduced cropland and improved basin ecosystems drive widespread CTP decrease. • Adapting the model to VIIRS enhances seamless CTP estimation at moderate-scale. Total phosphorus (TP) is non-optically active, thus TP concentration (CTP) estimation using remote sensing still exists grand challenge. This study developed a deep neural network model (DNN) for CTP estimation with synchronous in-situ measurements and MODIS-derived remote sensing reflectance (R rs) (N = 3916). Using DNN, the annual and intra-annual CTP spatial distributions of the Great Lakes since 2002 were reconstructed. Then, the reconstructions were correlated to nine potential factors, e.g., Chlorophyll-a, snowmelt, and cropland, to explain seasonal and long-term CTP variations. The results showed that DNN reliably estimated CTP from MODIS R rs , with R2, mean absolute error (MAE), root mean squared error (RMSE), mean absolute percentage error (MAPE), and root mean squared logarithmic error (RMSLE) of 0.83, 1.05 μg/L, 2.95 μg/L, 9.92%, and 0.13 on the test set. The near-surface CTP in the Great Lakes decreased significantly (p < 0.05) during 2002 − 2022, primarily attributed to cropland reduction, coupled with improvements in basin natural ecosystems. The sensitivity analysis verified the model robustness when confronted with input feature changes < 35%. This result along with the marginal difference between CTP derived from two sensors (R2 = 0.76, MAE = 2.12 μg/L, RMSE = 2.51 μg/L, MAPE = 11.52%, RMSLE = 0.24) suggested the model transferability from MODIS to VIIRS. This transformation facilitated optimal usage of MODIS-related archive and enhanced the continuity of CTP estimation at moderate resolution. This study presents a practical method for spatiotemporal reconstruction of CTP using remote sensing, and contributes to better understandings of driving factors behind CTP variations in the Great Lakes. [Display omitted] [ABSTRACT FROM AUTHOR]

Water-quality monitoring for small urban waterbodies by remote sensing gets to be difficult due to the coarse spatial resolution of remote-sensing imagery. The recently launched Sentinel-2 produces imagery with a spatial resolution of 10 × 10 m and a temporal resolution of 5 days. It provides an opportunity to conduct high-frequency water-quality monitoring for small waterbodies. Since illegal discharges are an important issue for urban water management, total phosphorous (TP), total nitrogen (TN), and chemical oxygen demand (COD) were chosen as the target water-quality parameters. TP, TN and COD, however, are non-optically active parameters. There are fairly limited previous studies on retrieving these parameters in comparison with optically active parameters, e.g. Chlorophyll-a etc. Based on the fact that non-optically active parameters may be highly correlated with optically active parameters, this study compared 255 possible Sentinel-2 imagery band compositions to identify the most appropriate ones for TP, TN and COD retrieval. Three machine-learning models, namely Random Forest (RF), Support Vector Regression (SVR) and Neural Networks (NN), were compared to seek the most robust ones for retrieving the above non-optically active parameters. Results showed that the most appropriate band (hereafter termed as ' B i n d e x ' for brevity) compositions for TP, TN, and COD retrieval were ' B 3 + B 4 + B 5 + B 6 + B 7 + B 8 ', ' B 3 + B 4 + B 5 \breAK + B 6 + B 7 + B 8 ', and ' B 2 + B 3 + B 5 + B 6 + B 7 + B 8 ' respectively. The coefficient of determination (R2) of TP, TN, and COD estimations by NN, RF and SVR was 0.94, 0.88, and 0.86, respectively. The retrieval performances of these non-optically active parameters were hence significantly improved by the optimized machine-learning models and imagery band selection. The developed models have limitations in applying to other areas, thus band selection and tuning parameters with new data are necessary for different areas. The water-quality mapping obtained from Sentinel-2 imagery provided a full spatial coverage of the water-quality characterization for the entire water surface, and helped identify illegal discharges to urban waterbodies. This study provides a new practical and efficient water-quality monitoring strategy for managing small waterbodies. [ABSTRACT FROM AUTHOR]

• A satellite-based urban three-source evapotranspiration (ET) model is developed. • A complementary relationship-based parameterization scheme for estimation of soil evaporation is proposed. • Three-source remote sensing model for urban areas outperformed the existing single-source urban ET models. Estimation of urban evapotranspiration (ET) is meaningful for evaluating the urban heat island effect and promoting urban water conservation. While multi-source satellite-based ET models for farmland and natural ecosystems have been developed, these models lack the object-oriented solutions for the urban areas. In view of this, the present study proposed a Three-source Remote sensing model for Urban areas (TRU) to simulate the urban ET, which can discriminate between urban soil evaporation, vegetation transpiration, and impervious surface evaporation. The new parameterization scheme proposed herein is characterized by the theory of complementary relationship to estimate the urban soil evaporation. Evaporation losses in the impervious areas were independently estimated in accordance with the distribution of urban land use and land cover (LULC). The developed TRU model was evaluated for 63 cloud-free test days in Tianjin, China by using 30 m Landsat Operational Land Imager (OLI)/Enhanced Thematic Mapper Plus (ETM+) images. The outcomes of this study revealed a root-mean-square-error (RMSE) of 40.2 W/m2 and a bias of 10.8 W/m2 compared with the two eddy correlation (EC) observations for the instantaneous latent heat flux (LE) simulation. Moreover, the RMSE and bias were 0.092 and −0.023, respectively in simulating the vegetation latent heat flux fraction (LE v /LE) when verified with respect to the stable water isotope measurements. Due to limited water availability, the impervious urban surface evaporation magnitude was found quite low with minimal variation on different test days. The results emphasize the importance of coupling LULC information in the urban ET modeling conceptualization. Certainly, the TRU model improved the ET simulation accuracy in urban areas and could facilitate the mapping of urban ET components at high spatial resolution. [ABSTRACT FROM AUTHOR]

• Different end-members affect the latent heat flux (LE) simulation. • End-member types do not shift the vegetation latent heat fraction (LE v /LE) simulation. • Different end-members do not shift the spatial patterns of LE and LE v /LE. • Actual end-members method (AEM) is highly unreliable. • Theoretical end-members estimation through the Zhang's method (TEM-Z) is recommended. Land surface temperature-fractional vegetation coverage (LST-f c) trapezoid framework based models have been widely developed, applied, and verified to estimate regional latent heat flux (LE) and its partitioning process. The basis of operating these models lies in the determination of end-members within the LST-f c trapezoid. However, the knowledge of how the end-members affect the LE and ratio of vegetation latent heat flux (LE v) to latent heat flux (LE v /LE) estimation is lacking in the literature. In this study, three widely used end-members determination methods, viz., determination of actual end-members within the scatter plot of LST-f c (AEM), determination of theoretical end-members through the Zhang's method (Zhang et al., 2005, TEM-Z), and determination of theoretical end-members through Long's method (Long and Singh. 2012b, TEM-L) were inter-compared. The outcome of this study was verified at the 11 Eddy Correlation (EC) observation stations for the 183 test days during 2015 using the Moderate-resolution Imaging Spectroradiometer (MODIS) products, revealing that different end-members combinations greatly alter the LE simulation. However, tuning of end-members do not shift the LE v /LE estimates, and spatial patterns of the LE and LE v /LE remained unchanged. Moreover, the performance of TEM-Z is better than the AEM, while the TEM-L provide relatively unstable performance during the 183 test days. Conclusively, this study endorses to adopt the TEM-Z method to determine the end-members. This study provides a clear understanding of the effect of different end-members to LE and LE v /LE simulation within the LST-f c trapezoid framework, and gives a clear insight for model selection, operation, and subsequent, guidelines for model improvement. [ABSTRACT FROM AUTHOR]

Dissolved oxygen (DO) is an effective indicator for water pollution. However, since DO is a non-optically active parameter and has little impact on the spectrum captured by satellite sensors, research on estimating DO by remote sensing at multiple spatiotemporal scales is limited. In this study, the support vector regression (SVR) models were developed and validated using the remote sensing reflectance derived from both Landsat and Moderate Resolution Imaging Spectroradiometer (MODIS) data and synchronous DO measurements (N = 188) and water temperature of Lake Huron and three other inland waterbodies (N = 282) covering latitude between 22–45 °N. Using the developed models, spatial distributions of the annual and monthly DO variability since 1984 and the annual monthly DO variability since 2000 in Lake Huron were reconstructed for the first time. The impacts of five climate factors on long-term DO trends were analyzed. Results showed that the developed SVR-based models had good robustness and generalization (average R 2 = 0.91, root mean square percentage error = 2.65%, mean absolute percentage error = 4.21%), and performed better than random forest and multiple linear regression. The monthly DO estimates by Landsat and MODIS data were highly consistent (average R 2 = 0.88). From 1984 to 2019, the oxygen loss in Lake Huron was 6.56%. Air temperature, incident shortwave radiation flux density, and precipitation were the main climate factors affecting annual DO of Lake Huron. This study demonstrated that using SVR-based models, Landsat and MODIS data could be used for long-term DO retrieval at multiple spatial and temporal scales. As data-driven models, combining spectrum and water temperature as well as extending the training set to cover more DO conditions could effectively improve model robustness and generalization. [Display omitted] • Machine learning were first used to estimate DO from Landsat and MODIS data. • The approach supported the longest and most frequent DO pattern reconstruction. • The DO models were validated in lakes between 22 and 43 °N with R 2 above 0.71. • Adding physical factor temperature to DO retrieval improved model generalization. • Impacts of five climate factors on annual DO trend were quantitatively clarified. [ABSTRACT FROM AUTHOR]

• Three types of two-source remote sensing ET models were verified at 21 EC stations across China. • The model discrepancies induced by different physical mechanism were identified. • TSEB provides the best accuracy in areas with high LAI and vegetation height. • PM-MU shows an advantage in the areas without vegetation. • M−PCACA provides the most stable performance on various land surfaces. Remote sensing (RS)-based two-source models have been widely used to separate soil evaporation (E) and vegetation transpiration (T) in ecosystems. Alternative model assumptions and physical mechanisms employed in different two-source models cause the difference in model performance. Understanding of the characteristics and limitations of alternative two-source models is critical to ensure the most effective applications in various climatic zones and underlying surfaces. This study investigated two types of land surface temperature (LST) decomposition based two-source models, namely Two-Source Energy Balance Model (TSEB) and Modified Pixel Component Arranging and Comparing Algorithm (M-PCACA); as well as a meteorological factors-based Penman-Monteith type model (PM-MU) in the partitioning of E and T. The three models were compared at 21 flux stations across China, which included seven types of underlying surfaces and one isotope station. The study demonstrated that all three models were applicable in a large scale study for latent heat flux (LE) estimation. Overall, M-PCACA performed the best with average root-mean-square errors (RMSE) of 38.7 W/m2, while PM-MU performs the worst with average RMSE of 61.4 W/m2. For ET partitioning, the two models based on LST decomposition (TSEB and M-PCACA) gave more reliable estimates compared with stable water isotope observations. In addition, verification results for the seven types of underlying surfaces showed that the TSEB gave the best accuracy in the areas with high leaf area index (LAI) and vegetation height (e.g. forest), while the PM-MU had an advantage in the areas without vegetation (e.g. deserts and water bodies). Due to the characteristics of the land surface temperature-vegetation coverage (LST-f c) trapezoidal framework, the M-PCACA model had the most stable model performance for various underlying surfaces and gave the best estimates in grassland, wetland, and farmland areas. Using a thorough discussion, this study identified the sources of discrepancies induced by different physical mechanisms of three remote sensing-based two-source models. This investigation provided insights to better understand the alternative two-source models and provides guidance for model application. [ABSTRACT FROM AUTHOR]

[Display omitted] Remote sensing provides full-coverage and dynamic water quality monitoring with high efficiency and low consumption. Deep learning (DL) has been progressively used in water quality retrieval because it efficiently captures the potential relationship between target variables and imagery. In this study, the multimodal deep learning (MDL) models were developed and rigorously validated using atmospherically corrected Landsat remote sensing reflectance data and synchronous water quality measurements for estimating long-term Chlorophyll- a (Chl- a), total phosphorus (TP), and total nitrogen (TN) in Lake Simcoe, Canada. Since TP and TN are non-optically active, their retrievals were based on the fact that they are closely related to the optically active constituents (OACs) such as Chl- a. We trained the MDL models with one in-situ measured data set (for Chl- a , N = 315, for TP and TN, N = 303), validated the models with two independent data sets (N = 147), and compared the model performances with several DL, machine learning, and empirical algorithms. The results indicated that the MDL models adequately estimated Chl- a (mean absolute error (MAE) = 32.57%, Bias = 10.61%), TP (MAE = 42.58%, Bias = −2.82%), and TN (MAE = 35.05%, Bias = 13.66%), and outperformed several other candidate algorithms, namely the progressively decreasing deep neural network (DNN), a DNN with trainable parameters similar to MDL but without splitting input features, the eXtreme Gradient Boosting, the support vector regression, the NASA Ocean Color two-band and three-band ratio algorithms, and another empirical algorithm of Landsat data in clear lakes. Using the MDL models, we reconstructed the historical spatiotemporal patterns of Chl- a , TP, and TN in Lake Simcoe since 1984, and investigated the effects of two water quality improvement programs. In addition, the physical mechanism and interpretability of the MDL models were explored by quantifying the contribution of each feature to the model outputs. The framework proposed in this study provides a practical method for long-term Chl- a , TP, and TN estimation at the regional scale. [ABSTRACT FROM AUTHOR]

[Display omitted] • An enhanced deep learning approach was proposed to remotely assess water quality. • The approach outperformed other candidate algorithms in water quality mapping. • Merged OLI/MSI data supports water quality detecting with repeat cycle <3 days. • Model physical interpretability was explored to split the estimating interactions. • Impacts of 12 natural and anthropogenic factors on water quality were clarified. Remote sensing has long been used for inland water quality monitoring. However, due to the complex correlation between water quality parameters (WQPs) and water optical properties, the interactions of WQPs, and the impacts of climate, using remote sensing reflectance (R rs) to adequately estimate WQPs is still a grand challenge. Deep learning has the potential in capturing the correlation among R rs , optically active constituents (OACs), and non-OACs, and is progressively used in remote sensing retrieval of inland water quality. In this study, the enhanced multimodal deep learning (EMDL) models were proposed for Chlorophyll-a, total phosphorous, total nitrogen, Secchi disk depth, dissolved organic carbon, and dissolved oxygen retrieval in Lake Simcoe (80 km north of Toronto, Canada). The EMDL models were developed and validated using the R rs data derived from the harmonized Landsat and Sentinel-2 images, synchronized water quality measurements, water surface temperature, and climate data (N = 1173). The performance of the EMDL models was compared to that of several other machine learning, deep learning, and empirical models. Using the developed EMDL models, the spatial distributions and long-term variations of the WQPs in Lake Simcoe from 2013 to 2019 were reconstructed. The impacts of 12 potential natural and anthropogenic factors on the water quality of the entire Lake Simcoe and its two most concerned estuaries were also quantitatively discussed. The results showed that the EMDL models produced satisfactory performance in estimation of the six WQPs, with the Slope being close to 1 (0.84–0.95), normalized mean absolute error ≤20.17%, and Bias ≤14.68%. The EMDL models had the potential to reconstruct the spatial patterns and time-series dynamics of water quality and effectively detect the gradients of spatial patterns. This study provides a novel approach to supporting the environmental management and identification of the affecting factors for the Lake Simcoe watershed. [ABSTRACT FROM AUTHOR]

• This study presents a systematically downscaling framework for GRACE-derived data. • Machine learning and remote sensing tools are used to present drought maps. • The efficiency of using ConvLSTM model is explored over previous methods. • The significant reasons for the water crisis and drought in Iran are investigated. • The results indicate that Iran has experienced water resource overexploitation. The current study proposes a new method to downscale the monthly GRACE-derived Terrestrial Water Storage Anomaly (TWSA) to 10 km spatial resolution over Iran using deep learning methods. First, the growing neural gas (GNG) method is utilized to cluster the TWSA data to find similar patterns. The Silhouette Chou Davies (SCD) validity measure, a combination of three robust cluster validity measures, is used to evaluate the performance of the GNG method. Then, the feasibility of deep learning Convolutional Long Short-Term Memory (ConvLSTM), shallow learning Feed Forward Neural Networks (FFNN), and Random Forest (RF) are examined in downscaling GRACE-derived TWSA using only remote sensing images. The results indicate that the deep learning method outperforms the RF and FFNN models by 7 % and 18 %, respectively. Then, the Ground Water Storage (GWS) data are isolated from TWSA, showing an agreement between the GRACE-derived GWS and ground-based measurements with R2 = 0.76. Additionally, the downscaled TWSA is used to generate annual drought frequency and change detection maps for the GWS from 2002 to 2016. Annual standardized precipitation index (SPI) drought frequency maps are also produced to gain deeper insight into the water resource scarcity of Iran. The results indicate that Iran has experienced water resource overexploitation since 2011 for agricultural activities, while meteorological drought is a trigger and intensifier for the water crisis in Iran. According to the results, Iran experienced exceptional drought in some regions, and the GWS has decreased all over the country. In addition, further analysis of Iran's water resources is provided until 2022. The current study calls for integrated sustainable water resources management in Iran; otherwise, irreversible water and environmental problems with higher frequencies will be expected in the near future. [ABSTRACT FROM AUTHOR]

The accurate estimation of coastal water quality parameters (WQPs) is crucial for decision-makers to manage water resources. Although various machine learning (ML) models have been developed for coastal water quality estimation using remote sensing data, the performance of these models has significant uncertainties when applied to regional scales. To address this issue, an ensemble ML-based model was developed in this study. The ensemble ML model was applied to estimate chlorophyll-a (Chla), turbidity, and dissolved oxygen (DO) based on Sentinel-2 satellite images in Shenzhen Bay, China. The optimal input features for each WQP were selected from eight spectral bands and seven spectral indices. A local explanation strategy termed Shapley Additive Explanations (SHAP) was employed to quantify contributions of each feature to model outputs. In addition, the impacts of three climate factors on the variation of each WQP were analyzed. The results suggested that the ensemble ML models have satisfied performance for Chla (errors = 1.7%), turbidity (errors = 1.5%) and DO estimation (errors = 0.02%). Band 3 (B3) has the highest positive contribution to Chla estimation, while Band Ration Index2 (BR2) has the highest negative contribution to turbidity estimation, and Band 7 (B7) has the highest positive contribution to DO estimation. The spatial patterns of the three WQPs revealed that the water quality deterioration in Shenzhen Bay was mainly influenced by input of terrestrial pollutants from the estuary. Correlation analysis demonstrated that air temperature (Temp) and average air pressure (AAP) exhibited the closest relationship with Chla. DO showed the strongest negative correlation with Temp, while turbidity was not sensitive to Temp, average wind speed (AWS), and AAP. Overall, the ensemble ML model proposed in this study provides an accurate and practical method for long-term Chla, turbidity, and DO estimation in coastal waters. • A machine learning (ML) based ensemble model was proposed for the estimation of water quality parameters. • The feature selection was conducted for the ML-based ensemble model. • The ML-based ensemble model improved the estimation accuracy of non-optical active constituent. • SHAP method was employed to quantify the contribution of the input features. • Climate factors driving mechanism was considered on interannunal dynamics of water quality. [ABSTRACT FROM AUTHOR]