Efficient and Regionally Transferable Snow Water Equivalent Estimation Using a Long Short‐Term Memory Network

[1] Accurate estimation of the characteristics of the winter snowpack is crucial for prediction of available water supply, flooding, and climate feedbacks. Remote sensing of snow has been most successful for quantifying the spatial extent of the snowpack, although satellite estimation of snow water equivalent (SWE), fractional snow covered area, and snow depth is improving. Here we show that GPS observations of vertical land surface loading reveal seasonal responses of the land surface to the total weight of snow, providing information about the stored SWE. We demonstrate that the seasonal signal in Scripps Orbit and Permanent Array Center (SOPAC) GPS vertical land surface position time series at six locations in the western United States is driven by elastic loading of the crust by the snowpack. GPS observations of land surface deformation are then used to predict the water load as a function of time at each location of interest and compared for validation to nearby Snowpack Telemetry observations of SWE. Estimates of soil moisture are included in the analysis and result in considerable improvement in the prediction of SWE.Citation: Ouellette, K. J., C. de Linage, and J. S. Famiglietti (2013), Estimating snow water equivalent from GPS vertical site-position observations in the western United States, Water Resour. Res., 49, 2508–2518, doi:10.1002/wrcr.20173.

Seasonal snow water equivalent (SWE) accumulation in California’s Sierra Nevada is primarily governed by a few orographically enhanced snowstorms. However, as air temperatures gradually rise, resulting in a shift from snow to rain, the governing processes determining SWE accumulation versus ablation become ambiguous. Using a network of 28 snow pillow measurements to represent an elevational and latitudinal gradient across the Sierra Nevada, we identify distributions of critical temperatures and corresponding storm and snowpack properties that describe how SWE accumulation varies across the range at an hourly timescale for water years 2010 through 2019. We also describe antecedent and prevailing conditions governing whether SWE accumulates or ablates during warm storms. Results show that atmospheric moisture regulates a temperature dependence of SWE accumulation. Conditions balancing precipitable water and snow formation requirements produce the most seasonal SWE, which was observed in the (low-elevation) northern and (middle-elevation) central Sierra Nevada. The high southern Sierra Nevada conservatively accumulates SWE with colder, drier air, resulting in less midwinter ablation. These differences explain a tendency for deep, low-density snowpacks to accumulate rather than ablate SWE during warm storms (having median temperatures exceeding 1.0°C), reflecting counteracting liquid storage and internal energy deficits. The storm events themselves in these cases are brief with modest moisture supplies or are otherwise followed immediately by ablation.

Electricity load forecasting is a crucial undertaking within all the deregulated markets globally. Among the research challenges on a global scale, the investigation of deep transfer learning (DTL) in the field of electricity load forecasting represents a fundamental effort that can inform artificial intelligence applications in general. In this paper, a comprehensive study is reported regarding day-ahead electricity load forecasting. For this purpose, three sequence-to-sequence (Seq2seq) deep learning (DL) models are used, namely the multilayer perceptron (MLP), the convolutional neural network (CNN) and the ensemble learning model (ELM), which consists of the weighted combination of the outputs of MLP and CNN models. Also, the study focuses on the development of different forecasting strategies based on DTL, emphasizing the way the datasets are trained and fine-tuned for higher forecasting accuracy. In order to implement the forecasting strategies using deep learning models, load datasets from three Greek islands, Rhodes, Lesvos, and Chios, are used. The main purpose is to apply DTL for day-ahead predictions (1–24 h) for each month of the year for the Chios dataset after training and fine-tuning the models using the datasets of the three islands in various combinations. Four DTL strategies are illustrated. In the first strategy (DTL Case 1), each of the three DL models is trained using only the Lesvos dataset, while fine-tuning is performed on the dataset of Chios island, in order to create day-ahead predictions for the Chios load. In the second strategy (DTL Case 2), data from both Lesvos and Rhodes concurrently are used for the DL model training period, and fine-tuning is performed on the data from Chios. The third DTL strategy (DTL Case 3) involves the training of the DL models using the Lesvos dataset, and the testing period is performed directly on the Chios dataset without fine-tuning. The fourth strategy is a multi-task deep learning (MTDL) approach, which has been extensively studied in recent years. In MTDL, the three DL models are trained simultaneously on all three datasets and the final predictions are made on the unknown part of the dataset of Chios. The results obtained demonstrate that DTL can be applied with high efficiency for day-ahead load forecasting. Specifically, DTL Case 1 and 2 outperformed MTDL in terms of load prediction accuracy. Regarding the DL models, all three exhibit very high prediction accuracy, especially in the two cases with fine-tuning. The ELM excels compared to the single models. More specifically, for conducting day-ahead predictions, it is concluded that the MLP model presents the best monthly forecasts with MAPE values of 6.24% and 6.01% for the first two cases, the CNN model presents the best monthly forecasts with MAPE values of 5.57% and 5.60%, respectively, and the ELM model achieves the best monthly forecasts with MAPE values of 5.29% and 5.31%, respectively, indicating the very high accuracy it can achieve.

Abstract. Despite decades of effort, there remains an inability to measure snow water equivalent (SWE) at high spatial resolutions using remote sensing. Passive gamma ray spectrometry is one of the only well-established methods to reliably remotely sense SWE, but airborne applications to date have been limited to observing kilometre-scale areal averages. Noting the increasing capabilities of unoccupied aerial vehicles (UAVs) and miniaturization of passive gamma ray spectrometers, this study tested the ability of a UAV-borne gamma spectrometer and concomitant UAV-borne lidar to quantify the spatial variability of SWE at high spatial resolutions. Gamma and lidar observations from a UAV (UAV-gamma and UAV-lidar) were collected over two seasons from shallow, wind-blown, prairie snowpacks in Saskatchewan, Canada, with validation data collected from manual snow depth and density observations. A fine-resolution (0.25 m) reference dataset of SWE, to test UAV-gamma methods, was developed from UAV-lidar snow depth and snow survey snow density observations. The ability of UAV-gamma to resolve the areal average and spatial variability of SWE was promising with appropriate flight characteristics. Survey flights flown at a velocity of 5 m s−1, altitude of 15 m, and line spacing of 15 m were unable to capture the average or spatial variability of SWE within the uncertainty of the reference dataset. Slower, lower, and denser flight lines at a velocity of 4 m s−1, altitude of 8 m, and line spacing of 8 m were able to successfully observe areal average SWE and its variability at spatial resolutions greater than 22.5 m. Using a combination of UAV-based gamma SWE and UAV-based lidar snow depth improved the spatial representation of SWE substantially and permitted estimation of SWE at a spatial resolution 0.25 m with a ± 14.3 mm error relative to the reference SWE dataset. UAV-borne gamma spectrometry to estimate SWE is a promising and novel technique that has the potential to improve the measurement of shallow prairie snowpacks, and when combined with UAV-borne lidar snow depths, can provide fine-resolution, high-accuracy estimates of prairie SWE. Research on optimal hardware, data processing, and interpolation techniques is called for to further improve this remote sensing product and explore its application in other environments.

Snow depth observations can be leveraged with data assimilation (DA) to improve estimation of snow density and snow water equivalent (SWE). A key consideration for mission and campaign design is how snow depth retrieval characteristics (including observation timing/frequency and sampling error) influence SWE accuracy and uncertainty in a DA framework. To quantify these effects, we implement a particle filter (PF) assimilation technique to assimilate depth and validate this approach against observed snow density and SWE at 49 snow telemetry sites across 9 years. We sample from continuous in situ snow depth records to test a range of measurement timing and sampling error scenarios representative of remote sensing capabilities. Assimilation reduces density bias by over 40% and SWE bias by over 70% across climate zones and in both wet and dry years. There is little incremental benefit to SWE accuracy when assimilating more than one depth observation near peak accumulation. SWE estimates are less sensitive to observation timing than sampling error. Alternatively, more frequent depth observations improve melt‐out date timing and reduce SWE uncertainty, a key consideration when evaluating the operational utility of DA. In matching depth observations, the PF mostly acts to increase model precipitation inputs, while not systematically shifting other parameter values or forcings across the climate zones represented with the study sites. This demonstrates that precipitation is the largest source of model error. With DA, density errors are still nontrivial (above 10%), illuminating the need for further improvements to modeled density to estimate SWE within specified error limits.

This article investigates the identification of hate speech on social media using machine learning and deep learning techniques. The research uses metrics such as F-measure, AUC-ROC, precision, accuracy, and recall assessing the effectiveness of various tactics. The findings indicate that deep learning models, particularly the bidirectional long short-term memory (BiLSTM) architecture, consistently outperform other methods in categorization tasks. The research emphasizes the importance of sophisticated neural network designs in identifying the intricacies of hostile and offensive content online. The study offers insights for promoting early identification and prevention of cyberbullying, improving secure and inclusive online environments. Future research may explore real-time detection systems, hybrid approaches, or the integration of complementary components to enhance and improve innovative technology in tackling this significant social issue. A sample tweet was annotated by specialists who categorize tweets as hate speech, offensive language, or neutral. The researchers applied shallow learning methodologies and integrated word embeddings like Word2Vec and GloVe to enhance the efficacy of deep learning models. The results indicate that BiLSTM surpasses shallow learning methods in detecting hate speech on Twitter, highlighting the efficacy of deep learning approaches in recognizing and tracking hate speech on social media platforms. When comparing different deep learning and machine learning models on different datasets, the results reveal that deep learning techniques are usually more effective. A reasonably high level of accuracy is achieved by KNN and SVM among classical algorithms, whereas Naïve Bayes performs the poorest. While deep learning approaches provide better results, tree-based models such as Random Forest and Decision Trees offer more consistent accuracy. Models based on neural networks, such as LSTM, CNN, and BI-LSTM, perform well, with LSTM-based methods excelling in particular. The most successful strategy for classification problems is the model presented, which obtains the greatest accuracy, precision, recall, F1-score of 95%. The research aids in the development of advanced tools and methodologies to mitigate hate speech on social media and foster positive online interactions. Future research may investigate alternative deep learning architectures, such as transformers, to enhance hate speech detection efficacy. The advancement of interpretable AI methodologies for identifying hate speech and delivering transparent forecasts might enhance user confidence and facilitate better content moderation decisions.

PurposeTo conduct a head-to-head comparison between deep learning (DL) and radiomics models across institutions for predicting microvascular invasion (MVI) in hepatocellular carcinoma (HCC) and to investigate the model robustness and generalizability through rigorous internal and external validation.MethodsThis retrospective study included 2304 preoperative images of 576 HCC lesions from two centers, with MVI status determined by postoperative histopathology. We developed DL and radiomics models for predicting the presence of MVI using B-mode ultrasound, contrast-enhanced ultrasound (CEUS) at the arterial, portal, and delayed phases, and a combined modality (B + CEUS). For radiomics, we constructed models with enlarged vs. original regions of interest (ROIs). A cross-validation approach was performed by training models on one center’s dataset and validating the other, and vice versa. This allowed assessment of the validity of different ultrasound modalities and the cross-center robustness of the models. The optimal model combined with alpha-fetoprotein (AFP) was also validated. The head-to-head comparison was based on the area under the receiver operating characteristic curve (AUC).ResultsThirteen DL models and 25 radiomics models using different ultrasound modalities were constructed and compared. B + CEUS was the optimal modality for both DL and radiomics models. The DL model achieved AUCs of 0.802–0.818 internally and 0.667–0.688 externally across the two centers, whereas radiomics achieved AUCs of 0.749–0.869 internally and 0.646–0.697 externally. The radiomics models showed overall improvement with enlarged ROIs (P < 0.05 for both CEUS and B + CEUS modalities). The DL models showed good cross-institutional robustness (P > 0.05 for all modalities, 1.6–2.1% differences in AUC for the optimal modality), whereas the radiomics models had relatively limited robustness across the two centers (12% drop-off in AUC for the optimal modality). Adding AFP improved the DL models (P < 0.05 externally) and well maintained the robustness, but did not benefit the radiomics model (P > 0.05).ConclusionCross-institutional validation indicated that DL demonstrated better robustness than radiomics for preoperative MVI prediction in patients with HCC, representing a promising solution to non-standardized ultrasound examination procedures.

Comparison of different automatic methods for estimating snow water equivalent

Evaluation of multispectral imaging for freeze damage assessment in strawberries using AI-based computer vision technology

<strong class="journal-contentHeaderColor">Abstract.</strong> An airborne gamma-ray remote sensing technique provides a strong potential to estimate reliable snow water equivalent (SWE) in forested environments where typical remote sensing techniques have large uncertainties. This study explores the utility of assimilating the temporally (up to four measurements during a winter period) and spatially sparse airborne gamma SWE observations into a land surface model to improve SWE estimates in forested areas in the northeastern U.S. Here, we demonstrate that the airborne gamma SWE observations add value to the SWE estimates from the Noah land surface model with multiple parameterization options (Noah-MP) via assimilation despite the limited number of the measurements. Improvements are witnessed during the snow accumulation period while reduced skills are seen during the snow melting period. The efficacy of the gamma data is greater for areas with lower vegetation cover fraction and topographic heterogeneity ranges, and it is still effective in reducing the SWE estimation errors for areas with higher topographic heterogeneity. The gamma SWE data assimilation (DA) also shows a potential of extending the impact of flight line-based measurements to adjacent areas without observations by employing a localization approach. The localized DA reduces the modeled SWE estimation errors for adjacent grid cells up to 32-km distances from the flight lines. The enhanced performance of the gamma SWE DA is evident when the results are compared to those from assimilating the existing satellite-based SWE retrievals from the Advanced Microwave Scanning Radiometer 2 (AMSR2) for the same locations and time periods. Although there is still room for improvement, particularly for the melting period, this study shows that the gamma SWE DA is a promising method to improve the SWE estimates in forested areas.

An airborne gamma-ray remote sensing technique provides a strong potential to estimate reliable snow water equivalent (SWE) in forested environments where typical remote sensing techniques have large uncertainties. This study explores the utility of assimilating the temporally (up to four measurements during a winter period) and spatially sparse airborne gamma SWE observations into a land surface model to improve SWE estimates in forested areas in the northeastern U.S. Here, we demonstrate that the airborne gamma SWE observations add value to the SWE estimates from the Noah land surface model with multiple parameterization options (Noah-MP) via assimilation despite the limited number of the measurements. Improvements are witnessed during the snow accumulation period while reduced skills are seen during the snow melting period. The efficacy of the gamma data is greater for areas with lower vegetation cover fraction and topographic heterogeneity ranges, and it is still effective in reducing the SWE estimation errors for areas with higher topographic heterogeneity. The gamma SWE data assimilation (DA) also shows a potential of extending the impact of flight line-based measurements to adjacent areas without observations by employing a localization approach. The localized DA reduces the modeled SWE estimation errors for adjacent grid cells up to 32-km distances from the flight lines. The enhanced performance of the gamma SWE DA is evident when the results are compared to those from assimilating the existing satellite-based SWE retrievals from the Advanced Microwave Scanning Radiometer 2 (AMSR2) for the same locations and time periods. Although there is still room for improvement, particularly for the melting period, this study shows that the gamma SWE DA is a promising method to improve the SWE estimates in forested areas.

COVID-19 is highly infectious and causes acute respiratory disease. Machine learning (ML) and deep learning (DL) models are vital in detecting disease from computerized chest tomography (CT) scans. The DL models outperformed the ML models. For COVID-19 detection from CT scan images, DL models are used as end-to-end models. Thus, the performance of the model is evaluated for the quality of the extracted feature and classification accuracy. There are four contributions included in this work. First, this research is motivated by studying the quality of the extracted feature from the DL by feeding these extracted to an ML model. In other words, we proposed comparing the end-to-end DL model performance against the approach of using DL for feature extraction and ML for the classification of COVID-19 CT scan images. Second, we proposed studying the effect of fusing extracted features from image descriptors, e.g., Scale-Invariant Feature Transform (SIFT), with extracted features from DL models. Third, we proposed a new Convolutional Neural Network (CNN) to be trained from scratch and then compared to the deep transfer learning on the same classification problem. Finally, we studied the performance gap between classic ML models against ensemble learning models. The proposed framework is evaluated using a CT dataset, where the obtained results are evaluated using five different metrics The obtained results revealed that using the proposed CNN model is better than using the well-known DL model for the purpose of feature extraction. Moreover, using a DL model for feature extraction and an ML model for the classification task achieved better results in comparison to using an end-to-end DL model for detecting COVID-19 CT scan images. Of note, the accuracy rate of the former method improved by using ensemble learning models instead of the classic ML models. The proposed method achieved the best accuracy rate of 99.39%.

We developed a multimodal ultrasound (US) deep learning (DL) fusion model to automatically classify early fibrosis in patients with chronic kidney disease (CKD). This prospective study included patients with CKD who underwent continuous gray-scale US, superb microvascular imaging, and strain elastography from May to November 2022. According to the pathological tubular atrophy and interstitial fibrosis score, patients were divided into minimal and mild groups (affected area ≤10% and 11 − 25% of the total cortical volume, respectively). The dataset was divided into training (70%) and test (30%) sets. A DL model combining the features of the three US modes was developed to predict early fibrosis in patients with CKD. We compared these findings with the area under the receiver operating characteristic curve (AUC) of the clinical model by analyzing the receiver operating characteristic curve in the test set. The AUC of single-mode DL based on gray-scale US, superb microvascular imaging, and strain elastography was 0.682, 0.745, and 0.648, respectively, while that of the multimodal US DL model was 0.86. The accuracy, specificity, and sensitivity of the multimodal US DL model were 0.779, 0.767, and 0.796, respectively, and the negative and positive predictive values were 0.842 and 0.706, respectively. The AUC of the multimodal US DL model was significantly better than that of the single-mode DL and clinical models. The DL algorithm developed using multimodal US images can effectively predict early fibrosis in patients with CKD with significantly greater accuracy than single-mode DL or clinical models.

We aimed at comparing intratumoral and peritumoral deep learning, radiomics, and fusion models in predicting KRAS mutations in rectal cancer using endorectal ultrasound imaging. This study included 304 patients with rectal cancer from Fujian Medical University Union Hospital. The patients were randomly divided into a training group (213 patients) and a test group (91 patients) at a 7:3 ratio. Radiomics and deep learning models were established using primary tumor and peritumoral images. In the optimally performing regions-of-interest, two fusion strategies, a feature-based and a decision-based model, were employed to build the fusion models. The Shapley additive explanation (SHAP) method was used to evaluate the significance of features in the optimal radiomics, deep learning, and fusion models. The performance of each model was assessed using the area under the receiver operating characteristic curve (AUC) and decision curve analysis (DCA). In the test cohort, both the radiomics and deep learning models exhibited optimal performance with a 10-pixel patch extension, yielding AUC values of 0.824 and 0.856, respectively. The feature-based DLRexpand10_FB model attained the highest AUC (0.896) across all study sets. In addition, the DLRexpand10_FB model demonstrated excellent sensitivity, specificity, and DCA. SHAP analysis underscored the deep learning feature (DL_1) as the most significant factor in the hybrid model. The feature-based fusion model DLRexpand10_FB can be employed to predict KRAS gene mutations based on pretreatment endorectal ultrasound images of rectal cancer. The integration of peritumoral regions enhanced the predictive performance of both the radiomics and deep learning models.

Deep learning breast cancer risk models demonstrate improved accuracy compared with traditional risk models but have not been prospectively tested. We compared the accuracy of a deep learning risk score derived from the patient's prior mammogram to traditional risk scores to prospectively identify patients with cancer in a cohort due for screening. We collected data on 119 139 bilateral screening mammograms in 57 617 consecutive patients screened at 5 facilities between September 18, 2017, and February 1, 2021. Patient demographics were retrieved from electronic medical records, cancer outcomes determined through regional tumor registry linkage, and comparisons made across risk models using Wilcoxon and Pearson χ2 2-sided tests. Deep learning, Tyrer-Cuzick, and National Cancer Institute Breast Cancer Risk Assessment Tool (NCI BCRAT) risk models were compared with respect to performance metrics and area under the receiver operating characteristic curves. Cancers detected per thousand patients screened were higher in patients at increased risk by the deep learning model (8.6, 95% confidence interval [CI] = 7.9 to 9.4) compared with Tyrer-Cuzick (4.4, 95% CI = 3.9 to 4.9) and NCI BCRAT (3.8, 95% CI = 3.3 to 4.3) models (P &lt; .001). Area under the receiver operating characteristic curves of the deep learning model (0.68, 95% CI = 0.66 to 0.70) was higher compared with Tyrer-Cuzick (0.57, 95% CI = 0.54 to 0.60) and NCI BCRAT (0.57, 95% CI = 0.54 to 0.60) models. Simulated screening of the top 50th percentile risk by the deep learning model captured statistically significantly more patients with cancer compared with Tyrer-Cuzick and NCI BCRAT models (P &lt; .001). A deep learning model to assess breast cancer risk can support feasible and effective risk-based screening and is superior to traditional models to identify patients destined to develop cancer in large screening cohorts.