Min Of Data Research Articles

Stratiform and Convective Rain Classification Using Machine Learning Models and Micro Rain Radar

Rain type classification into convective and stratiform is an essential step required to improve quantitative precipitation estimations by remote sensing instruments. Previous studies with Micro Rain Radar (MRR) measurements and subjective rules have been performed to classify rain events. However, automating this process by using machine learning (ML) models provides the advantages of fast and reliable classification with the possibility to classify rain minute by minute. A total of 20,979 min of rain data measured by an MRR at Das in northeast Spain were used to build seven types of ML models for stratiform and convective rain type classification. The proposed classification models use a set of 22 parameters that summarize the reflectivity, the Doppler velocity, and the spectral width (SW) above and below the so-called separation level (SL). This level is defined as the level with the highest increase in Doppler velocity and corresponds with the bright band in stratiform rain. A pre-classification of the rain type for each minute based on the rain microstructure provided by the collocated disdrometer was performed. Our results indicate that complex ML models, particularly tree-based ensembles such as xgboost and random forest which capture the interactions of different features, perform better than simpler models. Applying methods from the field of interpretable ML, we identified reflectivity at the lowest layer and the average spectral width in the layers below SL as the most important features. High reflectivity and low SW values indicate a higher probability of convective rain.

The faintest solar coronal hard X-rays observed with FOXSI

Context. Solar nanoflares are small impulsive events releasing magnetic energy in the corona. If nanoflares follow the same physics as their larger counterparts, they should emit hard X-rays (HXRs) but with a rather faint intensity. A copious and continuous presence of nanoflares would result in a sustained HXR emission. These nanoflares could deliver enormous amounts of energy into the solar corona, possibly accounting for its high temperatures. To date, there has not been any direct observation of such persistent HXRs from the quiescent Sun. However, the quiet-Sun HXR emission was constrained in 2010 using almost 12 days of quiescent solar off-pointing observations by the Reuven Ramaty High Energy Solar Spectroscopic Imager (RHESSI). These observations set 2σ upper limits at 3.4 × 10−2 photons s−1 cm−2 keV−1 and 9.5 × 10−4 photons s−1 cm−2 keV−1 for the 3–6 keV and 6–12 keV energy ranges, respectively. Aims. Observing faint HXR emission is challenging because it demands high sensitivity and dynamic range instruments. The Focusing Optics X-ray Solar Imager (FOXSI) sounding rocket experiment excels in these two attributes when compared with RHESSI. FOXSI completed its second and third successful flights (FOXSI-2 and -3) on December 11, 2014, and September 7, 2018, respectively. This paper aims to constrain the quiet-Sun emission in the 5–10 keV energy range using FOXSI-2 and -3 observations. Methods. To fully characterize the sensitivity of FOXSI, we assessed ghost ray backgrounds generated by sources outside of the field of view via a ray-tracing algorithm. We used a Bayesian approach to provide upper thresholds of quiet-Sun HXR emission and probability distributions for the expected flux when a quiet-Sun HXR source is assumed to exist. Results. We found a FOXSI-2 upper limit of 4.5 × 10−2 photons s−1 cm−2 keV−1 with a 2σ confidence level in the 5–10 keV energy range. This limit is the first-ever quiet-Sun upper threshold in HXR reported using ∼1 min observations during a period of high solar activity. RHESSI was unable to measure the quiet-Sun emission during active times due to its limited dynamic range. During the FOXSI-3 flight, the Sun exhibited a fairly quiet configuration, displaying only one aged nonflaring active region. Using the entire ∼6.5 min of FOXSI-3 data, we report a 2σ upper limit of ∼10−4 photons s−1 cm−2 keV−1 for the 5–10 keV energy range. Conclusions. The FOXSI-3 upper limits on quiet-Sun emission are similar to that previously reported, but FOXSI-3 achieved these results with only 5 min of observations or about 1/2600 less time than RHESSI. A possible future spacecraft using hard X-ray focusing optics like those in the FOXSI concept would allow enough observation time to constrain the current HXR quiet-Sun limits further, or perhaps even make direct detections. This is the first report of quiet-Sun HXR limits from FOXSI and the first science paper using FOXSI-3 observations.

Historic Spatial Patterns of Storm-Driven Compound Events in UK Estuaries

AbstractCompound estuarine flooding is driven by extreme sea-levels and river discharge occurring concurrently, or in close succession, and threatens low-lying coastal regions worldwide. We hypothesise that these drivers of flooding rarely occur independently and co-operate at sub-daily timescales. This research aimed to identify regions and individual estuaries within Britain susceptible to storm-driven compound events, using 27 tide gauges linked to 126 river gauges covering a 30-year record. Five methods were evaluated, based on daily mean, daily maximum, and instantaneous 15-min discharge data to identify extremes in the river records, with corresponding skew surges identified within a ‘storm window’ based on average hydrograph duration. The durations, relative timings, and overlap of these extreme events were also calculated. Dependence between extreme skew surge and river discharge in Britain displayed a clear east–west split, with gauges on the west coast showing stronger correlations up to 0.33. Interpreting dependence based on correlation alone can be misleading and should be considered alongside number of historic extreme events. The analyses identified 46 gauges, notably the Rivers Lune and Orchy, where there has been the greatest chance and most occurrences of river-sea extremes coinciding, and where these events readily overlapped one another. Our results were sensitive to the analysis method used. Most notably, daily mean discharge underestimated peaks in the record and did not accurately capture likelihood of compound events in 68% of estuaries. This has implications for future flood risk in Britain, whereby studies should capture sub-daily timescale and concurrent sea-fluvial climatology to support long-term flood management plans.

Measuring commodity market quality

In this paper, we identify the most suitable low-frequency proxies for analyzing commodity market quality. We use an 11-year sample of millisecond time-stamped order book data and examine the correlation of high-frequency liquidity and price efficiency measures with their low-frequency proxies measured with daily or 5-min Time-and-Sales (TAS) data. We find that for liquidity, the volatility-over-volume measures are the best proxies for bid–ask spread and price impact. The correlation of price efficiency measures with their daily-frequency counterparts is low. Moderately correlated proxies can be achieved by using 5-min data.

Automatic identification of tinnitus malingering based on overt and covert behavioral responses during psychoacoustic testing

Tinnitus, or ringing in the ears, is a prevalent condition that imposes a substantial health and financial burden on the patient and to society. The diagnosis of tinnitus, like pain, relies on patient self-report, which can complicate the distinction between actual and fraudulent claims. Here, we combined tablet-based self-directed hearing assessments with neural network classifiers to automatically differentiate participants with tinnitus (N = 24) from a malingering cohort, who were instructed to feign an imagined tinnitus percept (N = 28). We identified clear differences between the groups, both in their overt reporting of tinnitus features, but also covert differences in their fingertip movement trajectories on the tablet surface as they performed the reporting assay. Using only 10 min of data, we achieved 81% accuracy classifying patients and malingerers (ROC AUC = 0.88) with leave-one-out cross validation. Quantitative, automated measurements of tinnitus salience could improve clinical outcome assays and more accurately determine tinnitus incidence.

Multilayer Perceptron-Based Real-Time Intradialytic Hypotension Prediction Using Patient Baseline Information and Heart-Rate Variation.

Intradialytic hypotension (IDH) is a common side effect that occurs during hemodialysis and poses a great risk for dialysis patients. Many studies have been conducted so far to predict IDH, but most of these could not be applied in real-time because they used only underlying patient information or static patient disease information. In this study, we propose a multilayer perceptron (MP)-based IDH prediction model using heart rate (HR) information corresponding to time-series information and static data of patients. This study aimed to validate whether HR differences and HR slope information affect real-time IDH prediction in patients undergoing hemodialysis. Clinical data were collected from 80 hemodialysis patients from 9 September to 17 October 2020, in the artificial kidney room at Yeungnam University Medical Center (YUMC), Daegu, South Korea. The patients typically underwent hemodialysis 12 times during this period, 1 to 2 h per session. Therefore, the HR difference and HR slope information within up to 1 h before IDH occurrence were used as time-series input data for the MP model. Among the MP models using the number and data length of different hidden layers, the model using 60 min of data before the occurrence of two layers and IDH showed maximum performance, with an accuracy of 81.5%, a true positive rate of 73.8%, and positive predictive value of 87.3%. This study aimed to predict IDH in real-time by continuously supplying HR information to MP models along with static data such as age, diabetes, hypertension, and ultrafiltration. The current MP model was implemented using relatively limited parameters; however, its performance may be further improved by adding additional parameters in the future, further enabling real-time IDH prediction to play a supporting role for medical staff.

The effect of general anesthesia on the test-retest reliability of resting-state fMRI metrics and optimization of scan length.

Resting-state functional magnetic resonance imaging (rs-fMRI) has been known as a powerful tool in neuroscience. However, exploring the test–retest reliability of the metrics derived from the rs-fMRI BOLD signal is essential, particularly in the studies of patients with neurological disorders. Here, two factors, namely, the effect of anesthesia and scan length, have been estimated on the reliability of rs-fMRI measurements. A total of nine patients with drug-resistant epilepsy (DRE) requiring interstitial thermal therapy (LITT) were scanned in two states. The first scan was performed in an awake state before surgery on the same patient. The second scan was performed 2 weeks later under general anesthesia necessary for LITT surgery. At each state, two rs-fMRI sessions were obtained that each one lasted 15 min, and the effect of scan length was evaluated. Voxel-wise rs-fMRI metrics, including the amplitude of low-frequency fluctuation (ALFF), the fractional amplitude of low-frequency fluctuation (fALFF), functional connectivity (FC), and regional homogeneity (ReHo), were measured. Intraclass correlation coefficient (ICC) was calculated to estimate the reliability of the measurements in two states of awake and under anesthesia. Overall, it appeared that the reliability of rs-fMRI metrics improved under anesthesia. From the 15-min data, we found mean ICC values in awake state including 0.81, 0.51, 0.65, and 0.84 for ALFF, fALFF, FC, and ReHo, respectively, as well as 0.80, 0.59, 0.83, and 0.88 for ALFF, fALFF, FC, and ReHo, respectively, under anesthesia. Additionally, our findings revealed that reliability increases as the function of scan length. We showed that the optimized scan length to achieve less variability of rs-fMRI measurements was 3.1–7.5 min shorter in an anesthetized, compared to a wakeful state.

Individualized diagnosis of preclinical Alzheimer’s Disease using deep neural networks

The early diagnosis of Alzheimer’s Disease (AD) plays a central role in the treatment of AD. Particularly, identifying the preclinical AD (pAD) stage could be crucial for timely treatment in the elderly. However, screening participants with pAD requires a series of psychological and neurological examinations. Thus, an efficient diagnostic tool is needed. Here, we recruited 91 elderly participants and collected 1 min of resting-state electroencephalography data to classify participants as normal aging or diagnosed with pAD. We used deep neural networks (Deep ConvNet, EEGNet, EEG-TCNet, and cascade CRNN) in the within- and cross-subject paradigms for classification and found individual variations of classification accuracy in the cross-subject paradigm. Further, we proposed an individualized diagnostic strategy to identify neurophysiological similarities across participants and the proposed approach considering individual characteristics improved the diagnostic performance by approximately 20%. Our findings suggest that considering individual characteristics would be a breakthrough in diagnosing AD using deep neural networks.

Realized higher-order moments spillovers between commodity and stock markets: Evidence from China

We construct daily realized volatility, skewness, and kurtosis using 5-min data of eight Chinese commodity futures and the Chinese stock market index from March 26, 2018 to October 22, 2020, then analyse the dynamic spillovers of realized moments among these markets. The results show that the spillover effects between commodity and stock markets intensify during shock periods such as ‘trade disputes between China and the United States’ and ‘COVID-19’. Volatility spillovers are relatively stronger than spillovers in skewness or spillovers in kurtosis; however, spillovers in higher-order moments seem to contain additional information. Shocks from the silver market influence realized moments of other markets. Soybean, corn, aluminium, and oil markets are affected by other markets. The contribution of wheat as a net transmitter to the system of spillovers between stock and commodity markets is only observed at higher-order realized moments. The results from OLS and quantile regressions show that the total spillovers are generally affected by the US stock market, economic uncertainties, and the COVID-19 outbreak.

The size of good and bad volatility shocks does matter for spillovers

Based on the rationale that the propagation of stock volatility shocks within a system can be affected by the size of shocks, we apply a tail-based approach of spillovers based on the variance decomposition of a quantile vector autoregression model. The analysis involves the decomposition of the realized variance into positive and negative realized semivariances using 5-min data on six major stock market indices from the US, Eurozone, UK, Japan, China, and India for the period February 14, 2000 - September 30, 2021. The results show that the propagation of volatility shocks within the system is not only shaped by the sign of the shocks (good versus bad volatility) but also by the shock size. For each good and bad volatility spillover, we detect a heterogeneity resulting from the difference in the size of spillovers between the upper and middle quantiles and thus reveal a relative intensity effect, as measured by the Relative Intensity of Shock Spillover (RISS) measure which we propose herein. This points to the necessity of going beyond studying spillovers of average shocks and employing tail-based models capable of uncovering the heterogeneous and intensity effects of the shock size. Otherwise, these features remain hidden, leading to suboptimal inferences and policy implications.

Evaluating the potential of H8/AHI geostationary observations for monitoring vegetation phenology over different ecosystem types in northern China

• Different strategies for compositing daily EVI2 from 10-min AHI data were assessed. • Mean value compositing using data around solar noon minimized anisotropic effects. • Composited AHI daily EVI2 time series have 5.3 times more observations than MODIS. • Phenology dates were successfully retrieved from composited AHI EVI2 time series. • AHI can reduce uncertainty in phenology monitoring with more valid data and less gaps. Sensors onboard the new generation of geostationary (GEO) satellites launched over past few years have much improved radiometric and temporal resolutions. As one of the latest GEO sensors, the Japanese Himawari-8 Advanced Himawari Imager (hereafter H8/AHI) has temporal resolution of 10 minutes and spectral bands similar to MODIS. Observations from H8/AHI therefore have the potential to improve the precision of vegetation phenology monitoring by significantly shortening the compositing period. In this study, we evaluated methods for compositing daily 2-band Enhanced Vegetation Index (EVI2) from 10-minutes H8/AHI observations at six study sites in northern China encompassing forest, cropland, and grassland. After screening the cloudy observations by a physics-based cloud detection algorithm, key phenological metrics were retrieved from the composited AHI EVI2. H8/AHI 10-min raw EVI2 showed significant diurnal variations that are mainly associated with sun angle variations. By calculating the daily average EVI2 values in different time intervals, we concluded that the mean value using a fixed time window (10:30–13:30) centered around local solar noon effectively reduced the anisotropy effect from the diurnal sun-angle variations, leading to stable EVI2 time series suitable for extracting phenological metrics. On average, the amount of the composited H8/AHI EVI2 time-series data was 5.3 times more than the Terra/Aqua combined MODIS EVI2 product per year and the average size (width) of the gaps in H8/AHI EVI2 time series was 75% smaller than that of the MODIS. Our results demonstrated the promising capability of the new generation GEO satellite for generating time series vegetation index with fewer cloud-induced gaps and higher temporal resolution, hence improving the monitoring of global vegetation phenology and ecosystem responses to climate change.

Quad-kernel deep convolutional neural network for intra-hour photovoltaic power forecasting

Photovoltaic (PV) power is highly stochastic and volatile, and PV power forecasting is a key technology to guarantee the safe and economic operation of high-penetration renewable power systems. To improve the accuracy of PV power forecasting, a quad-kernel deep convolutional neural network (QK_CNN) model is proposed to perform intra-hour PV power forecasting for the next four timesteps: four CNNs with different kernel sizes are used to extract different local cross features between sequence elements of four timesteps; a single-kernel CNN is used to further feature extraction of these features, and then the target sequence forecasting results are obtained; global maximum pooling method is used to simplify the feature extraction process and improve model learning efficiency. Operation data from a 26.52 kW PV plant in CentralAustralia is selected as the experimental data. Compared with single-kernel CNN and hybrid models (CNN_LSTM) on 5, 10, and 15 min of resolution data, respectively, the proposed model shows better forecasting performance and is able to explain 96 % to 98 % of the total variation of the forecasted PV power. All these demonstrates that the CNNs with specific design have great potential to handle the task of PV power forecasting as well.

Rapid Robot-Aided On-site Testing of Fume Cupboards.

Although containment testing of fume cupboards (FC) according to the standards EN 14175-3 (2019) or ANSI/ASHRAE 110 (2016) is well established for type testing, its application is currently much less accepted and practised for evaluating containment on-site. Few of the several million FC in the market have been tested at installation and commissioning, and even less undergo verification of containment during their service life in the laboratories. Several reasons have led to this unsafe situation. To address this challenge, a new concept has been developed to allow for rapid on-site testing of FC to gain knowledge as to the functional efficiency as well as to safety aspects for the operator. The concept consists of a movable robot-aided test equipment that can be installed quickly to the FC in running labs. Multiple sensors detect the tracer gas isopropanol. Within a test run of only 10-min data is collected to quantify containment at the sash opening and to determine purge efficiency. The method reveals impact from interfering effects such as draughts, air distribution, and movements and from equipment installed, and is a tool for the optimization of operating conditions of a lab. This article presents an advanced alternative to the existing containment tests, particularly for on-site testing. The method assesses not only proper operation of the FC in its environment, but also the suitability of a FC for a given use under aspects of health and safety evaluation.

Impact of renewable electricity on utility finances: Assessing merit order effect for an Indian utility

Declining levelized costs of renewable energy have become a driving force in supporting renewable energy in India; the levelized cost of wind and solar has plummeted to between 3.3¢/kWh and 4¢/kWh. However, limited research exists on the impacts of renewables on the finances of Indian distribution utilities. The present study examines the financial impact of incremental penetration of the hydro, solar, and wind alternative in the generation mix of the distribution company of the Indian state of Madhya Pradesh. Using real-time 15-min interval data and a merit-order-dispatch model, the study quantifies the merit-order-effect and assess the potential savings from renewables. The results indicate that depending upon the penetration level under demand growth cases, renewable integration could lower the power purchase cost of Indian utilities by up to 11%. Wind produces most savings between 0.11¢/kWh and 2.71¢/kWh followed by solar −0.17¢/kWh and 2.56¢/kWh and hydro −0.32¢/kWh and 2.05¢/kWh. The savings will increase with rising electricity demand and plummeting costs of renewables. Integrating moderate levels of renewables no longer presents economic-environment trade-offs and can simultaneously meet multiple policy goals of energy affordability and environmental sustainability.

A new approach to estimating oxygen off-gas fraction and dynamic alpha factor in aeration systems using hybrid machine learning and mechanistic models

The aeration process is the largest energy consumer in wastewater treatment plants (WWTP), and the optimization of the process based on computational models can offer significant savings for the plant. Recent theoretical developments have revealed that many of the parameters commonly assumed as constants in aeration modelling have a dynamic nature; however, there still lacks a universal way to model these factors accurately. This work proposed a new framework and optimal time scale for building hybrid machine learning-mechanistic oxygen transfer rate models. The data used for the modelling were from existing sensors in a full-scale WWTP supplemented by a two-month field sampling campaign for collecting 15 min of dynamic data, including off-gas air fraction. Two machine learning (ML) models were built: the first set was a novel hybrid ML-mechanistic model developed to estimate the off-gas aeration fraction from wastewater parameters and calculate dynamic alpha per the off-gas method. The second ML model directly estimated dynamic alpha from wastewater process and operation parameters. The performance of the ML models was compared with recalibrated published regression-based models. Results showed that the hybrid ML-mechanistic alpha model (NSE = 0.96 and RMSE = 0.03) were more accurate than the direct ML alpha (NSE = 0.67 and RMSE = 0.07) and recalibrated regression (NSE < 0 and RMSE = 0.18 to 0.28) models, and therefore could be efficiently applied to predict dynamic alpha factor of an activated sludge plant. In addition to typical surfactant indicators such as MLSS and COD, other parameters like temperature and aeration ammonia concentration were substantial in dynamic off-gas and alpha estimation.

FlexLip: A Controllable Text-to-Lip System.

The task of converting text input into video content is becoming an important topic for synthetic media generation. Several methods have been proposed with some of them reaching close-to-natural performances in constrained tasks. In this paper, we tackle a subissue of the text-to-video generation problem, by converting the text into lip landmarks. However, we do this using a modular, controllable system architecture and evaluate each of its individual components. Our system, entitled FlexLip, is split into two separate modules: text-to-speech and speech-to-lip, both having underlying controllable deep neural network architectures. This modularity enables the easy replacement of each of its components, while also ensuring the fast adaptation to new speaker identities by disentangling or projecting the input features. We show that by using as little as 20 min of data for the audio generation component, and as little as 5 min for the speech-to-lip component, the objective measures of the generated lip landmarks are comparable with those obtained when using a larger set of training samples. We also introduce a series of objective evaluation measures over the complete flow of our system by taking into consideration several aspects of the data and system configuration. These aspects pertain to the quality and amount of training data, the use of pretrained models, and the data contained therein, as well as the identity of the target speaker; with regard to the latter, we show that we can perform zero-shot lip adaptation to an unseen identity by simply updating the shape of the lips in our model.

Classification of Cassini’s Orbit Regions as Magnetosphere, Magnetosheath, and Solar Wind via Machine Learning

Several machine learning algorithms and feature subsets from a variety of particle and magnetic field instruments on-board the Cassini spacecraft were explored for their utility in classifying orbit segments as magnetosphere, magnetosheath or solar wind. Using a list of manually detected magnetopause and bow shock crossings from mission scientists, random forest (RF), support vector machine (SVM), logistic regression (LR) and recurrent neural network long short-term memory (RNN LSTM) classification algorithms were trained and tested. A detailed error analysis revealed a RNN LSTM model provided the best overall performance with a 93.1% accuracy on the unseen test set and MCC score of 0.88 when utilizing 60 min of magnetometer data (|B|, Bθ, Bϕ and BR) to predict the region at the final time step. RF models using a combination of magnetometer and particle data, spanning H+, He+, He++ and electrons at a single time step, provided a nearly equivalent performance with a test set accuracy of 91.4% and MCC score of 0.84. Derived boundary crossings from each model’s region predictions revealed that the RNN model was able to successfully detect 82.1% of labeled magnetopause crossings and 91.2% of labeled bow shock crossings, while the RF model using magnetometer and particle data detected 82.4 and 74.3%, respectively.

A Multi-market Comparison of the Intraday Lead-Lag Relations Among Stock Index-Based Spot, Futures and Options.

Using 1-min data, we explore the dynamic variation of the intraday lead–lag relations between stock indices and their derivatives through a comprehensive study with broader coverage of research objectives and methodologies. This paper provides explicit evidence that the futures and options exhibit price leadership over the spot market, and the options is ahead of the futures on most trading days in all three markets. This paper also reports a new finding that the relation between the derivative and its underlying index reverses when the index return has a significantly larger mean value, and the reversal phenomenon is also observed in the relations between the futures and the options, which enriches the empirical results of intraday lead–lag relations. Moreover, these conclusions still hold under the impact of extreme events, e.g., the outbreak of the Covid-19. Finally, we construct a pair trading strategy based on the intraday lead–lag relationships, which can get better performance than the corresponding spot index. Our findings can potentially help regulators understand the price discovery process between the index and its derivatives, and also be of great value for timely adjustment of investors intraday trading strategies.

Estimation of the Hurst Parameter in Spot Volatility

This paper contributes in three stages in a logic of the cognitive process: we firstly propose a new estimation of Hurst exponent by changing frequency method which is purely mathematical. Then we want to check if the new Hurst is efficient, so we prove the advantages of this new Hurst in asymptotic variance in the perspective compared with other two Hurst estimator. However, a purely mathematical game is not enough, a good estimation should be proven by reality, so we apply the new Hurst estimator into truncated and non-truncated spot volatility which fills the gap of previous literatures using 5-min price data (Source: Wind Financial Terminal) of 10 Chinese A-share industry indices from 1 January 2005 until 31 December 2020.

Imaging evolution of Cascadia slow-slip event using high-rate GPS

The slip history of short-term slow slip event (SSE) is typically inferred from daily Global Positioning System (GPS) data, which, however, cannot image the sub-daily processes, leaving the underlying mechanisms of SSEs elusive. To address the temporal resolution issue, we attempted to employ the kinematic subdaily GPS analysis, which has never been applied to SSE studies because its signal-to-noise ratio has been believed too low. By carefully post-processing sub-daily positions to remove non-tectonic position fluctuation, our 30-min kinematic data clearly exhibits the transient motion of a few mm during one Cascadia SSE. A spatiotemporal slip image by inverting the 30-min data exhibits a multi-stage evolution; it consists of an isotropic growth of SSE followed by an along-strike migration and termination within the rheologically controlled down-dip width. This transition at the slip growth mode is similar to the rupture growth of regular earthquakes, implying the presence of common mechanical factors behind the two distinct slip phenomena. The comparison with a slip inversion of the daily GPS demonstrates the current performance and limitation of the subdaily data in the SSE detection and imaging. Better understanding of the non-tectonic noise in the kinematic GPS analysis will further improve the temporal resolution of SSE.