Imaging the spatio-temporal variations in deep tremor activity using cluster analysis techniques

2021 ◽  
Author(s):  
Anca Opris ◽  
Sumanta Kundu ◽  
Takahiro Hatano
Keyword(s):  
Machine Learning ◽  
Cluster Analysis ◽  
Regional Scale ◽  
Temporal Variations ◽  
Machine Learning Algorithms ◽  
Temporal Behaviour ◽  
Slow Earthquakes ◽  
Short Period ◽  
Spatial Correlation Structure ◽  
Spatio Temporal

<div>More than 15 years of seismic observations on slow earthquakes are available for the Nankai and Cascadia regions, due to the high density of seismic stations and constant improvements. It was observed that deep tremor activity exhibits highly non-Poissonian behaviour, consisting of short-period bursts separated by long periods of inactivity, as well as significant spatial variations throughout a tectonic region (Obara, 2011). Tremor activity in these regions has shown episodic behaviour with different recurrence interval. Modelling the space-time variations can help the unified understanding of the phenomenon. Catalogues with more than 30.000 (Idehara et al., 2014) and 130000 LFE’s (Mizuno et al, 2019) are available for the world tremor databese. If we consider LFE’s source as a spatial correlation structure which is evolving in time, in order to reveal the characteristics of this structure, we used the Grassberger Procaccia algorithm to calculate the combined correlation dimension (Tosi et al.,2008) of tremor activity (Cc (r, τ)), at both local and regional scale. The integral representation is shown as contour map (facilitating the possibility of using machine learning algorithms based on image processing for identifying the characteristic image of each tremor patch). Thus, implementing machine learning methods for LFE cluster analysis is required. After performing the cluster analysis, we could identify the specific spatio-temporal behaviour of each of the tremor patches in the studied regions, not just the features which were described in previous studies, such as recurrence intervals for short-term slow slip events (Idehara et al., 2014), tremor migration (for monitoring purposes), but also new features which could be used for forecasting.</div><div> </div><div><img src="https://contentmanager.copernicus.org/fileStorageProxy.php?f=gepj.0b8b5af57a0068928921161/sdaolpUECMynit/12UGE&app=m&a=0&c=4087202a58071c68d39e6d5f305ba0b4&ct=x&pn=gepj.elif&d=1" alt=""></div>

scholarly journals Multilayer Soil Moisture Mapping at a Regional Scale from Multisource Data via a Machine Learning Method

2019 ◽  
Vol 11 (3) ◽  
pp. 284 ◽  
Author(s):  
Linglin Zeng ◽  
Shun Hu ◽  
Daxiang Xiang ◽  
Xiang Zhang ◽  
Deren Li ◽  
...  
Keyword(s):  
Machine Learning ◽  
Soil Moisture ◽  
Regional Scale ◽  
Remotely Sensed ◽  
Temporal Variations ◽  
Training Data ◽  
Estimation Accuracy ◽  
Learning Approaches ◽  
Remotely Sensed Data ◽  
Deep Soil

Soil moisture mapping at a regional scale is commonplace since these data are required in many applications, such as hydrological and agricultural analyses. The use of remotely sensed data for the estimation of deep soil moisture at a regional scale has received far less emphasis. The objective of this study was to map the 500-m, 8-day average and daily soil moisture at different soil depths in Oklahoma from remotely sensed and ground-measured data using the random forest (RF) method, which is one of the machine-learning approaches. In order to investigate the estimation accuracy of the RF method at both a spatial and a temporal scale, two independent soil moisture estimation experiments were conducted using data from 2010 to 2014: a year-to-year experiment (with a root mean square error (RMSE) ranging from 0.038 to 0.050 m3/m3) and a station-to-station experiment (with an RMSE ranging from 0.044 to 0.057 m3/m3). Then, the data requirements, importance factors, and spatial and temporal variations in estimation accuracy were discussed based on the results using the training data selected by iterated random sampling. The highly accurate estimations of both the surface and the deep soil moisture for the study area reveal the potential of RF methods when mapping soil moisture at a regional scale, especially when considering the high heterogeneity of land-cover types and topography in the study area.

scholarly journals Regional Mapping of Groundwater Potential in Ar Rub Al Khali, Arabian Peninsula Using the Classification and Regression Trees Model

2021 ◽  
Vol 13 (12) ◽  
pp. 2300
Author(s):  
Samy Elmahdy ◽  
Tarig Ali ◽  
Mohamed Mohamed
Keyword(s):  
Machine Learning ◽  
Regional Scale ◽  
Regression Trees ◽  
Classification And Regression Trees ◽  
Groundwater Potential ◽  
Machine Learning Algorithms ◽  
Conditioning Factors ◽  
Potential Mapping ◽  
Classification And Regression ◽  
Groundwater Potential Mapping

Mapping of groundwater potential in remote arid and semi-arid regions underneath sand sheets over a very regional scale is a challenge and requires an accurate classifier. The Classification and Regression Trees (CART) model is a robust machine learning classifier used in groundwater potential mapping over a very regional scale. Ten essential groundwater conditioning factors (GWCFs) were constructed using remote sensing data. The spatial relationship between these conditioning factors and the observed groundwater wells locations was optimized and identified by using the chi-square method. A total of 185 groundwater well locations were randomly divided into 129 (70%) for training the model and 56 (30%) for validation. The model was applied for groundwater potential mapping by using optimal parameters values for additive trees were 186, the value for the learning rate was 0.1, and the maximum size of the tree was five. The validation result demonstrated that the area under the curve (AUC) of the CART was 0.920, which represents a predictive accuracy of 92%. The resulting map demonstrated that the depressions of Mondafan, Khujaymah and Wajid Mutaridah depression and the southern gulf salt basin (SGSB) near Saudi Arabia, Oman and the United Arab Emirates (UAE) borders reserve fresh fossil groundwater as indicated from the observed lakes and recovered paleolakes. The proposed model and the new maps are effective at enhancing the mapping of groundwater potential over a very regional scale obtained using machine learning algorithms, which are used rarely in the literature and can be applied to the Sahara and the Kalahari Desert.

scholarly journals Combining Regression and Clustering for Financial Analysis

2020 ◽  
Vol 2 (1) ◽  
pp. 109-118
Author(s):  
Andreea Ioana Chiriac
Keyword(s):  
Artificial Intelligence ◽  
Machine Learning ◽  
Cluster Analysis ◽  
Business Processes ◽  
Financial Analysis ◽  
Two Dimensions ◽  
Machine Learning Algorithms ◽  
Financial Ratio ◽  
Research Perspective ◽  
Systems Learning

Abstract Artificial Intelligence is used in business through machine learning algorithms. Machine learning is a part of computer science focused on computer systems learning to perform a specific task without using explicit instructions, relying on patterns and inference instead. Though it might seem like we’ve come a long way in the last ten years, which is true from a research perspective, the adoption of AI among corporations is still relatively low. Over time it became possible to automate more tasks and business processes than ever before. The benefit of using artificial intelligence is that does not require to program every step of the process, predicting at each step what could happen and how to resolve it. The algorithms decide for themselves in each case how the problems should be solved, based on the data that is used. I apply Python language to create a synthetic feature vector that allows visualizations in two dimensions for EDIBTA financial ratio. I use Mean-Square Error in order to evaluate the success, having the optimal parameters. In this section, I also mentioned about the purpose, goals, and applications of cluster analysis. I indicated about the basics of cluster analysis and how to do it and also did a demonstration on how to use K-Means.

Evaluation of Sea fog Detection Accuracy Based on Geostationary Satellite Image Using Machine Learning

2020 ◽  
Author(s):  
NaKyeong Kim ◽  
Suho Bak ◽  
Minji Jeong ◽  
Hongjoo Yoon
Keyword(s):  
Machine Learning ◽  
Remote Sensing ◽  
Satellite Remote Sensing ◽  
Geostationary Satellite ◽  
Machine Learning Algorithms ◽  
Detection Accuracy ◽  
Sea Fog ◽  
Short Period ◽  
Wide Range ◽  
Fog Detection

<p><span>A sea fog is a fog caused by the cooling of the air near the ocean-atmosphere boundary layer when the warm sea surface air moves to a cold sea level. Sea fog affects a variety of aspects, including maritime and coastal transportation, military activities and fishing activities. In particular, it is important to detect sea fog as they can lead to ship accidents due to reduced visibility. Due to the wide range of sea fog events and the lack of constant occurrence, it is generally detected through satellite remote sensing. Because sea fog travels in a short period of time, it uses geostationary satellites with higher time resolution than polar satellites to detect fog. A method for detecting fog by using the difference between 11 μm channel and 3.7 μm channel was widely used when detecting fog by satellite remote sensing, but this is difficult to distinguish between lower clouds and fog. Traditional algorithms are difficult to find accurate thresholds for fog and cloud. However, machine learning algorithms can be used as a useful tool to determine this. In this study, based on geostationary satellite imaging data, a comparative analysis of sea fog detection accuracy was conducted through various methods of machine learning, such as Random Forest, Multi-Layer Perceptron, and Convolutional Neural Networks.</span></p>

Spatio-temporal cluster analyses of landslides in Italy at national and regional scale

2020 ◽  
Author(s):  
Marj Tonini ◽  
Kim Romailler ◽  
Gaetano Pecoraro ◽  
Michele Calvello
Keyword(s):  
Cluster Analysis ◽  
Point Processes ◽  
Statistical Significance ◽  
Regional Scale ◽  
Regional Analysis ◽  
Scan Statistics ◽  
Economic Losses ◽  
Campania Region ◽  
Temporal Cluster ◽  
Spatio Temporal

<p><strong>Keywords:</strong> Landslides, FraneItalia, cluster analysis, spatio-temporal point process</p><p>In Italy landslides pose a significant and widespread risk, resulting in a large number of casualties and huge economic losses. Landslide inventories are critical to support investigations of where and when landslides have happened and may occur in the future, i.e. to establish reliable correlations between triggering factors and landslide occurrences. To deal with this issue, statistical methods originally developed for spatio-temporal stochastic point processes can be useful for identifying correlations between events in space and time and detecting a significant excess of cases within large landslide datasets.</p><p>In the present study, the authors propose an approach to analyze and visualize spatio-temporal clusters of landslides occurred in Italy in the period 2010-2017, considering the weather warning zones as territorial units. Besides, a regional analysis was conducted in Campania region considering the municipalities as territorial units. Data on landslide occurrences derived from the FraneItalia catalog, an inventory retrieved from online Italian news. The database contains 8931 landslides, grouped in 4231 single events and 938 areal events (records referring to multiple landslides triggered by the same cause in the same geographic area). Analyses were performed both annually, considering each year individually, and globally, considering the entire frame period. We applied the spatio-temporal scan statistics permutation model (STPSS, integrated in SaTScan<sup>TM</sup> software), which allowed detecting clusters’ location and estimating their statistical significance. STPSS is based on cylindrical moving windows which scan the area across the space and in time counting the number of observed and expected occurrences and computing the likelihood ratio. The statistical inference (p-value) is evaluated by Monte Carlo sampling and finally the most likely clusters in the real and randomly generated datasets are compared.</p><p>Although more detailed analyses are required for the determination of cause-effect relationships among landslides and other variables, some relations with the local topographic and meteorological conditions can already be argued. At national scale, spatio-temporal clusters of landslides are mainly recurrent in two zones: the area enclosing Liguria Region – Northern Tuscany at north-west and the area between Abruzzo and Molise regions at centre-east. During the year, landslide clusters are particularly abundant between October and March. as most of the events in the FraneItalia catalog are rainfall-induced, strongly influenced by seasonal rainfall patterns. Concerning the regional analysis, most of the clusters are located in the Lattari mountains, the Pizzo d’Alvano massif and the Picentini mountains, areas highly susceptible to landslide occurrence due to geomorphological factors.</p><p>In conclusion, the application of spatio-temporal cluster analysis at various scale allowed the identification of frame periods with greater landslide activity. The question of whether this increase in activity depends climate conditions or topographic factors is still open and request further investigations.</p><p>REFERENCES</p><p>Calvello, M., Pecoraro, G. FraneItalia: a catalog of recent Italian landslides. <em>Geoenvironmental Disasters</em>. 5: 13 (2018)</p><p>Tonini, M. & Cama, M. Spatio-temporal pattern distribution of landslides causing damage in Switzerland. <em>Landslides</em> 16 (2019)</p>

scholarly journals An improved sea ice detection algorithm using MODIS: application as a new European sea ice extent indicator

2021 ◽  
Vol 15 (6) ◽  
pp. 2803-2818
Author(s):  
Joan Antoni Parera-Portell ◽  
Raquel Ubach ◽  
Charles Gignac
Keyword(s):  
Sea Ice ◽  
Regional Scale ◽  
Ice Cover ◽  
Detection Algorithm ◽  
Temporal Variations ◽  
The Arctic ◽  
Sea Ice Extent ◽  
Monitoring Tools ◽  
Spatio Temporal ◽  
Ice Detection

Abstract. The continued loss of sea ice in the Northern Hemisphere due to global warming poses a threat to biota and human activities, evidencing the necessity of efficient sea ice monitoring tools. Aiming at the creation of an improved sea ice extent indicator covering the European regional seas, the new IceMap500 algorithm has been developed to classify sea ice and water at a resolution of 500 m at nadir. IceMap500 features a classification strategy built upon previous MODIS sea ice extent algorithms and a new method to reclassify areas affected by resolution-breaking features inherited from the MODIS cloud mask. This approach results in an enlargement of mapped area, a reduction of potential error sources and a better delineation of the sea ice edge, while still systematically achieving accuracies above 90 %, as obtained by manual validation. Swath maps have been aggregated at a monthly scale to obtain sea ice extent with a method that is sensitive to spatio-temporal variations in the sea ice cover and that can be used as an additional error filter. The resulting dataset, covering the months of maximum and minimum sea ice extent (i.e. March and September) over 2 decades (from 2000 to 2019), demonstrates the algorithm's applicability as a monitoring tool and as an indicator, illustrating the sea ice decline at a regional scale. The European sea regions located in the Arctic, NE Atlantic and Barents seas display clear negative trends in both March (−27.98 ± 6.01 × 103 km2yr−1) and September (−16.47 ± 5.66 × 103 km2yr−1). Such trends indicate that the sea ice cover is shrinking at a rate of ∼ 9 % and ∼ 13 % per decade, respectively, even though the sea ice extent loss is comparatively ∼ 70 % greater in March.

Implications of methane emissions in biogeochemical budgeting: A study from a eutrophic tropical lake of South India

2020 ◽  
Author(s):  
Revathy Das ◽  
Appukuttan Pillai Krishnakumar ◽  
Krishnan AnoopKrishnan ◽  
Vivekanandan Nandakumar
Keyword(s):  
Plant Biomass ◽  
Regional Scale ◽  
Temporal Variations ◽  
Physical Factors ◽  
Growing Seasons ◽  
Littoral Zones ◽  
Flux Measurements ◽  
The Mean ◽  
Spatio Temporal ◽  
Lake Systems

<p>Greenhouse gases (GHGs), especially, methane (CH<sub>4</sub>) emissions from the littoral zones of the lakes play an important role in regional biogeochemical budgets. Only a few studies are available in literature highlighting the direct flux measurements of CH<sub>4   </sub>from the aquatic systems. In the present study, an attempt has been made to quantify the spatio-temporal variations of CH<sub>4</sub> efflux and the key physical factors controlling the emission rate, from the vegetated littoral zones of lake Vellayani (5.55Km<sup>2</sup>), located in the urbanized area of Thiruvananthapuram city, Kerala, South-West India. CH<sub>4</sub> efflux were collected from different vegetations in littoral zones, using a static chamber, during the peak growing seasons from March to October in 2016 and further analyses were carried out by using Gas Chromatograph (PE Clarus 500, PerkinElmer, Inc.). The mean efflux rate of CH<sub>4   </sub>from the emergent plant species (Phragmites australis and Typha spp.) was 114.4 mg CH<sub>4</sub> m<sup>-2</sup>h<sup>-1</sup>; while, in the floating leaved species (Nymphaea spp. and Nelumbo Spp.), it <sub>  </sub>was   observed to be 32.6 mgCH<sub>4</sub> m<sup>-2</sup>h<sup>-1</sup>. The results reveal that CH<sub>4</sub> efflux in the zone of emergent vegetation was significantly higher than the floating-leaved zone indicating the importance of plant biomass and standing water depths for the spatial variations of CH<sub>4 </sub>efflux. However, no significant temporal variations were noticed in the physical factors during the peak growing seasons. These results indicate that the vegetated littoral zones of lake, especially the emergent plant zones were supersaturated with CH<sub>4</sub>, facilitating the production of carbon for CH<sub>4</sub> emission<sub>,</sub> but also enable the release of CH<sub>4 </sub>by the diffusion from the intercellular gas lacunas. We conclude that the atmospheric CH<sub>4</sub> emissions will be affected by the growth of exotic species in the lake systems and may be the reason for enhancing the climate warming in local/regional scale.</p>

scholarly journals Correcting for Systematic Underestimation of Topographic Glacier Aerodynamic Roughness Values From Hintereisferner, Austria

2021 ◽  
Vol 9 ◽  
Author(s):  
Joshua R. Chambers ◽  
Mark W. Smith ◽  
Thomas Smith ◽  
Rudolf Sailer ◽  
Duncan J. Quincey ◽  
...  
Keyword(s):  
Laser Scanning ◽  
Regional Scale ◽  
Laser Scanner ◽  
Turbulent Energy ◽  
Temporal Variations ◽  
Glacier Surface ◽  
Order Of Magnitude ◽  
Spatio Temporal ◽  
Aerodynamic Roughness ◽  
Elevation Model

Spatially-distributed values of glacier aerodynamic roughness (z0) are vital for robust estimates of turbulent energy fluxes and ice and snow melt. Microtopographic data allow rapid estimates of z0 over discrete plot-scale areas, but are sensitive to data scale and resolution. Here, we use an extensive multi-scale dataset from Hintereisferner, Austria, to develop a correction factor to derive z0 values from coarse resolution (up to 30 m) topographic data that are more commonly available over larger areas. Resulting z0 estimates are within an order of magnitude of previously validated, plot-scale estimates and aerodynamic values. The method is developed and tested using plot-scale microtopography data generated by structure from motion photogrammetry combined with glacier-scale data acquired by a permanent in-situ terrestrial laser scanner. Finally, we demonstrate the application of the method to a regional-scale digital elevation model acquired by airborne laser scanning. Our workflow opens up the possibility of including spatio-temporal variations of z0 within glacier surface energy balance models without the need for extensive additional field data collection.

Sign in / Sign up

Export Citation Format

Share Document