Observations of an extreme atmospheric river storm with a diverse sensor network

Hatchett, Benjamin J.; Cao, Qiqi; Dawson, Phillip B.; Ellis, C. J.; Hecht, Chad W.; Kawzenuk, Brian; Osborne, Tashiana; Lancaster, Jeremy T.; Wilson, Anna M.; Anderson, M.; Dettinger, Michael D.; Kalansky, Julie; Kaplan, Michael L.; Lettenmaier, Dennis P.; Oakley, Nina S.; Ralph, M.; Reynolds, David W.; White, Allen B.; Sierks, Michael; Sumargo, Edwin

doi:10.1002/essoar.10502163.1

Cited by 5 publications

(7 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Based on observations and anecdotal reports from the operational meteorology community, NCFRs are relatively common in northern California (e.g., Blier, 2003; Blier et al, 2005; Jorgensen et al, 2003; King et al, 2009; White et al, 2003). Although a catalog of NCFR events has never been compiled (whether causing significant landscape response or not), we can point to another recent and nearby NCFR on 14 February 2019 (Hatchett et al, 2020) as evidence of their occurrence and consequence in the central Sierra region. The 14 February 2019 event also had radar reflectivity in excess of 50 dBZ in a narrow band (Figure 9), and traversed through the two counties (Calaveras and Amador) located northwest of Groveland, producing 10‐min rainfall intensities approaching 60 mm/hr (MesoWest, 2019).…”

Section: Discussionmentioning

confidence: 88%

Linking Mesoscale Meteorology With Extreme Landscape Response: Effects of Narrow Cold Frontal Rainbands (NCFR)

et al. 2020

Self Cite

View full text Add to dashboard Cite

Landscapes evolve in response to prolonged and/or intense precipitation resulting from atmospheric processes at various spatial and temporal scales. Whereas synoptic (large-scale) features (e.g., atmospheric rivers and hurricanes) govern regional-scale hydrologic hazards such as widespread flooding, mesoscale features such as thunderstorms or squall lines are more likely to trigger localized geomorphic hazards such as landslides. Thus, to better understand relations between hydrometeorological drivers and landscape response, a knowledge of mesoscale meteorology and its impacts is needed. Here we investigate the extreme geomorphic response associated with one type of mesoscale meteorological feature, the narrow cold frontal rainband (NCFR). Resulting from low-level convergence and shallow convection along a cold front, NCFRs are narrow bands of high-intensity rainfall that occur in midlatitude areas of the world. Our study examines an NCFR impacting the Sierra Nevada foothills (California, USA) that initiated over 500 landslides, mobilized~360,000 metric tons of sediment to the fluvial system (as much as 16 times the local annual sediment yield), and severely damaged local infrastructure and regional water transport facilities. Coupling geomorphological field investigations with meteorological analyses, we demonstrate that precipitation associated with the NCFR was both intense (maximum 15 min intensity of 70 mm/hr) and localized, resulting in a highly concentrated band of shallow landsliding. This meteorological phenomenon likely plays an important role in landscape evolution and hazard initiation. Other types of mesoscale meteorological features also occur globally and offer new avenues for understanding the effects of storms on landscapes. Plain Language Summary Major storms can cause extreme and hazardous landscape disturbances, but links between storm conditions and landscape response such as erosion and landslides remain poorly constrained. This is partly due to the lack of attention generally given to the finer-scale details of storms. We examined one type of atmospheric feature that is common in western North America (as well as in other regions), the narrow cold frontal rainband, and studied its effects on the landscape. In 2018, one such event in the Tuolumne River watershed, California, caused more than 500 landslides in a narrow area, moving more sediment in one day than the river would normally transport in a year. We find that landscape change, including potentially hazardous events such as landslides, can be driven primarily by fine-scale rainfall patterns rather than by the larger-scale storm conditions. More integration between weather and landscape scientists can advance knowledge of how storms influence landscapes and produce hazards, especially during extreme events.

show abstract

Section: Discussionmentioning

confidence: 88%

Linking Mesoscale Meteorology With Extreme Landscape Response: Effects of Narrow Cold Frontal Rainbands (NCFR)

et al. 2020

Self Cite

View full text Add to dashboard Cite

show abstract

“…The GCN layer is derived from Spektral. 4 Calculations are done with support of Numpy 5 and table formatting with Pandas. 6 Furthermore, to reduce the overall training time, the models are trained on a dedicated server with two Intel Xeon CPUs (3.2 GHz), 256 GB RAM and a Nvidia Quadro RTX3956000 (24 GB) GPU.…”

Section: Software and Computermentioning

confidence: 99%

“…Spread across a large geographical region, a set of sensors can then form a sensor network used for data collection and analysis [2], in particular considering large-scale time series data. Example domains where such real-world sensor data is analyzed include, e. g., traffic [3], weather [4] and seismology [5], in particular regarding time series regression and classification, e. g., [6,7].…”

Section: Introductionmentioning

confidence: 99%

Graph neural networks for multivariate time series regression with application to seismic data

Bloemheuvel

Hoogen

Jozinović

et al. 2022

Int J Data Sci Anal

View full text Add to dashboard Cite

Machine learning, with its advances in deep learning has shown great potential in analyzing time series. In many scenarios, however, additional information that can potentially improve the predictions is available. This is crucial for data that arise from e. g., sensor networks that contain information about sensor locations. Then, such spatial information can be exploited by modeling it via graph structures, along with the sequential (time series) information. Recent advances in adapting deep learning to graphs have shown potential in various tasks. However, these methods have not been adapted for time series tasks to a great extent. Most attempts have essentially consolidated around time series forecasting with small sequence lengths. Generally, these architectures are not well suited for regression or classification tasks where the value to be predicted is not strictly depending on the most recent values, but rather on the whole length of the time series. We propose TISER-GCN, a novel graph neural network architecture for processing, in particular, these long time series in a multivariate regression task. Our proposed model is tested on two seismic datasets containing earthquake waveforms, where the goal is to predict maximum intensity measurements of ground shaking at each seismic station. Our findings demonstrate promising results of our approach—with an average MSE reduction of 16.3%—compared to the best performing baselines. In addition, our approach matches the baseline scores by needing only half the input size. The results are discussed in depth with an additional ablation study.

show abstract

“…Spread across a large geographical region, a set of sensors can then form a sensor network used for data collection and analysis [2], in particular considering large-scale time series data. Example domains where such real-world sensor data is analyzed include, e. g., traffic [3], weather [4] and seismology [5], regarding time series anomaly detection, forecasting, segmentation, as well as regression and classification. Exemplary standard techniques to tackle such problems are Gaussian processes [6], ARIMA [7] and XGBoost [8].…”

Section: Introductionmentioning

confidence: 99%

Graph Neural Networks for Multivariate Time Series Regression with Application to Seismic Data

Bloemheuvel,

Hoogen,

Jozinović

et al. 2022

Preprint

View full text Add to dashboard Cite

Machine learning, with its advances in Deep Learning has shown great potential in analysing time series in the past. However, in many scenarios, additional information is available that can potentially improve predictions, by incorporating it into the learning methods. This is crucial for data that arises from e. g., sensor networks that contain information about sensor locations. Then, such spatial information can be exploited by modeling it via graph structures, along with the sequential (time) information. Recent advances in adapting Deep Learning to graphs have shown promising potential in various graph-related tasks. However, these

show abstract

Observations of an extreme atmospheric river storm with a diverse sensor network

Cited by 5 publications

References 0 publications

Linking Mesoscale Meteorology With Extreme Landscape Response: Effects of Narrow Cold Frontal Rainbands (NCFR)

Linking Mesoscale Meteorology With Extreme Landscape Response: Effects of Narrow Cold Frontal Rainbands (NCFR)

Graph neural networks for multivariate time series regression with application to seismic data

Graph Neural Networks for Multivariate Time Series Regression with Application to Seismic Data

Contact Info

Product

Resources

About