High-performance computing for the simulation of dust storms

doi:10.1016/j.compenvurbsys.2009.08.002

Computers, Environment and Urban Systems

Volume 34, Issue 4, July 2010, Pages 278-290

https://doi.org/10.1016/j.compenvurbsys.2009.08.002 Get rights and content

Abstract

Dust storm is a primary natural hazard that impact human activities. Previous work has simulated dust events with Eta model. This paper reports our effort to migrate a Dust Regional Atmospheric Model (DREAM) to the Nonhydrostatic Mesoscale Model (NMM), core of the Weather Research and Forecasting (WRF) system’s (WRF–NMM)-based dust model. The process is as follows: (1) the DREAM model is modified to fit into a high-performance computing environment by parallelizing the dust model based on the WRF–NMM weather forecasting model. We couple a dust module to the WRF–NMM model and parallelize the serial program using Message Passing Interface (MPI). (2) The performance of the parallel dust simulation model is tested using the southwestern United States as the experiment area and a dust event during 7 January 2008 for the case study. (3) The new dust model based on WRF–NMM is implemented and tested on a Linux cluster with 28 computing nodes and over 200 CPU cores. The new model can leverage high-performance computing clusters to reduce the execution time while at the same time enhancing the simulation resolution. As tested, when simulating a 72 h period, the WRF–NMM dust model took 1.8 h when using a resolution of 1/9 of a degree grid space, when 36 CPU cores were used for high-performance computing. The results show that the parallelized version of the dust simulation model can get a maximum speedup about 13.3. We also built a transformation component to export the model output to geographic information system (GIS) software through a Web Map Service (WMS).

Introduction

Wind-blown dust is an important environmental issue. In regions with high soil erosion, dust behaves as a pollutant that significantly reduces the air quality (Nichovic, Kallos, Papadopoulos, & Kakaliagou, 2001). Dust storms directly affect visibility and have a significant impact on daily commercial and military operations near desert regions (Barnum et al., 2004). In addition to the mineral composition of the particles themselves, dust can carry a variety of other particles, including bacteria, fungi, and chemical contaminants, all of which can adversely affect human health and the environment (Taylor, 2002). Dust can also cause either a warming or a cooling of the Earth’s surface depending on the characteristics of the dust, concentration, vertical distribution in the atmospheric column, particle size distribution and mineralogy, and external variables including the albedo and temperature of the underlying surface (Harrison, Kohfeld, Roelandt, & Claquin, 2001). Particulate matter (PM) of 10 μm diameter or less, and especially of 2.5 μm or less, is exceptionally detrimental. Studies have suggested that breathing fine particles spewed by vehicles, factories, and power plants can trigger heart attacks and worsen respiratory disease in vulnerable people (e.g. patients with chronic pneumonia) (Kaiser, 2005).

Using the Global Ozone Chemistry Aerosol Radiation Transport (GOCART) model, Taylor (2002) identified 10 main dust sources: (1) the Salton Sea (California, US), (2) Patagonia (South America), (3) the Altipläno (South America), (4) the Sahel region (Africa), (5) the Sahara Desert (Africa), (6) the Namibian desert lands (Africa), (7) the Indus Valley (Asia), (8) the Taklimakan Desert (Asia), (9) the Gobi Desert (Asia), and (10) the Lake Eyre basin (Australia). Every year, Sahelian dust storms in Africa, which can rise up to 4 km above ground level, launch massive amounts of dust (estimated from 500 million to over 1 billion tons) into the atmosphere (Taylor, 2002). Wind erosion is a major agent of land degradation over large areas of inland Australia (Butler, Hogarth, & McTainsh, 1996). The first dust source in the map is located in southwestern United States. In the arid and semi-arid southwestern US, including Arizona, New Mexico, and Texas, wind-blown desert dust turns out to be one of the key sources of PM pollution. Multi-state dust events occur occasionally in this region (Yin, Nickovic, Barbaris, Chandy, & Sprigg, 2005). One of the worst desert dust storms in recent years in the southwestern US was observed during the period of 15–17 December 2003 (Yin et al., 2005). The major effects of that dust storm were seen in eastern New Mexico and western Texas, where cities were shrouded in the dust and visibilities dropped to as low as 1/4 mile at some locations.

The impact of dust on human health has increased the need for better understanding and eventually prediction of the atmospheric dust cycle (Nichovic et al., 2001). A deeper understanding of the relative magnitude of the various potential effects of dust will depend on the development and application of global models of the dust cycle and its interaction with other components of the Earth system (Harrison et al., 2001). NASA funded the Public Health Applications in Remote Sensing (PHAiRS) (Mahler, Thome, Yin, & Sprigg, 2006) project, which aims to address this problem by using remote-sensing products to support public health decision-making. As part of PHAiRS, a model for simulating desert dust cycles, known as the Dust Regional Atmospheric Modeling (DREAM-Eta) system, is employed to forecast dust events in the southwestern US.

The resolution of the simulations of dust storms needs to be increased to the ZIP-code level within the southwestern US in the PHAiRS project in order to support public health decision-making. This requirement will make the model more computationally intensive. High-performance computing can harness high-performance hardware such as multi-CPU-core computers or clusters to enhance computing capacity and reduce execution time for scientific models. High-performance computing is used to enhance the computing capacity of the dust simulation models in this paper. A parallel version of the dust simulation model is urgently needed to take full advantage of the efficiency enhancements offered by computing clusters and by high-performance computing. Furthermore, in order to share the dust simulation results with the public and to assess the environmental impact of dust storms, an information system is needed. Geographic information system (GIS) is an effective tool that can meet that need. GIS has the ability to overlay locations of people especially sensitive to dust live, such as schools and hospitals, onto the dust model output. However, such a function requires interoperability between the dust simulation model and the GIS software. This paper is a report of our research and development progress.

The principle of dust simulation models is introduced in Section 2. Section 3 discusses the use of high-performance computing support for dust simulation models. A dust model is coupled into WRF–NMM model. The dust event of 7 January 2008 is used as a case study to test the dust simulation model running in the high-performance computing environment, as described in Section 4. Section 5 concludes the work the paper has done and further research in the future.

Section snippets

Dust storm and dust simulation model

Many scientific models have been developed to simulate dust movement and its effects on the atmosphere. The United States Air Force Weather Agency (AFWA) supported the development of a dust forecast mode called the Community Aerosol and Radiation Model for Atmospheres (CARMA) (Barnum et al., 2004). The version of CARMA developed for the daily forecasting of dust has been modified to assimilate meteorological forecast data from the Penn State fifth generation Mesoscale Meteorology Model (MM5) (

Parallel computing and limitations of previous model

Parallel processing is a computing technique that emphasizes the feasible exploitation of available concurrency in a computational process (Hazra, 1995). Parallel computers can be classified into two kinds by the main memory structure: shared memory and distributed memory. The shared-memory parallel computer shares memory among all processing nodes in a single address space. Distributed memory refers to memory being logically distributed or physically distributed, in which case the processing

Study area description

The dust event on 7 January 2008 is used as the case study for the WRF–NMM dust simulation model. The study area is the southwestern United States as shown in Fig. 7. The center of the model domain is latitude 34.02 and longitude −108.9. The grid space is 0.073424° (approximately 10 km) in the X direction and 0.072729° (approximately 10 km) in the Y direction. There are 151 grid boxes in the X direction and 219 grid boxes in the Y direction.

There are three types of input data for the dust

Conclusion

This paper reports a new version of a dust simulation model developed to enhance the resolution and response time of dust prediction. The old serial computing DREAM-Eta model is replaced by the parallel computing WRF–NMM dust model, in which the dust simulation model is coupled into the latest version of the weather forecasting model WRF–NMM developed by NCEP. The new model can leverage high-performance computing clusters to reduce the execution time while at the same time enhancing the

Acknowledgements

This research is supported by NASA through Grants NNX07AD99G and SMD-08-0768, and by FGDC through Grant G09AC00103, and by the National Basic Research Program of China (973 program) through Grant 2009CB723906.

References (33)

B.H. Barnum et al.
Forecasting dust storms using the CARMA-dust model and MM5 weather data
Environmental Modelling and Software
(2004)
S.R.M. Barros et al.
On the parallelization of global spectral weather models
Parallel Computing
(1994)
H.J. Butler et al.
A source-based model for describing dust concentrations during wind erosion events: An initial study
Environmental Software
(1996)
J.H. Christensen
The Danish Eulerian hemispheric model – A three-dimensional air pollution model used for the Arctic
Atmospheric Environment
(1997)
J. Drake et al.
Introduction to the special issue on parallel computing in climate and weather modeling
Parallel Computing
(1995)
G.A. Grell et al.
Application of a multiscale, coupled MM5/chemistry model to the complex terrain of the VOTALP valley campaign
Atmospheric Environment
(2000)
G.A. Grell et al.
Fully coupled “online” chemistry within the WRF model
Atmospheric Environment
(2005)
S.P. Harrison et al.
The role of dust in climate changes today, at the last glacial maximum and in the future
Earth-Science Reviews
(2001)
P.R. Woodward et al.
Parallel computation of turbulent fluid flows with the piecewise-parabolic method
Parallel Computational Fluid Dynamics
(2006)
D. Yin et al.
Modeling wind-blown desert dust in the southwestern United States for public health warning: A case study
Atmospheric Environment
(2005)

D. Yin et al.

The impact of using different land cover data on wind-blown desert dust modeling results in the southwestern United States

Atmospheric Environment

(2007)

P. Alpert et al.

A dust prediction system with TOMS initialization

Monthly Weather Review

(2002)

R.A. Anthes et al.

Development of hydrodynamic models suitable for air pollution and other meso-meteorological studies

Monthly Weather Review

(1978)

R.L. Carpenter et al.

Application of the Piecewise Parabolic Method (PPM) to meteorological modeling

Monthly Weather Review

(1990)

Casas, M., Badia, R. M., & Labarta, J. (2008). Automatic analysis of speedup of MPI applications. In Proceedings of the...

T.K. Hazra

Parallel computing

IEEE Potentials

(1995)

Cited by (44)

Deep learning-based downscaling of tropospheric nitrogen dioxide using ground-level and satellite observations
2021, Science of the Total Environment
Air quality is one of the major issues within an urban area that affect people's living environment and health conditions. Existing observations are not adequate to provide a spatiotemporally comprehensive air quality information for vulnerable populations to plan ahead. Launched in 2017, TROPOspheric Monitoring Instrument (TROPOMI) provides a high spatial resolution (~5 km) tropospheric air quality measurement that captures the spatial variability of air pollution, but still limited by its daily overpass in the temporal dimension and relatively short historical records. Integrating with the hourly available AirNOW observations by ground-level discrete stations, we proposed and compared two deep learning methods that learn the relationship between the ground-level nitrogen dioxide (NO₂) observation from AirNOW and the tropospheric NO₂ column density from TROPOMI to downscale the daily NO₂ to an hourly resolution. The input predictors include the locations of AirNOW stations, AirNOW NO₂ observations, boundary layer height, other meteorological status, elevation, major roads, and power plants. The learned relationship can be used to produce NO₂ emission estimates at the sub-urban scale on an hourly basis. The two methods include 1) an integrated method between inverse weighted distance and a feed forward neural network (IDW + DNN), and 2) a deep matrix network (DMN) that maps the discrete AirNOW observations directly to the distribution of TROPOMI observations. We further compared the accuracies of both models using different configurations of input predictors and validated their average Root Mean Squared Error (RMSE), average Mean Absolute Error (MAE) and the spatial distribution of errors. Results show that DMN generates more reliable NO₂ estimates and captures a better spatial distribution of NO₂ concentrations than the IDW + DNN model.
The importance of spatio-temporal infrastructure assessment:Evidence for 5G from the Oxford–Cambridge Arc
2020, Computers, Environment and Urban Systems
The roll-out of 5G infrastructure can provide enhanced high capacity, low latency communications enabling a range of new use cases. However, to deliver the improvements 5G promises, we need to understand how to enhance capacity and coverage, at reasonable cost, across space and over time. In this paper, we take a spatio-temporal simulation modeling approach, using industry-standard engineering models of 5G wireless networks, to test how different infrastructure strategies perform under scenarios of uncertain future demand. We use coupled open-source models to analyze a UK growth corridor, a system-of-cities comprising 7 urban areas, known as the Oxford-Cambridge Arc. We find that population growth has a marginal impact on total demand for 5G (up to 15%), as the main factor driving demand is the increase in per user data consumption resulting mainly from video. Additionally, the results suggest only limited justification for deploying 5G based purely on the need for more capacity. Strategies which reuse existing brownfield Macro Cell sites are enough to meet future demand for Enhanced Mobile Broadband, except in the densest urban areas. While spatio-temporal analysis of infrastructure is common in some sectors (e.g. transport, energy and water), there has been a lack of open analysis of digital infrastructure. This study makes a novel contribution by providing an open and reproducible spatio-temporal assessment of different 5G technologies at a time when 5G is starting to roll-out around the world.
Adaptive energy-aware scheduling method in a meteorological cloud
2019, Future Generation Computer Systems
Citation Excerpt :
While most researchers focus on the DAG of parallel applications, we solve the problem by considering the relation between the number of assigned resources and the execution time. Parallel meteorological models include WRF [17], the High-resolution Regional Model (HRM) [13], Consortium for Small-Scale Model (COSMO) [13], Nonhydrostatic Mesoscale Model (NMM) [14], etc. Such models involve typical moldable parallel tasks [15,16] that can only be configured in the beginning, and the assigned resources are exclusively kept until all sub-tasks have been finished.
This paper focuses on the energy-aware scheduling problem of moldable non-linear parallel tasks in a meteorological cloud. Such a meteorological Cloud mainly provides computing resources for the execution of meteorological models, such as Weather Research and Forecasting model (WRF). In a meteorological Cloud, the parallelism of tasks (i.e., meteorological models) can only be configured in the beginning, and the assigned resources retained exclusively until all sub-tasks have been finished. For the scheduling of those tasks, one key challenge is how to reduce the average energy consumption while guaranteeing others requirements of such tasks. We address this challenge by considering simultaneously the deadlines of tasks, the energy consumption, the system load, and the non-linear speedup of parallel tasks when we make the scheduling decision. Specifically, we propose an adaptive energy-aware scheduling method called ASSD, that is based on the Dynamic Voltage and Frequency Scaling (DVFS) model of computing resources and the speedup of tasks under different parallelisms. We evaluate our method via simulations on a meteorological cloud. Our results show that the proposed method not only increases the number of completed tasks but also significantly reduces the average energy consumption.
Utilizing Cloud Computing to address big geospatial data challenges
2017, Computers, Environment and Urban Systems
Big Data has emerged with new opportunities for research, development, innovation and business. It is characterized by the so-called four Vs: volume, velocity, veracity and variety and may bring significant value through the processing of Big Data. The transformation of Big Data's 4 Vs into the 5th (value) is a grand challenge for processing capacity. Cloud Computing has emerged as a new paradigm to provide computing as a utility service for addressing different processing needs with a) on demand services, b) pooled resources, c) elasticity, d) broad band access and e) measured services. The utility of delivering computing capability fosters a potential solution for the transformation of Big Data's 4 Vs into the 5th (value). This paper investigates how Cloud Computing can be utilized to address Big Data challenges to enable such transformation. We introduce and review four geospatial scientific examples, including climate studies, geospatial knowledge mining, land cover simulation, and dust storm modelling. The method is presented in a tabular framework as a guidance to leverage Cloud Computing for Big Data solutions. It is demostrated throught the four examples that the framework method supports the life cycle of Big Data processing, including management, access, mining analytics, simulation and forecasting. This tabular framework can also be referred as a guidance to develop potential solutions for other big geospatial data challenges and initiatives, such as smart cities.
Building Model as a Service to support geosciences
2017, Computers, Environment and Urban Systems
Citation Excerpt :
The code-level parallelization of enabling a model to run in parallel using the parallelization technologies such as MPICH27 is well studied and beyond the scope of this paper. However, if a model is already MPI-enabled, such as the dust model NMM-dust (Xie et al., 2010), such a model can be incorporated into MaaS, and a cloud-based HPC cluster will be provisioned to support running the model. This is discussed in Section 5.4.
Modeling is a fundamental methodology for simulating the past, understanding the present and predicting the future of the geospatial systems and phenomena. However, modeling in the geospatial science poses several challenges, including complex model setup, repetition in model setup, requirement for large, scalable computing resources, and management of a large amount of model output. To address these challenges, we propose Model as a Service (MaaS) by leveraging the latest advancement of cloud computing. MaaS enables various geoscience models to be published as services, and these services can be accessed through a simple web interface. MaaS automates the processes of configuring machines, setting up and running models, and managing model outputs. The computing resources are automatically provisioned by MaaS in a cloud environment. A proof-of-concept MaaS prototype is presented using a global climate change model (ModelE). Experimental results show that the MaaS prototype significantly simplifies model setup, accelerates model simulation and enhances model output by providing a web-based, on-demand, scalable modeling environment.
A view-dependent spatiotemporal saliency-driven approach for time varying volumetric data in geovisualization
2016, Computers, Environment and Urban Systems
Geospatial datasets from satellite observations and model simulations are becoming more accessible. These spatiotemporal datasets are relatively massive for visualization to support advanced analysis and decision making. A challenge to visualizing massive geospatial datasets is identifying critical spatial and temporal changes reflected in the data while maintaining high interactive rendering speed, even when data are accessed remotely. We propose a view-dependent spatiotemporal saliency-driven approach that facilitates the discovery of regions showing high levels of spatiotemporal variability and reduces the rendering intensity of interactive visualization. Our method is based on a novel definition of data saliency, a spatiotemporal tree structure to store visual saliency values, as well as a saliency-driven view-dependent level-of-detail (LOD) control. To demonstrate its applicability, we have implemented the approach with an open-source remote visualization package and conducted experiments with spatiotemporal datasets produced by a regional dust storm simulation model. The results show that the proposed method may not be outstanding in some specific situations, but it consistently performs very well across different settings according to different criteria.

View all citing articles on Scopus

View full text

High-performance computing for the simulation of dust storms

Abstract

Introduction

Section snippets

Dust storm and dust simulation model

Parallel computing and limitations of previous model

Study area description

Conclusion

Acknowledgements

Environmental Modelling and Software

Parallel Computing

Environmental Software

Atmospheric Environment

Parallel Computing

Atmospheric Environment

Atmospheric Environment

Earth-Science Reviews

Parallel Computational Fluid Dynamics

Atmospheric Environment

Atmospheric Environment

A dust prediction system with TOMS initialization

Monthly Weather Review

Development of hydrodynamic models suitable for air pollution and other meso-meteorological studies

Monthly Weather Review

Application of the Piecewise Parabolic Method (PPM) to meteorological modeling

Monthly Weather Review

Parallel computing

IEEE Potentials