Engineering topics

Keller, Christoph A.

Publications and source records attributed to Keller, Christoph A..

13 records

NASA NTRS2019

Accelerated Simulation of Air Pollution Using NVIDIA RAPIDS

Atmospheric chemistry models are a central tool to study and forecast the impact of air pollution on the environment, vegetation, and human health. However, the numerical simulation of chemical kinetics is computationally expensive due to the stiffness of the system of ordinary differential equations that describes atmospheric chemistry. Here we present an alternative approach to the computation of atmospheric chemistry based on machine learning. Our training data set is produced using the NASA Goddard Earth Observing System (GEOS) model with GEOS-Chem chemistry, run on the NASA Center for Climate Simulation (NCCS) Discover supercomputing cluster on 384 Intel Xeon Haswell cores. This model spends more than 50% of total run time on solving atmospheric chemistry. The data set contains as input features the air pollution concentrations before solving the differential equations, together with some key physical parameters such as temperature and sun intensity. As target variables we define the air pollution concentrations after solving the differential equations. Using Dask-cuDF and Dask-XGBoost on the NVIDIA RAPIDS platform on 8 Tesla V100 GPUs, we generate from this training set gradient boosted decision tree models that can reproduce the simulation of chemical kinetics. We do this on the NCCS Advanced Data Analytics Platform (ADAPT) science cloud environment. Our application takes full advantage of recent advances in Dask-XGBoost, such as multi-node and multi-GPU scaling for distributed training with large data sets. The increase in training data size enabled by this is critical to capture the full range of chemical environments encountered across the globe and all annual seasons.The boosted tree models offer good predictability and show many of the features of the full chemistry reference simulation. Further improvements can be achieved through mass balance considerations and by accounting for error correlations. We incorporate the boosted tree models into the GEOS reference model using XGBoost's C API. This enables a seamless integration of the GPU trained models into GEOS-Chem, which is written in Fortran and optimized for use in a massively parallel CPU environment. We show the benefits of this approach and discuss the potential speedup of this machine learning accelerated atmospheric chemistry model.

Keller, Christoph A.↗

NASA NTRS2019

Machine Learning Application to Atmospheric Chemistry Modeling

Atmospheric chemistry models are a central tool to study the impact of chemical constituents on the environment, vegetation and human health. These models split the atmosphere in a large number of grid-boxes and consider the emission of compounds into these boxes and their subsequent transport, deposition, and chemical processing. The chemistry is represented through a series of simultaneous ordinary differential equations, one for each compound. Given the difference in life-times between the chemical compounds (milli-seconds for O (sup 1) D (Deuterium) to years for CH4) these equations are numerically stiff and solving them consists of a significant fraction of the computational burden of a chemistry model. We have investigated a machine learning approach to emulate the chemistry instead of solving the differential equations numerically. From a one-month simulation of the GEOS-Chem model we have produced a training dataset consisting of the concentration of compounds before and after the differential equations are solved, together with some key physical parameters for every grid-box and time-step. From this dataset we have trained a machine learning algorithm (regression forest) to be able to predict the concentration of the compounds after the integration step based on the concentrations and physical state at the beginning of the time step. We have then included this algorithm back into the GEOS-Chem model, bypassing the need to integrate the chemistry. This machine learning approach shows many of the characteristics of the full simulation and has the potential to be substantially faster. There are a wide range of application for such an approach - generating boundary conditions, for use in air quality forecasts, chemical data assimilation systems, etc. We discuss speed and accuracy of our approach, and highlight some potential future directions for improving it.

Keller, Christoph A.↗

NASA NTRS2019

Atmospheric Chemistry Modeling and Air Quality Forecasting Using Machine Learning

Atmospheric chemistry models are a central tool to study the impact of chemical constituents on the environment, vegetation and human health. These models split the atmosphere in a large number of grid-boxes and consider the emission of compounds into these boxes and their subsequent transport, deposition, and chemical processing. The chemistry is represented through a series of simultaneous ordinary differential equations, one for each compound. Given the difference in life-times between the chemical compounds (milli-seconds for O1D to years for CH4) these equations are numerically stiff and solving them consists of a significant fraction of the computational burden of a chemistry model.We have investigated a machine learning approach to emulate the chemistry instead of solving the differential equations numerically. From a one-month simulation of the GEOS-Chem model we have produced a training dataset consisting of the concentration of compounds before and after the differential equations are solved, together with some key physical parameters for every grid-box and time-step. From this dataset we have trained a machine learning algorithm (regression forest) to be able to predict the concentration of the compounds after the integration step based on the concentrations and physical state at the beginning of the time step. We have then included this algorithm back into the GEOS-Chem model, bypassing the need to integrate the chemistry.This machine learning approach shows many of the characteristics of the full simulation and has the potential to be substantially faster. There are a wide range of application for such an approach - generating boundary conditions, for use in air quality forecasts, chemical data assimilation systems, etc. We discuss speed and accuracy of our approach, and highlight some potential future directions for improving it.

Keller, Christoph A.↗

NASA NTRS2019

Atmospheric Chemistry Modeling Using Machine Learning

Atmospheric chemistry models are a central tool to study the impact of chemical constituents on the environment, vegetation and human health. These models split the atmosphere in a large number of grid-boxes and consider the emission of compounds into these boxes and their subsequent transport, deposition, and chemical processing. The chemistry is represented through a series of simultaneous ordinary differential equations, one for each compound. Given the difference in life-times between the chemical compounds (milli-seconds for O1D to years for CH4) these equations are numerically stiff and solving them consists of a significant fraction of the computational burden of a chemistry model. We have investigated a machine learning approach to emulate the chemistry instead of solving the differential equations numerically. From a one-month simulation of the GEOS-Chem model we have produced a training dataset consisting of the concentration of compounds before and after the differential equations are solved, together with some key physical parameters for every grid-box and time-step. From this dataset we have trained a machine learning algorithm (regression forest) to be able to predict the concentration of the compounds after the integration step based on the concentrations and physical state at the beginning of the time step. We have then included this algorithm back into the GEOS-Chem model, bypassing the need to integrate the chemistry. This machine learning approach shows many of the characteristics of the full simulation and has the potential to be substantially faster. There are a wide range of application for such an approach - generating boundary conditions, for use in air quality forecasts, chemical data assimilation systems, etc. We discuss speed and accuracy of our approach, and highlight some potential future directions for improving it.

Keller, Christoph A.↗

NASA NTRS2018

The Central Role of Air Quality Observations in NASA's GEOS Composition Forecasting Model

The NASA GEOS composition forecast model (GEOS-CF) provides global, high-resolution (25 km) air quality forecasts in near-real time. This system combines the operational GEOS-5 weather forecasting model with the state-of-the-science GEOS-Chem chemistry module to provide detailed chemical analysis of a wide range of air pollutants such as ozone, carbon monoxide, nitrogen oxides, and fine particulate matter (PM2.5). The resolution of the forecasts is the highest compared to current, publicly-available global composition forecasts.Air quality observations are an indispensable tool to evaluate the model's ability to capture the strong temporal and spatial gradients of air pollutants across the globe. We show how comparisons against near-real time observations available through OpenAQ (www.openaq.org) demonstrate the model's overall success in reproducing surface concentrations of ozone, nitrogen dioxide, and PM2.5. This analysis also helps identifying current limitations of the model, for example over South America. The model-observation mismatches are most likely caused by uncertainties in the emissions data. Using the example of Rio de Janeiro, we show how the model skill can be improved by using local, high-resolution emission inventories in combination with air quality data.

Keller, Christoph A.↗

NASA NTRS2018

Air Quality Forecasts Using the NASA GEOS Composition Model

No abstract available

Keller, Christoph A.↗

NASA NTRS2018

NASA's GEOS Composition Model Assessment of PM2.5 During Wildfires: Inferring the Impact of PM2.5 Exposure on Adverse Respiratory & Cardiovascular Conditions

Particulate matter pollution is a mixture of solid and liquid droplets floating in the air that can lead to reduced air quality and increased adverse health impact. Fine particulate matter (PM2.5) can be emitted into the air from anthropogenic sources such as the burning of fossil fuels, motor vehicles, and powerplant emissions. Exposure to PM2.5 can aggravate pre-existing respiratory and cardiovascular conditions. When PM2.5 is inhaled it can cause damage to the lungs such as reduced lung function and shortness of breath. After being inhaled PM2.5 can enter the bloodstream and cause harm to the heart. One major natural source of PM2.5 exposure is from wildfire smoke. The particulates within the smoke from the wildfires can spread from the initial source region, potentially impacting communities both near and far. During and after wildfire events, PM2.5 levels can exceed the WHO air quality guidelines (10 m.g/m^3 annual mean; 25 m.g/m^3 daily mean), becoming hazardous to an individual's health. Global models can be used to simulate the emission and transport of these particulates and subsequently they can be valuable to air quality forecasting in highly polluted areas. The NASA Goddard Earth Observing System (GEOS) Composition Forecast (GEOS-CF) system has been used to produce near-real time air quality forecasts of atmospheric composition at a high global resolution of 25x25 km2. The GEOS-CF system utilizes the GEOS weather forecast model coupled with GEOS-Chem (version 11) chemistry module to provide analyses and forecasts of various toxic air pollutants, including PM2.5. The GEOS-CF simulated high levels of PM2.5 (40 m.g/m^3 to 250 m.g/m^3 ), exceeding the WHO guidelines, during multiple recent regional and global wildfire seasons, including the 2017 Washington State and Northern and Southern California wildfire seasons. Furthermore, the GEOS-CF simulated PM2.5 applied to a human health assessment model, BenMAP (The Environmental Benefits Mapping and Analysis Program, version 1.3), estimates the impact on adverse respiratory health conditions due to PM2.5 exposure from wildfires. The GEOS-CF predicted PM2.5 during the wildfire season with the corresponding BenMAP results provides an assessment of the human health impact of PM2.5 exposure.

Saunders, Emily↗

NASA NTRS2018

Machine Learning Application to Atmospheric Chemistry Modeling

Atmospheric chemistry is a high-dimensionality, large-data problem and thus may be suited to machine-learning algorithms. We show here the potential of a random forest regression algorithm to replace the gas-phase chemistry solver in the GEOS-Chem chemistry model. In this proof-of-concept study, we used one month of model output to train random forest regression models to predict the concentrations of each long-lived chemical species after integration based upon the physical and chemical conditions before the chemical integration. The choice of prediction type has a strong impact on the skill of the regression model. We find best results from predicting the change in concentration for very long-lived species and the absolute concentration for shorter lived species. The skill of the machine learning algorithm is further improved by using a family approach for NO and NO2 rather than treating them independently.By replacing the numerical integrator with the random forest algorithm and running this model for one month, we find that the model is able to reproduce many of the features of the reference chemistry simulation. Replacing the integration methodology with a machine learning algorithm has the potential to be substantially faster. There are a wide range of applications for such an approach, e.g. to generate boundary conditions, for use in air quality forecasts or chemical data assimilation systems, etc.

Keller, Christoph A.↗

NASA NTRS2018

GEOS-CF Model Assessment of PM2.5 During Wildfires: Inferring the Impact of PM2.5 Exposure on Adverse Respiratory & Cardiovascular Conditions

Exposure to fine particulate matter (PM2.5) can aggravate pre-existing respiratory and cardiovascular conditions. When PM2.5 is inhaled it can cause damage to the lungs such as reduced lung function and shortness of breath. After being inhaled PM2.5 can enter the bloodstream and cause harm to the heart. One major natural source of PM2.5 exposure is from wildfire smoke. The particulates within the smoke from the wildfires can spread from the initial source region, potentially impacting communities both near and far. During and after wildfire events, PM2.5 levels can exceed the WHO air quality guidelines (10 gm^3 annual mean; 25 gm^3 daily mean), becoming hazardous to an individual's health. Global models can be used to simulate the emission and transport of these particulates and subsequently they can be valuable to air quality forecasting in highly polluted areas. The NASA Goddard Earth Observing System (GEOS) version 5 Composition Forecast (GEOS-CF) system has been used to produce near-real time air quality forecasts of atmospheric composition at a high global resolution of 25 km. The GEOS-CF system utilizes the GEOS weather forecast model coupled with GEOS-Chem (version 11) chemistry module to provide analyses and forecasts of various toxic air pollutants, including PM2.5. The GEOS-CF simulated high levels of PM2.5 (40 gm^3 to 250 gm^3 ), exceeding the WHO guidelines, during multiple recent regional and global wildfire seasons, including the 2017 Seattle, WA and Los Angeles, CA wildfire seasons, and biomass burning events in India. Furthermore, the GEOS-CF simulated PM2.5 applied to a human health assessment model, BenMAP (The Environmental Benefits Mapping and Analysis Program, version 1.3), estimates the impact on adverse respiratory health conditions due to PM2.5 exposure from wildfires. The GEOS-CF predicted PM2.5 during the wildfire season with the corresponding BenMAP results provides an assessment of the human health impact of PM2.5 exposure.

Saunders, Emily↗

NASA NTRS2018

Air Quality Forecasts Using the NASA GEOS Model

Knowland, K. Emma↗

NASA NTRS2018

Air Quality Modeling Using the NASA GEOS-5 Multispecies Data Assimilation System

The NASA Goddard Earth Observing System (GEOS) data assimilation system (DAS) has been expanded to include chemically reactive tropospheric trace gases including ozone (O3), nitrogen dioxide (NO2), and carbon monoxide (CO). This system combines model analyses from the GEOS-5 model with detailed atmospheric chemistry and observations from MLS (O3), OMI (O3 and NO2), and MOPITT (CO). We show results from a variety of assimilation test experiments, highlighting the improvements in the representation of model species concentrations by up to 50% compared to an assimilation-free control experiment. Taking into account the rapid chemical cycling of NO2 when applying the assimilation increments greatly improves assimilation skills for NO2 and provides large benefits for model concentrations near the surface. Analysis of the geospatial distribution of the assimilation increments suggest that the free-running model overestimates biomass burning emissions but underestimates lightning NOx emissions by 5-20%. We discuss the capability of the chemical data assimilation system to improve atmospheric composition forecasts through improved initial value and boundary condition inputs, particularly during air pollution events. We find that the current assimilation system meaningfully improves short-term forecasts (1-3 day). For longer-term forecasts more emphasis on updating the emissions instead of initial concentration fields is needed.

Keller, Christoph A.↗

NASA NTRS2018

Air Quality Forecasts Using the NASA GEOS Model

We provide an introduction to a new high-resolution (0.25 degree) global composition forecast produced by NASA's Global Modeling and Assimilation office. The NASA Goddard Earth Observing System version 5 (GEOS-5) model has been expanded to provide global near-real-time forecasts of atmospheric composition at a horizontal resolution of 0.25 degrees (~25 km). Previously, this combination of detailed chemistry and resolution was only provided by regional models. This system combines the operational GEOS-5 weather forecasting model with the state-of-the-science GEOS-Chem chemistry module (version 11) to provide detailed chemical analysis of a wide range of air pollutants such as ozone, carbon monoxide, nitrogen oxides, and fine particulate matter (PM2.5). The resolution of the forecasts is the highest resolution compared to current, publically-available global composition forecasts. Evaluation and validation of modeled trace gases and aerosols compared to surface and satellite observations will be presented for constituents relative to health air quality standards. Comparisons of modeled trace gases and aerosols against satellite observations show that the model produces realistic concentrations of atmospheric constituents in the free troposphere. Model comparisons against surface observations highlight the model's capability to capture the diurnal variability of air pollutants under a variety of meteorological conditions. The GEOS-5 composition forecasting system offers a new tool for scientists and the public health community, and is being developed jointly with several government and non-profit partners. Potential applications include air quality warnings, flight campaign planning and exposure studies using the archived analysis fields.

Keller, Christoph A.↗

NASA NTRS2018

Atmospheric Chemistry Modeling Using a Regression Forest Model

Atmospheric chemistry is central to many environmental issues such as air pollution, climate change, and stratospheric ozone loss. Chemistry Transport Models (CTM) are a central tool for understanding these issues, whether for research or for forecasting. These models split the atmosphere in a large number of grid-boxes and consider the emission of compounds into these boxes and their subsequent transport, deposition, and chemical processing. The chemistry is represented through a series of simultaneous ordinary differential equations, one for each compound. Given the difference in life-times between the chemical compounds (milli-seconds for O1D to years for CH4) these equations are numerically stiff and solving them consists of a significant fraction of the computational burden of a CTM. We have investigated a machine learning approach to solving the differential equations instead of solving them numerically. From an annual simulation of the GEOS-Chem model we have produced a training dataset consisting of the concentration of compounds before and after the differential equations are solved, together with some key physical parameters for every grid-box and time-step. From this dataset we have trained a machine learning algorithm (regression forest) to be able to predict the concentration of the compounds after the integration step based on the concentrations and physical state at the beginning of the time step. We have then included this algorithm back into the GEOS-Chem model, bypassing the need to integrate the chemistry. This machine learning approach shows many of the characteristics of the full simulation and has the potential to be substantially faster. There are a wide range of application for such an approach - generating boundary conditions, for use in air quality forecasts, chemical data assimilation systems, centennial scale climate simulations etc. We discuss our approches' speed and accuracy, and highlight some potential future directions for improving this approach.

Keller, Christoph A.↗