Dataset Information

A multiphase CMAQ version 5.0 adjoint.

ABSTRACT: We present the development of a multiphase adjoint for the Community Multiscale Air Quality (CMAQ) model, a widely used chemical transport model. The adjoint model provides location- and time-specific gradients that can be used in various applications such as backward sensitivity analysis, source attribution, optimal pollution control, data assimilation, and inverse modeling. The science processes of the CMAQ model include gas-phase chemistry, aerosol dynamics and thermodynamics, cloud chemistry and dynamics, diffusion, and advection. Discrete adjoints are implemented for all the science processes, with an additional continuous adjoint for advection. The development of discrete adjoints is assisted with algorithmic differentiation (AD) tools. Particularly, the Kinetic PreProcessor (KPP) is implemented for gas-phase and aqueous chemistry, and two different automatic differentiation tools are used for other processes such as clouds, aerosols, diffusion, and advection. The continuous adjoint of advection is developed manually. For adjoint validation, the brute-force or finite-difference method (FDM) is implemented process by process with box- or column-model simulations. Due to the inherent limitations of the FDM caused by numerical round-off errors, the complex variable method (CVM) is adopted where necessary. The adjoint model often shows better agreement with the CVM than with the FDM. The adjoints of all science processes compare favorably with the FDM and CVM. In an example application of the full multiphase adjoint model, we provide the first estimates of how emissions of particulate matter (PM_2.5) affect public health across the US.

SUBMITTER: Zhao S

PROVIDER: S-EPMC7745733 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:The Community Multiscale Air Quality (CMAQ) model is a comprehensive multipollutant air quality modeling system developed and maintained by the US Environmental Protection Agency's (EPA) Office of Research and Development (ORD). Recently, version 5.1 of the CMAQ model (v5.1) was released to the public, incorporating a large number of science updates and extended capabilities over the previous release version of the model (v5.0.2). These updates include the following: improvements in the meteorological calculations in both CMAQ and the Weather Research and Forecast (WRF) model used to provide meteorological fields to CMAQ, updates to the gas and aerosol chemistry, revisions to the calculations of clouds and photolysis, and improvements to the dry and wet deposition in the model. Sensitivity simulations isolating several of the major updates to the modeling system show that changes to the meteorological calculations result in enhanced afternoon and early evening mixing in the model, periods when the model historically underestimates mixing. This enhanced mixing results in higher ozone (O3) mixing ratios on average due to reduced NO titration, and lower fine particulate matter (PM2.5) concentrations due to greater dilution of primary pollutants (e.g., elemental and organic carbon). Updates to the clouds and photolysis calculations greatly improve consistency between the WRF and CMAQ models and result in generally higher O3 mixing ratios, primarily due to reduced cloudiness and attenuation of photolysis in the model. Updates to the aerosol chemistry result in higher secondary organic aerosol (SOA) concentrations in the summer, thereby reducing summertime PM2.5 bias (PM2.5 is typically underestimated by CMAQ in the summer), while updates to the gas chemistry result in slightly higher O3 and PM2.5 on average in January and July. Overall, the seasonal variation in simulated PM2.5 generally improves in CMAQv5.1 (when considering all model updates), as simulated PM2.5 concentrations decrease in the winter (when PM2.5 is generally overestimated by CMAQ) and increase in the summer (when PM2.5 is generally underestimated by CMAQ). Ozone mixing ratios are higher on average with v5.1 vs. v5.0.2, resulting in higher O3 mean bias, as O3 tends to be overestimated by CMAQ throughout most of the year (especially at locations where the observed O3 is low); however, O3 correlation is largely improved with v5.1. Sensitivity simulations for several hypothetical emission reduction scenarios show that v5.1 tends to be slightly more responsive to reductions in NO x (NO + NO2), VOC and SO x (SO2 + SO4) emissions than v5.0.2, representing an improvement as previous studies have shown CMAQ to underestimate the observed reduction in O3 due to large, widespread reductions in observed emissions.

Project description:Many scientific problems such as classifier training or medical image reconstruction can be expressed as minimization of differentiable real-valued cost functions and solved with iterative gradient-based methods. Adjoint algorithmic differentiation (AAD) enables automated computation of gradients of such cost functions implemented as computer programs. To backpropagate adjoint derivatives, excessive memory is potentially required to store the intermediate partial derivatives on a dedicated data structure, referred to as the "tape". Parallelization is difficult because threads need to synchronize their accesses during taping and backpropagation. This situation is aggravated for many-core architectures, such as Graphics Processing Units (GPUs), because of the large number of light-weight threads and the limited memory size in general as well as per thread. We show how these limitations can be mediated if the cost function is expressed using GPU-accelerated vector and matrix operations which are recognized as intrinsic functions by our AAD software. We compare this approach with naive and vectorized implementations for CPUs. We use four increasingly complex cost functions to evaluate the performance with respect to memory consumption and gradient computation times. Using vectorization, CPU and GPU memory consumption could be substantially reduced compared to the naive reference implementation, in some cases even by an order of complexity. The vectorization allowed usage of optimized parallel libraries during forward and reverse passes which resulted in high speedups for the vectorized CPU version compared to the naive reference implementation. The GPU version achieved an additional speedup of 7.5 ± 4.4, showing that the processing power of GPUs can be utilized for AAD using this concept. Furthermore, we show how this software can be systematically extended for more complex problems such as nonlinear absorption reconstruction for fluorescence-mediated tomography.

Dataset Information

A multiphase CMAQ version 5.0 adjoint.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets