Development and application of the WRFPLUS-Chem online chemistry adjoint and WRFDA-Chem assimilation system

Abstract

Abstract. Here we present the online meteorology and chemistry adjoint and tangent linear model, WRFPLUS-Chem, which incorporates modules to treat boundary layer mixing, emission, aging, dry deposition, and advection of black carbon aerosol. We also develop land surface and surface layer adjoints to account for coupling between radiation and vertical mixing. Model performance is verified against finite difference derivative approximations. A second order checkpointing scheme is created to reduce computational costs and enable simulations longer than six hours. The adjoint is coupled to WRFDA-Chem, in order to conduct a sensitivity study of anthropogenic and biomass burning sources throughout California during the 2008 Arctic Research of the Composition of the Troposphere from Aircraft and Satellites (ARCTAS) field campaign. A cost function weighting scheme was devised to increase adjoint sensitivity robustness in future inverse modeling studies. Results of the sensitivity study show that, for this domain and time period, anthropogenic emissions are over predicted, while wildfire emissions are under predicted. We consider the diurnal variation in emission sensitivities to determine at what time sources should be scaled up or down. Also, adjoint sensitivities for two choices of land surface model indicate that emission inversion results would be sensitive to forward model configuration. The tools described here are the first step in conducting four-dimensional variational data assimilation in a coupled meteorology-chemistry model, which will potentially provide new constraints on aerosol precursor emissions and their distributions. Such analyses will be invaluable to assessments of particulate matter health and climate impacts.

Wildfire Detection Probability of MODIS Fire Products under the Constraint of Environmental Factors: A Study Based on Confirmed Ground Wildfire Records

“…It highlights the necessity of the serious attention to local environmental conditions as well as the cooperation with ground inspection and patrols for instantaneous wildfire assessment and design of fire prevention policies and measures in a region. Additionally, the MODIS fire products are widely used as fire-location inputs for various dynamic global vegetation models (DGVMs) and general circulation models (GCMs) while the fire related uncertainties have already been discussed [68,69]. For example, Veira et al [69] applied MOD14A1 data to the GCM ECHAM6-HAM2 to simulate global patterns in the wildfire-caused emissions, and found that important omission biases may be introduced by the uncertainties in MODIS fire products.…”

WRF‐TEB: Implementation and Evaluation of the Coupled Weather Research and Forecasting (WRF) and Town Energy Balance (TEB) Model

“…WRF‐TEB is developed using WRF‐CMake version 4.1.5 (Riechert & Meyer, 2019a, 2020; Table 1) as it adds CMake (Kitware Inc., 2019a) support to the latest versions of WRF to simplify the configuration and build process of WRF and WPS (WRF Preprocessing System). Although, WRF‐CMake version 4.1.5 does not include support for WRF‐Chem (Grell et al., 2005), WRF‐DA (Huang et al., 2009), WRFPLUS (Guerrette & Henze, 2015), or WRF‐Hydro (Gochis et al., 2018), its benefits may outweigh these limitations to model developers, code maintainers, and end‐users wishing to build WRF, as it includes: robust incremental rebuilds, dependency analysis of Fortran code, flexible library dependency discovery, integrated support for shared (Open Multi‐Processing; OpenMP) and distributed (Message Passing Interface; MPI) memory, support for automated testing using continuous integration (CI), and availability of experimental prebuilt binary releases for Linux, macOS, and Windows from the project's GitHub page or through the integration with GIS4WRF (Meyer & Riechert, 2019a), a QGIS (QGIS Development Team, 2019) toolkit for preprocessing and postprocessing, visualizing, and running simulations in WRF. Here we refer to both the physical model and the software (i.e., WRF‐CMake) as WRF, unless highlighting specific software features.…”

Source Attribution of Societal Impacts of PM2.5 Pollution from Regional to Hemispheric Scales

“…Adjoint models trace back influences (sensitivities) from a predefined receptor area to each source location at preceding timesteps, backward in time. While adjoint versions have been developed for a number of CTMs (Elbern and Schmidt, 1999;Guerrette and Henze, 2015;Hakami et al, 2005;Hakami et al, 2007;Henze et al, 2007;Martien et al, 2006;Mallet and Sportisse, 2005;Menut et al, 2000;Menut, 2003;Sandu et al, 2005;Zhao et al, 2020), most health-based applications of the adjoint method have been limited to GEOS-Chem and CMAQ models Hakami, 2013a, 2013b;Koo et al, 2013;Pappin et al, 2015Nawaz et al, 2021;Qu et al, 2022). GEOS-Chem is a global model, and as such most of the health-based applications of its adjoint have been at the global scale, while higher-resolution GEOS-Chem simulations have been applied at regional-to-continental scales in the U.S. and elsewhere in the world.…”