Development of UI-WRF-Chem (v1.0) for the MAIA satellite mission: case demonstration
Bibliographic record
Abstract
Abstract. The Multi-Angle Imager for Aerosols (MAIA) satellite mission, to be jointly implemented by NASA and the Italian Space Agency with an expected 2026 launch, aims to study how different types of particulate matter (PM) pollution affect human health. The investigation will primarily focus on a discrete set of globally distributed Primary Target Areas (PTAs) containing major metropolitan cities, and will integrate satellite observations, ground observations, and chemical transport model (CTM) outputs to generate maps of near-surface total and speciated PM within the PTAs. In addition, the MAIA investigation will provide satellite measurements of aerosols over a set of Secondary Target Areas (STAs), which are useful for studying air quality more broadly. For the CTM, we have developed a Unified Inputs (of initial and boundary conditions) for WRF-Chem (UI-WRF-Chem) modeling framework to support the MAIA satellite mission. These developments include: (1) application of NASA GEOS FP and MERRA-2 data to provide both meteorological and chemical initial and boundary conditions for performing WRF-Chem simulations of air quality at a fine spatial resolution for both forecast and reanalysis modes; (2) a stand-alone emission preprocessor that ingests both global and regional anthropogenic emission inventories as well as fire emissions; (3) application of MODIS land data to improve land surface properties such as land cover type; (4) application of GLDAS and NLDAS data to constrain surface soil properties such as soil moisture; (5) development of a new soil NOx emission scheme – the Berkeley Dalhousie Iowa Soil NO Parameterization (BDISNP). Here, we illustrate the model improvements because of these developments over four target areas: Beijing in China, CHN-Beijing (STA); Rome in Italy, ITA-Rome (PTA); Los Angeles in the U.S., USA-Angeles (PTA) and Atlanta in the U.S., USA-Atlanta (PTA). UI-WRF-Chem is configured as 2 nested domains using an outer domain (D1) and inner domain (D2) with a 12 km and 4 km spatial resolution, respectively. For each target area, we first run a suite of sensitivity simulations to test the model sensitivity to different options of physics schemes and then select the optimal combination of physics schemes based on evaluation of model simulated meteorology. For the inner domain (D2), we have chosen to turn off the traditional Grell 3D ensemble (G3D) cumulus scheme. We conduct a case study over USA-Atlanta for June 2022 to demonstrate the impacts of cumulus scheme on precipitation and subsequent surface PM2.5 concentration. Our results show that keeping the G3D cumulus scheme on results in higher precipitation and lower PM2.5 than the simulation with the G3D cumulus scheme off. Compared with surface observations of precipitation and PM2.5 concentration, the sensitivity simulation with the G3D scheme off shows better performance than keeping it on. We focus on two dust intrusion events over CHN-Beijing and ITA-Rome, which occurred in March 2018 and June 2023, respectively. We carry out a suite of sensitivity simulations using UI-WRF-Chem by excluding chemical boundary conditions or including MERRA-2 chemical boundary conditions. Our results show that using MERRA-2 data to provide chemical boundary conditions can help improve model simulation of surface PM concentration and AOD. Some of the target areas have also experienced significant changes in land cover and land use over the past decade. Our case study over CHN-Beijing in July 2018 investigates the impacts of improved land surface properties with timely MODIS land data on capturing the urban heat island phenomena. Model-simulated surface skin temperature shows better agreement with MODIS observed land surface temperature. The updated soil NOx emission scheme leads to higher NO2 vertical column density (VCD) in rural areas over CHN-Beijing target area, which matches better with TROPOMI observed NO2 VCD. This in turn affects the simulation of surface nitrate concentration. Lastly, we conduct a case study over USA-LosAngeles to tune the dust emissions. This gives an example to show the fine-tuning work we do over each target area to investigate the problem specific to that target area as we continue evaluating and improving model performance.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.004 | 0.003 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.004 | 0.001 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.010 | 0.005 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".