Improving dust simulations in WRF-Chem v4.1.3 coupled with the GOCART aerosol module

Abstract

Abstract. In this paper, we rectify inconsistencies that emerge in the Weather Research and Forecasting model with chemistry (WRF-Chem) v3.2 code when using the Goddard Chemistry Aerosol Radiation and Transport (GOCART) aerosol module. These inconsistencies have been reported, and corrections have been implemented in WRF-Chem v4.1.3. Here, we use a WRF-Chem experimental setup configured over the Middle East (ME) to estimate the effects of these inconsistencies. Firstly, we show that the old version underestimates the PM2.5 diagnostic output by 7 % and overestimates PM10 by 5 % in comparison with the corrected one. Secondly, we demonstrate that submicron dust particles' contribution was incorrectly accounted for in the calculation of optical properties. Therefore, aerosol optical depth (AOD) in the old version was 25 %–30 % less than in the corrected one. Thirdly, we show that the gravitational settling procedure, in comparison with the corrected version, caused higher dust column loadings by 4 %–6 %, PM10 surface concentrations by 2 %–4 %, and mass of the gravitationally settled dust by 5 %–10 %. The cumulative effect of the found inconsistencies led to the significantly higher dust content in the atmosphere in comparison with the corrected WRF-Chem version. Our results explain why in many WRF-Chem simulations PM10 concentrations were exaggerated. We present the methodology for calculating diagnostics we used to estimate the impacts of introduced code modifications. We share the developed Merra2BC interpolator, which allows processing Modern-Era Retrospective Analysis for Research and Applications, version 2 (MERRA-2) output for constructing initial and boundary conditions for chemical species and aerosols.

How Does a Pinatubo‐Size Volcanic Cloud Reach the Middle Stratosphere?

“…The modification of the gravitational settling numerical scheme is discussed in detail in Ukhov et al. (2021).…”

Effects of Variable Eruption Source Parameters on Volcanic Plume Transport: Example of the 23 November 2013 Paroxysm of Etna

“…In Figure 9, AI is visualized from Level-2 Swath and 50 × 50 km resolution, over 1 h time coverage around 12:00 UTC on 23 November 2013. A modeled quantity that could be qualitatively compared with the OMPS/AI is the columnar mass density, which is obtained by considering in each (x, y) grid point the following columnar quantity [53]:…”

Evaluation of Nine Operational Models in Forecasting Different Types of Synoptic Dust Events in the Middle East

“…These results agree with the model's performance against MODIS observations, indicating the positive feedback of data assimilation in the forecasting process. On the other hand, the AOD forecasts from other models such as EMA-RegCM4 and NMMB-BSC highly departed from the AERONET measurements (Table 3), justifying the complexity of various parameters such as size distribution, in model AOD estimates [56].…”

“…That study showed that the CAMS model represented better the AOD values, likely related to MODIS data assimilation, although all models underestimated the high AODs compared to Terra-MODIS retrievals, especially near the dust source. There are some difficulties in calculating AOD by dust forecasting models, with the uncertainties mostly related to the complex refractive index (scattering and absorbing properties of dust), as well as to the size distribution (relative partitioning of the fine and coarse modes) [56]. So, for satisfactory predictions of the dust-AOD, scattering, absorption processes and particle size distribution should be accurately parameterized in the model, as they have different responses to AOD values.…”

“…So, for satisfactory predictions of the dust-AOD, scattering, absorption processes and particle size distribution should be accurately parameterized in the model, as they have different responses to AOD values. Ukhov et al [56] revealed this challenging issue and the effect of fine particles in dust AOD calculations for the WRF-Chem model. Huneeus et al [28] compared the outputs of 15 global aerosol models in the AeroCom project.…”