Biomass burning (BB) plays a key role in the biosphere–atmosphere interaction. It is a major source of trace gases and aerosols that alters the atmosphere and the water cycle. Additionally, these emissions are often related to other detrimental impacts including biodiversity loss in fire-sensitive biomes, increase of respiratory diseases, and massive economic losses. BB emissions are used as inputs in models that estimate air quality and the effect of fires on Earth’s climate. Hence, an accurate estimation of BB emissions is paramount. While BB emissions spread over most of the global vegetated areas, the integration of orbital remote sensing and modelling is the most effective approach to estimate them from regional to global scales. BB emission estimation follows the relationship between burned biomass and the emission factor (EF - mass emitted of a given species, for example carbon dioxide, per mass of dry matter burned). The burned biomass can be estimated using two approaches: (i) based on the relationship among burned area, above-ground biomass, and combustion completeness; or (ii) based on fire radiative power (FRP), a quantitative measurement that is directly related to the rate of burned biomass and is estimated to each active fire detected by several orbital sensors such as the Moderate Resolution Imaging Spectroradiometer (MODIS) sensor. EF values, which are Land Use and Land Cover (LULC) based, are required to estimate BB emissions independently on the approach adopted to estimate the burned biomass. Although novel approaches to improve the accuracy of BB emissions have been developed, the impact of EF values on the final estimated emissions remains uncertain. We have evaluated the impact of the EFs on the final estimate of fine particulate matter (PM2.5) emitted from BB in the Brazilian Amazon during a nineteen years’ time series (2002-2020) by running the PREP-CHEM-SRC emissions preprocessor tool under four EF scenarios: the tool original EF values based on the work of Andreae and Merlet (2001), the average EF values recently updated by Andreae (2019), and the minimum and maximum EF values also proposed by this author. The minimum (maximum) EF values were defined as the average EF value for each LULC class minus (plus) one standard deviation. The PM2.5 emissions were estimated at the spatial resolution of 0.1º using the FRP approach implemented on PREP-CHEM-SRC (3BEM_FRP model) having MODIS active fires as input, since this approach requires fewer inputs and the impact of the EFs over the emissions would be more evident. Our results showed that the annual average PM2.5 emission in the Amazon varied by 163% between the four EF scenarios (from1,426 Gg and 3,747 Gg), while the scenario based on the average values was the closest to the one based on PREP-CHEM-SRC original EF values (2,582 Gg and 2,213 Gg, respectively – an increase of 17%). These results contribute to the better understanding of how this single parameter impacts on the estimation of BB emissions.

A physics suite, which includes the aerosol-aware double moment Thompson-Eidhammer microphysics scheme (TH-E MP), the scale-aware and aerosol-aware Community Convective Cloud (C3) parameterization, and the MYNN-EDMF boundary layer and shallow cloud scheme, is under development at NOAA. We recently implemented a simple approach to improve the aerosol representation in the UFS. Sea salt, dust, biomass burning, and anthropogenic aerosol emissions have been embedded as CCPP-compliant subroutines. The prognostic emissions of sea-salt, and organic carbon are combined to represent the “water friendly” aerosol emission, while the prognostic emissions of dust are used to represent “ice friendly” aerosol emission for TH-E MP. With this implementation, we previously examined the aerosol indirect feedback when using the TH-E scheme in the global UFS forecast with C768 (~13km) horizontal resolution and 127 vertical levels. There are significant cloud-radiation responses to the aerosol differences, and the severely positive precipitation bias over Europe and North America was significantly alleviated when applying this aerosol emission method for indirect feedback. Here we add the indirect feedback using the C3 convective parameterization. C3 is a new collaborative development which adds several features from the currently operational SAS scheme to the Grell-Freitas parameterization. This study indicates that aerosol-physics interactions using a very simple and computationally highly efficient approach have significant impacts on the numerical weather prediction in the global UFS applications.

Fires drive global ecosystem change, impacting carbon dynamics, atmospheric composition, biodiversity, and human well-being. Biomass burning, a major outcome of fires, significantly contributes to greenhouse gas and aerosol emissions. Among these, fine particulate matter (PM2.5) is particularly concerning due to its adverse effects on air quality and health, and its substantial yet uncertain role in Earth’s energy balance. Variability in emission factors (EFs) remains a key source of uncertainty in emission estimates. This study evaluates PM2.5 emission sensitivity to EFs variability in Brazil’s Amazon and Cerrado biomes over 2002–2023 using the 3BEM_FRP model implemented in the PREP-CHEM-SRC tool. We updated the EFs with values and uncertainty ranges from Andreae (2019), which reflect a more comprehensive literature review than earlier datasets. The results reveal that the annual average PM2.5 emissions varied by up to 162% in the Amazon (1213 Gg yr−1 to 3172 Gg yr−1) and 184% in the Cerrado (601 Gg yr−1 to 1709 Gg yr−1). The Average peak emissions at the grid-cell level reached 5688 Mg yr−1 in the “Arc of Deforestation” region under the High-end EF scenario. Notably, the PM2.5 emissions from Amazon forest areas increased over time despite shrinking forest cover, indicating that Amazonian forests are becoming more vulnerable to fire. In the Cerrado, savannas are the primary land cover contributing to the total PM2.5 emissions, accounting for 64% to 80%. These findings underscore the importance of accurate, region-specific EFs for improving emission models and reducing uncertainties.

Abstract When the negatively buoyant air in the cloud downdrafts reaches the surface, it spreads out horizontally, producing cold pools. A cold pool can trigger new convective cells. However, when combined with the ambient vertical wind shear, it can also connect and upscale them into large mesoscale convective systems (MCS). Given the broad spectrum of scales of the atmospheric phenomenon involving the interaction between cold pools and the MCS, a parameterization was designed here. Then, it is coupled with a classical convection parameterization to be applied in an atmospheric model with an insufficient spatial resolution to explicitly resolve convection and the sub‐cloud layer. A new scalar quantity related to the deficit of moist static energy detrained by the downdrafts mass flux is proposed. This quantity is subject to grid‐scale advection, mixing, and a sink term representing dissipation processes. The model is then applied to simulate moist convection development over a large portion of tropical land in the Amazon Basin in a wet and dry‐to‐wet 10‐days period. Our results show that the cold pool edge parameterization improves the organization, longevity, propagation, and severity of simulated MCS over the Amazon and other different continental areas.

Background The record number of wildfires in the United States in recent years has led to an increased focus on developing tools to accurately forecast their impacts at high spatial and temporal resolutions. Aims The Warn-on-Forecast System for Smoke (WoFS-Smoke) was developed to improve these forecasts using wildfire properties retrieved from satellites to generate smoke plumes in the system. Methods The WoFS is a regional domain ensemble data assimilation and forecasting system built around the concept of creating short-term (0–6 h) forecasts of high impact weather. This work extends WoFS-Smoke by ingesting data from the GOES-16 satellite at 15-min intervals to sample the rapidly changing conditions associated with wildfires. Key results Comparison of experiments with and without GOES-16 data show that ingesting high temporal frequency data allows for wildfires to be initiated in the model earlier, leading to improved smoke forecasts during their early phases. Decreasing smoke plume intensity associated with weakening fires was also better forecast. Conclusions The results were consistent for a large fire near Boulder, Colorado and a multi-fire event in Texas, Oklahoma, and Arkansas, indicating a broad applicability of this system. Implications The development of WoFS-Smoke using geostationary satellite data allows for a significant advancement in smoke forecasting and its downstream impacts such as reductions in air quality, visibility, and potentially properties of severe convection.

Plume height plays a vital role in wildfire smoke dispersion and the subsequent effects on air quality and human health. In this study, we assess the impact of different plume rise schemes on predicting the dispersion of wildfire air pollution and the exceedances of the National Ambient Air Quality Standards (NAAQS) for fine particulate matter (PM2.5) during the 2020 western United States wildfire season. Three widely used plume rise schemes (Briggs, 1969; Freitas et al., 2007; Sofiev et al., 2012) are compared within the Community Multiscale Air Quality (CMAQ) modeling framework. The plume heights simulated by these schemes are comparable to the aerosol height observed by the Multi-angle Imaging SpectroRadiometer (MISR) and Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations (CALIPSO). The performance of the simulations with these schemes varies by fire case and weather conditions. On average, simulations with higher plume injection heights predict lower aerosol optical depth (AOD) and surface PM2.5 concentrations near the source region but higher AOD and PM2.5 in downwind regions due to the faster spread of the smoke plume once ejected. The 2-month mean AOD difference caused by different plume rise schemes is approximately 20 %-30 % near the source regions and 5 %-10 % in the downwind regions. Thick smoke blocks sunlight and suppresses photochemical reactions in areas with high AOD. The surface PM2.5 difference reaches 70 % on the West Coast of the USA, and the difference is lower than 15 % in the downwind regions. Moreover, the plume injection height affects pollution exceedance (> 35 mu gm(-3)) predictions. Higher plume heights generally produce larger downwind PM2.5 exceedance areas. The PM2.5 exceedance areas predicted by the three schemes largely overlap, suggesting that all schemes perform similarly during large wildfire events when the predicted concentrations are well above the exceedance threshold. At the edges of the smoke plumes, however, there are noticeable differences in the PM2.5 concentration and predicted PM2.5 exceedance region. For the whole period of study, the difference in the total number of exceedance days could be as large as 20 d in northern California and 4 d in the downwind regions. This disagreement among the PM2.5 exceedance forecasts may affect key decision-making regarding early warning of extreme air pollution episodes at local levels during large wildfire events.

<p>Significant problems in numerical weather prediction modeling systems appear when the horizontal grid-spacing is between 20 km and 1 km and when deep convection is important. These scales are usually termed “Gray Scales”.  Techniques have been developed so that the behavior of the convective parameterization changes with the horizontal grid spacing of the model; such parameterizations are said to become “scale-aware”. Commonly used techniques involve applying a scaling approach to smoothly transition from parameterized to resolved convection. These are similar to an elegantly simple mathematical method originally developed by Arakawa et al. (2011), which scales the convective tendencies in dependence on the convective area fraction. Here we show that the scaling approaches are flawed, since they fail to consider the fact that the impacts of deep convection on those scales are not limited to one grid box, and usually – because of the scaling -- leads to light precipitation covering too much area, as we have previously shown in HRRR simulations. Any scaling approach is especially flawed in areas of light forcing (such as daytime heating) and in the tropics, when the explicit microphysics parameterization is not yet producing precipitation. We will show examples of these problems and discuss possible solutions as applied to NOAA’s new RRFS storm-scale modeling system.</p>

Biomass burning (BB) emissions negatively impact the biosphere and human lives. Orbital remote sensing and modelling are used to estimate BB emissions on regional to global scales, but these estimates are subject to errors related to the parameters, data, and methods available. For example, emission factors (mass emitted by species during BB per mass of dry matter burned) are based on land use and land cover (LULC) classifications that vary considerably across products. In this work, we evaluate how BB emissions vary in the PREP-CHEM-SRC emission estimator tool (version 1.8.3) when it is run with original LULC data from MDC12Q1 (collection 5.1) and newer LULC data from MapBiomas (collection 6.0). We compare the results using both datasets in the Brazilian Amazon and Cerrado biomes during the 2002–2020 time series. A major reallocation of emissions occurs within Brazil when using the MapBiomas product, with emissions decreasing by 788 Gg (−1.91% year−1) in the Amazon and emissions increasing by 371 Gg (2.44% year−1) in the Cerrado. The differences identified are mostly associated with the better capture of the deforestation process in the Amazon and forest formations in Northern Cerrado with the MapBiomas product, as emissions in forest-related LULCs decreased by 5260 Gg in the Amazon biome and increased by 1676 Gg in the Cerrado biome. This is an important improvement to PREP-CHEM-SRC, which could be considered the tool to build South America’s official BB emission inventory and to provide a basis for setting emission reduction targets and assessing the effectiveness of mitigation strategies.

Smoke from the 2018 Camp Fire in Northern California blanketed a large part of the region for two weeks, creating poor air quality in the “unhealthy” range for millions of people. The NOAA Global System Laboratory’s HRRR-Smoke model was operating experimentally in real time during the Camp Fire. Here, output from the HRRR-Smoke model is compared to surface observations of PM2.5 from AQS and PurpleAir sensors as well as satellite observation data. The HRRR-Smoke model grid at 3-km resolution successfully simulated the evolution of the plume during the initial phase of the fire (8-10 November 2018). Stereoscopic satellite plume height retrievals were used to compare with model output (for the first time, to the authors’ knowledge), showing that HRRR-Smoke is able to represent the complex 3D distribution of the smoke plume over complex terrain. On 15-16 November, HRRR-Smoke was able to capture the intensification of PM2.5 pollution due to a high pressure system and subsidence that trapped smoke close to the surface; however, HRRR-Smoke later underpredicted PM2.5 levels due to likely underestimates of the fire radiative power (FRP) derived from satellite observations. The intensity of the Camp Fire smoke event and the resulting pollution during the stagnation episodes make it an excellent test case for HRRR-Smoke in predicting PM2.5 levels, which were so high from this single fire event that the usual anthropogenic pollution sources became insignificant. The HRRR-Smoke model was implemented operationally at NOAA/NCEP in December 2020, now providing essential support for smoke forecasting as the impact of US wildfires continues to increase in scope and magnitude.

This paper summarizes the current challenges in climate and weather research and provides suggestions for future research directions in global observing systems, in modelling and prediction, and in academic environment and education systems.

This study aims to assess the contribution to prediction performance of dynamically downscaling seasonal precipitation for the Brazilian Northeast region (NEB) produced in February with a global climate model for the austral autumn (March-April-May). This season includes the peak precipitation period of a large portion of this region. To achieve this aim, 32 years of hindcasts for the period 1979–2010 were produced with a regional model known as the Brazilian developments on the Regional Atmospheric Modeling System (BRAMS) run at 30 km spatial resolution, which was unidirectional nested to a global model run at approximately 200 km spatial resolution. The used global model was the Atmospheric General Circulation Models of the Brazilian Center for Weather Forecast and Climate Studies (CPTEC/AGCM), which provided initial and lateral boundary conditions for BRAMS. Hindcast ensembles of 10 members were produced with both the global and the regional models, simulating an operational prediction environment using persisted sea surface temperature (SST) anomalies observed in the previous January of each of the 32 hindcast years as oceanic boundary conditions. A comparative deterministic and probabilistic performance assessment of the two models was conducted. The deterministic assessment included the investigation of the mean error (bias) of the produced seasonal precipitation hindcasts, the degree of linear association between the observed and predicted precipitation anomalies measuring phase error through the correlation of the ensemble means with the corresponding observations, and amplitude error investigated through the ratio between the hindcast ensemble mean and observed variances. The probabilistic assessment was performed investigation the reliability, resolution, and discrimination characteristics of tercile probability predictions for the above normal and below normal categories, and also for probabilistic predictions for the occurrence of the symmetric event positive (or negative) precipitation anomaly. The two investigated modeling approaches showed similar performance, with an indication for dynamically downscaled predictions to have reduced amplitude error and slightly improved reliability and resolution compared to the predictions produced with the global model.

Over the last decades, several real-time smoke prediction systems have been developed worldwide for air quality forecast to support decision making to control and manage anthropogenic pollutantions and smoke impacts. In Portugal, smoke modelling, as well as air quality forecast, has been developed by the research group GEMAC at the University of Aveiro, Portugal. However, the current forecast system does not integrate wildfire emissions. The ability from modelling to predict the behaviour of fire smoke in rural areas is an effective way to improve the efficiency of air quality and to prevent public health consequences. From 1980 to 2017, 4.4 Mha of cumulative burned area in rural fires, accounting for roughly half of Portugal's continental area, causing damage to infrastructure and lives. Hence, the forecast of smoke emissions has become of vital importance. There is a wide variety of models available to simulate fire-smoke phenomena. Nonetheless, it is necessary to consider computational aspects, resources, and goals to choose a suitable model to fit the purpose. In the ongoing FIRESMOKE project, developed by the GEMAC research group from University of Aveiro in Portugal, and GMAI from the Center for Weather Forecasting and Climate Studies in Brazil, the meteorological weather forecast model BRAMS-SFIRE is implemented to be part of the new version 5.6.2 of BRAMS. BRAMS-SFIRE model was coupled to simulate a broad integration between the surface fire fluxes and the atmospheric environment and presented a good accuracy in terms of the physics of the atmosphere and fire interaction. This project aims to improve the SFIRE model to include the crown fire behaviour. The goal is the incorporation of some formulations of the “crown fire potential” by linking models of surface and crown fire behaviour from Scott and Reinhardt (2001) and injecting fire smoke into the chemistry module of BRAMS. These developments are part of a whole system for forecasting and monitoring forest fire smoke emissions that incorporates other anthropogenic and biogenic sources of air pollution, to provide a public access service of atmospheric scope, over the domain of continental Portugal.

Large land extensions are subjected to environmental degradation and land-use changes by fires annually. In tropical regions, such as South America, the global demand for commodities leads to the conversion of natural vegetation into agricultural land-uses. With the new orbital fire products that have improved in spatial and temporal resolution, it is now possible to better understand fire properties at large scales, such as fire radiative power (FRP), fire spread, heat flux, and fire life cycle. This study aims to integrate polar-orbit and geostationary satellites' fire-related products to estimate biomass burning (BB) emissions on a continental-scale to better monitor smoke plumes in near-real-time (NRT) while using the Brazilian developments on the Regional Atmospheric Modeling System (BRAMS). The total aerosol optical depth at 550 nm channel from the model during the South American 2020 fire season (September-November) showed a good agreement with Modern-Era Retrospective analysis for Research and Applications (MERRA-2) and Copernicus Atmosphere Monitoring Service (CAMS) Global Near-Real-Time reanalyzes (R = 0.97, p < 0.05, student t-test). Moreover, the new improvements made in the Brazilian Biomass Burning Emission Model with fire radiative power (3BEM-FRP) better represent smoke emissions of large fire events, such as the Pantanal 2020 fires.

A new open biomass burning inventory is presented that relies on the fire radiative power data from the Visible Infrared Imaging Radiometer Suite (VIIRS) on board the Suomi NPP satellite. This VIIRS-based Fire Emission Inventory (VFEI) provides emission data from early 2012 to 2019 for more than 40 species of gases and aerosols at spatial resolutions of around 500 m. We found that VFEI produces similar results when compared to other major inventories in many regions of the world. Additionally, we conducted regional simulations using VFEI with the Weather Research and Forecasting (WRF) model with chemistry (WRF-Chem) for Southern Africa (September 2016) and North America (July–August 2019). We compared aerosol optical depth (AOD) from the model against two observational datasets: the MODIS Multi-Angle Implementation of Atmospheric Correction (MAIAC) product and AErosol RObotic NETwork (AERONET) stations. Results showed good agreement between both simulations and the datasets, with mean AOD biases of around +0.03 for Southern Africa and −0.01 for North America. Both simulations were not only able to reproduce the AOD magnitudes accurately, but also the inter-diurnal variations of smoke concentration. In addition, we made use of the airborne data from the ObseRvations of Aerosols above CLouds and their intEractionS (ORACLES; Southern Africa) and the Fire Influence on Regional to Global Environments Experiment and Air Quality (FIREX-AQ; North America) campaigns to evaluate the simulations. In Southern Africa, results showed correlations higher than 0.77 when comparing carbon monoxide and black carbon. In North America, correlations were lower and biases higher. However, this is because the model was not able to reproduce the timing, shape, and location of individual plumes over complex terrain (Rocky Mountains) during the FIREX-AQ campaign period.

Convective parameterization is the long-lasting bottleneck of global climate modelling and one of the most difficult problems in atmospheric sciences. Uncertainty in convective parameterization is the leading cause of the widespread climate sensitivity in IPCC global warming projections. This paper reviews the observations and parameterizations of atmospheric convection with emphasis on the cloud structure, bulk effects, and closure assumption. The representative state-of-the-art convection schemes are presented, including the ECMWF convection scheme, the Grell scheme used in NCEP model and WRF model, the Zhang-MacFarlane scheme used in NCAR and DOE models, and parameterizations of shallow moist convection. The observed convection has self-suppression mechanisms caused by entrainment in convective updrafts, surface cold pool generated by unsaturated convective downdrafts, and warm and dry lower troposphere created by mesoscale downdrafts. The post-convection environment is often characterized by "diamond sounding" suggesting an over-stabilization rather than barely returning to neutral state. Then the pre-convection environment is characterized by slow moistening of lower troposphere triggered by surface moisture convergence and other mechanisms. The over-stabilization and slow moistening make the convection events episodic and decouple the middle/upper troposphere from the boundary layer, making the state-type quasi-equilibrium hypothesis invalid. Right now, unsaturated convective downdrafts and especially mesoscale downdrafts are missing in most convection schemes, while some schemes are using undiluted convective updrafts, all of which favour easily turned-on convection linked to double-ITCZ (inter-tropical convergence zone), overly weak MJO (Madden-Julian Oscillation) and precocious diurnal precipitation maximum. We propose a new strategy for convection scheme development using reanalysis-driven model experiments such as the assimilation runs in weather prediction centres and the decadal prediction runs in climate modelling centres, aided by satellite simulators evaluating key characteristics such as the lifecycle of convective cloud-top distribution and stratiform precipitation fraction.

We use the Weather Research and Forecasting with Chemistry (WRF-Chem) model with new implementations of GOES-16 wildfire emissions and plume rise based on fire radiative power (FRP) to interpret aerosol observations during the 2019 NASA-NOAA FIREX-AQ field campaign and perform model evaluations. We compare simulated aerosol concentrations and optical properties against observations of black carbon aerosol from the NOAA Single Particle Soot Photometer (NOAA-SP2), organic aerosol from the CU High-Resolution Aerosol Mass Spectrometer (HR-AMS), and aerosol backscatter coefficients from the high-spectral-resolution lidar (HSRL) system. This study focuses on the Williams Flats fire in Washington, which was repeatedly sampled during four science flights by the NASA DC-8 (3–8 August 2019). The emissions and plume-rise methodologies are implemented following NOAA's operational High-Resolution Rapid Refresh coupled with Smoke (HRRR-Smoke) forecasting model. In addition, new GOES-16 FRP-based diurnal cycle functions are developed and incorporated into WRF-Chem. The FIREX-AQ observations represented a diverse set of sampled environments ranging from fresh/aged smoke from the Williams Flats fire to remnants of plumes transported over long distances. The Williams Flats fire resulted in significant aerosol enhancements during 3–8 August 2019, which were substantially underestimated by the standard version of WRF-Chem. The simulated black carbon (BC) and organic carbon (OC) concentrations increased between a factor of 92–125 (BC) and a factor of 28–78 (OC) with the new implementation compared to the standard WRF-Chem version. These increases resulted in better agreement with the FIREX-AQ airborne observations for BC and OC concentrations (particularly for fresh smoke sampling phases) and aerosol backscatter coefficients. The model still showed a low bias in simulating the aerosol loadings observed in aged plumes from Williams Flats. WRF-Chem with the FRP-based plume rise simulated similar plume heights to the standard plume-rise model in WRF-Chem. The simulated plume heights (for both versions) compared well with estimated plume heights using the HSRL measurements. Therefore, the better agreement with observations was mainly driven by the higher emissions in the FRP-based version. The model evaluations also highlighted the importance of accurately accounting for the wildfire diurnal cycle and including adequate representation of the underlying chemical mechanisms, both of which could significantly impact model forecasting performance.

This research begins the process of creating an ensemble-based forecast system for smoke aerosols generated from wildfires using a modified version of the National Severe Storms Laboratory (NSSL) Warn-on-Forecast System (WoFS). The existing WoFS has proven effective in generating short-term (0-3 h) probabilistic forecasts of high-impact weather events such as storm rotation, hail, severe winds, and heavy rainfall. However, it does not include any information on large smoke plumes generated from wildfires that impact air quality and the surrounding environment. The prototype WoFS-Smoke system is based on the deterministic High-Resolution Rapid Refresh-Smoke (HRRR-Smoke) model. HRRR-Smoke runs over a continental United States (CONUS) domain with a 3-km horizontal grid spacing, with hourly forecasts out to 48 h. The smoke plume injection algorithm in HRRR-Smoke is integrated into the WoFS forming WOFS-Smoke so that the advantages of the rapidly cycling, ensemble-based WoFS can be used to generate short-term (0-3 h) probabilistic forecasts of smoke. WoFS-Smoke forecasts from three wildfire cases during 2020 show that the system generates a realistic representation of wildfire smoke when compared against satellite observations. Comparison of smoke forecasts with radar data show that forecast smoke reaches higher levels than radar-detected debris, but exceptions to this are noted. The radiative effect of smoke on surface temperature forecasts is evident, which reduces forecast errors compared to experiments that do not include smoke.

Injections of wildfire smoke plumes into the free troposphere impact air quality, yet model forecasts of injections are poor. Here, we use aircraft observations obtained during the 2019 western US wildfires (FIREX-AQ) to evaluate a commonly used smoke plume rise parameterization in two atmospheric chemistry-transport models (WRF-Chem and HRRR-Smoke). Observations show that smoke injections into the free troposphere occur in 35% of plumes, whereas the models forecast 59–95% indicating false injections in the simulations. False injections were associated with both models overestimating fire heat flux and terrain height, and with WRF-Chem underestimating planetary boundary layer height. We estimate that the radiant fraction of heat flux is 0.5 to 25 times larger in models than in observations, depending on fuel type. Model performance was substantially improved by using observed heat flux and boundary layer heights, confirming that models need accurate heat fluxes and boundary layer heights to correctly forecast plume injections.