Recommendations on statistics and benchmarks to assess photochemical model performance

Emery, Chris; Liu, Zhen; Russell, Armistead G.; Odman, M. Talat; Yarwood, Greg; Kumar, Naresh

doi:10.1080/10962247.2016.1265027

Cited by 431 publications

(316 citation statements)

References 57 publications

Supporting

Mentioning

284

Contrasting

Unclassified

Order By: Relevance

“…While daytime ozone statistics were fairly good domain wide and along the western coast of Lake Michigan, IOA 97.4% and 79.7%, respectively, ozone was biased low in the coastal west region (−4.4 ppb, or a normalized mean bias of −8.58%). This is within the ±15% criteria recommended in Emery et al (). But more importantly, peak ozone concentrations are underpredicted during the high ozone events on 2 June, 10–13 June, and 15–17 June 2017.…”

Section: Discussionsupporting

confidence: 85%

Sensitivity of Meteorological Skill to Selection of WRF‐Chem Physical Parameterizations and Impact on Ozone Prediction During the Lake Michigan Ozone Study (LMOS)

et al. 2020

View full text Add to dashboard Cite

Ozone concentrations in excess of health‐based standards occur along the coastline of Lake Michigan. A complex pattern of ozone precursor emissions interfaces with a complex meteorological environment, presenting a challenge for air quality management and simulation. Precursors are transported into a shallow, stable boundary layer over the lake. This is followed by ozone formation and transport back onshore through a combination of synoptic and lake breeze winds. In this study, we use measurements during the Lake Michigan Ozone Study 2017 (LMOS) to quantitatively evaluate the Weather Research and Forecasting with Chemistry (WRF‐Chem) model at 4 km horizontal resolution for key features of high ozone episodes over Southern Lake Michigan, with a focus on meteorological performance. WRF‐Chem showed good performance and successful reproduction of meteorological fields and clouds. Lake breeze model skill was inconsistent, with both good and poor performance depending on site and day. The combination of Noah land surface model and High‐Resolution Rapid Refresh meteorology gave the best performance with the mean bias of −0.5 °C for temperature, −0.6 °C for dewpoint temperature, and −0.3 m/s for wind speed along the western coast of Lake Michigan during the daytime. For ozone, WRF‐Chem was biased low (−4.4 ppb mean bias for daytime ozone) and underestimated hourly peak ozone. In some cases, ozone bias can be attributed to transport and lake breeze errors. Average ozone concentration showed minor (<2 ppb) sensitivity to changes to meteorology initial and boundary conditions or the land surface model.

show abstract

Section: Discussionsupporting

confidence: 85%

Sensitivity of Meteorological Skill to Selection of WRF‐Chem Physical Parameterizations and Impact on Ozone Prediction During the Lake Michigan Ozone Study (LMOS)

et al. 2020

View full text Add to dashboard Cite

show abstract

“…Additional model performance statistics are provided in Table S1. Model performance generally met benchmarks proposed by Emery et al ().…”

Section: Resultsmentioning

confidence: 57%

Modeling NH₄NO₃ Over the San Joaquin Valley During the 2013 DISCOVER‐AQ Campaign

et al. 2018

View full text Add to dashboard Cite

The San Joaquin Valley (SJV) of California experiences high concentrations of particulate matter NHNO during episodes of meteorological stagnation in winter. A rich data set of observations related to NHNO formation was acquired during multiple periods of elevated NHNO during the Deriving Information on Surface Conditions from Column and Vertically Resolved Observations Relevant to Air Quality (DISCOVER-AQ) field campaign in SJV in January and February 2013. Here NHNO is simulated during the SJV DISCOVER-AQ study period with the Community Multiscale Air Quality (CMAQ) model, diagnostic model evaluation is performed using the DISCOVER-AQ data set, and integrated reaction rate analysis is used to quantify HNO production rates. Simulated NO generally agrees well with routine monitoring of 24-hr average NO, but comparisons with hourly average NO measurements in Fresno revealed differences at higher time resolution. Predictions of gas-particle partitioning of total nitrate (HNO + NO) and NHx (NH + NH) generally agree well with measurements in Fresno, although partitioning of total nitrate to HNO is sometimes overestimated at low relative humidity in afternoon. Gas-particle partitioning results indicate that NHNO formation is limited by HNO availability in both the model and ambient. NH mixing ratios are underestimated, particularly in areas with large agricultural activity, and additional work on the spatial allocation of NH emissions is warranted. During a period of elevated NHNO, the model predicted that the OH + NO pathway contributed 46% to total HNOproduction in SJV and the NO heterogeneous hydrolysis pathway contributed 54%. The relative importance of the OH + NO pathway for HNO production is predicted to increase as NOx emissions decrease.

show abstract

“…As shown in Figure , model predictions of indoor ammonium concentrations, obtained using Equation (), tend to be above the 1:1 line for ammonium concentrations lower than 1 μg/m 3 , and below the 1:1 line for ammonium concentrations higher than 1 μg/m 3 . The NMB is −42%, higher than the proposed criterion for 24‐hours average ammonium (±30%), as shown in Table . r is 0.68, higher than the proposed criterion of 0.4.…”

Section: Resultsmentioning

confidence: 81%

“…The other parameter is the correlation coefficient ( r ) defined as:

r = \frac{\sum [()(P_{i} - \bar{P}) ()(O_{i} - \bar{O})]}{\sqrt{\sum {()(P_{i} - \bar{P})}^{2}} \sqrt{\sum {()(O_{i} - \bar{O})}^{2}}}

where

\bar{P}

and

\bar{Q}

are the mean values of model simulation and observation, respectively. Emery et al argued that a larger number of simulation‐observation pairs tends to result in better NMB. Correlation coefficient ( r ) is therefore needed to address variability over the entire range of the simulation‐observation pairs.…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Redistribution of PM_2.5‐associated nitrate and ammonium during outdoor‐to‐indoor transport

2019

View full text Add to dashboard Cite

Nitrate and ammonium ions are major constituents of outdoor PM2.5. Human exposure to these ions occurs primarily indoors. To assess the adverse outcomes from exposure to them, it is necessary to quantify the relationships between outdoor and indoor PM2.5 nitrate and ammonium. The relationships for the two semi‐volatile ions are more complex than those of non‐volatile PM2.5 constituents (eg, sulfate, elemental carbon). This study presents a mechanistic description of their outdoor‐indoor relationships that incorporates a dynamic gas‐particle partitioning and key parameters such as the pH and water content of PM2.5. Compared to measurements of nitrate and ammonium, the model has normalized mean biases of −9% and −42% and correlation coefficients of 0.95 and 0.68 for nitrate and ammonium, respectively. This suggests satisfactory agreement for nitrate, but less strong for ammonium. Sensitivity analysis on key parameters indicates that the model generally works well across a range of values typical of indoor settings. The model's performance is sensitive to pH and water content in PM2.5, which control the gas‐particle partitioning process. Indoor PM2.5 tends to be more acidic than outdoor PM2.5, raising potential health concern. The model provides insights in exposure assessment, source apportionment, and health‐composition attribution of indoor PM2.5.

show abstract

Recommendations on statistics and benchmarks to assess photochemical model performance

Cited by 431 publications

References 57 publications

Sensitivity of Meteorological Skill to Selection of WRF‐Chem Physical Parameterizations and Impact on Ozone Prediction During the Lake Michigan Ozone Study (LMOS)

Sensitivity of Meteorological Skill to Selection of WRF‐Chem Physical Parameterizations and Impact on Ozone Prediction During the Lake Michigan Ozone Study (LMOS)

Modeling NH₄NO₃ Over the San Joaquin Valley During the 2013 DISCOVER‐AQ Campaign

Redistribution of PM_2.5‐associated nitrate and ammonium during outdoor‐to‐indoor transport

Contact Info

Product

Resources

About

Recommendations on statistics and benchmarks to assess photochemical model performance

Cited by 431 publications

References 57 publications

Sensitivity of Meteorological Skill to Selection of WRF‐Chem Physical Parameterizations and Impact on Ozone Prediction During the Lake Michigan Ozone Study (LMOS)

Sensitivity of Meteorological Skill to Selection of WRF‐Chem Physical Parameterizations and Impact on Ozone Prediction During the Lake Michigan Ozone Study (LMOS)

Modeling NH4NO3 Over the San Joaquin Valley During the 2013 DISCOVER‐AQ Campaign

Redistribution of PM2.5‐associated nitrate and ammonium during outdoor‐to‐indoor transport

Contact Info

Product

Resources

About

Modeling NH₄NO₃ Over the San Joaquin Valley During the 2013 DISCOVER‐AQ Campaign

Redistribution of PM_2.5‐associated nitrate and ammonium during outdoor‐to‐indoor transport