Impact of physical parameterizations and initial conditions on simulated atmospheric transport and CO<sub>2</sub> mole fractions in the US Midwest

Abstract

Abstract. Atmospheric transport model errors are one of the main contributors to the uncertainty affecting CO2 inverse flux estimates. In this study, we determine the leading causes of transport errors over the US upper Midwest with a large set of simulations generated with the Weather Research and Forecasting (WRF) mesoscale model. The various WRF simulations are performed using different meteorological driver datasets and physical parameterizations including planetary boundary layer (PBL) schemes, land surface models (LSMs), cumulus parameterizations and microphysics parameterizations. All the different model configurations were coupled to CO2 fluxes and lateral boundary conditions from the CarbonTracker inversion system to simulate atmospheric CO2 mole fractions. PBL height, wind speed, wind direction, and atmospheric CO2 mole fractions are compared to observations during a month in the summer of 2008, and statistical analyses were performed to evaluate the impact of both physics parameterizations and meteorological datasets on these variables. All of the physical parameterizations and the meteorological initial and boundary conditions contribute 3 to 4 ppm to the model-to-model variability in daytime PBL CO2 except for the microphysics parameterization which has a smaller contribution. PBL height varies across ensemble members by 300 to 400 m, and this variability is controlled by the same physics parameterizations. Daily PBL CO2 mole fraction errors are correlated with errors in the PBL height. We show that specific model configurations systematically overestimate or underestimate the PBL height averaged across the region with biases closely correlated with the choice of LSM, PBL scheme, and cumulus parameterization (CP). Domain average PBL wind speed is overestimated in nearly every model configuration. Both planetary boundary layer height (PBLH) and PBL wind speed biases show coherent spatial variations across the Midwest, with PBLH overestimated averaged across configurations by 300–400 m in the west, and PBL winds overestimated by about 1 m s−1 on average in the east. We find model configurations with lower biases averaged across the domain, but no single configuration is optimal across the entire region and for all meteorological variables. We conclude that model ensembles that include multiple physics parameterizations and meteorological initial conditions are likely to be necessary to encompass the atmospheric conditions most important to the transport of CO2 in the PBL, but that construction of such an ensemble will be challenging due to ensemble biases that vary across the region.

Characterization of Regional‐Scale CO2 Transport Uncertainties in an Ensemble with Flow‐Dependent Transport Errors

“…Previous studies have mostly focused on the first class of transport errors and quantified the uncertainties using different transport models (e.g., Engelen et al, 2002;Gurney et al, 2002) or perturbed model physics (e.g., Díaz-Isaac et al, 2018). Previous studies have mostly focused on the first class of transport errors and quantified the uncertainties using different transport models (e.g., Engelen et al, 2002;Gurney et al, 2002) or perturbed model physics (e.g., Díaz-Isaac et al, 2018).…”

“…Transport errors can be grossly divided into two components: (1) transport errors due to errors in the transport model, including errors in model parameters, and (2) transport errors due to errors in meteorological initial and boundary conditions. Previous studies have mostly focused on the first class of transport errors and quantified the uncertainties using different transport models (e.g., Engelen et al, 2002;Gurney et al, 2002) or perturbed model physics (e.g., Díaz-Isaac et al, 2018). However, even with a perfect transport model, there can exist significant transport errors due to errors in meteorological initial and boundary conditions combined with intrinsic atmospheric error growth.…”

Seasonal Characteristics of Model Uncertainties From Biogenic Fluxes, Transport, and Large‐Scale Boundary Inflow in Atmospheric CO2 Simulations Over North America

“…Our selection of the transport members was based on empirical evidence from Díaz‐Isaac et al () that a calibrated small‐sized ensemble could recover the model errors built from a large‐sized ensemble (such as a 45‐member ensemble). This 10‐member transport ensemble appeared to be biased in Figure .…”

“…Combining these two methods likely cover transport uncertainty due to model physics and dynamics. Here we varied the land surface models (LSMs) and planetary boundary layer (PBL) schemes in the WRF‐Chem model package based on the findings of Díaz‐Isaac et al (), we chose three combinations: (1) Mellor‐Yamada Nakanishi and Niino Level 2.5 (MYNN 2.5) PBL scheme with Noah LSM, (2) Mellor‐Yamada‐Janjic PBL scheme with RUC LSM, (3) Yonsei University PBL scheme with five‐layer thermal diffusion LSM. In addition, three random dynamical perturbations were then applied to each physics scheme using a SKEBS.…”

“…In this case, a considerable number of observations fall into the outer bins, which leads to “U‐shaped” rank histograms with scores greater than 1. Díaz‐Isaac et al () used rank histograms to define the optimal selection of members among 45 to produce the most accurate representation of transport model errors. Here we use a simpler approach by starting with a small‐size ensemble and gradually adding members to the ensemble until the flatness of the histogram is acceptable.…”