Assessing the Iterative Finite Difference Mass Balance and 4D‐Var Methods to Derive Ammonia Emissions Over North America Using Synthetic Observations

Abstract

We evaluate two inverse modeling methods by conducting inversion experiments using the GEOS‐Chem chemical transport model and its adjoint. We simulate synthetic NH3 column density as observed by the Cross‐track Infrared Sounder over North America to test the ability of the iterative finite difference mass balance (IFDMB) and the four‐dimensional variational assimilation (4D‐Var) methods to recover known NH3 emissions. Comparing to the more rigorous 4D‐Var method, the IFDMB approach requires 3–4 times lower computational cost and yields similar or smaller errors (12–17% vs 17–26%) in the top‐down inventories at 2° × 2.5° resolution. These errors in IFDMB‐derived emission estimates are amplified (53–62%) if compared to the assumed true emissions at 0.25° × 0.3125° resolution. When directly conducting inversions at 0.25° × 0.3125°, the IFDMB consistently exhibits larger errors (44–69% vs 30–45%) than the 4D‐Var approach. Analysis of simulated differences in NH3 columns and in NH3 emissions suggests stronger misalignments at the finer resolution, since the local column is more strongly influenced by spatial smearing from neighboring grids. Adjoint calculations indicate that the number of adjacent grids needed to account for most (>65%) of the emission contributions to the local columnar NH3 abundance over an NH3 source site increases from ~1 at 2° × 2.5° to ~10 at 0.25° × 0.3125°, leading to increased errors especially in IFDMB. Applying inversion results at 2° × 2.5° to update the a priori emissions at 0.25° × 0.3125° could improve the accuracy of IFDMB inversions and reduce the computational cost of 4D‐Var.

Ammonia measurements from space with the Cross-track Infrared Sounder: characteristics and applications

“…Additional applications such as using the CrIS CFPR products for model inversions and data assimilation (e.g. Li et al, 2019) are currently being explored, which take advantage of the averaging kernels and error covariance matrix provided in the CrIS-retrieved product (e.g. observation operator) to provide top-down constraints on the ammonia emissions.…”

“…The output sensitivity and error parameter characterization are key for utilizing CrIS observations in air quality model applications such as data assimilation, data fusion, and model-based emission inversions (e.g. Li et al, 2019). It is also important that, as done first in the TES NH 3 retrieval (Shephard et al, 2011), the CFPR algorithm uses only three a priori ammonia profiles.…”

Do alternative inventories converge on the spatiotemporal representation of spring ammonia emissions in France?

“…This approach has been debated by Turner et al [2012], who found that non-local sources contribute substantially to columns of short-lived species. Li et al [2019] evaluated the ability of both a mass-balance approach and a variational assimilation to recover known NH 3 emissions at different spatial resolutions. At a 2°x2.5° resolution, they found that both methods yielded similar values.…”

High-resolution Hybrid Inversion of IASI Ammonia Columns to Constrain U.S. Ammonia Emissions Using the CMAQ Adjoint Model

“…Bottom-up inventories calculate the NH3 emissions based on estimated activity levels and corresponding emission factors, both of which are subject to high uncertainties, particularly for agricultural sources, the major contributor (Cooter et al, 2012;McQuilling and Adams, 2015). Several studies have utilized NH3 column density retrieved from IASI (Clarisse et al, 2009;Van Damme et 70 al., 2015b) or the Atmospheric Infrared Sounder (AIRS; (Warner et al, 2016)) as well as the inferred surface mixing ratio of NH3 from the Cross-track Infrared Sounder (CrIS; (Shephard and Cady-Pereira, 2015;Shephard et al, 2019)) to characterize the spatiotemporal distribution of NH3. These satellite measurements are useful for supplementing emission inventories to identify and quantify underestimated or missing emission hotspots, especially in intensive agricultural zones (Van Damme et al, 2018;Dammers et al, 2019;Clarisse et al, 2019).…”

“…The adjoint-based inversion method calculates the gradients of a cost function analytically and searches the solution using a steepest-descent optimization algorithm through iterating (Brasseur and Jacob, 2017). By testing the performance of the inverse modeling method using artificial observational data, Li et al (2019) proposed that a two-step optimization process, which combines the 90 iterative mass balance (IMB) method and the 4D-Var method, can further reduce the computational cost. The IMB method assumes a linear relationship between the NH3 column density and local NH3 emission and searches the emission scaling factors iteratively until the simulated NH3 column density converges to the observations.…”