# Adjoint evaluation using CVM

**URL:** <https://forum.cmascenter.org/t/adjoint-evaluation-using-cvm/6129>\
**Category:** CMAQ-Adjoint\
**Created:** [January 15, 2026, 3:26am UTC](https://forum.cmascenter.org/t/adjoint-evaluation-using-cvm/6129 "2026-01-15T03:26:27Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![Hujie\_0608](https://avatars.discourse-cdn.com/v4/letter/h/e495f1/32.png) [@Hujie\_0608](https://forum.cmascenter.org/u/Hujie_0608)\
**Post date:** [January 15, 2026, 3:26am UTC](https://forum.cmascenter.org/t/adjoint-evaluation-using-cvm/6129/1 "2026-01-15T03:26:28Z")

</div>

Hi,

I am using the CVM to evaluate the adjoint sensitivities. It appears that the generated _SGRID_ files represent concentration changes resulting from perturbations in species concentrations. I would like to ask whether it is possible to calculate ozone concentration changes associated with NO₂ emission perturbations using CVM. If so, which components or settings should be modified to obtain this result? Thank you.

Best regards,

Hu Jie

---

<div class="post-metadata">

**Author:** ![zhaoshunliu](https://avatars.discourse-cdn.com/v4/letter/z/5e9695/32.png) [@zhaoshunliu](https://forum.cmascenter.org/u/zhaoshunliu)\
**Post date:** [January 15, 2026, 2:17pm UTC](https://forum.cmascenter.org/t/adjoint-evaluation-using-cvm/6129/2 "2026-01-15T14:17:15Z")

</div>

Hi Jie,

Unfortunately, CVM in its current form won’t give you sensitivities to emissions; the part dealing with emission processing is not converted to CVM yet. If you ever consider doing it yourself, please let me know and we can talk more.

Best,  
Shunliu

---

<div class="post-metadata">

**Author:** ![Hujie\_0608](https://avatars.discourse-cdn.com/v4/letter/h/e495f1/32.png) [@Hujie\_0608](https://forum.cmascenter.org/u/Hujie_0608)\
**Post date:** [January 16, 2026, 6:30am UTC](https://forum.cmascenter.org/t/adjoint-evaluation-using-cvm/6129/3 "2026-01-16T06:30:51Z")

</div>

Hi Shunliu,

Thank you for your response. I used the adjoint model and the CVM to compare sensitivity results using the benchmark dataset. The receptor grid was set to **C1R1L1** by default. For the adjoint simulation, I defined the cost function as the **ASO4J concentration at the last hour** , and for the CVM test I perturbed the **ASO4J concentration**  **at the first hour** as well.

I then compared the outputs from **`lgrid_2008099.ncf`** and **`SGRID_2008099.ncf`** at the grid **C1R1L1** , but the results differ substantially. I would like to ask whether there is any mistake with my setup or if additional settings are required to make the adjoint and CVM results comparable.

Thank you very much for your help.

Best regards,

Hu Jie