Hello Alicia,
thanks for your response.
I had another look at the code. Your interpretation of the code section from lines 516 to 679 (specifically the “land” section starting on lines 576) is correct. While PX LSM and therefore CMAQ M3DRY do not expect any land cell where VEGF zero, and while rcut, rwetsfc, and rgndc are all calculated even though there is no vegetation, they ultimately don’t impact the rci calculation on lines 670 - 675 since vegcr is 0, so rci (inverted rc) is indeed just calculated from the deposition to soil (dry, wet, and/or snow-covered). My suggestion on Friday that additional coding might be needed in that section to handle this situation was not correct.
You are also correct that by modifying line 420 to ensure that the calculation of moisture parameters affecting the soil and cuticular resistances occurs for all land cells (even when VEGF is zero) will then allow the code to compute these resistances (for your scenario, most importantly the effect of moisture on soil resistance) in lines 516 - 579.
That said, there may be parts of the soil resistance calculations (e.g. the rgnd equations in lines 635 and 639) where the assumptions made in the WRF PX LSM / CMAQ M3DRY framework and the WRF NOAH LSM are inconsistent. Specifically, there may be inconsistencies about the depth of the first soil layer for which WRF soil moisture is calculated (variable SOIM1), which in turn would mean that the rgnd equation may need to look different for WRF NOAH LSM SOIM1 than WRF PX LSM SOIM1. This would be similar to this discussion about windblown dust and its dependency on soil moisture assumptions in the WRF LSM. Similarly, there is a comment on line 625 that “Canopy height is assumed to be 10 * z0 according to PX-LSM“. I do not know whether this assumption would be consistent with the assumptions in the NOAH LSM.
Ultimately, the code modification you envision should functionally work to have M3DRY compute dry deposition velocities for all grid cells for your met fields, but some aspects of these computations may be affected by inconsistencies in assumptions between WRF PX LSM and WRF NOAH LSM.