Isolate the effects of individual WRF parameterization schemes

Determine the individual contributions of the microphysics, cumulus, planetary boundary-layer, surface-layer, radiation, and related parameterization schemes to simulated rainfall skill in the WRF simulations of extreme rainfall over Kerala by conducting controlled experiments that vary one parameterization at a time.

Background

The twelve experiments alter multiple physical parameterizations simultaneously. Although the consistent performance of the six YSU–MM5 configurations suggests that boundary-layer and surface-layer treatment is the dominant control for the studied event, the experimental design does not permit definitive attribution to any single scheme.

The authors explicitly state that specific scheme contributions cannot be distinguished and that controlled one-at-a-time experiments are needed to separate nonlinear interactions among physics packages from the effects of individual parameterizations.

References

Because multiple physical schemes were changed concurrently during the trials, it is impossible to definitively credit the performance of any one configuration to a single parameterization. Interactions between the microphysics, cumulus, planetary boundary-layer, surface-layer, land-surface, and radiation schemes are reflected in the model response.

— Boundary-layer scheme selection outweighs microphysics in WRF simulations of the August 2018 Kerala extreme rainfall event  (2609.09709 - Kumar et al., 9 Sep 2026) in Section 3.2, page 13; Discussion, Section 4, pages 27–28; Conclusions, Section 5, page 29

Because the spread in RMSE across these six configurations is less than 6 per cent and the spread in correlation less than 0.04, no single member of the group can be identified as definitively superior on the present sample of 585 point-days.

— Boundary-layer scheme selection outweighs microphysics in WRF simulations of the August 2018 Kerala extreme rainfall event  (2609.09709 - Kumar et al., 9 Sep 2026) in Section 3.1, page 12; Discussion, Section 4, pages 25–26; Conclusions, Section 5, page 29

A direct test is available: WRF accumulates the convective and grid -scale components separately, so partitioning the simulated rainfall between them would give the convective fraction explicitly, and would establish whether that fraction falls at heavy thresholds as the argument requires.

— Cumulus parameterization controls simulated rainfall more than microphysics, but not at heavy thresholds: a 3 * 2 factorial WRF study over the Western Himalaya  (2610.17896 - Kumar et al., 13 Oct 2026) in Section 4.1, Why the cumulus scheme dominates