We are happy to read that this reviewer also understood and appreciated the goals of this manuscript.

As the review points out, the focus here is on enhanced temporal data output. As mentioned in the abstract, this paper grew out of a project focused on parameterizing and characterizing the impact of subgrid heterogeneity on ESM behavior and performance, so we fully appreciate the importance of the reviewer’s first two concerns. We note in the abstract that this manuscript enables a necessary step towards achieving that aim, but we agree with the reviewer that the manuscript will be strengthened if we broaden our discussion of spatial scales.

In response to the first request, we have added the following text:

The spatial scales of individual grid cells in ESM simulations included in the most recent Coupled Model Intercomparison Project (CMIP6) typically range from 50 km to 250 km, with models run at resolutions finer than 50 km eligible for participation in the High-Resolution Model Intercomparison Project (HighResMIP; Haarsma et al., 2016). These resolutions suggest that the footprint sampled from in situ observations (ranging from cm-scale soil moisture probes to wind- and height-dependent flux tower sampling fetches on the order of hundreds of meters) is substantially smaller than individual ESM grid cells. This suggests that, when possible, observational comparisons should be made against sub-grid tiles representing fractional areas of differing land use types. However, saving tile-specific high-frequency data is likely not feasible for most modelling centres. Given that reality, the data request outlined here will enable previously impossible assessment of grid-cell mean behaviour throughout the diurnal cycle. Future work motivated by the CLASP project can extend these lines of inquiry to issues centred on sub-grid spatial heterogeneity, or to comparisons with global storm-resolving efforts like those of Stevens et al. (2019).

In response to the second point, archive requirements for 1-degree information were provided just to give readers a sense of the storage space needed to meet this data request. To make this clearer, we added the phase “(for reference)” after “Assuming a 1-degree grid.”

In response to the third point, we have reformatted Figure 2 so the panels are arranged vertically, allowing them to be bigger so the lines (even the slimmest ones with the weakest correlations) are more visible. This also addresses a point made by Reviewer #3.