Fix discontinuous water runoff

[LenkaNovak]

This is likely to be fixed by applying dss to precipitation sources.

[LenkaNovak]

We checked that this is already dss'd

applying

Ws_cpy = deepcopy(Y_lnd.Ws)
dss_buffer = ClimaCore.Spaces.create_dss_buffer(Ws_cpy)
ClimaCore.Spaces.weighted_dss!(Ws_cpy, dss_buffer)

resulted in no change (i.e. sum(Y_lnd.Ws .- Ws_cpy) = 0).

Note, most of the grid imprinting is seen over areas which are not integrated over land.

[LenkaNovak]

Some discontinuity expected due to 0M instantaneously translating onto the land variables.

There is also additional discontinuity due to plotting (here blown up to show range of [-1e-5, 0] of the same col. integrated rain field, same snapshot (after 100d) using both methods).

ClimaCorePlots (nonconservative interpolation):

ClimaCoreTempestRemap (conservative interpolation to RLL, where a lot of the discontinuity within elements goes away):

Conclusion: rain-derived variables behave within expected bounds. Thanks, @trontrytel, for your thoughts.