k_soil_root uses incorrect soil conductance

### Brief summary of bug

PHS calculates k_soil_root for each soil layer based on hk_l. However, hk_l is defined at the interface between two adjoining soil layers. For PHS the relevant value is not at the interface, but just within the given soil layer.

### General bug information

**CTSM version you are using:** release-clm5.0.30

**Does this bug cause significantly incorrect results in the model's science?** Yes, but answer changes expected to be relatively small.

**Configurations affected:** use_hydrstess=.true.

### Details of bug

Here the value of hk_l should be replaced with an alternative variable representing hk for a single soil layer:
https://github.com/ESCOMP/CTSM/blob/ad9944e57bf420482268d02fe3d08a3966ce23d6/src/biogeophys/PhotosynthesisMod.F90#L2836
Such a variable would need to be calculated within SoilWaterMovementMod. As such it may be appropriate to rename hk_l to hk_interface

As it stands, hk_l effectively averages together the hk values from soil layers i and i+1, whereas PHS needs hk based just on soil layer i. This bug can cause some (minor) incorrect dynamics in the vertical distribution of soil water, but effects on downstream variables like photosynthesis, transpiration, soilwater_10cm, and TWS are all expected to be relatively small.