hk_l used before properly initialized

### Brief summary of bug

The layer hydraulic conductivity is used in PhotosynthesisHydraulicStress at first time step before it is calculated in compute_hydraulic_properties.

### General bug information

**CTSM version you are using:** release-clm5.0.17

**Does this bug cause significantly incorrect results in the model's science?** No

**Configurations affected:** Cold starts

### Details of bug

hk_l is used in PhotosynthesisHydraulicStress in this line:

soil_conductance = min(hksat(c,j),hk_l(c,j))/(1.e3*r_soil)

hk_l is calculated later in the driver sequence in compute_hydraulic_properties. In a cold start, hk_l is initialized to 0.0, so soil conductance will be zero on the first time step. Later in the first time step, hk_l is properly calculated and thus for the second time step forward the soil_conductance is correct.
One possible solution to this is to call compute_hydraulic_properties to calculate hk_l based on the cold start soil moisture.

### Important details of your setup / configuration so we can reproduce the bug

No runs done, Sean, Gordon, and I just looked at the code.