Aquifer water state variable not output to history when use_groundwater_irrigation is true

[swensosc]

In WaterStateType, the history output for wa_col is only active when use_aquifer_layer is true. It should also be output when use_groundwater_irrigation is true. This is not straightforward to implement, because use_groundwater_irrigation is a private parameter in IrrigationMod, so some way of passing the value of use_groundwater_irrigation to WaterStateType is needed. Until then, users running with active groundwater irrigation will not have history output for WA (the water in the aquifer).

[billsacks]

This does seem a bit tricky. Looking through the initialization ordering - especially in clm_instInit, it looks like it might work to move the initialization of soil_water_retention_curve and irrigation_inst to immediately before the call to water_inst%Init. Then we could add a function in IrrigationMod to return the value of use_groundwater_irrigation, which would get passed down through water_inst%Init similarly to what is done for use_aquifer_layer. Based on the arguments to create_soil_water_retention_curve (none) and to irrigation_inst%init, it looks like it should be safe to do this rearrangement.