message about both GPP=0 and prediction time period for metab_night >24hr day

[aappling-usgs]

You get different messages on predict_metab depending on whether your metab_night inputs are for <=24 hours or >24 hours. Unify these.

For values within the day_start=12, day_end=36 range:

dat <- data_metab("3", "15")
metab(specs=specs(mm_name("night"), day_start=12, day_end=34), data=dat)
for metab_night, GPP estimates are 0 because they're for nighttime only

For values out of the 12-36 range:

dat <- data_metab("3", "15")
metab(specs=specs(mm_name("night"), day_start=10, day_end=36), data=dat)
daily metabolism predictions are for the period from 10 to 34 hours on each date
 (differs from the model-fitting range of 10 to 36 hours)

Make it so that both of the messages ("for metab_night, GPP estimates are 0..." and "daily metabolism predictions are for the period...") appear when both conditions apply.

But see also #327, which could render the message "for metab_night, GPP estimates are 0 because they're for nighttime only" obsolete.