the model currently has like 1e-3 coefficient for fixed effect precipitation, whereas it has 1e-1 coefficient for the unobserved random effect. my guess is this is because the southern catchments all have massive amount of precipiation compared to everywhere else but i dont know.
try:
- remove the southern catchments
- add other effects
- make precipitation random effect
the model currently has like 1e-3 coefficient for fixed effect precipitation, whereas it has 1e-1 coefficient for the unobserved random effect. my guess is this is because the southern catchments all have massive amount of precipiation compared to everywhere else but i dont know.
try: