Is the specification of GAMLSS using the DAGUM distributions, leaving sigma and nu constants, the following?
eq.mu_1 <- VALUE ~ s(LONGITUDE,LATITUDE)+s(CONSTRU