lme (nlme) error: "system is computationally singular"

Question

I am starting a mixed-model selection analysis with lme in R. But when I start to fit the full model:

fullmodOver <-  lme(rrF ~ Old_N_dep+Overstory_old+Altitude+MAT+MAP+EIV_R+EIV_F+Overstory_diff+SCA_old+SCA_diff, data = regdata, random = ~ 1|PlotNR, method = "REML")

I get the following error:

Error in na.fail.default(list(rrF = c(-0.0380393284017694, -0.0110047996186618,  : 
  missing values in object

Therefore I delete all missing value of variables with the following:

regdata <- regdata[!apply(regdata[,c("Overstory_old","Old_N_dep","Altitude","MAT","MAP","EIV_R","EIV_F","Overstory_diff","SCA_old","SCA_diff")], 1, anyNA),]

When I re-fit the model I get again an error:

Error in solve.default(estimates[dimE[1] - (p:1), dimE[2] - (p:1), drop = FALSE]) : 
  system is computationally singular: reciprocal condition number = 7.92883e-17

This looks similar to what described here, where - if I understand correctly - the dependent variable is perfectly correlated to one of the dependent variabes. When I try to fit the same model with lmer it doesn't give me an error, but I get results with coefficients of parameters that are very close to zero, like it happens in the stackoverflow question I linked to you before. However, my dependent and indipendent variables in the full model don't seems correlated. Initial_survey and EIV_N (higly correlated with Old_N_dep and EIV_R, respectively) are excluded from the main formula:

Probably good to mention (I don't know if it's relevant): after checking for collinearity and before being included in the model, variables where scaled with the following code (variables considered in the model are all numerical):

###z-normalization of numerical variables
regdata<-regdata%>%
  mutate(Initial_survey = as.numeric(Initial_survey))%>%
  mutate(MAP = as.numeric(MAP))%>%
  mutate(Altitude=as.numeric(Altitude))
regdata$mNdep<-scale(regdata$mNdep)
regdata$mMAT<-scale(regdata$mMAT)
regdata$mMAP<-scale(regdata$mMAP)
regdata$Altitude<-scale(regdata$Altitude)
regdata$MAT<-scale(regdata$MAT)
regdata$MAP<-scale(regdata$MAP)
regdata$Old_N_dep<-scale(regdata$Old_N_dep)
regdata$EIV_L_new<-scale(regdata$EIV_L_new)
regdata$Overstory_new<-scale(regdata$Overstory_new)
regdata$SCA_new<-scale(regdata$SCA_new)
regdata$EIV_L_old<-scale(regdata$EIV_L_old)
regdata$EIV_R<-scale(regdata$EIV_R)
regdata$EIV_N<-scale(regdata$EIV_N)
regdata$EIV_F<-scale(regdata$EIV_F)
regdata$Overstory_old<-scale(regdata$Overstory_old)
regdata$SCA_old<-scale(regdata$SCA_old)
regdata$Overstory_diff<-scale(regdata$Overstory_diff)
regdata$SCA_diff<-scale(regdata$SCA_diff)
regdata$EIV_L_diff<-scale(regdata$EIV_L_diff)
regdata$Initial_survey<-scale(regdata$Initial_survey)

Do you have any clue on what I am missing here? Thanks.

PS: Probably you might want to take a look at the dataset but I don't know how to show you that here, (it's a quite big dataset). But I think I might show it to you if it's necessary.