runGAMFusion <- function(i) {            # loop over dates

  #Adding layer LST to the raster stack

  pos<-match(LST_month,layerNames(s_raster)) #Find column with the current month for instance mm12

  r1<-raster(s_raster,layer=pos)             #Select layer from stack

  layerNames(r1)<-"LST"

  s_raster<-addLayer(s_raster,r1)            #Adding current month

  data_v <- ghcn.subsets[[i]][ind.testing, ]    #Testing/validation dataset using input sampling

  #  STEP 1 - LST 10 year monthly averages

  #names(d)[pos]<-c("dailyTmax")

  names(d)[pos]<-y_var_name

  names(x)[pos]<-y_var_name

  #names(x)[pos]<-c("dailyTmax")

  #data_s$y_var<-data_s$LSTD_bias

  ###BEFORE GAM prediction the data object must be transformed to SDF

  coords<- data_v[,c('x_OR83M','y_OR83M')]

  coordinates(data_v)<-coords

  proj4string(data_v)<-CRS  #Need to assign coordinates...

  coords<- data_s[,c('x_OR83M','y_OR83M')]

  coordinates(data_s)<-coords

  proj4string(data_s)<-CRS  #Need to assign coordinates..

  nv<-nrow(data_v)

  ###GAM PREDICTION

  #data_s$y_var<-data_s$dailyTmax  #This shoudl be changed for any variable!!!

  #data_v$y_var<-data_v$dailyTmax

  data_v$y_var<-data_v[[y_var_name]]

  data_s$y_var<-data_s[[y_var_name]]

  #Model and response variable can be changed without affecting the script

  formula1 <- as.formula("y_var ~ s(lat) + s(lon) + s(ELEV_SRTM)", env=.GlobalEnv)

  formula2 <- as.formula("y_var~ s(lat,lon)+ s(ELEV_SRTM)", env=.GlobalEnv)

  formula3 <- as.formula("y_var~ s(lat) + s (lon) + s (ELEV_SRTM) +  s (Northness)+ s (Eastness) + s(DISTOC)", env=.GlobalEnv)

  formula4 <- as.formula("y_var~ s(lat) + s (lon) + s(ELEV_SRTM) + s(Northness) + s (Eastness) + s(DISTOC) + s(LST)", env=.GlobalEnv)

  formula5 <- as.formula("y_var~ s(lat,lon) +s(ELEV_SRTM) + s(Northness,Eastness) + s(DISTOC) + s(LST)", env=.GlobalEnv)

  formula6 <- as.formula("y_var~ s(lat,lon) +s(ELEV_SRTM) + s(Northness,Eastness) + s(DISTOC) + s(LST)+s(LC1)", env=.GlobalEnv)

  formula7 <- as.formula("y_var~ s(lat,lon) +s(ELEV_SRTM) + s(Northness,Eastness) + s(DISTOC) + s(LST)+s(LC3)", env=.GlobalEnv)

  formula8 <- as.formula("y_var~ s(lat,lon) +s(ELEV_SRTM) + s(Northness,Eastness) + s(DISTOC) + s(LST) + s(LC1,LC3)", env=.GlobalEnv)

  mod1<- try(gam(formula1, data=data_s))

  mod2<- try(gam(formula2, data=data_s)) #modified nesting....from 3 to 2

  mod3<- try(gam(formula3, data=data_s))

  mod4<- try(gam(formula4, data=data_s))

  mod5<- try(gam(formula5, data=data_s))

  mod6<- try(gam(formula6, data=data_s))

  mod7<- try(gam(formula7, data=data_s))

  mod8<- try(gam(formula8, data=data_s))

#   mod1<- try(gam(formula1, data=data_s))

#   mod2<- try(gam(formula2, data=data_s)) #modified nesting....from 3 to 2

#   mod3<- try(gam(y_var~ s(lat) + s (lon) + s (ELEV_SRTM) +  s (Northness)+ s (Eastness) + s(DISTOC), data=data_s))

#   mod4<- try(gam(y_var~ s(lat) + s (lon) + s(ELEV_SRTM) + s(Northness) + s (Eastness) + s(DISTOC) + s(LST), data=data_s))

#   mod5<- try(gam(y_var~ s(lat,lon) +s(ELEV_SRTM) + s(Northness,Eastness) + s(DISTOC) + s(LST), data=data_s))

#   mod6<- try(gam(y_var~ s(lat,lon) +s(ELEV_SRTM) + s(Northness,Eastness) + s(DISTOC) + s(LST)+s(LC1), data=data_s))

#   mod7<- try(gam(y_var~ s(lat,lon) +s(ELEV_SRTM) + s(Northness,Eastness) + s(DISTOC) + s(LST)+s(LC3), data=data_s))

#   mod8<- try(gam(y_var~ s(lat,lon) +s(ELEV_SRTM) + s(Northness,Eastness) + s(DISTOC) + s(LST) + s(LC1,LC3), data=data_s))

#   

  #Added

  #tmax_predicted=themolst+daily_delta_rast-bias_rast #Final surface?? but daily_rst

  j=nmodels+1

  for (j in 1:nmodels){

      y_var_fit= mod$fit

      #results_RMSE_f[j+3]<- sqrt(sum((y_var_fit-data_s$y_var)^2)/ns)

      results_RMSE_f[j+3]<-sqrt(mean(mod$residuals^2,na.rm=TRUE))

      #results_MAE_f[j+3]<-sum(abs(y_var_fit-data_s$y_var))/ns

      results_MAE_f[j+3]<-mean(abs(mod$residuals),na.rm=TRUE)

      if (predval==1) {

        ##Model assessment: specific diagnostic/metrics for GAM

        name<-paste("mod",j,sep="")  #modj is the name of The "j" model (mod1 if j=1) 

        mod<-get(name)               #accessing GAM model ojbect "j"

        s_sgdf<-as(s_raster,"SpatialGridDataFrame") #Conversion to spatial grid data frame

        rpred<- predict(mod, newdata=s_sgdf, se.fit = TRUE) #Using the coeff to predict new values.

        y_pred<-rpred$fit

        raster_pred<-r1

        layerNames(raster_pred)<-"y_pred"

        values(raster_pred)<-as.numeric(y_pred)

        data_name<-paste("predicted_mod",j,"_",dates[[i]],sep="")

        raster_name<-paste("GAM_",data_name,out_prefix,".rst", sep="")

        writeRaster(raster_pred, filename=raster_name,overwrite=TRUE)  #Writing the data in a raster file format...(IDRISI)

        #writeRaster(r2, filename=raster_name,overwrite=TRUE)  #Writing the data in a raster file format...(IDRISI)

        pred_sgdf<-as(raster_pred,"SpatialGridDataFrame") #Conversion to spatial grid data frame

        #rpred_val_s <- overlay(raster_pred,data_s)             #This overlays the kriged surface tmax and the location of weather stations

        rpred_val_s <- overlay(pred_sgdf,data_s)             #This overlays the kriged surface tmax and the location of weather stations

        rpred_val_v <- overlay(pred_sgdf,data_v)             #This overlays the kriged surface tmax and the location of weather stations

        pred_mod<-paste("pred_mod",j,sep="")

        #Adding the results back into the original dataframes.

        data_s[[pred_mod]]<-rpred_val_s$y_pred

        data_v[[pred_mod]]<-rpred_val_v$y_pred  

        #Model assessment: RMSE and then krig the residuals....!

        res_mod_s<- data_s$y_var - data_s[[pred_mod]]           #Residuals from kriging training

        res_mod_v<- data_v$y_var - data_v[[pred_mod]]           #Residuals from kriging validation

      }

      if (predval==0) {

        y_mod<- predict(mod, newdata=data_v, se.fit = TRUE) #Using the coeff to predict new values.

        pred_mod<-paste("pred_mod",j,sep="")

        #Adding the results back into the original dataframes.

        data_s[[pred_mod]]<-as.numeric(mod$fit)

        data_v[[pred_mod]]<-as.numeric(y_mod$fit)

        #Model assessment: RMSE and then krig the residuals....!

        res_mod_s<- data_s$y_var - data_s[[pred_mod]]           #Residuals from kriging training

        res_mod_v<- data_v$y_var - data_v[[pred_mod]]           #Residuals from kriging validation

      }

      ####ADDED ON JULY 20th

      res_mod<-res_mod_v

      #RMSE_mod <- sqrt(sum(res_mod^2)/nv)                 #RMSE FOR REGRESSION STEP 1: GAM  

      RMSE_mod<- sqrt(mean(res_mod^2,na.rm=TRUE))

      #MAE_mod<- sum(abs(res_mod),na.rm=TRUE)/(nv-sum(is.na(res_mod)))        #MAE from kriged surface validation

      MAE_mod<- mean(abs(res_mod), na.rm=TRUE)

      #ME_mod<- sum(res_mod,na.rm=TRUE)/(nv-sum(is.na(res_mod)))                    #ME, Mean Error or bias FOR REGRESSION STEP 1: GAM

      ME_mod<- mean(res_mod,na.rm=TRUE)                            #ME, Mean Error or bias FOR REGRESSION STEP 1: GAM

      #R2_mod<- cor(data_v$y_var,data_v[[pred_mod]])^2              #R2, coef. of var FOR REGRESSION STEP 1: GAM

      R2_mod<- cor(data_v$y_var,data_v[[pred_mod]], use="complete")^2

      data_v[[name2]]<-as.numeric(res_mod_v)

      data_s[[name2]]<-as.numeric(res_mod_s)

  mod_obj<-list(mod1,mod2,mod3,mod4,mod5,mod6,mod7,mod8,mod9a,mod9b)

  names(mod_obj)<-c("mod1","mod2","mod3","mod4","mod5","mod6","mod7","mod8","mod9a","mod9b")

  #results_list<-list(data_s,data_v,tb_metrics1,tb_metrics2)

  results_list<-list(data_s,data_v,tb_metrics1,tb_metrics2,mod_obj)

  names(results_list)<-c("data_s","data_v","tb_metrics1","tb_metrics2","mod_obj")

  save(results_list,file= paste(path,"/","results_list_metrics_objects_",dates[i],out_prefix,".RData",sep=""))