##################    Data preparation for interpolation   #######################################

############################ Covariate production for a given tile/region ##########################################

#This script produces covariates raster for a a specfied study area.                             

# It requires the following inputs:                                                              

# 1)list of modis tiles or shape file with region outline            

# 2)input names for global covariates:

# -SRTM CGIAR 1 km

# -Canopy heihgt (Simard)

# -land cover concensus (Jetz lab)

# -MODIS LST: mean and obs

#3) The output is a multiband file in tif format with projected covariates for the processing region/tile.             

#AUTHOR: Benoit Parmentier                                                                       

#DATE: 01/28/2013                                                                                 

#PROJECT: NCEAS INPLANT: Environment and Organisms --TASK#363--   

##Comments and TODO:

#This script is meant to be for general processing tile by tile or region by region.

#We must decide on a local projection. This can best set up from the tile/region extent: for now use

#- do not keep output in memory??

##################################################################################################

###Loading R library and packages   

library(RPostgreSQL)

library(sp)                                             # Spatial pacakge with class definition by Bivand et al.

library(spdep)                                          # Spatial pacakge with methods and spatial stat. by Bivand et al.

library(rgdal)                                          # GDAL wrapper for R, spatial utilities

library(gtools)

library(rasterVis)

library(graphics)

library(grid)

library(lattice)

### Parameters and arguments

##Paths to inputs and output

in_path <- "/home/parmentier/Data/IPLANT_project/Venezuela_interpolation/Venezuela_01142013/input_data/"

lc_path<-"/home/layers/data/land-cover/lc-consensus-global"

elev_path<-"/home/layers/data/terrain/dem-cgiar-srtm-1km-tif"

infile2<-"modis_sinusoidal_grid_world.shp"

infile3<-"countries_sinusoidal_world.shp"

infile_reg_outline=""  #input region outline defined by polygon

                   #for the processing tile/region? This is a group fo six tiles for now.

#### Functions used in the script  ###

create_modis_tiles_region<-function(modis_grid,tiles){

  #This functions returns a subset of tiles from the modis grdi.

  #Arguments: modies grid tile,list of tiles

  #Output: spatial grid data frame of the subset of tiles

  h_list<-lapply(tiles,substr,start=2,stop=3) #passing multiple arguments

  v_list<-lapply(tiles,substr,start=5,stop=6) #passing multiple arguments

}

create_polygon_from_extent<-function(reg_ref_rast){

  #This functions returns polygon sp from input rast

  #Arguments: input ref rast

  #Output: spatial polygon

  set1f <- function(x){rep(1, x)}

  tmp_rast <- init(reg_ref_rast, fun=set1f, overwrite=TRUE)

  reg_outline_poly<-rasterToPolygons(tmp_rast)

  return(reg_outline_poly)

}

  #Arguments: raster name of the file,reference file with

  #Output: spatial grid data frame of the subset of tiles

  layer_rast<-raster(raster_name)

  new_proj<-proj4string(layer_rast)                  #Extract coordinates reference system in PROJ4 format

  #layer_projected_rast<-projectRaster(from=layer_crop_rast,crs=proj4string(reg_outline),method="ngb")

  layer_projected_rast<-projectRaster(from=layer_crop_rast,to=reg_ref_rast,method="ngb")

  return(layer_projected_rast)

}

mosaic_raster_list<-function(mosaic_list,out_names,out_path){

  #This functions returns a subset of tiles from the modis grid.

  #Arguments: modies grid tile,list of tiles

  #Output: spatial grid data frame of the subset of tiles

  #Note that rasters are assumed to be in the same projection system!!

  rast_list<-vector("list",length(mosaic_list))

  for (i in 1:length(mosaic_list)){  

    # read the individual rasters into a list of RasterLayer objects

    # this may be changed so that it is not read in the memory!!!

    input.rasters <- lapply(as.character(mosaic_list[[i]]), raster)

    mosaiced_rast<-input.rasters[[1]]

    for (k in 2:length(input.rasters)){

      mosaiced_rast<-mosaic(mosaiced_rast,input.rasters[[k]], fun=mean)

      #mosaiced_rast<-mosaic(mosaiced_rast,raster(input.rasters[[k]]), fun=mean)

    }

    data_name<-paste("mosaiced_",sep="") #can add more later...

    raster_name<-paste(data_name,out_names[i],".tif", sep="")

    writeRaster(mosaiced_rast, filename=file.path(out_path,raster_name),overwrite=TRUE)  

    #Writing the data in a raster file format...  

    rast_list[[i]]<-file.path(out_path,raster_name)

  }

  return(rast_list)

}

###########################################################

############ Main body: BEGIN--START OF THE SCRIPT ###################

##### STEP 1: Reading region or tile information to set the study or processing region

modis_grid<-readOGR(".", filename)     #Reading shape file using rgdal library

world_countries<-readOGR(".", filename)     #Reading shape file using rgdal library

if (infile_reg_outline!=""){

  filename<-sub(".shp","",infile_reg_outline)   #Removing the extension from file.

  reg_outline<-readOGR(".", filename)

}

if (infile_reg_outline==""){

  reg_outline<-create_modis_tiles_region(modis_grid,list_tiles_modis) #problem...this does not 

                                                                      #align with extent of modis LST!!!

#1) LST climatology: project, mosaic

tmp_str2<- mixedsort(list.files(pattern=pat_str2))

pat_str1 <- glob2rx(paste("mean","*.tif",sep=""))

tmp_str1<- mixedsort(list.files(pattern=pat_str1))

#add lines using grep to select tiles...

#list_date_names<-as.character(0:11)

#lsit_date_names<-month.abb

new_list<-vector("list",length(mosaic_list))

  j<-grep(list_date_names[i],mosaic_list,value=FALSE)

  names(mosaic_list)[j]<-list_date_names[i]

out_rastnames<-paste(list_date_names,out_rastnames,sep="")

nobs_m_list<-mosaic_raster_list(mosaic_list,out_rastnames,out_path)

plot(stack(nobs_m_list))

##Now mosaic for mean: should reorder files!!

pat_str1 <- glob2rx(paste("mean","*.tif",sep=""))

tmp_str1<- mixedsort(list.files(pattern=pat_str1))

out_rastnames<-paste("_lst_","mean",out_suffix,sep="")

list_date_names<-c("jan","feb","mar","apr","may","jun","jul","aug","sep","oct","nov","dec")

mosaic_list<-split(tmp_str1,list_date_names)

new_list<-vector("list",length(mosaic_list))

for (i in 1:length(list_date_names)){

  j<-grep(list_date_names[i],mosaic_list,value=FALSE)

  names(mosaic_list)[j]<-list_date_names[i]

  new_list[i]<-mosaic_list[j]

}

mosaic_list<-new_list

out_rastnames<-paste(list_date_names,out_rastnames,sep="")

mean_m_list<-mosaic_raster_list(mosaic_list,out_rastnames,out_path)

plot(stack(mean_m_list))

#Use this as ref file for now?? Ok for the time being: this will need to change to be a processing tile.

ref_rast<-raster(mean_m_list[[1]]) 

# +proj=sinu +lon_0=0 +x_0=0 +y_0=0 +a=6370997 +b=6370997 +units=m +no_defs for ref_rast

#"+proj=sinu +lon_0=0 +x_0=0 +y_0=0 +a=6371007.181 +b=6371007.181 +units=m +no_defs" ??

#reassign proj from modis tile to raster? there is a 10m diff in semi-axes...(a and b)

#min_val<-(-15+273.16) #if values less than -15C then screen out (note the Kelvin units that will need to be changed later in all datasets)

#LST[LST < (min_val)]<-NA

##2) Crop and reproject Canopy height data

canopy_name<-file.path(in_path,infile5)

#new_proj<-proj4string(canopy)                  #Assign coordinates reference system in PROJ4 format

#region_temp_projected<-spTransform(modis_tiles,CRS(new_proj))     #Project from WGS84 to new coord. system

#canopy_crop_rast<-crop(canopy, region_temp_projected) #crop using the extent from teh region tile

#canopy_projected_rast<-projectRaster(from=canopy_crop_rast,crs=proj4string(modis_tiles),method="ngb")

#s_raster<-dropLayer(s_raster,pos)

#CANHEIGHT[is.na(CANHEIGHT)]<-0

CANHEIGHT<-create_raster_region(canopy_name,ref_rast)

#3) Creating elev, aspect, slope from STRM

#region_temp_projected<-spTransform(modis_tiles,CRS(new_proj))     #Project from WGS84 to new coord. system

#SRTM_crop_rast<-crop(SRTM, region_temp_projected) #crop using the extent from teh region tile

#SRTM_projected_rast<-projectRaster(from=SRTM,crs=proj4string(modis_tiles),method="ngb")

#ELEV_SRTM<-raster(s_raster,layer=pos)             #Select layer from stack on 10/30

#s_raster<-dropLayer(s_raster,pos)

#ELEV_SRTM[ELEV_SRTM <0]<-NA

#extent...For the time being just use sinusoidal projection.

###calculate slope and aspect

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

pos<-match("slope",names(terrain_rast)) #Find column with name "value"

r2<-raster(terrain_rast,layer=pos)             #Select layer from stack

E<-sin(r1)

Nw<-sin(r2)*cos(r1)   #Adding a variable to the dataframe

Ew<-sin(r2)*sin(r1)   #Adding variable to the dataframe.

#4) LCC land cover

setwd(lc_path)

#lc_name<-"con_1km_class_1.tif"

lc_list<-list.files(pattern="con_1km_class_.*.tif")

#lc<-raster(file.path(lc_path,lc_names))

for (i in 1:length(lc_list)){

  lc_name<-lc_list[[i]]

  raster_name<-paste(data_name,out_suffix,".tif", sep="")

  writeRaster(lc_reg, filename=file.path(out_path,raster_name),overwrite=TRUE)  

  lc_reg_list[[i]]<-file.path(out_path,raster_name)

}

setwd(out_path)

LC12<-raster(lc_reg_s,layer=nlayers(lc_reg_s)) #this is open water

LC_mask[LC_mask==100]<-NA

LC_mask <- LC_mask > 100

#5) DISTOC, distance to coast: Would be useful to have a distance to coast layer ready...

#tab_freq<-freq(rc)

#Modify at a later stage:

#raster<-"DISTOC_VE_01292013.rst"  

#ocean_rast[ocean_rast==0]<-NA

#load DISTOC produced from IDRISI: automate the process later

distoc_reg<-raster(file.path(in_path,"DISTOC_VE_01292013.rst"))

#6) X-Y coordinates and LAT-LONG: do not keep in memory?

xy <-coordinates(r1)  #get x and y projected coordinates...

CRS_interp<-proj4string(r1)

y <-init(r1,v="y")

lon <-x

lat <-lon

lon <-setValues(lon,xy_latlon[,1]) #longitude for every pixel in the processing tile/region

lat <-setValues(lat,xy_latlon[,2]) #latitude for every pixel in the processing tile/region

rm(r1)

##Step 3: combine covariates in one stack for the next work flow stage

#Create a stack in tif format...

rnames<-c("x","y","lon","lat","N","E","N_w","E_w","elev","slope","aspect","CANHEIGHT","DISTOC")

names(r)<-rnames

s_raster<-r

#Add landcover layers

lc_names<-c("LC1","LC2","LC3","LC4","LC5","LC6","LC7","LC8","LC9","LC10","LC11","LC12")

names(lc_reg_s)<-lc_names #assign land cover names

s_raster<-addLayer(s_raster, lc_reg_s)

lst_s<-stack(c(as.character(mean_m_list),as.character(nobs_m_list)))

lst_names<-c("mm_01","mm_02","mm_03","mm_04","mm_05","mm_06","mm_07","mm_08","mm_09","mm_10","mm_11","mm_12",

             "nobs_01","nobs_02","nobs_03","nobs_04","nobs_05","nobs_06","nobs_07","nobs_08",

             "nobs_09","nobs_10","nobs_11","nobs_12")

names(lst_s)<-lst_names

s_raster<-addLayer(s_raster, lst_s)

covar_names<-c(rnames,lc_names,lst_names)

names(s_raster)<-covar_names

#Write out stack of number of change 

data_name<-paste("covariates_",out_region_name,"_",sep="")

raster_name<-paste(data_name,out_suffix,".tif", sep="")

s_raster<-mask(s_raster,LC_mask,filename=raster_name,

               overwrite=TRUE,NAflag=-999,bylayer=FALSE,bandorder="BSQ")

################### END OF SCRIPT #####################