View source for Digital terrain analysis with RHydro

This is an example using the DEM supplied with the package. Load the package and the DEM:

  library(RHydro)
  data(huagrahuma.dem)

Fill depression and look at the difference between the original and filled map:
 
  dem.filled <- sinkfind(huagrahuma.dem, cellsize=25, degree=0.1)
  library(lattice)
  levelplot(dem.filled)
  
  difference <- dem.filled - huagrahuma.dem
  max(difference)
  min(difference)
  levelplot(difference)

Topographic index calculation:

  topidx <- atb(dem.filled, cellsize=25)
  class(topidx)
  str(topidx)
  levelplot(topidx$atb)
  levelplot(topidx$area)

If you want, you can compare with the analysis on the non-filled DEM that contains obvious artefacts 

  old <- atb(huagrahuma.dem, cellsize=25)
  levelplot(old$area)
  levelplot(topidx$area - old$area)

Have a look at the potential outlet:

  outlet(topidx$area,c(28,8),2)

We know from manual catchment delineation (on a topographic map) that the area of the catchment is around 2.4 km2, so the following cells are likely to contain the outlet: c(28,8); c(29,8); c(29,9). Just try them out with the subcatch() function and take the one you like.

  catchment <- subcatch(dem.filled, c(29,8)) 
  image(catchment)
  catchment[catchment == 0] <- NA

  flowlength <- flowlength(dem.filled, c(29,8))

Headwater cells are identified on the basis of a topographic index
or an accumulation area threshold. This is an “intuitive” choice:

  rivers <- river(dem.filled,topidx$atb,topidx$area,cellsize=25,thatb=12.35,tharea=10000)
  image(rivers)

Play around with the values of thatb and tharea and look at the impact…

Mask the rivers and topographic index maps outside the catchment:

  rivers <- rivers * catchment
  topidxmap <- topidx$atb * catchment

Eliminate the NA values from the topidxmap:

  topidxarray <- as.vector(topidxmap)
  topidxarray <- topidxarray[!is.na(topidxarray)]

And make topographic index classes which can be used with topmodel:

  classes <- make.classes(topidxarray,16)
  plot(classes)

The object ''classes'' can be used as input for [[topmodel]]