Hydrological data import - Revision history

Wbuytaert at 20:04, 25 June 2008

2008-06-25T20:04:11Z

Wbuytaert at 10:23, 10 June 2008

2008-06-10T10:23:48Z

Wbuytaert at 22:22, 7 March 2008

2008-03-07T22:22:07Z

Wbuytaert at 22:20, 7 March 2008

2008-03-07T22:20:17Z

New page

This example is primarily intended to be used with the topmodel() function. However, it can be useful for other analysis such as baseflow separation.

In the example, all input files are supposed to be (.txt) files. (Each decent spreadsheet converts easily to text files with "save as"). All inputs are given in meters per time unit

==Rainfall==

In this case we have the rainfall data from a tipping bucket gauge. This means that data are stored as the moments of the tipping of the bucket. The resolution of the gauges is 0.2 mm, so each entry represents 0.2 mm of rainfall. An input file, say rain.txt could look like this:

24/05/2001 00:52:46
24/05/2001 01:07:37
24/05/2001 01:22:51
24/05/2001 01:27:00
25/05/2001 01:05:17
...

The code we use to import these data into R is:

rain <- read.table("rain.txt")
rain <- paste(rain[,1],rain[,2])
rain <- strptime(as.character(rain), format = "%d/%m/%Y %H:%M:%S")
rain <- as.POSIXct(rain, tz="GMT")

NOTE: be sure to set the time zone correctly. Importing data from a region without summer time on a computer configured for a time zone with summer time is problematic. You can change the time zone in R with:

Sys.putenv(TZ="GMT")

Then we transform the data to rainfall per time unit, say, per 15 min:

rain <- rain+15*60
rain <- cut(c(trunc(rain[1],"hour"),rain), "15 min")
rain <- tapply(rep(0.2,length(rain)),rain,"sum")
rain[1] <- rain[1]-0.2

Note the first line of this code block: we shift all the entries 15 min. to the future. This makes sure that the rainfall associated with each time step is the rainfall that occurred over the past time interval (and not the coming interval).

==Discharge==

Discharge is easier to import and does not require much preprocessing. The same goes for rainfall which is already a time series. For instance, if the input file is:

04/09/2001 15:30 0.0492
04/09/2001 16:00 0.0495
04/09/2001 16:30 0.0503
04/09/2001 17:00 0.0504
04/09/2001 17:30 0.0500
...

Import these data:

input <- read.table("discharge.txt")
time <- paste(input[,1],input[,2])
time <- strptime(as.character(time), format = "%d/%m/%Y %H:%M")
time <- as.POSIXct(time, tz="GMT")
level <- input[,3]
discharge <- data.frame(time,level)
rm(input,time,level)

Strictly, we do not need the time information in the discharge and rain objects, but it is handy to have them around, for instance to check that both series are synchronous.

← Older revision		Revision as of 20:04, 25 June 2008
Line 1:		Line 1:
−	~~[[Category:Projects]]~~
−	~~[[Category:Hydrology in R]]~~

−	~~This example is primarily intended to be used with the topmodel() function. However, it can be useful for other analysis such as baseflow separation.~~
−
−	~~In the example, all input files are supposed to be (.txt) files. (Each decent spreadsheet converts easily to text files with "save as"). All inputs are given in meters per time unit~~
−
−	~~==Rainfall==~~
−
−	In this case we have the rainfall data from a tipping bucket gauge. This means that data are stored as the moments of the tipping of the bucket. The resolution of the gauges is 0.2 mm, so each entry represents 0.2 mm of rainfall. An input file, say rain.txt could look like this:
−
−	~~24/05/2001 00:52:46~~
−	~~24/05/2001 01:07:37~~
−	~~24/05/2001 01:22:51~~
−	~~24/05/2001 01:27:00~~
−	~~25/05/2001 01:05:17~~
−	~~...~~
−
−	~~The code we use to import these data into R is:~~
−
−	~~rain <- read.table("rain.txt")~~
−	~~rain <- paste(rain[,1],rain[,2])~~
−	~~rain <- strptime(as.character(rain), format = "%d/%m/%Y %H:%M:%S")~~
−	~~rain <- as.POSIXct(rain, tz="GMT")~~
−
−	NOTE: be sure to set the time zone correctly. If you are working on data from another time zone it may be advisable to set the environment variable accordingly, e.g. to avoid problems with daylight saving time. You can set the time zone in R with:
−
−	~~Sys.putenv(TZ="GMT")~~
−
−	~~Then we transform the data to rainfall per time unit, say, per 15 min:~~
−
−	~~rain <- rain+15*60~~
−	~~rain <- cut(c(trunc(rain[1],"hour"),rain), "15 min")~~
−	~~rain <- tapply(rep(0.2,length(rain)),rain,"sum")~~
−	~~rain[1] <- rain[1]-0.2~~
−
−	Note the first line of this code block: we shift all the entries 15 min. to the future. This makes sure that the rainfall associated with each time step is the rainfall that occurred over the past time interval (and not the coming interval).
−
−	~~==Discharge==~~
−
−	~~Discharge is easier to import and does not require much preprocessing. The same goes for rainfall which is already a time series. For instance, if the input file is:~~
−
−	~~04/09/2001 15:30 0.0492~~
−	~~04/09/2001 16:00 0.0495~~
−	~~04/09/2001 16:30 0.0503~~
−	~~04/09/2001 17:00 0.0504~~
−	~~04/09/2001 17:30 0.0500~~
−	~~...~~
−
−	~~Import these data:~~
−
−	~~input <- read.table("discharge.txt")~~
−	~~time <- paste(input[,1],input[,2])~~
−	~~time <- strptime(as.character(time), format = "%d/%m/%Y %H:%M")~~
−	~~time <- as.POSIXct(time, tz="GMT")~~
−	~~level <- input[,3]~~
−	~~discharge <- data.frame(time,level)~~
−	~~rm(input,time,level)~~
−
−	~~Strictly, we do not need the time information in the discharge and rain objects, but it is handy to have them around, for instance to check that both series are synchronous.~~

← Older revision		Revision as of 22:22, 7 March 2008
Line 1:		Line 1:
		+	[[Category:Projects]]
		+	[[Category:Hydrology in R]]
		+
	This example is primarily intended to be used with the topmodel() function. However, it can be useful for other analysis such as baseflow separation.		This example is primarily intended to be used with the topmodel() function. However, it can be useful for other analysis such as baseflow separation.

@@ Line 24: / Line 24: @@
      rain <- as.POSIXct(rain, tz="GMT")
-NOTE: be sure to set the time zone correctly. Importing data from a region without summer time on a computer configured for a time zone with summer     time is problematic. You can change the time zone in R with:
+NOTE: be sure to set the time zone correctly. If you are working on data from another time zone it may be advisable to set the environment variable accordingly, e.g. to avoid problems with daylight saving time. You can set the time zone in R with:
      Sys.putenv(TZ="GMT")