Aggregation of time-series data

Aggregation of time-series data and start/end time selection. This function provides the option to select the temporal step size for aggregation of a single time series origination from an is.trex-compliant object. Additionally, the user can define the start and end time of the series and select the function used for aggregation.

agg.data(input,
 time.agg = 60*24,
 start = "2012-07-28 00:00",
 end = "2012-10-29 00:00",
 FUN = "mean",
 unit = 60,
 na.rm = TRUE,
 df = FALSE)

Arguments

input: An is.trex-compliant time series from tdm_cal.sfd outputs (e.g., X$sfd.mw$sfd).
time.agg: Numeric, the aggregation time in minutes (default = 60).
start: Character string, the start time for the series. Format has to be provided in "UTC" (e.g. "2012-05-28 00:00" or Year-Month-Day Hour:Minute). Starting time should not be earlier than the start of the series. If not provided the entire series is considered.
end: character string, the end time for the series. Format has to be provided in "UTC" (e.g. "2012-06-28 00:50" or Year-Month-Day Hour:Minute). Starting time should be earlier than the end time and the end time should not exceed the length of the series. If not provided the entire series is considered.
FUN: Quoted function name to compute the summary statistics which can be applied to all data subsets (see aggregate; including "sum", "mean", "median", "sd", "se", "min", "max").
unit: Numeric, the minutes in which a velocity unity is provided (e.g., $cm^3 cm^{-2} h^{-1} = 60$) for summation (FUN = “sum”; default = 60).
na.rm: Logical; if TRUE missing values are removed (default = TRUE).
df: Logical; if TRUE, output is provided in a data.frame format with a timestamp and a value column. If FALSE, output is provided as a zoo vector object (default = FALSE).

Value

A zoo object or data.frame in the appropriate format for other functionalities.

Details

Time series are often derived at variable resolutions. This function provides the option to aggregate (homogenize) time steps with standard FUN statistics. When applying this function to calculate summed sap flow values (e.g., $cm^3 cm^{-2} d^{-1}$) one needs to include the velocity unit, as the summation is dependent upon the minimum timestep of the time series (e.g., $cm^3 cm^{-2} h^{-1}$, unit = 60).

Examples

if (FALSE) {
#aggregate SFD values to mean hourly and daily sums

raw   <- example.data(type="doy")

input <- is.trex(raw,tz="GMT",time.format="%H:%M",
                  solar.time=TRUE,long.deg=7.7459,ref.add=FALSE,df=FALSE)

input[which(input<0.4)]<-NA

k.input<-tdm_dt.max(dt.steps(input,time.int=15,
               max.gap=180,decimals=10),methods=c("mw"))

sfd.input<-tdm_cal.sfd(k.input,make.plot=FALSE,
                  df=FALSE,wood="Coniferous")$sfd.mw$sfd

# means
output.1hmean <- agg.data(sfd.input,
                       time.agg=60,
                       start="2012-07-28 00:00",
                       end="2012-08-29 00:00",
                       FUN="mean",
                       na.rm=TRUE,
                       df=FALSE)
output.6hmean <- agg.data(sfd.input,
                         time.agg=60*6,
                         start="2012-07-28 00:00",
                         end="2012-08-29 00:00",
                         FUN="mean",
                         na.rm=TRUE,
                         df=FALSE)
plot(output.1hmean,col="cyan")
lines(output.6hmean,col="black")

# daily sums
output.dsum<-agg.data(sfd.input,
                      time.agg=60*24,
                      start="2012-07-28 00:00",
                      end="2012-10-29 00:00",
                      FUN="sum",
                      unit=60,
                      na.rm=TRUE,
                      df=FALSE)
plot(output.dsum)
points(output.dsum,pch=16)
}