| lars {seas} | R Documentation |
Read and write data from the LARS-WG stochastic weather generator file formats; also convert to a format for HELP
# read synthetic or observed *.st file read.lars(stfile, year.offset=0) # write observed climate data (*.st and/or *.sr) write.lars(x,stfile,datfile,site,lat,lon,alt) # experimental functions (may not work great; or at all!) lars2help(infile, outfile, year.offset, site) write.lars.scenario(file, x1, x2, name="anomaly")
stfile |
file name with ‘*.st’ extension; this is a
‘site file’ for LARS-WG which contains meta-data for the
climate data, and has the location of the the climate data file; for
write.lars, if this variable is NA or
FALSE, this file will not be written (however,
datfile must be defined) |
datfile |
file name with either ‘*.sr’ or ‘*.dat’
extension; contains climate data, as described by stfile;
this does not need to be set if stfile is defined, as this
datum is found in the ‘st’ file |
file |
file name with a ‘*.sce’ extension; this is a ‘scenario’ file with absolute and relative changes of climate data |
infile |
input file |
outfile |
output file |
x |
data.frame of climate data |
x1 |
same as x |
x2 |
same as x |
year.offset |
offset of years between what is contained in the
data files and what is needed in R to produce a reasonable
‘Date’; this is required, for example, if
synthetic data are produced that start from an arbitrary year
‘1’ but represent climate from the year ‘2000’ |
site |
same as ‘[SITE]’ in ‘st’ file; if missing,
this will try to read from attr(x$name); this is the same as
a ‘region’ for HELP |
name |
scenario name |
lat |
same as ‘LAT’ in ‘st’ file; if missing, this
will try to be read from attr(x$latitude) |
lon |
same as ‘LON’ in ‘st’ file; if missing, this
will try to be read from attr(x$longitude) |
alt |
same as ‘ALT’ in ‘st’ file; if missing, this
will try to be read from attr(x$elevation) |
These functions interface with the LARS-WG files (Version 4.0), which is a stochastic weather generator by Mikhail Semenov.
The climate data files used with LARS-WG have two parts: (1)~a ‘site file’ with a ‘st’ extension, containing the meta-data; and (2)~a data file with a ‘*.sr’ or ‘*.dat’ extension, containing all the data. The variable names are translated according to the following table:
| seas | LARS-WG |
year | ‘YEAR’ |
yday | ‘JDAY’ |
t_min | ‘MIN’ |
t_max | ‘MAX’ |
preicp | ‘RAIN’ |
solar | ‘RAD’ |
sun | ‘SUN’ |
pet | ‘PET’ |
To write climate data from R to a LARS-WG file, the data.frame
names need to match those in the seas-side of the table.
Data exported from write.lars always has legal (according to
the Gregorian calendar) and increasing sequence of days (even if there
are gaps in x$date). Missing data values are written as
-99.
Synthetically generated data from LARS-WG use a 365-day calendar, and
may need to be converted to a Gregorian calendar, which can be done
using conv365toGregorian.
lars2help and write.lars.scenario are experimental
functions to translate data between LARS and HELP (see
write.help for more info).
M.W. Toews
LARS-WG was can be downloaded for academic and research uses from
http://www.rothamsted.bbsrc.ac.uk/mas-models/larswg.php
Semenov, M.A. and Barrow, E.M. (1997) “Use of a stochastic weather generator in the development of climate change scenarios”. Climate Change: 35, 397–414. doi: 10.1023/A:1005342632279
write.help, read.sdsm,
summerland example synthetic data,
conv365toGregorian
stfile <- system.file("data/summerland.st",package="seas")
print(stfile)
summ <- read.lars(stfile,year.offset=1960)
head(summ)
str(summ)
# plot temperature
summ$t_mean <- rowMeans(summ[,c("t_min","t_max")])
plot.seas.temp(summ)
# plot solar radiation
plot.seas.var(summ,"solar")
# plot precipitation
summ.ss <- seas.sum(summ)
image(summ.ss)
plot(seas.norm(summ.ss))