The climdexInput
class consists of all the data necessary to compute
the climdex indices. Users will not need to modify any of the slots in this
class. That being said, users may want or need to repurpose this data for
further analysis. The following description of the data is aimed at that
audience.
The data
slot contains time series' of daily data of equal length for
each of the provided variables. Missing days have been replaced with NA.
The dates
slot is the corresponding series of dates (of type PCICt)
for the daily data.
The quantiles
slot contains quantiles used for computing the
tn/tx 10/90p indices, w/csdi, r95ptot, and r99ptot. If precipitation data
is supplied, the 'prec' member contains the 95th and 99th percentile values
for precipitation within the base period. For tmin and tmax, if present each
will have a corresponding member in the slot. Within each of these, there
will be an 'inbase' and 'outbase' member, corresponding to thresholds to be
used within the base period (inbase) and outside the base period (outbase).
The 'inbase' member consists of one percentile for each day of the year,
computed using an n-day (default is 5-day) running window surrounding that
day. These percentiles are computed for at least the 10th and 90th
percentile of the data. For the 'outbase' member, given n years
of data to use as the base period, there are n * (n - 1) sets of daily
quantiles of the same type as those in 'inbase'.
To ease computation of monthly and annual data, date.factors
contains date factors which group data into annual and monthly time
buckets. They are of the same length as the time series and can be reused
for computation of any annual or monthly aggregates.
The climdexInput class also includes NA masks for both monthly
and annual as parts of the namasks
slot. Each of these masks consist
of a vector of numbers of the same length as the monthly or annual output
data. The values used are 1 to signify that the data meets the QC criteria,
and NA to signify it does not. Years with more than (by default) 15 days
missing, and months with more than (by default) 3 days missing, are
considered to be of poor quality and are masked here with NA. These
thresholds can be set when instantiating the object, and are stored in the
max.missing.days
slot.
The base.range
slot contains vector of type PCICt containing the
first and last day included in the baseline.
The northern.hemisphere
slot contains a boolean indicating whether
the data came from the northern hemisphere. If FALSE, data is assumed to
have come from the southern hemisphere. This is used when computing growing
season length; if the data is from the southern hemisphere, growing season
length is the growing season starting in the beginning of July of the year
indicated, running to the end of June of the following year.
The max.missing.days
slot is a vector consisting of 'annual'
(the number of days that can be missing in a year) and 'monthly' (the
number of days that can be missing in a month. If one month in a year fails
the test, the corresponding year will be omitted.