incorrect identification of leap years; multiple implementations.

[dlebauer]

The logic to test if a year is a leap year has been implemented perhaps a dozen times independently within the PEcAn code base, including once as the function is.leap.
2. there are many cases where year %% 4 i== 0 is used although it should be year %% 400 == 0 | (year %% 4 == 0 & year %% 100 != 0) (is.leap and leap_year implement this correctly).

1. The lubridate package has a function leap_year that returns T/F. We already use this function in many places and I suggest that this is a good solution
2. (just a comment / food for thought) testing if a year is a leap year is often followed by calculating time steps (e.g. days or seconds/year). lubridate also provides convenience functions for this that would be less error prone than declaring seconds / year or days / month constants repeatedly in the code.

TODO

• bugfix: All of the invalid uses of year %% 4 i== 0 should be replaced by a valid function (leap_year)
• enhancement: deprecate is.leap and replace it with lubridate::leap_year
• documentation add some notes about existing 'helper' functions to the model template README

[robkooper]

Other handy functions to add:

days_in_year <- function(year) {
  ifelse(lubridate:: leap_year(year), 366, 365)
}

hours_in_year <- function(year, every=24) {
  24/every * ifelse(lubridate:: leap_year(year), 366, 365)
}

now we can simply call `24/6 * days_in_year(1999)`` to get the value for every 6 hours.

[ashiklom]

Closed by #1623.