Herbarium-Derived Phenological Data in North America

We present infrastructure for developing large-scale and long-term
phenological datasets across multiple herbaria, as well as a sample
dataset that has been acquired from the digital archives of 440 distinct
herbaria across North America and further processed to evaluate
phenological status. This dataset contains 2,319,672 specimen records of
plants collected while reproductively active. These data have been
modified to explicitly codify the observed phenological status of each
specimen at the time of collection, and to remove specimens for which
information essential to assessing their phenology or the corresponding
climate conditions in the year and location of collection were
missing.  As different collectors have used distinct taxonomic
schema over space and time in documenting the specimens being collected,
these data were also rectified into a single unified taxonomic schema to
ensure that consistent taxon names were used throughout the
dataset.  Further, this data has been united with long-term and
annual climate conditions in the year and location of collection, as
derived from PRISM climate data
(https://www.prism.oregonstate.edu/).  To date, this data
includes 2,319,672  specimens across 25,429 plant taxa. However,
this represents a living dataset that will continue to be updated as
digitization efforts proceed and additional digital specimen records
become available.