Deriving Nutrient Criteria for the Benchmark Population

Deriving nutrient criteria from the benchmark population begins with a
population of nutrient values from the dataset from the proposed rule
and identified as “allstr00.txt” in the associated data files and
provided in docket at proposal 1/2010.  This data set contains data from
Florida’s STORET database and FDEP’s GWIS database, with data
screened to meet data quality objectives based on QA rule 62-160.  The
data set contains approximately 280,000 observations at 9,401 sites.

The following steps were followed to identify the final population of
benchmark sites from which the percentiles are calculated:

All records from allstr00.txt were merged with the file containing
associated WBID identifier codes and nutrient regions
(“sites01.final.txt”, provided by FDEP).  The unique site identifier
code (“ORG_Station”) was used for this merge.

All samples with TN or TP data were retained.

All records were merged by the unique site identifier code with land
cover information in the riparian corridor (“ldi.corridor.txt”). 
All samples with a Landscape Development Index (LDI) >2 were flagged for
exclusion (“Exclude = 1”).  Sites with no riparian corridor LDI
value were also excluded.

All records were merged by the unique site identifier code to a summary
of “best professional judgment” by Florida DEP
(“s.bpj.site.txt”).  This summary was based on expert opinion of
district biologists, site visits by FDEP staff, digital ortho-quad
aerial surveys, FDEP staff discussions, and EPA analysis of LDI.   Sites
identified in this screen were flagged (“Exclude = 1”).  Sites with
no documentation to indicate a best professional judgment summary were
also excluded.

Sites in WBIDs  matching WBID identifier codes on EPAs  303(d) list
(“s.303d.txt”) based on impairments due to nutrients and dissolved
oxygen were flagged (“Exclude =1”).

Sites in WBIDs with a watershed landscape development index, LDI >3,
were flagged (“s.ldi.txt”) (“Exclude = 1”).

Sites in WBIDs with average nitrate/nitrite concentrations
(“s.nox.new.txt”) > 0.35 mg/L were flagged (“Exclude = 1”).

Sites in WBIDs with an average stream condition index (SCI) scores
(“s.sci.wbid.txt”) <40 were flagged (“Exclude = 1”).

The following steps were followed to derive the criteria, based on this
benchmark data set:

A new unique identifier was created for each sample based on the WBID
and the year in which the sample was collected.  

Final distributions for each of the 5 nutrient regions were computed as
the geometric mean values of TN and TP for each WBID-Year.   

Only WBID-Years calculated using at least 4 samples were included in the
final distributions.

