Benthic index (TBBI) outcome by bay segment and year
Arguments
- benthicdata
raw benthic monitoring data, e.g.
tbeptools::benthicdata
Value
A data.frame with columns yr, bay_segment,
TBBI (median station-level score for that bay segment/year), and
outcome (0-1, 1 = best)
Details
Uses anlz_tbbiscr to get station-level
TBBI scores (0-100), filtered to the same stations
anlz_tbbimed uses for its bay segment grades
(FundingProject == "TBEP", ProgramID == 4, and
TBBI between 0 and 100, for the 7 bay segments
anlz_tbbimed scores), then takes the median
station score for each bay segment/year. outcome comes directly
from that continuous median via util_outcome (type =
"continuous", from = c(73, 87)), TBBI's own grade breakpoints
(Degraded below 73, Intermediate 73-87, Healthy above 87). This is the
same clamped-breakpoint-window treatment used for the Nekton Index's
from = c(32, 46).
This bypasses anlz_tbbimed entirely, which
instead grades a bay segment/year Poor/Fair/Good from the
proportion of its stations falling in each of the
Degraded/Intermediate/Healthy categories. This is a compound rule on
those proportions, not a discretized version of a single continuous
statistic, so there's no direct continuous equivalent of it the way
there is for TBNI's or PEL/TEL's breakpoints. Scoring the median of the
raw station values instead is a related but distinct measure. It can
disagree with anlz_tbbimed's category in some
edge cases (e.g.
a bay segment split between many Healthy and a few very Degraded
stations can have a middling median while still tripping the proportion
rule's Degraded >= 0.2 condition for "Poor").
Examples
anlz_sed_tbbi(tbeptools::benthicdata)
#> # A tibble: 222 × 4
#> yr bay_segment TBBI outcome
#> <dbl> <chr> <dbl> <dbl>
#> 1 1995 BCB 87.1 1
#> 2 1996 BCB 84.1 0.793
#> 3 1997 BCB 78.2 0.368
#> 4 1998 BCB 83.6 0.755
#> 5 1999 BCB 80.3 0.523
#> 6 2000 BCB 85.4 0.885
#> 7 2001 BCB 80.3 0.522
#> 8 2002 BCB 78.7 0.406
#> 9 2003 BCB 84.4 0.811
#> 10 2004 BCB 85.3 0.879
#> # ℹ 212 more rows
